Zoomorphology (2005) 124: 9–26 DOI 10.1007/s00435-005-0107-1
O R I GI N A L A R T IC L E
Iain C. Wilkie Æ Mary McKew M. Daniela Candia Carnevali
Functional morphology of the compass-rotular ligament of Echinus esculentus (Echinodermata: Echinoida): a non-mutable collagenous component of Aristotle’s lantern Received: 21 November 2003 / Accepted: 20 October 2004 / Published online: 2 February 2005 Springer-Verlag 2005
Abstract This paper provides the first detailed account of the histological and ultrastructural organisation of an echinoderm ligament that is non-mutable. Each of the five compass ossicles on the aboral side of the lantern of regular echinoids is linked to an underlying rotular ossicle by a compass-rotular ligament (CRL). The structure and anatomical relations of the CRL of Echinus esculentus L. were examined by light microscopy and scanning and transmission electron microscopy, and its responsiveness to neuroactive agents was observed in stress relaxation tests. The CRL consists of (1) thick collagen fibres that form an outer crossed fibre lattice and an internal system of parallel suspensory fibres and (2) regions composed mainly of bundles of ca. 12 nm microfibrils. The CRL is sparsely cellular, most cell bodies belonging to heterogeneous vesicle-containing cells that undergo apoptosis-like cytoplasmic fragmentation and produce linear aggregations of cell fragments. Coelothelia investing the CRL show intense secretory activity reminiscent of lamellar body production by mammalian mesothelia and are the source of bundles of cell processes that penetrate deeply into the CRL, branch sparingly and terminate between the thick fibres. Isolated CRL preparations are unresponsive to seawater containing 1 mM acetylcholine or 100 mM K+, but show a transient increase in stress relaxation rate when treated with 0.56 M KCl. The effects of solutions of other alkali metal chlorides suggest that the latter response is due to the direct action of KCl on extracellular components of the CRL and is not nervously mediated. I. C. Wilkie (&) Æ M. McKew Department of Biological and Biomedical Sciences, Glasgow Caledonian University, 70 Cowcaddens Road, Glasgow, G4 0BA, Scotland E-mail:
[email protected] Tel.: +141-3318515 Fax: +141-3313208 M. D. Candia Carnevali Dipartimento di Biologia ‘‘Luigi Gorini’’, Universita` degli Studi di Milano, Via Celoria 26, 20133 Milano, Italy
Keywords Apoptosis Æ Collagen Æ Mechanical properties Æ Mesothelium Æ Microfibrils
Introduction Collagenous connective tissue is a ubiquitous structural material in the bodies of most Metazoa. Amongst the best studied is that possessed by Echinodermata that has the apparently unique capacity to undergo rapid, nervously mediated changes in passive mechanical properties and is known as ‘mutable collagenous tissue’ (MCT). There is an extensive literature on the organisation (from the gross to the molecular levels), biomechanics and physiology of MCT (reviewed by Wilkie 1996, 2002), more recent contributions to which highlight its structural complexity (Trotter et al. 1996, 1999; Thurmond et al. 1997), demonstrate that it has the additional capacity for active force generation (Birenheide et al. 2000; Motokawa et al. 2004) and suggest that its variable tensility depends on the secretion from effector cells of organic molecules that modulate interactions between extracellular components (Koob et al. 1999; Szulgit and Shadwick 2000; Trotter et al. 2000; Tipper et al. 2003). As a result of this preoccupation with MCT, the question of whether or not conventional, i.e. nonmutable, collagenous tissue occurs in the Echinodermata was neglected, even though a detailed comparison of such tissue and MCT might help to pinpoint the structural and physiological correlates of variable tensility in the latter. There is limited evidence that a few echinoderm structures consist of non-mutable collagenous tissue, viz the ‘non-autotomy tendons’ of ophiuroid intervertebral muscles, which, unlike the ‘autotomy tendons’, never undergo rupture during arm detachment (Wilkie and Emson 1987); certain ligaments in the ophiuroid mouth-frame that are unresponsive to elevated [K+] (Candia Carnevali et al. 1994); and the central spine ligament of cidaroid echinoids, which,
10
unlike that of diadematoid echinoids (Motokawa 1983), lacks possible effector cells and is unresponsive to acetylcholine (del Castillo et al. 1995). However, information on all these structures is scant and the reported evidence for their non-mutability is entirely qualitative. The compass-rotular ligament (CRL) is one of several collagenous structures in the echinoid lantern that have attracted little attention from echinoderm biologists. In a preliminary study, it was found that the CRL in two echinoid species is insensitive to physicochemical treatments that change consistently the passive mechanical properties of mutable collagenous structures, although the significance of this observation was obscured by what appeared to be interspecific discrepancies in the effects of other agents (Wilkie et al. 1995). A recent examination of the structure and physiology of the CRL of four Mediterranean echinoids provided more evidence that the ligament does not consist of MCT (Wilkie et al. 2003). The present paper presents the results of a morphological and experimental investigation of the CRL of the NE Atlantic echinoid Echinus esculentus Linne´, 1758, which supports the view that it also is non-mutable.
Materials and methods
specimens in 2% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) containing 1.4% NaCl in vials that were surrounded by ice for 1 h, then taken out of the ice and left at 4C for a further 2 h. The specimens were washed overnight in buffer at 4C, post-fixed in 1% OsO4 in buffer for 2 h at room temperature, washed in distilled water then 25% ethanol, block-stained with 2% uranyl acetate in 25% ethanol for 2 h in the dark and embedded in araldite. These specimens were not decalcified. Ultrathin sections were mounted on carbon-coated grids, stained with uranyl acetate and lead citrate, and observed in a JEOL 100CX transmission electron microscope. Useful information was also obtained from material processed using seawater as a substitute for cacodylate buffer, some of which was decalcified with EDTA after secondary fixation. In some cases, 1% tannic acid was added to the primary fixative to improve the visualisation of extracellular components. After fixation, some specimens were block-stained for proteoglycans using the method of Erlinger (1995): they were washed in 0.2 M acetate buffer (pH 5.6) and stained overnight in 1% cupromeronic blue (Seigaku) in 0.2 M acetate buffer with 0.3% MgCl2; they were then washed in the previous solution without the dye, stained for 1 h in 0.5% NaWO4 in 0.2 M acetate buffer with 0.3% MgCl2, and left overnight in 0.5% NaWO4 in 30% ethanol.
Animals Specimens of E. esculentus of test diameter 56–102 mm that were collected in the Firth of Clyde, Scotland, by scuba divers were purchased from the Specimen Supply Department, University Marine Biological Station, Millport, transported to Glasgow Caledonian University and maintained in seawater aquaria at 8C.
Mechanical tests Isolated preparations of the CRL were obtained from lanterns exposed by removing the top half of the test. The compass elevator muscles were severed and each compass was transected about half way along its length
Light microscopy Specimens each consisting of a compass and rotula linked by a CRL were excised, fixed in neutral buffered formalin, decalcified with 3% nitric acid in 70% ethanol or with saturated EDTA (pH adjusted to 7.0 with KOH), and embedded in ‘Paraplast’. Sections 8 lm thick were stained with Milligan’s trichrome (Humason 1979), examined with an Olympus BH-2 microscope and photographed using an Olympus DP-50 digital camera and analySIS software. Electron microscopy For scanning electron microscopy, soft tissues were removed partly or completely from excised compass-rotula specimens with bleach. The preparations were then washed in water and ethanol, air-dried, coated with gold and viewed in a Philips 500 SEM. For transmission electron microscopy, several protocols were tried, none of which provided completely satisfactory preservation of cytological detail. The best results were obtained by fixing excised compass-rotula
Fig. 1 Diagrammatic aboral view of echinoid lantern with components of some segments omitted to show anatomical relations (adapted from Ma¨rkel 1979). cd Compass depressor ligament, cm compass elevator muscle, co compass ossicle, CRL compass-rotular ligament, ja jaw, oe location of oesophagus, pm protractor muscle, pg perignathic girdle (edge of test), rm retractor muscle
11
(see Fig. 1). The pharyngeal levator tendons and the ligaments connecting the rotulae to the underlying jaws were then cut and the five preparations, each consisting of one rotula and a half compass linked by a CRL, were lifted away and placed in seawater. The mechanical behaviour of a CRL was recorded by holding the rotula horizontally in a clamp attached to a manipulator and by linking the free end of the compass via a heart-clip and silver chain to a rigidly fixed isometric force transducer (Fig. 2A), care being taken to minimise disturbance to the CRL. At the start of the test, the manipulator was used to lower the clamp, which subjected the CRL to rotational strain, until a force (i.e. the stretch resistance of the CRL) of 8–10 gf was registered, at which point the position of the clamp was fixed. The recorded force always then decayed, a phenomenon known as ‘stress relaxation’, which results from the viscous flow or sliding apart of components within a structure under tension. As illustrated in Fig. 2B, the rate of stress relaxation, represented by the slope of the force-time curve, was initially rapid but decelerated to a low constant value. Preparations were allowed to relax undisturbed for 5 min at room temperature (19–24C) and then the effect of chemical agents on the rate of
Fig. 2A, B Experimental procedures. A Diagrammatic lateral view of preparation and apparatus. cl Clamp, it isometric transducer, ma manipulator. Other lettering as in previous figure. B Diagram of recording of stress relaxation test. Response index=DF+3/DF 3
stress relaxation was determined by pipetting solutions over the ligaments very gently in order to avoid mechanical artefacts. Effects were quantified by calculating the ‘response index’: force decay during the 3 min after the start of treatment expressed as a proportion of that during the 3 min before the start of treatment (Fig. 2B). Student’s t-tests or Mann–Whitney tests were applied to the results using the GraphPad InStat programme.
Results Gross anatomy, anatomical relations and mobility of the compass-rotular joint and ligament The rotulae interconnect the upper edges of the lantern jaws at mobile joints that allow the jaws to spread apart during feeding activity (Fig. 1) (see Ma¨rkel 1979; Candia Carnevali et al. 1991). Each rotula is surmounted by an elongated and curved compass ossicle. The adaxial (i.e. oesophagus-facing) end of the compass is hook-like and connected to the rotula by the CRL. Surface features of the ossicles at the compassrotular joint are illustrated in Fig. 3. The compass hook is located in an indentation in the adaxial end of the rotula (Fig. 3A, C). The aboral edge of the latter forms an overhang with an undulating margin. The floor of the crevice below the overhang consists of a series of radial ridges and grooves (Fig. 3C, D). The extreme oral-abaxial surface of the compass hook facing the rotula is sharply pointed and has on either side a shallow depression which is an attachment area for ligament fibres (Fig. 3E, F). The surface stereom (i.e. calcite meshwork) of both the compass and the rotula in the region of the joint is almost exclusively of the labyrinthic type (see Candia Carnevali et al. 1991), the only exception being short bands of fascicular stereom near the lateral ends of the rotular overhang. These are attachment areas for coelomic septa (Fig. 3D). It is significant that there is no microperforate or imperforate stereom, types that are present on articular bearing surfaces in the lantern and elsewhere in echinoderms (Candia Carnevali et al. 1991). This correlates with the fact that there is no direct contact between the compass and rotula, the compass being suspended within the adaxial indentation of the rotula by the CRL (Figs. 1, 3A, 4). In the intact lantern, the coelomic space between each compass and its underlying rotula, which is known as the ‘adapical coelomic cavity’ (Stauber 1993), is delimited laterally and adaxially by a coelomic septum consisting of two coelothelia separated by finely fibrous connective tissue. The CRL is a specialised region of the lower adaxial portion of this septum, in which the connective tissue layer is enormously expanded and highly modified in structure. The concave aboral surface of the CRL, which floors the adapical coelomic cavity, includes
12
a number of radially orientated ridges and grooves (Fig. 5C–F). With regard to the mobility of the compass-rotular joint, it was observed that, when anaesthetised lanterns
were turned completely upside-down in seawater, compasses that were attached by only the CRL flopped loosely downwards (i.e. in the anatomically aboral direction) by about 90C and were wafted back and forth
13 b Fig. 3A–F Scanning electron micrographs of rotula and compass ossicle of E. esculentus. A Adaxial side of partly digested specimen showing anatomical relations between compass, rotula and fibrous components of CRL (asterisks). B Right side of CRL shown in A, illustrating crossed-fibre arrangement of lattice fibres (arrow). C Adaxial side of rotula. In intact lantern, suspensory fibres of CRL are inserted into underside of overhang (arrowheads). Note ridges and grooves on floor of crevice below overhang. D Lateral extremity of crevice in adaxial side of rotula. Surface stereom is labyrinthic, except for band of fascicular stereom indicating attachment area of coelomic septum (arrow). E, F Compass hook. Lateral (E) and oral (F) views showing depressions for attachment of CRL fibres (stars)
by perturbation of the seawater. However, when the hook end of compasses was grasped with forceps, pulling either radially or transversely without rotation, or pulling vertically up or down, produced very little displacement. The compass-rotular joint is thus very lax with regard to rotational but not uniaxial movements. Microarchitecture and ultrastructure of CRL Extracellular components Thick fibres In histological sections of the CRL, its most prominent components are thick, basophilic fibres that are stained strongly by aniline blue and toluidine blue
Fig. 4 General anatomy of compass-rotular joint. Semithin sagittal section stained with toluidine blue; oesophagus is beyond left margin. Note contrast between suspensory fibres (white asterisk), which are stained heavily, and weakly stained oral sparse region (black asterisk). ac Adapical coelomic cavity, thick arrow insertion of suspensory fibres into underside of rotular overhang, thin arrows coelothelia. Other lettering as in previous figures
(Figs. 4, 5A–F). In the outer adaxial layer of the CRL these form an oblique crossed fibre lattice connecting the lateral surfaces of the compass hook to the adaxial projections of the rotula (Figs. 3B, 5D). In the inner part of the ligament, the fibres form parallel arrays most of which are orientated obliquely and connect the compass hook to more aborally located attachment points on the rotula, including the underside of the adaxial rotular overhang. The compass hook is thus suspended from the rotula by the latter fibres (Figs. 4, 5A–D). The lattice and suspensory fibres comprise parallel aggregations of densely packed, cross-banded collagen fibrils with a periodicity of 59–68 nm (mean±SD 62.1±3.3, n=6) (Fig. 6A–D). After treatment with cupromeronic blue, the adjacent fibrils within each fibre are seen to be connected by short filaments 7–14 nm in diameter with an electron-dense outer layer and lighter core. These filaments are attached to specific sites on the surface of the collagen fibrils, there being two such binding sites within each D-period (Fig. 6C). The gaps between the collagen fibres (i.e. fibril aggregations) contain variable numbers of microfibrils that are hollow, have a diameter of 9.4–14 nm (mean maximum±SD observed in eight different regions of the CRL was 11.7±1.3 nm) and show usually no periodic structure (Fig. 6E). In the interfibre gaps they are scattered sparsely or form discrete parallel bundles 310– 480 nm in diameter (Fig. 6B,D). Sparse regions Between the coelothelium flooring the adapical coelomic cavity and the suspensory fibres, the ligament consists of finely filamentous material that includes some thicker, widely separated basophilic fibres (Fig. 5A, E, F). The filamentous material of this ‘aboral sparse region’ includes widely separated bundles (diameter 180–280 nm) of tightly packed 12 nm microfibrils, narrow and loose bundles of collagen fibrils that are roughly parallel to the microfibril bundles, isolated microfibrils and isolated collagen fibrils (Fig. 7A, B). Below the thick lattice and suspensory fibres, there is a voluminous ‘oral sparse region’. The bulk of this region consists of a loose array of very fine, widely separated and roughly parallel fibres that are stained weakly by the aniline blue of Milligan’s trichrome or by toluidine blue in semithin sections (Figs. 4, 5A, G, H). In the electron microscope, these fibres are seen to be microfibril bundles 280–420 nm in diameter (Fig. 7C, E). Although consisting predominantly of typical 12 nm microfibrils, the bundles include regions of tightlypacked fibrils with a diameter of 21–24 nm, a beaded appearance and a periodicity of 55–62 nm (Fig. 7D). Such regions were found most often near the suspensory fibres where the microfibril bundles merge with looser arrangements of microfibrils (Fig. 7E). Commonly present within the microfibril bundles are thick, electrondense super-fibrils with a diameter of 24–48 nm and no discernible periodicity, which often have an undulating profile and sometimes appear to coalesce to form even thicker fibrils up to ca. 100 nm in diameter (Fig. 7F).
14
15 b Fig. 5A–H Histology of CRL. A–F Stained with Milligan’s trichrome: connective tissue blue; cellular elements pink-red. A Sagittal section; oesophagus beyond left margin. B, C lines indicating approximate level of sections in B and C respectively, sf suspensory fibres, arrow insertion of suspensory fibres into underside of rotular overhang, upper asterisk aboral sparse region, lower asterisk oral sparse region. B Frontal section; oesophagus beyond top margin. lf lattice fibres, asterisk oral sparse region. C Frontal section of left side of CRL shown in B but at more aboral level near upper edge of CRL. D approximate level of section in D, ic inner coelothelium, oc outer epithelium, asterisks aboral grooves (continuous with adapical coelomic cavity), arrows ridges. D Transverse section through one side of CRL, showing ridges and furrows on aboral edge (arrow). E More magnified view of part of D illustrating thickened inner coelothelium (ic) on groove floor and coelothelial incursion (arrow) into CRL. F Transverse section showing thickened coelothelium on groove floor (ic) and coelothelial incursion (arrow) that branches near suspensory fibres. G, H Semithin araldite sections of oral sparse region (see Fig. 4) cut in sagittal plane and stained with toluidine blue. Note fine, weakly stained fibres, lines of granules (arrows), heavily stained cells and cell apparently undergoing fragmentation (fc)
Isolated microfibrils are scattered between the microfibril bundles, but the oral sparse region seems to be devoid of collagen fibrils except near lattice or suspensory fibres or the outer coelothelium. Cellular components Intraligamental cells Nucleated cell bodies are distributed sparsely within the CRL and are especially scarce amongst the thick fibres (see e.g. Fig. 5E, F). In semithin araldite sections of the sparse regions, some large cells (longest dimension up to 25 lm) appear to be undergoing cytoplasmic fragmentation and others are small (ca. 5 lm), oval or triangular in outline, and heavily stained (Fig. 5G, H). The sparse regions also contain abundant linear aggregations of granules aligned roughly parallel to the fine fibres. Examination of ultrathin sections indicated that most intraligamental cell bodies have similar cytological features. They possess a Golgi apparatus, rough endoplasmic reticulum, mitochondria and many membranebounded vesicles of variable size and with heterogeneous contents. The cytoplasm between the organelles of these cells tends to have a rather electron-dense, granular appearance and the nucleus has a roughly oval or less regular profile (up to 3.4·7.3 lm) and electron-dense peripheral and central chromatin clumps with an irregular and diffuse edge (Figs. 6A, 8A–C). These heterogeneous vesicle-containing cells (HVCs) are never surrounded by an external lamina. Close to coelothelia or coelothelial extensions (described below) HVCs are roughly fusiform with irregularly shaped processes (Fig. 8A). Elsewhere within the ligament, i.e. between the lattice and suspensory fibres and within the aboral and oral sparse regions, their outline is less attenuated and more oval (Figs. 6A, 8B, D, E). The abundant granules seen in semithin araldite sections correspond to membrane-bounded bodies that
have an irregular to roughly circular profile and a wide range of sizes (diameter 0.2–3 lm) (Figs. 7C, 8C, D). These are clearly cell fragments and not transverse sections of parallel clusters of ramifying cell processes, since no longitudinal sections of such processes were observed in any of numerous semithin and ultrathin sections of the whole CRL cut in different anatomical planes. Larger fragments contain mitochondria and membranebounded vesicles with variable dimensions and contents, although a particular type of ‘grey vesicle’ can be recognised that has amorphous, moderately electrondense contents, a regularly circular profile and maximum diameter of 110–125 nm (Fig. 8H). The cell fragments are derived from HVCs. Figure 8C shows part of a typical HVC that appears to have released small and large cytoplasmic fragments, and Fig. 8F, G shows cells in which almost all the cytoplasm has divided into fragments. The nuclear morphology of the latter two cells differs from that of most HVCs observed in this investigation. In the cell shown in Fig. 8F, the chromatin has reduced electron density and is detached from the nuclear membrane; and in Fig. 8G, the nuclear membrane appears to be disarranged and there is an electron-lucent vesicle within the nucleus. Coelothelia Other cellular elements in the CRL are derived from the outer coelothelium on the adaxial and adoral surfaces of the ligament and the inner coelothelium that floors the adapical coelomic cavity. The CRL contains bundles of processes belonging to cell bodies in these coelothelia. Outer coelothelial incursions into the ligament are relatively scarce, but the inner coelothelium of the aboral grooves gives rise to more abundant tracts of acidophilic and sometimes varicose processes that pass through the aboral sparse region as far as the lattice and suspensory fibre systems (Fig. 5D–F). Branching of the tracts was observed near or within these fibre systems (Fig. 5F). The ultrastructure of the aboral grooves is illustrated in Fig. 9. Peritoneocytes in the aboral grooves have a single cilium surrounded by a collar of 10–13 radially arranged cytoplasmic lamellae (Fig. 9A, B). These cells are involved in intense secretory activity. Their cytoplasm contains much rough endoplasmic reticulum, many mitochondria and many membrane-bounded vesicles that tend to cluster apically, are of variable size and have contents of generally low electron density (Fig. 9A–C). Some large vesicles contain multiple concentric membrane layers that generate small vesicles with an irregular rounded profile up to 120 nm in diameter and moderately electron-dense contents. These small vesicles are released into large cell membrane blebs that are detached to form multivesicular bodies up to 2.5 lm in diameter (Fig. 9C). The lumen of the coelomic grooves contains numerous other membrane-bounded vesicles that may be derived from peritoneocytes directly or indirectly via multivesicular bodies (Fig. 9A–C). Membrane-bounded vesicles are also released by the
16
17 b Fig. 6A–E Ultrastructure of thick fibres. A Horizontal section of lattice fibres, showing bundles of collagen fibrils cut transversely (thin arrow) and nearly longitudinally (thick arrow). hvc Cell containing heterogeneous vesicles. B Sagittal section including parts of two suspensory fibres (sf) and bundles of microfibrils (arrowheads) in the gap between them. Note absence of cellular elements. C Horizontal section of collagen fibrils stained with cupromeronic blue to reveal proteoglycans which appear as short filaments attached to specific sites on fibrils. Double filaments suggest there are two binding sites within each fibril D-period (arrowheads). D More magnified view of part of B, showing bundles of microfibrils (arrowheads) in gap between two suspensory fibres (arrows). E Transverse and longitudinal sections of microfibril bundles. Some microfibrils have electron-lucent core (arrowheads)
inner coelothelium outside the aboral grooves and by the outer coelothelium (not illustrated). The bundles of cell processes that emerge from the inner coelothelium of the aboral grooves and penetrate the underlying ligament are ensheathed by a basal lamina (Fig. 9D). The bundles include peritoneocyte processes and processes containing electron-dense granules, at least some of the latter belonging to granule-containing perikarya located in the coelothelium (Fig. 9A). The bundles terminate abruptly between lattice and suspensory fibres, sometimes close to HVCs or electronopaque bodies (Fig. 9D), and are connected to adjacent collagen fibrils via the basal lamina. Within these terminals, the endings of individual cell processes are appressed closely against the inner surface of the basal lamina, and some are dilated and contain a single large vesicle enclosing moderately electron dense, flocculent material (Fig. 9E). No evidence was found that individual cell processes branch away from the confines of the basal lamina-enclosed bundles, i.e. no solitary granule-containing processes were observed in any region of the CRL.
is isosmotic with seawater. In six ligaments, this was followed after 18–46 s by a small, though distinct, increase in the rate of stress relaxation. The significance of this finding was explored by comparing the effect of KCl and other alkali metal chlorides on non-anaesthetised and anaesthetised ligaments. After excision, preparations from one group of animals were left for at least 30 min in SW alone and preparations from a separate group of animals were left in SW containing 0.1% propylene phenoxetol (1-phenoxy propane-2-ol: Nipa Laboratories). During the subsequent stress relaxation tests, the non-anaesthetised preparations were flooded with SW or 0.56 M solutions of NaCl, KCl, RbCl or CsCl, and the anaesthetised preparations with the same media containing 0.1% propylene phenoxetol. Fig. 10A shows that KCl, RbCl and CsCl, but not NaCl, caused a significant increase in the rate of stress relaxation of non-anaesthetised preparations. Fig. 10B shows that in the anaesthetised preparations, the effects of KCl, RbCl and CsCl persisted, and that NaCl also increased significantly the rate of stress relaxation. Rb+ and Cs+ ions block potassium channels and thereby cause cell membrane depolarisation in a variety of organisms. To make sure that the response of the CRL to RbCl and CsCl was not a consequence of such activity, the effect of 0.56 M CsCl was compared with that of two other potassium channel blockers: tetraethylammonium acetate (20 mM in natural seawater) and barium chloride (20 mM in sulphate-free artificial seawater). These did not affect significantly the rate of stress relaxation of the CRL (Fig. 11A), nor, in a separate experiment, did a solution of 0.38 M BaCl2 alone (Fig. 11B).
Discussion
Effects of chemical agents on mechanical behaviour
Structure of the CRL
Initial experiments were conducted on 25 preparations each of which was treated sequentially with 1 mM acetylcholine (ACh), normal seawater (SW) and seawater containing 100 mM K+ (100KSW) for 3 min each in that or the reverse order. Visual inspection of the recordings revealed no discernible change in the rate of stress relaxation following exposure to ACh or 100KSW. This result was confirmed quantitatively in another batch of 30 preparations each of which was treated with only ACh, 100KSW or SW. The response index of the SW-treated ligaments was 0.66±0.27 (±SD, n=10), indicating that the rate of stress relaxation declined during the course of the control tests. There was no significant difference between the response index of SW-treated CRLs and those of ACh- or 100KSW-treated ligaments, which were 0.50±0.24 (n=10; t18=1.3709; P=0.1873) and 0.61±0.32 (n=10; t18=0.3686; P=0.7167), respectively. To check if the CRL might be responsive to a higher [K+], 10 ligaments were treated with 0.56 M KCl, which
Extracellular components The extracellular compartment of the CRL is dominated by two fibre-forming components: collagen fibrils and microfibrils. Most of the collagen fibrils are assembled into the thick parallel aggregations of the lattice and suspensory fibre systems and the thinner fibres that reinforce the ridges in the aboral surface of the ligament and underlie the coelothelia. Staining with cupromeronic blue revealed that within the fibres proteoglycans are attached to the individual collagen fibrils at specific binding sites, as demonstrated in mammals and other echinoderms (Scott 1991; Erlinger et al. 1993; Nishimura et al. 1996). The proteoglycan-collagen interaction of the CRL resembles particularly that of mutable ligaments in the arm of the crinoid Antedon bifida (Pennant 1777) in that the ultrastructure of the cupromeronic blue-stained deposits is identical and in there being two proteoglycan binding sites within each D-period of the collagen fibrils (Erlinger et al. 1993).
18
19 b Fig. 7A–F Ultrastructure of sparse regions. A Horizontal section of aboral sparse region. Representative view that includes isolated microfibrils (single arrowhead), bundle of microfibrils (double arrowheads) and loose aggregations of collagen fibrils (arrow). B More magnified view of part of A, with bundle of microfibrils and cross-banded collagen fibril. C–F Oral sparse region. C Representative view that includes isolated microfibrils (single arrowhead), bundles of microfibrils (double arrowheads) and cell fragments (arrow). No collagen fibrils are present. D Part of microfibril bundle that includes thicker, beaded fibrils with periodicity 55–62 nm. E Suspensory fibre (sf) and adjacent sparse region which includes three parallel microfibril bundles (arrowheads) and merger between one of these and microfibrils at outer edge of suspensory fibre (arrow). F Part of microfibril bundle showing single (arrow) and fused (arrowheads) microfibrils
The microfibrils of the CRL exhibit an unusual organisation. Hollow, sometimes beaded, microfibrils 10–15 nm in diameter are ubiquitous in echinoderm connective tissue (both mutable and non-mutable) and often form, as in the CRL, parallel aggregations (i.e. fibres) up to ca. 0.5 lm in diameter (del Castillo et al. 1995; Wilkie 1996; Wilkie et al. 1998, 2000). However, the oral sparse region, which represents a large proportion of the total CRL volume, is unique in consisting of an array of microfibril aggregations with no associated collagen fibrils. The microfibril aggregations include many curved fibrils with a diameter of ca. 20– 100 nm, and less common parallel assemblages of 21– 24 nm fibrils with an axial periodicity of 55–62 nm. The latter may result from the merging of pairs of 10–15 nm microfibrils, and the thicker curved fibrils from the fusion of larger numbers of microfibrils. Such a process accompanies the ageing of the zonular fibres of the eye of Cavia porcellus (Linne´ 1758), in which 12 nm microfibrils combine to form 25 nm fibrils with 56 nm periodicity (Hanssen et al. 2001). The incorporation of microfibrils into densely packed, banded fibres with a periodicity of ca. 50 nm is also an age-related phenomenon in human cartilage (Keene et al. 1997). The hollow, beaded 12 nm microfibrils of mammals are elastic and composed partly of the glycoprotein fibrillin (Sherratt et al. 2001). The only echinoderm microfibrils that were examined in detail are those of holothurian dermis, which also demonstrate long-range elasticity and contain fibrillin or fibrillin-like molecules (Thurmond and Trotter 1996; Thurmond et al. 1997). Heterogeneous vesicle-containing cells Heterogeneous vesicle-containing cells (HVCs) are common in echinoderm connective tissue. They are thought to be pluripotential and have the capacity to switch to a fibroblastic or a fibroclastic phenotype (Heinzeller and Welsch 1994; Wilkie 1996). In the CRL, these cells are the source of the remarkable aggregations of aligned cell fragments. The process by which these fragments are generated resembles that associated with apoptosis (programmed cell death), in which the cytoplasm is divided into membrane-bounded globules, known as
apoptotic bodies, containing intact organelles (Wyllie et al. 1980; Kerr et al. 1994). However, apoptosis involves characteristic changes in nuclear morphology including chromatin compaction and segregation into large, sharpedged masses lying against the nuclear membrane, followed by budding of the nucleus (Wyllie et al. 1980; Robertson and Thomson 1982; Kerr et al. 1994; Mire and Venable 1999), which were not seen in any cells of the CRL. Furthermore, apoptotic bodies are usually removed rapidly by phagocytosis, whereas we found no evidence that this is the normal fate of HVC fragments. The nucleus of some fragmenting cells resembles that of cells undergoing necrosis (injury-induced degeneration): the nuclear contents have greatly reduced electron density and the nuclear membrane has ruptured (Wyllie et al. 1980). These may be damaged or senescent HVCs. It is possible that the very electron-dense bodies observed near terminals of coelothelial incursions, and which were also seen in the CRL of Paracentrotus lividus (Lamarck 1816) (Wilkie et al. 2003), are senescent cells that are being eliminated by another mechanism. Since nothing similar to the linearly arranged cell fragments was reported previously as occurring in the connective tissue of echinoderms or other animals, their functional significance can only be speculated upon. The presence in them of ‘grey vesicles’ suggests one role might be the transport throughout the ligament of a secretory product necessary for the functioning of the microfibril aggregations or the electron-lucent (and presumably fluid-rich) material between these aggregates.
Coelothelia The monociliated peritoneocytes of the inner and outer coelothelia are involved in intense secretory activity, which is a commonly encountered feature of such choanocyte-like cells (also known as ‘flagellated collar cells’) in echinoderms (see e.g. Bachmann and Goldschmid 1978; Heinzeller and Welsch 1994; Bonasoro et al. 1995; Candia Carnevali et al. 1995) and of coelomic epithelia in general, including mammalian mesothelia (Welsch 1995). The last mentioned cells synthesise ‘lamellar bodies’ consisting of concentric arrangements of lipid bilayers, which after extrusion into the peritoneal cavity, form multivesicular structures or expand into a variety of vesicles with electron-lucent contents (Dobbie and Anderson 1996). There are striking similarities between lamellar bodies and the large intracellular vesicles in the peritoneocytes (mesothelial cells) of the aboral CRL grooves that also contain concentric membrane layers and give rise to multivesicular structures that are released into the groove lumen (a coelomic space like the peritoneal cavity). It seems likely that these secretory phenomena are homologous and that Dobbie (1996) was correct to infer that lamellar body-secreting tissues represent a biological system ‘‘of supreme evolutionary antiquity’’. The confinement of a specialised coelothelium to the grooves in the aboral surface of the CRL indicates that
20
21 b Fig. 8A–H Intraligamental cellular components. A Heterogeneous vesicle-containing cell (HVC) from region close to coelothelial incursion into CRL. B HVC in lattice fibre system, showing Golgi apparatus (arrow). C Edge of HVC in sparse region, which appears to be releasing cytoplasmic fragments (arrowheads). Note abundant mitochondria (mi) and nearby row of cell fragments (arrow). D Cell bodies and fragments in gaps between lattice fibres (fi). E Cell in sparse region with Golgi apparatus (arrow), mitochondria (arrowheads) and heterogeneous vesicles. F Fragmenting cell with abnormal nuclear morphology (n). G Fragmenting cell with possibly necrotic nucleus (n). H Examples of small (large arrowhead) and large cell fragments. Large fragment contains mitochondrion (arrow), ‘grey’ vesicles (small arrowheads) and various other vesicles
these are a primary morphological adaptation and lessens the likelihood that the ridge and groove topology is an adaptation for increasing surface area or represents transient folding of the aboral surface of the CRL. The coelothelium at the bottom of the grooves consists mainly of densely packed, columnar peritoneocytes which are likely to generate strong ciliary currents. The grooves may therefore serve to channel and direct these currents, thereby optimising the distribution of secretory products, respiratory gases or other materials over the lining of the adapical coelomic cavity. The groove epithelium is the source of large tracts of granule-containing cell processes that penetrate deeply into the ligament, though separated from its extracellular compartment by a basal lamina. It is notable that the tracts branch and terminate only near or within the lattice and suspensory fibre systems, which indicates that their function could be related specifically to these components. The frequent proximity of bundle terminals to HVCs suggests they might have a role in regulating the activities of these cells. It is highly unlikely that they are juxtaligamental cell processes, which are present in every echinoderm structure confirmed by physiological methods to consist of mutable collagenous tissue (MCT). Juxtaligamental cells control directly the tensile properties of MCT and invariably contain large electron-dense, membrane-bounded granules that are spherical to ellipsoid in shape and up to 300·700 nm in size (Wilkie 1996, 2002; Trotter et al. 2000). While similar granules are present in some cell processes in the inner coelothelial extensions into the CRL, the latter differ from juxtaligamental elements in two important ways. First, MCT is densely permeated by single, or small bundles of, granule-containing cellular elements. Neither small bundles nor single granule-containing cell processes were found in any region of the CRL. Indeed, large parts of the CRL, particularly the suspensory fibres, are devoid of any cellular elements. Second, in all echinoderm classes except the Holothuroidea, neither juxtaligamental cell bodies nor their processes are separated from the MCT extracellular matrix by a basal lamina, whether entire perikarya and their processes are located within the MCT, as in the compass depressor ligaments of certain echinoids (Wilkie et al. 1992), or whether their perikarya are outside, and only their
processes are inside, the MCT, as in ophiuroid ligaments (Wilkie 1979). Effects of chemical agents on the CRL Isolated CRL preparations displayed no detectable response to 1 mM ACh or 100KSW, sensitivity to these and other neuroactive agents being a diagnostic feature of MCT-containing structures (Wilkie 1996). For example, with regard to echinoid lantern-associated components, 1 mM ACh and 100KASW stiffen reversibly the compass depressor ligament and peristomial membrane of P. lividus (Wilkie et al. 1992, 1993). A 0.56 M solution of KCl alone did, however, cause a small but statistically significant increase in the rate of stress relaxation of the CRL. This same effect was produced in unanaesthetised preparations by 0.56 M RbCl and CsCl, though not NaCl. Elevated [K+]o, Rb+ ions and Cs+ ions can inhibit K+ efflux across cell membranes and thereby cause depolarisation, high [K+]o by abolishing the K+ concentration gradient across the membrane, and Rb+ and Cs+ by blocking K+ channels (see e.g. Butt et al. 1990; Alkadhi and Simples 1991; Van Driessche and De Wolf 1991; Zanello and Barrantes 1992). It is therefore possible that the response of the CRL to these agents could result from the stimulation of neural pathways and/or effector cells that modulate its mechanical properties. However, the persistence of the destiffening action of KCl, RbCl and CsCl in the presence of anaesthetic indicates that their effect does not depend on neural conduction. Furthermore, the ineffectiveness of tetraethylammonium ions and BaCl2 at concentrations that would be expected to inhibit a range of different K+ channel types (Nelson and Quayle 1995), is evidence that the response to RbCl and CsCl is not mediated by K+ channel blockade of either neurons or effector cells. If the destiffening caused by 0.56 M KCl, RbCl and CsCl is not cellularly mediated, it must be due to their direct action on the extracellular matrix of the CRL. It was established, primarily on the basis of evidence derived from experiments on cell-dead holothurian dermis, that the tensile properties of echinoderm collagenous tissue depend partly on interactions between inorganic ions, particularly monovalent and divalent cations, and fixed charges on extracellular macromolecules. The exact role of the inorganic ions in these interactions is as yet unclear, but could include (1) stabilising the conformation of intermolecular binding sites, (2) partially shielding, and thereby determining the degree of electrostatic interaction between, the fixed charges, and, in the case of divalent cations, (3) acting as ionic crossbridges (Trotter et al. 1997). The absence of divalent cations from the single-salt solutions used in our investigation could therefore have contributed to their destiffening effect, a conjecture supported by the finding that Ca2+ depletion reduces significantly the breaking strength of the CRL of P.
22
23 b Fig. 9A–E Coelothelia and coelothelial incursions. Horizontal sections. A Inner coelothelium of aboral groove. Note peritoneocytes (pe), cilia surrounded by cytoplasmic lamellae (arrows) and membrane-bounded secretory products (arrowheads). ac Adapical coelomic cavity, cp bundles of granule-containing cell processes. B Inner coelothelium. Coelomic cavity (ac) contains many membrane-bounded secretory products (arrowheads) that are released from peritoneocytes. ci cilia, arrow membrane bleb containing vesicles. C Edge of peritoneocyte that contains rough endoplasmic reticulum (er) and large vesicle enclosing concentric membrane layers (arrow) that give rise to small vesicles. Arrowhead possibly earlier stage of secretory process. Note large detached vesicle containing small vesicles. D Terminal of coelothelial incursion (arrow) between suspensory fibres, which abuts electron-dense body that may be effete or apoptotic HVC. Coelothelial incursion is ensheathed by basal lamina (arrowheads). E Terminal of coelothelial incursion between suspensory fibres. Note basal lamina (arrow) and terminal vesicles (stars)
increase) due to the osmotic lysis of effector cells and the consequent release of chemical factors that modulate directly interactions between matrix macromolecules (Shadwick and Pollock 1988; Trotter and Koob 1995; Wilkie et al. 1999). It must therefore be concluded that the CRL of E. esculentus is not mutable. This is consistent with the results of a preliminary ultrastructural analysis of the CRL of P. lividus and physiological experiments on the CRL of this and three other echinoid species, which elicited no evidence for the presence of MCT (Wilkie et al. 2003). A possible functional explanation for the absence of mutability was suggested by the observations on the
lividus (Wilkie et al. 2003). However, the varying potency of the four monovalent cations implies that Ca2+ depletion cannot be the only factor. The order of effectiveness was Na+
