Proc. Natl. Acad. Sci. USAVol. 82, pp. 1555-1557, March 1985Neurobiology

Nerves in the spine of a sea urchin: A neglected division of theechinoderm nervous system

(electrophysiology/echinold spines/nerve fibers/sensory neurites/calcium spikes)

DAVID S. SMITH*, DEIDRE BRINKt, AND Jose DEL CASTILLOt*Department of Zoology, University of Oxford, South Parks Road, Oxford OX1 3PS, England; and tLaboratory of Neurobiology, Medical Sciences Campus,University of Puerto Rico, Boulevard del Valle 201, Old San Juan, PR 00901

Communicated by Theodore H. Bullock, November 5, 1984

ABSTRACT Electrical stimulation of the primary spines(>10 cm long) of the tropical sea urchin Diadema antlarumelicits graded compound action potentials that are conductedat a constant speed of m27 cm/sec. Ion substitution experi-ments suggest that these are due to the summation of calciumspikes. Structural studies have revealed the presence of up to21 regularly disposed nerves within the spine shaft, each nervebundle including >1000 neurites in the basal region, narrow-ing to slender groups of processes near the spine tip. Theneurites in each nerve range in diameter from <0.1-2 ,um.Most appear to be distal processes of presumed sensoryperikarya situated at the level of the tissue cone surroundingthe spine base, or more proximally, although some neuritesmay arise from perikarya near the spine tip. These nervetracts are thought to correspond to the nerve fibers describedby Hamann [(1887) Jena Z. Naturwiss. 21, 114-176] almost acentury ago in spines of Centrostephanus longispinus and,thus, to represent a long-neglected region of the echinoidnervous system.

In the last of a classic series of papers devoted to theanatomy and histology of echinoderms published almost acentury ago, Hamann (1) reported en passant that whilestudying decalcified preparations of the sea urchinCentrostephanus longispinus, he detected structures that hebelieved to be nerve fibers, ascending the spine from thenerve ring that surrounds each spine base. Thus, in a briefpassage accompanied by a single figure, Hamann providedthe first suggestion of a region of the echinoid nervoussystem that, until now, has not been further investigated.Although Hamann's observation is cited briefly by Hyman(2) and by Smith (3), the spinal nerves are not mentioned bySmith (4) or Cobb (5) or in the valuable reviews of theechinoderm nervous system by Pentreath and Cobb (6, 7),We were led to the recognition of a system of nerves in the

spines of the tropical sea urchin Diadema antillarum byelectrophysiological observations. While studying someproperties of the muscles that move the primary spines (8),we found that electrical stimulation of a spine tip elicitssynchronized convergent movement of the surroundingspines-comparable with the response to tactile stimulationdescribed by Bullock (9). We concluded that, after electricalstimulation, a wave of excitation is propagated, in someway, along the spine. This was confirmed by the observationthat graded compound action potentials, conducted in bothdirections, are elicited by electrical stimulation of isolatedspines. These action potentials were studied in a standardpreparation: a segment of a freshly isolated spine, 10 cm ormore long, cut well above the base to ensure absence ofmuscle or subjacent nerve ring.

FIG. 1. Compound action potential recorded from a primaryspine of Diadema immersed in paraffin oil. Average of 10 consecu-tive traces obtained with a Nicolet digital oscilloscope.

The action potentials were recorded with a pair of wickelectrodes, connected via agar bridges to Ag/AgCl pellets orcalomel half-cells, making contact with the spines eithersuspended in air or immersed in paraffin oil. Electricalstimuli 5-10 msec long were applied via a pair of stimulatingelectrodes, and another nonpolarizable electrode was usedto ground the preparation. The recording electrodes wereconnected directly to a Tektronix 5A22N differential ampli-fier, 1 Mfl input resistance, set to a bandwidth of 0.2-300Hz. The output of the amplifier was fed into the input of aNicolet digital oscilloscope with which 5-10 consecutiveaction potentials were averaged to improve the signal-to-noise ratio.The peak-to-peak amplitudes of the action potentials re-

corded from spines suspended in air were between 50 and100 puV, while those recorded from spines immersed inparaffin oil were often B-5 times larger and far more stable.The size of the recorded potentials did not increase appreci-ably on interposing a high impedance preamplifier or onincreasing the bandwidth of the main amplifier. Occasion-ally, simple diphasic responses were recorded, but generallythe recorded waveforms were as illustrated in Fig. 1.Conduction velocity was measured in six spines by re-

cording the action potentials at four different distances fromthe stimulating electrodes. In each instance, over distancesof at least 10 cm along the spine, it was found that conduc-tion takes place at a rate of -27 cm/sec (at 21'C-230C).Experiments have been carried out to assess the influence

of the ionic composition of the ambient medium on theelectrical activity displayed by the spine. In each experi-ment, 20 different spines were tested. Each freshly cut spinewas placed directly over the electrode array, and a controlaction potential was recorded. The spine was then immersedfor 15 min in artificial sea water with modified ionic composi-tion or in natural sea water to which a range of cations ororganic compounds were added. Reversibility of the effectswas checked by reimmersion of the spine in natural sea

1555

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 82 (1985)

b 7,..

'I

0.

N

iNf 4 ' ; '4 ZffeO'4 --~~~~~~~-4. ~ ~ ~ ~ L

M > > At W St S: e .

2 9 S <, Ad,&X~~~~~~~~5, !,4 k'

>s W~~~~f r.i..A,t

-~~~~~~~~~~~~ ~ ~ ~ ~ ~ ~ ~ A

FIG. 2. (a) Scanning electron micrograph of a transversely fractured Diadema spine, after removal of all soft tissue by Clorox treatment.Note the radial calcified flanges (arrows) and channels between them that accommodate the nerves in the intact spine, in the positions indicated(*). (Bar = 0.25 mm.) (b) Light micrograph of a 1-Am transverse section of a decalcified and plastic-embedded Diadema spine. Section wastaken -1 cm distal to the tissue cone investing the spine articulation. About one-third of the spine perimeter is included, together with 6 of thecirclet of regularly placed nerve tracts (arrows) extending along the spine between the radial calcified flanges (cf. a). (Bar = 0.25 mm.) (c)Material as in b, but representing a low magnification transmission electron micrograph including a portion of a single nerve tract. Note theextracellular basal laminar sheet (arrow) defining the tract and suspending it from the surface epithelium (cf. b). Tract includes a large numberof transversely sectioned neurites further illustrated in d. (Bar = 5 Aum.) (d) Transmission electron micrograph of a region within a Diademaspine nerve tract, at higher magnification. Note the various diameters of the neurites, in which longitudinally oriented microtubtles are the mostconspicuous inclusion. (Bar = 250 nm.)

water for 15 min prior to recording. The following observa-tions, which will be described in more detail elsewhere,suggest that the action potentials are produced by theactivation of voltage-dependent calcium channels. (t0 Theaction potentials are not blocked by tetrodotoxin (0.5pug/ml). (it) The action potentials continue to be generated inartificial sea water in which NaCl is replaced by an osmoti-cally equivalent amount of sucrose. (iii) The action poten-tials are greatly diminished in size, or completely abolished,in Ca-free medium. (iv) The amplitude of the action poten-tials increases when the calcium concentration of the artifi-cial sea water increases; however, the electrical activity ofmost preparations is blocked when the Ca concentration isdoubled. (v) The action potentials are abolished by 2-5 mMCo2+, Cd2+, and La3' ions. Calcium can be replaced bySr2+, but not by Ba2+ ions, which block electrical activity.(vW) The action potentials are blocked by the organic Ca-channel blocker Bepridil (W-2799 CERM) at a concentrationof 2 mM.

We have examined the structure of Diadema spines inquest of the excitable components inferred from the electri-cal findings. In principle, the action potentials might begenerated within the epithelial cell layer investing the spineand continuous with that of the general body surface, or byneurites situated elsewhere within the spine shaft. Disrup-tion of the epidermal spine covering was carried out eitherby drying the spine surface with paper tissue (which be-comes stained with epidermal pigment), or by gently scrap-ing the spine surface with a scalpel. Neither procedureprevented the conduction of the action potentials, militatingagainst the spreading of excitation by an epithelial cell-to-cell mechanism.We have observed a regular array of slender nerves,

deeply situated within the spine and thus. not accessible tosurface abrasion, which are the most likely substrates for theaction potentials. Since the rigid skeleton of the spine,composed mainly of calcium carbonate with a substantialamount of magnesium carbonate (10), precludes conven-

1556 Neurobiology: Smith et al.

I

)Io1%

lb

40

Proc. Natl. Acad. Sci. USA 82 (1985) 1557

tional sectioning, we have examined the organization of thespine in two complementary ways: (i) by scanning electronmicroscopy of spines from which cellular structures wereremoved and (it) by light and transmission electron micros-copy of decalcified spines from which crystalline materialwas removed prior to embedding and sectioning. For scan-ning electron microscopy preparations, spines were treatedwith 25% (vol/vol) Clorox. Decalcification, accomplished byHamann (1) by chromic acid treatment, was carried out here,after glutaraldehyde fixation, with 0.1 M EGTA/EDTA, atroom temperature.A scanning electron micrograph of a transversely frac-

tured spine, after Clorox treatment, is shown in Fig. 2a, anda sector of a 1-,gm section of a plastic embedded spine isshown in Fig. 2b. The former reveals the complex crystallineframework of the spine shaft, including the fenestratedtubular core and the regularly placed radial flanges, betweenwhich lie the spine nerves. Up to 21 nerves are present in thespines examined, and their regular disposition is shown inFig. 2b. This light micrograph illustrates the epitheliuminvesting the spine, together with thq basal laminar systemflanking each radial crystalline flange in the intact spine andforming mesenteries extending from beneath the epitheliumto the spine core and en route accommodating the cylindricaltracts of neurites that are situated at -40% of the distancebetween the spine surface and the tubular core.

Despite their depth within the spine, the fact that thenerves lie within radial prolongations of the epithelial basallaminar system indicates that they are part of the generalbasi-epithelial nerve plexus.

Fig. 2c represents a survey transmission electron micro-graph of a portion of a single nerve, including the ac-companying basal lamina sheath. A high magnification field,including a few transversely sectioned neurites within anerve, is illustrated in Fig. 2d.

Fig. 2b-d represents material within the proximal region ofthe spine but at least 1 cm above the apex of the tissue coneof muscle and "ligament" that invests the spine base and itsarticulation with the test (8). From the base, each spinetapers gradually to a pointed tip, and the nerve tractsbecome correspondingly smaller in diameter along the spine.In the proximal regions illustrated, each nerve bundle is35-50 gm in diameter and contains >1000 neurites. Most ofthe latter are in the <0.1- to 0.3-,um diameter range, butsome much larger profiles, up to =2 gtm, are included. Wehave traced some of these slender nerve processes to cellbodies situated beneath the surface epithelium near the apexof the tissue cone. Preliminary fine structural studies suggestthat at least 20 neurites may stem from a single perikaryon atthis level. It is possible, however, that some of the spineneurites arise from cell bodies situated more proximally,possibly within the nerve ring encircling each spine baseabove the test (6, 7). In addition, some profiles within thespinal nerve tracts may represent proximal processes arisingfrom cell bodies located near the spine tip.The nerve cell processes (neurites) comprising the spine

nerves conform to other neural components of echinoderms(7, 8) (i) in their small diameter; (ii) in their simple cytoplas-mic organization, primarily involving oriented microtubules;(iih) in the absence of accompanying glial processes; and (iv)in the presence of local dilatations or "varicosities" seen inlongitudinal sections. The neurites arranged in compact

cylindrical tracts in Diadema spines are of unusual length-up to 20 cm in processes traversing the longest primaryspines. Previous accounts (7) of echinoderm nervous sys-tems have not revealed distal nerve processes comparablewith the putative sensory neurites described here.Although Hamann (1) did not further illustrate his state-

ment that nerves ascend the spines of the sea urchin hestudied, we suggest that our work on Diadema substantiallysupports his proposal. Bullock (9) has investigated the"spine convergence" response previously considered byvon Uexkull (11) and also described here. While Bullock'sexperimental work was particularly concerned with thespine-to-spine pathway, which he was able clearly to distin-guish from a coelenterate-type nerve net at the level of thetest epithelium, the present findings are directly relevant tosome of his observations. Thus, Bullock noted that "veryslight tactile stimulation of the side of the spine, sometimeseven near its tip, is a strong stimulus for movement ofneighboring spines."While the neural pathway functionally linking the array of

spines has yet to be determined, the present anatomical andelectrophysiological findings are consistent with the sensitiv-ity of spines noted in several genera by Bullock (9) anddocument an intrinsic portion of the basi-epithelial plexuswithin the spine shaft. Whether this system is of generaloccurrence has yet to be ascertained, but documentation ofthis presumably sensory system in Diadema is of physiologi-cal interest.

In the absence of structural or other evidence of effectorcells along the spine shaft, we conclude that they areprimarily sensory in function, although what type of stimulithey receive is undetermined. As such, they are of interest tocomparative physiologists because of their length and regu-larity of distribution reflecting the radial symmetry of thespine shaft and their spatial separation from non-nervoustissue.

We are grateful to Mr. F. McKenzie for collecting the Diademaand to Wallace Laboratories for a gift of Bepridil (W-2799 CERM).This work was supported by National Institutes of Health GrantsNS-07464 and K6N-14938 (to J.d.C.).

1. Hamann, 0. (1887) Jena Z. Naturwiss. 21, 114-176.2. Hyman, L. H. (1955) The Invertebrates: Echinodermata

(McGraw-Hill, New York), Vol. 4.3. Bullock, T. H. & Horridge, G. A., eds. (1965) Structure and

Function in the Nervous Systems of Invertebrates (Freeman,San Francisco), pp. 1519-1558.

4. Smith, J. E. (1966) in Physiology of Echinodermata, ed.Boolootian, R. A. (Interscience, New York), pp. 503-511.

5. Cobb, J. L. S. (1982) in Echinoderms: Proceedings of theInternational Conference at Tampa Bay, ed. Lawrence, J. M.(Balkana, Rotterdam, The Netherlands), pp. 409-412.

6. Pentreath, V. W. & Cobb, J. L. S. (1972) Biol. Rev. 47,363-392.

7. Pentreath, V. W. & Cobb, J. L. S. (1982) in Electrical Con-duction and Behaviour in "Simple" Invertebrates, ed.Shelton, G. A. B. (Clarendon, Oxford), pp. 440-472.

8. Smith, D. S., Wainwright, S. A., Baker, J. & Cayer, M. L.(1981) Tissue Cell 13, 299-320.

9. Bullock, T. H. (1965) Am. Zool. 5, 545-562.10. Raup, D. M. (1966) in Physiology of Echinodermata, ed.

Boolootian, R. A. (Interscience, New York), pp. 379-395.11. von Uexkull, J. (1900) Z. Biol. 39, 73-112.

Neurobiology: Smith et al.