Current Biology 19, 1264–1269, August 11, 2009 ª2009 Elsevier Ltd All rights reserved DOI 10.1016/j.cub.2009.05.063

Report

Centralization of the DeuterostomeNervous System Predates Chordates

Marc Nomaksteinsky,1,2 Eric Rottinger,3

Heloıse D. Dufour,1,2,5 Zoubida Chettouh,1,2 Chris J. Lowe,4

Mark Q. Martindale,3 and Jean-Francois Brunet1,2,*1Department of Biology, Ecole Normale Superieure,75005 Paris, France2CNRS UMR 8542, 75005 Paris, France3Kewalo Marine Laboratory, University of Hawaii, Honolulu,HI 96813, USA4Department of Organismal Biology and Anatomy,University of Chicago, Chicago, IL 60637, USA

Summary

The origin of the chordate central nervous system (CNS) is

unknown. One theory is that a CNS was present in the firstbilaterian and that it gave rise to both the ventral cord of

protostomes and the dorsal cord of deuterostomes (re-viewed in [1]). Another theory (reviewed in [2]) proposes

that the chordate CNS arose by a dramatic process of dors-alization and internalization from a diffuse nerve net coex-

tensive with the skin of the animal, such as enteropneustworms (Hemichordata, Ambulacraria) are supposed to

have [3]. We show here that juvenile and adult enteropneustworms in fact have a bona fide CNS, i.e., dense agglomera-

tions of neurons associated with a neuropil, forming twocords, ventral and dorsal. The latter is internalized in the

collar as a chordate-like neural tube. Contrary to previous

assumptions, the greater part of the adult enteropneustskin is nonneural, although elements of the peripheral

nervous system (PNS) are found there. We use molecularmarkers to show that several neuronal types are anatomi-

cally segregated in the CNS and PNS. These neuroanatom-ical features, whatever their homologies with the chordate

CNS, imply that nervous system centralization predates theevolutionary separation of chordate and hemichordate line-

ages.

Results and Discussion

Hemichordates, together with echinoderms, are the sistergroup of chordates [4]. They include enteropneust wormsand pterobranches [5]. The nervous system of enteropneusts,whose body plan is probably less derived than that of the echi-noderms, plays a major role in hypotheses about the origin ofthe chordate central nervous system (CNS). These hypoth-eses, however, are based on divergent interpretations of theneuroanatomy of enteropneusts. Some authors [6, 7], fol-lowing the description by Bateson [8] and others (reviewed in[9]), surmise that enteropneusts are equipped with a CNScomposed of ‘‘cords,’’ as in spinal cord (of vertebrates) orventral nerve cord (e.g., of annelids). Because enteropneusts

*Correspondence: jfbrunet@biologie.ens.fr5Present address: Howard Hughes Medical Institute and Laboratory of

Molecular Biology, University of Wisconsin–Madison, Madison, WI 53706,

USA

have two cords, one ventral and one dorsal [9], opinions varyas to which is homologous to the chordate CNS (reviewed in[2]). However, following a more recent tradition initiated byBullock [3, 10] and Knight-Jones [11], the currently proposedview dismisses these ‘‘cords’’ as mere ‘‘conducting tracts’’[11, 12] or ‘‘through-conducting neurite bundles’’ [2] lackingneuronal cell bodies [12]. From this viewpoint, enteropneustworms have no bona fide CNS [12], only an ‘‘epidermal nervenet’’ coextensive with the skin of the animal and possiblyhomologous, as a whole, to the vertebrate CNS [11, 13]. This‘‘nerve net,’’ also described by Bateson [8], is thought toconsist of neuronal cell bodies densely distributed amongepithelial cells throughout the entire body of the animal andwhose projections form a fiber layer underneath, ubiquitousalbeit thickened at the level of the ‘‘cords.’’ The conflictbetween the ‘‘centralized’’ and ‘‘diffuse’’ views of the enterop-neust nervous system becomes particularly acute concerningthe collar cord, a subepidermal continuation of the trunk dorsalcord in the collar or mesosome (Figure 1A). The internalizationof the dorsal cord at the level of the collar occurs by ingression(in Saccoglossus [14]) or invagination (in ptychoderids [9, 15])and was originally proposed to be homologous to neurulationin chordates [16]. ‘‘Giant nerve cells’’ have been described asspecific to the collar cord [3, 11, 17–19], reinforcing its status asa CNS. However, the existence of these cells has been ques-tioned [20], and the collar cord has been deemed a ‘‘submergedstrip’’ of otherwise unremarkable epithelium [10].

The coexistence of these two theories, which invite diver-gent evolutionary scenarios [2], is only allowed by the scarcityof histological data. To help resolve the issue, we examinedthe neuroanatomy of juvenile and adult enteropneusts byin situ hybridization (ISH) and immunofluorescence with a setof pan-neuronal and neuron-type-specific probes.

As a model species, we chose Ptychodera flava, an enterop-neust with indirect development in the family Ptychoderidae.We first cloned and examined the expression of a P. flavahomolog of Elav (HuC/D), which codes for an RNA-bindingprotein expressed in neurons [21] (see Figure S1 available on-line). On transverse sections through the collar or mesosomeof adult animals, the only prominent signal was in the collarcord (Figure 1B). At higher magnification and when counter-stained with DAPI and anti-a-tubulin immunohistochemistry,the Elav signal appeared confined to a bilateral ventral clusterof cells in the floor of the cord, on either side of a narrow medialgap and separated from the lumen by Elav2 cells (Figure 1C).The much thinner roof of the cord did not express Elav. Asecond pan-neuronal gene, Synaptotagmin I, which encodesa protein involved in neurotransmitter release [22] (Figure S1),was expressed in the same pattern described for Elav(Figure 1D). The ventral concentrations of Elav+/Synaptotag-min+ cells overlaid a thick mass of fibers devoid of Elav/Synap-totagmin signal but strongly labeled with anti-a-tubulin(Figures 1C and 1D). This fiber layer, although previouslydescribed as made exclusively of longitudinally coursing axons[3, 20], contained instead many dorsoventrally or randomlyoriented fibers, some emerging from the overlying cell bodies,as would be expected of a neuropil (Figures 1E and 1F). Theoverall arrangement of cells and fibers of the collar cord was

Evolution of the Central Nervous System1265

Figure 1. Structure of the Collar Cord of Ptychodera flava

(A) Simplified schematic of an enteropneust worm sectioned along the

midline (adapted from [9]). Only the ectodermic and endodermic walls are

represented, not the mesoderm filling the coelomic cavities or the heart or

kidney.

(B) Transverse section through the collar of P. flava (plane of section indi-

cated in A) hybridized with Elav. A signal is visible in the collar cord (red

arrowhead) but, at this magnification, not in the surface epithelium of the

collar (black arrowheads).

(C) Transverse section through the collar cord of P. flava treated by

combined in situ hybridization for Elav, immunofluorescence for acetylated

a-tubulin, and DAPI nuclear stain. The collar cord is connected to the

surface ectoderm by the dorsal mesentery (black arrowhead). The neuropil

of the cord is continuous through the dorsal mesentery with the nerve fiber

layer of the skin (red arrowhead). The Elav signal is restricted to the cellular

portion of the ventral wall of the cord.

(D) Transverse section through the collar cord of P. flava hybridized with

Synaptotagmin I. Inset: low-magnification view of the section through the

collar. As for Elav, a signal is visible only in the cord.

(E) Sagittal section through the collar cord stained with DAPI and immuno-

fluorescence for acetylated a-tubulin. Numerous fibers are visible, forming

a neuropil. White arrowhead and inset: thick layer of cross-reactive material

between the cellular layer and neuropil, which might correspond to Spen-

gel’s ‘‘reticular membrane’’ [17].

(F) Transverse section through the floor of the collar cord stained with Mal-

lory’s trichrome, showing processes oriented perpendicular to the main axis

of the cord. The arrowhead points to the same structure as in (E).

(G and H) Side-by-side comparison of a transverse section through an E5.5

chicken hindbrain hybridized with ELAVL4/HuD (G) and a section through

the P. flava collar cord hybridized with Elav (H). Note the similar relative

thus reminiscent of the vertebrate neural tube (compareFigures 1G and 1H), with equivalents of a lumen, a mantle layer(or gray matter) made of neuronal bodies, and a marginal layer(or white matter) made of nerve fibers. The collar cord also hasstructures that correspond topologically to a neuron-freedorsal ‘‘roof plate,’’ a ventral ‘‘floor plate,’’ and a ventricularzone.

Caudal to the collar, the cord opens to the outside througha neuropore, and its fibrous layer continues in a superficialposition (i.e., no longer internalized) in the form of a ‘‘dorsalcord’’ [9]. Similarly, the population of neurons in the floor ofthe collar cord continues along the dorsal midline of the trunk,although now superficially located and overlying the fibers(Figures 2A and 2B; Figure S2). On the ventral midline is anequivalent, if slightly larger, concentration of neurons over-lying the fibers of the ‘‘ventral cord’’ (Figures 2A and 2C;Figure S2). Just behind the collar, the so-called ‘‘peripharyng-eal nerve ring’’ [11] or ‘‘circumesophageal tract’’ [2] connectingthe ventral cord with the dorsal cord shows the same histolog-ical structure (Figures 2A and 2B; Figure S2), demonstratingthat it also is a nerve center. Rostral to the anterior neuropore,the collar cord continues in a superficial position and widensto cover the dorsal half of the proboscis stem, where bothneuronal and fiber layers are thickest (Figures 2A, 2D, and2E). Further rostrally, this neural plate-like structure extendsventrally to fully encircle the stem just behind the proboscisproper (Figure 2D, inset). Outside of the cords and the neuralplate of the proboscis stem, which together form the enterop-neust CNS, Elav+ cells can be found, interspersed withnonneuronal cells, throughout the proboscis (Figure 2F) ata density lower than 2 neurons per 100 nonneuronal cellsand in the inner folds of the collar (Figure S2). Few neuronsare visible on the outer surface of the collar (Figure 2G) or inthe trunk epithelium (Figure 2H) (less than 2 neurons per1000 nonneuronal cells). We propose that these scatteredneurons in the skin constitute a peripheral nervous system(PNS), as found in cephalochordates (e.g., [23, 24]) or urochor-dates (e.g., [25, 26]). Finally, neurons are scattered in the endo-dermal lining of the pharynx (Figure 2I).

Juvenile specimens of Saccoglossus kowalevskii have beenreported, based on Elav expression, to lack any concentrationof neurons [13]. In contrast, ISH with Elav on the collar region ofan adult specimen of Saccoglossus showed a concentration ofElav+ cells at the same position as in P. flava (compareFigure 2J with Figure 1C) (see also Figure S3). Thus, in agree-ment with classical drawings [9], the collar cord has essentiallythe same structure across genera, the major yet superficialdifference being the morphology of the lumen—open in pty-choderids, collapsed in harrimaniids—possibly owing to thedifferent modes of neurulation. Outside of the collar cord, theanatomy of the nervous system of S. kowalevskii was verysimilar to that of P. flava (Figure S3).

To determine whether this condensed nervous systemsomehow arises from an early, more diffuse state, we exam-ined metamorphic specimens of P. flava towed from theplankton (Figures 3A and 3A0). ISH with Elav on whole-mount

positions of a ‘‘roof plate’’ (red asterisk), lumen, ‘‘floor plate’’ (yellow

asterisk), ventricular layer (black arrowhead), mantle layer (where the Elav

signal resides), and fibrous marginal zone (blue arrowhead) (only starting

to form in the chicken embryo).

cc, collar cord; dc, dorsal cord; gs, gill slit; lu, lumen; m, mouth; np, neuropil;

ph, pharynx; vc, ventral cord. Scale bars represent 100 mm in (B)–(D), (G),

and (H); 50 mm in (E) and (F).

Current Biology Vol 19 No 151266

Figure 2. Structure of the P. flava CNS and PNS as Revealed by Elav Expres-

sion

(A) Sagittal section through the proboscis, collar, and anterior trunk of an en-

teropneust, hybridized with Elav. For orientation, compare with Figure 1A.

Planes of section or areas depicted in (B)–(I) are outlined in red. At this

magnification, only the CNS is visible, not the PNS elsewhere in the skin.

The dorsal and ventral trunk cords—visible on the transverse sections in

(B) and (C)—are too narrow to capture sagittally on any meaningful length

at this scale. The deep indentation in the proboscis is a fixation artifact.

(B–D) Transverse sections through the junction between collar and trunk (B),

trunk (C), and proboscis stem (D). The cords and anterior neural plate are

indicated with red arrowheads. The peripharyngeal cord only occasionally

meets the plane of section in (B). Inset in (D): transverse section of the

proboscis stem immediately behind the proboscis proper. At this level,

embryos and on sections revealed a discrete signal in thedorsal and ventral cords (Figures 3B–3D) showing that, fromthe earliest stage of metamorphosis, the nervous system ofP. flava has a centralized component. How can one reconcilethe cord system of ptychoderids and adult harrimaniids withthe diffuse Elav expression described in juvenile S. kowalevskii[13]? A possible answer is that the development of harrima-niids is misleadingly called direct and would be better charac-terized as indirect but ‘‘abbreviated’’ [9]: several structuresof the hatching animal (apical tuft, telotroch, and postanaltail) are later lost, together with its swimming behavior. Simi-larly, the homogeneous distribution of neurons reported in

the dorsal neural plate completely encircles the stem. Black arrowheads

in (D) point to the collar folds that cover the proboscis stem.

(E) Sagittal section through the collar and proboscis stem. The neuropil is

thickest in the stem.

(F–I) High-magnification views of the epithelium of the proboscis (F), collar

(G), trunk (H), and pharynx (I). The only neurons visible in (G) and (H) (outside

the trunk ventral cord) are indicated by red arrowheads.

(J) Transverse section thought the collar cord of an adult S. kowalevskii,

hybridized with Elav. Despite the slightly different shape of the cord and

thinner layer of neurons, the overall structure is the same as in P. flava. No

neurons are visible in the thick skin above the collar cord. The black arrow-

head indicates the dorsal mesentery.

cc, collar cord; ce, collar epidermis; cf, collar fold; dph, digestive pharynx; n,

neuropore; np, neuropil; ph, pharynx; ps, proboscis stem; rph, respiratory

pharynx; vc, ventral cord. Scale bars represent 200 mm in (A)–(D); 100 mm

in (E)–(J).

Figure 3. Structure of the Nervous System of P. flava at Metamorphosis

(A and A0) Tornaria larvae of P. flava at the beginning of metamorphosis (A)

and at the juvenile stage (A0). The telotroch of the larva (blue arrowhead)

gives way to a pigmented band in the juvenile (blue arrowhead).

(B) Dorsal view of a metamorphic larva at a stage comparable to (A), hybrid-

ized with Elav, showing the dorsal cord interrupted at the level of the telo-

troch (blue arrowhead).

(C and D) Transverse sections through the collar (C) and trunk (D) of a juve-

nile individual at a stage comparable to (A0), hybridized with Elav. The invag-

inating collar cord (C) and the dorsal and ventral cords (D) are indicated by

red arrowheads.

in, intestine; m, mouth; ph, pharynx. Scale bars represent 500 mm in (A) and

(B); 100 mm in (C) and (D).

Evolution of the Central Nervous System1267

Figure 4. Distribution of Neuronal Types in the

Enteropneust PNS and CNS

(A) Immunofluorescence with an anti-serotonin

(5-HT) antibody combined with DAPI stain on

a transverse section of the collar cord. No 5-HT

cell is visible, whereas numerous 5-HT fibers

are found in the marginal layer.

(B) Immunofluorescence with an anti-serotonin

antibody on the PNS of the inner folds of the

collar. A few positive fibers are seen in the fiber

layer (red bracket). Inset: close-up of a seroto-

nergic neuron.

(C) Immunofluorescence with an anti-GABA

antibody on the neural plate of the proboscis

stem. Many positive fibers are found in the thick

fiber layer (red bracket). Inset: close-up of a

GABAergic neuron.

(D–I) Transverse sections thought the collar cord

(D and G–I) or proboscis stem (E and F), hybrid-

ized with VAChT (D and F), Elav (E and G),

Drg11 (H), and Hb9 (I), showing the respective

abundance and position of presumptive moto-

neurons and sensory neurons. Insets in (D) and

(I): close-up of a VAChT+ neuron and an Hb9+

neuron, respectively. Insets in (E) and (F): low-

magnification view of a transverse section

through the anterior proboscis stem, a detail of

which is shown in the main panel.

Scale bars represent 100 mm.

S. kowalevskii juveniles could represent a short-lived larvalnervous system suited for the brief motile phase of theanimal’s existence, with no stronger connection to the adultCNS than the ciliated bands of ptychoderids’ tornaria larvae[7, 27]. In line with this, a dorsal midline concentration ofElav-positive cells has been reported in late S. kowalevskiijuveniles [13]. The widespread ectodermal expression of ante-roposterior patterning genes, which in some cases is interrup-ted or modified on the dorsal midline [13], as well as dorsoven-tral patterning genes [28] should be reexamined in relationshipto the newly described CNS. Finally, it is possible that part ofthe juvenile—or larval—diffuse nervous system of S. kowalev-skii is fated to become the PNS of the adult, especially in theproboscis.

We then asked whether different types of neurons reside inspecific regions of the CNS or PNS. Histologically, we failedto identify any giant nerve cell [17, 18] in the collar cord, inagreement with a previous report that these cells are not readilydistinguishable in Ptychoderidae [18]. We then examined theoccurrence of the neurotransmitter phenotypes cholinergic,serotonergic, and GABAergic with a VAChT probe—codingfor the vesicular acetylcholine transporter (Figure S1)—andantibodies against 5-hydroxytryptamine (5-HT, serotonin) andGABA, respectively. Numerous giant 5-HT neurons, in theabsence of 5-HT fibers, have been recently reported in thecollar cord of another ptychoderid, Glossobalanus berkeleyi[19]. The arrangement was the opposite in P. flava, whose collarcord was rich in 5-HT fibers but devoid of 5-HT cells (Figure 4A).However, 5-HT cells were abundant in the PNS (Figure 4B;Figure S4). By far the highest density of GABAergic neuronswas found in the neural plate at the base of the proboscis(Figure 4C), but some were also detected in the ventral anddorsal cords and in the PNS of the proboscis (Figure S4).Finally, VAChT was expressed in a few cells per section in thecords (Figure 4D; Figure S5) and in a dense dorsal cluster inthe anterior neural plate, at the level where it surrounds the

proboscis stem (Figures 4E and 4F). No VAChT expressionwas detected in the PNS.

We then cloned and examined the expression of twoneuron-type-specific transcription factors, Drg11 and Hb9/Mnr2 (Figure S1), chosen for their marked level of neuron-type specificity in chordates. In vertebrates, Drg11 is largelyspecific for somatic sensory neurons of the CNS and PNS[29, 30]. Inside the vertebrate nervous system, Hb9 is specificfor somatic motoneurons [31] and a subclass of interneurons[32] in the spinal cord and hindbrain. It is also expressed insomatic motoneurons in larval urochordates [33] and possiblycephalochordates [34], as well as in protostomes [35, 36]. InP. flava, Dgr11+ neurons and Hb9+ neurons were found in thecollar cord (Figures 4G–4I), proboscis stem, and ventral anddorsal cords (Figure S5). An occasional Drg11+ neuron (datanot shown) but no Hb9+ neurons were found in the PNS ofthe proboscis or collar folds. Although more markers areneeded to better understand the neuroanatomy of enterop-neusts, two notions already emerge from our analysis. First,the phenotypes that we have diagnosed are anatomicallysegregated: Drg11+, Hb9+, and cholinergic neurons are prefer-entially or exclusively located in the CNS, whereas 5-HTneurons are restricted to the PNS, further arguing for thedifferent nature of these two parts of the nervous system.Second, if cell types are used as an argument for organ-levelhomology [37], then the entire cord system or CNS of P. flavawould correspond to the spinal and hindbrain levels of thevertebrate CNS, where Hb9+ and Dgr11+ cells are confined(notwithstanding a midbrain and diencephalic expression ofDrg11 apparently specific to teleosts [38]).

Conclusions

The nervous system of enteropneust worms is much morelocalized than previously believed. In one of the last histolog-ical studies published to date ([10], reviewed in [3]), the‘‘astounding’’ abundance of neurons said to reside throughout

Current Biology Vol 19 No 151268

the adult animal’s skin were never ‘‘distinguishable withcertainty’’ and were deduced from ‘‘strong presumptiveevidence.’’ The second and last study [11], although referringto a ‘‘richly nervous epidermis [possibly] homologous withthe neural plate of vertebrates,’’ in fact did not show anyneurons in the collar or trunk outside the cords. We haveshown that these neurons do exist but are relatively rare andwould be best described as a PNS. Apart from them, thenervous system of enteropneusts consists of cord-like massesof neurons overlying a neuropil. Moreover, the dorsal cord isinternalized at the level of the collar in a manner indistinguish-able from the chordate neural tube. A CNS has also beendescribed in pterobranches in the form of a ganglion situatedat a position similar to that of the enteropneusts’ collar cord[39–41], suggesting that whether pterobranches are the sistergroup of enteropneusts [5] or are nested in their midst [42], thepossession of a CNS is primitive for hemichordates.

The present findings suggest that it is implausible that theenteropneust skin is homologous to the chordate CNS andreopen the possibility that some portions of its nerve cordsare, instead. The data provide no reason to favor the ventralover the collar cord, or vice versa, as homologs of the chordateCNS (as variously suggested in [6, 7, 19, 43]), because theneuron types identified so far are present throughout thecord system. It is notable that the collar cord, a bona fide neuralcenter, and its anterior continuation on the proboscis stemconstitute a spatial impediment to the evolution of a new mouthin chordates, as required by the dorsoventral inversion hypoth-esis [6, 44]. At this stage of our knowledge, it remains possiblethat what is conserved, constituting a kind of ‘‘deep homology’’[45] of neural development between hemichordates and chor-dates, and possibly protostomes as well [1], is not the CNS it-self but the ontogenetic module giving rise to it, which wouldhave been deployed at different sites of the body—sometimesat multiple sites—in different lineages [46].

Experimental Procedures

Animal Collection

Adult P. flava were collected at low tide on Paiko Beach on the southern

shore of Oahu (HI, USA), and metamorphic specimens were towed from

the plankton 2 miles south of Honolulu. Animals were treated for immunoflu-

orescence or in situ hybridization as detailed in Supplemental Experimental

Procedures.

Accession Numbers

Sequences described herein have been deposited at GenBank with the

accession numbers GQ223115 (Elav), GQ229033 (Drg11), GQ229034

(Hb9), GQ229035 (Synaptotagmin), and GQ229036 (VAChT).

Supplemental Data

Supplemental Data include Supplemental Experimental Procedures and

five figures and can be found with this article online at http://www.cell.

com/current-biology/supplemental/S0960-9822(09)01234-2.

Acknowledgments

We thank A. Hejnol and N. Meyer for valuable advice at the beginning of this

work, R. Chock for organizing the plankton tows, R. de Rosa for help with

the orthology analyses, and C. Goridis for critical reading of the manuscript.

This work was supported by grants from Centre National de la Recherche

Scientifique, Fondation pour la Recherche Medicale, and Agence Nationale

pour la Recherche (to J.-F.B.).

Received: May 7, 2009

Revised: May 25, 2009

Accepted: May 26, 2009

Published online: June 25, 2009

References

1. Arendt, D., Denes, A.S., Jekely, G., and Tessmar-Raible, K. (2008). The

evolution of nervous system centralization. Philos. Trans. R. Soc.

Lond. B Biol. Sci. 363, 1523–1528.

2. Holland, N.D. (2003). Early central nervous system evolution: An era of

skin brains? Nat. Rev. Neurosci. 4, 617–627.

3. Bullock, T.H., and Horridge, G.A. (1965). Hemichordata. Structure and

Function of the Nervous System of Invertebrates (San Francisco: W.H.

Freeman), pp. 1567–1577.

4. Bourlat, S.J., Juliusdottir, T., Lowe, C.J., Freeman, R., Aronowicz, J.,

Kirschner, M., Lander, E.S., Thorndyke, M., Nakano, H., Kohn, A.B.,

et al. (2006). Deuterostome phylogeny reveals monophyletic chordates

and the new phylum Xenoturbellida. Nature 444, 85–88.

5. Halanych, K.M. (1995). The phylogenetic position of the pterobranch

hemichordates based on 18S rDNA sequence data. Mol. Phylogenet.

Evol. 4, 72–76.

6. Nubler-Jung, K., and Arendt, D. (1996). Enteropneusts and chordate

evolution. Curr. Biol. 6, 352–353.

7. Salvini-Plawen, L.V. (1998). The urochordate larva and archichordate

organization: Chordate origin and anagenesis revisited. J. Zool. Syst.

Evol. Res. 36, 129–145.

8. Bateson, W. (1886). Continued account of the later stages in the devel-

opment of Balanoglossus Kowalevskii and of the morphology of the en-

teropneusta. Q. J. Microsc. Sci. 26, 511–534.

9. Dawydoff, C. (1948). Classe des Enteropneustes. In Traite de Zoologie,

Volume XI, P.-P. Grasse, ed. (Paris: Librairies de l’Academie de Mede-

cine), pp. 369–453.

10. Bullock, T.H. (1946). The anatomical organization of the nervous system

of Enteropneusta. Q. J. Microsc. Sci. 86, 55–112.

11. Knight-Jones, E.W. (1952). On the nervous system of Saccoglossus

cambrensis (Enteropneusta). Philos. Trans. R. Soc. Lond. B Biol. Sci.

236, 315–354.

12. Gerhart, J., Lowe, C., and Kirschner, M. (2005). Hemichordates and the

origin of chordates. Curr. Opin. Genet. Dev. 15, 461–467.

13. Lowe, C.J., Wu, M., Salic, A., Evans, L., Lander, E., Stange-Thomann, N.,

Gruber, C.E., Gerhart, J., and Kirschner, M. (2003). Anteroposterior

patterning in hemichordates and the origins of the chordate nervous

system. Cell 113, 853–865.

14. Bateson, W. (1884). The early stages in the development of Balanoglos-

sus (sp. incert.). Q. J. Microsc. Sci. 24, 208–236.

15. Morgan, T.H. (1894). The development of Balanoglossus. J. Morphol. 9,

1–86.

16. Bateson, W. (1886). The ancestry of the chordata. Q. J. Microsc. Sci. 26,

535–571.

17. Spengel, J.W. (1893). Die Enteropneusten des Golfes von Neapel. Fauna

and Flora des Golfes von Neapel 18.

18. Bullock, T.H. (1944). The giant nerve fibre system in Balanoglossids.

J. Comp. Neurol. 80, 355–367.

19. Brown, F.D., Prendergast, A., and Swalla, B.J. (2008). Man is but a worm:

Chordate origins. Genesis 46, 605–613.

20. Cameron, C.B., and Mackie, G.O. (1996). Conduction pathways in the

nervous system of Sacoglossus sp. (Enteropneusta). Can. J. Zool. 74,

15–19.

21. Pascale, A., Amadio, M., and Quattrone, A. (2008). Defining a neuron:

Neuronal ELAV proteins. Cell. Mol. Life Sci. 65, 128–140.

22. Rizo, J., and Rosenmund, C. (2008). Synaptic vesicle fusion. Nat. Struct.

Mol. Biol. 15, 665–674.

23. Bone, Q. (1961). The organization of the atrial nervous system of amphi-

oxus (Branchiostoma Lanceolatum (Pallas)). Philos. Trans. R. Soc.

Lond. B Biol. Sci. 243, 241–269.

24. Benito-Gutierrez, E., Illas, M., Comella, J.X., and Garcia-Fernandez, J.

(2005). Outlining the nascent nervous system of Branchiostoma floridae

(amphioxus) by the pan-neural marker AmphiElav. Brain Res. Bull. 66,

518–521.

25. Pasini, A., Amiel, A., Rothbacher, U., Roure, A., Lemaire, P., and Darras,

S. (2006). Formation of the ascidian epidermal sensory neurons: Insights

into the origin of the chordate peripheral nervous system. PLoS Biol. 4,

e225.

Evolution of the Central Nervous System1269

26. Mackie, G.O., and Singla, C.L. (2003). The capsular organ of Chelyo-

soma productum (Ascidiacea: Corellidae): A new tunicate hydrody-

namic sense organ. Brain Behav. Evol. 61, 45–58.

27. Morgan, T.H. (1891). The growth and metamorphosis of Tornaria.

J. Morphol. 5, 407–458.

28. Lowe, C.J., Terasaki, M., Wu, M., Freeman, R.M., Jr., Runft, L., Kwan, K.,

Haigo, S., Aronowicz, J., Lander, E., Gruber, C., et al. (2006). Dorsoven-

tral patterning in hemichordates: Insights into early chordate evolution.

PLoS Biol. 4, e291.

29. Saito, T., Greenwood, A., Sun, Q., and Anderson, D.J. (1995). Identifica-

tion by differential RT-PCR of a novel paired homeodomain protein

specifically expressed in sensory neurons and a subset of their CNS

targets. Mol. Cell. Neurosci. 6, 280–292.

30. Rebelo, S., Reguenga, C., Osorio, L., Pereira, C., Lopes, C., and Lima, D.

(2007). DRG11 immunohistochemical expression during embryonic

development in the mouse. Dev. Dyn. 236, 2653–2660.

31. Tanabe, Y., William, C., and Jessell, T.M. (1998). Specification of motor

neuron identity by the MNR2 homeodomain protein. Cell 95, 67–80.

32. Wilson, J.M., Hartley, R., Maxwell, D.J., Todd, A.J., Lieberam, I.,

Kaltschmidt, J.A., Yoshida, Y., Jessell, T.M., and Brownstone, R.M.

(2005). Conditional rhythmicity of ventral spinal interneurons defined

by expression of the Hb9 homeodomain protein. J. Neurosci. 25,

5710–5719.

33. Dufour, H.D., Chettouh, Z., Deyts, C., de Rosa, R., Goridis, C., Joly, J.-S.,

and Brunet, J.-F. (2006). Pre-craniate origin of cranial motoneurons.

Proc. Natl. Acad. Sci. USA 103, 8727–8732.

34. Ferrier, D.E., Brooke, N.M., Panopoulou, G., and Holland, P.W. (2001).

The Mnx homeobox gene class defined by HB9, MNR2 and amphioxus

AmphiMnx. Dev. Genes Evol. 211, 103–107.

35. Thor, S., and Thomas, J.B. (2002). Motor neuron specification in worms,

flies and mice: Conserved and ‘lost’ mechanisms. Curr. Opin. Genet.

Dev. 12, 558–564.

36. Denes, A.S., Jekely, G., Steinmetz, P.R., Raible, F., Snyman, H., Prud’-

homme, B., Ferrier, D.E., Balavoine, G., and Arendt, D. (2007). Molecular

architecture of annelid nerve cord supports common origin of nervous

system centralization in bilateria. Cell 129, 277–288.

37. Arendt, D. (2005). Genes and homology in nervous system evolution:

Comparing gene functions, expression patterns, and cell type molec-

ular fingerprints. Theory Biosci. 124, 185–197.

38. McCormick, L.J., Hutt, J.A., Hazan, J., Houart, C., and Cohen, J. (2007).

The homeodomain transcription factor drg11 is expressed in primary

sensory neurons and their putative CNS targets during embryonic

development of the zebrafish. Gene Expr. Patterns 7, 289–296.

39. M’Intosh, W.C. (1887). Report on Cephalodiscus dodecalophus M’In-

tosh, a new type of polyzoa. Challenger Reports 20, 1–48.

40. Rehkamper, G., Welsch, U., and Dilly, P.N. (1987). Fine structure of the

ganglion of Cephalodiscus gracilis (Pterobranchia, Hemichordata).

J. Comp. Neurol. 259, 308–315.

41. Sato, A., Bishop, J.D., and Holland, P.W. (2008). Developmental biology

of pterobranch hemichordates: History and perspectives. Genesis 46,

587–591.

42. Cannon, J.T., Rychel, A.L., Eccleston, H., Halanych, K.M., and Swalla,

B.J. (2009). Molecular phylogeny of hemichordata, with updated status

of deep-sea enteropneusts. Mol. Phylogenet. Evol. 52, 17–24.

43. Nubler-Jung, K., and Arendt, D. (1999). Dorsoventral axis inversion: En-

teropneust anatomy links invertebrates to chordates turned upside

down. J. Zool. Syst. Evol. Res. 37, 93–100.

44. Lacalli, T.C. (2008). Basic features of the ancestral chordate brain: A pro-

tochordate perspective. Brain Res. Bull. 75, 319–323.

45. Shubin, N., Tabin, C., and Carroll, S. (2009). Deep homology and the

origins of evolutionary novelty. Nature 457, 818–823.

46. Gerhart, J. (2000). Inversion of the chordate body axis: Are there alterna-

tives? Proc. Natl. Acad. Sci. USA 97, 4445–4448.