Abstract Neochildia fusca is a member of the taxonAcoela, a group of flatworms that, according to some recent molecular phylogenetic analyses, are distinct fromother flatworms and constitute a basal branch with a sister taxon relationship to the rest of the Bilateria. Inthis paper, we analyze early neural development in thisspecies and report the sequence and expression of twoPit-Oct-Unc (POU) genes, NeocBrn-1 and NeocBrn-3.Homologs of these highly conserved genes play a role inneural fate determination in vertebrates, Drosophila andCaenorhabditis elegans. Acoels, including Neochildia,have a unique invariant pattern of early cleavage calledduet spiral cleavage. In subsequent cell divisions descen-dants of the first three micromere duets form an outerlayer of epidermal and neural progenitors surroundingthe meso/endoderm progenitors, which are themselvesdescended from the macromere duet 4A, B and the micromere duet 4a, b. Organ formation begins at mid-embryonic stages with the epidermal primordium adopt-ing a ciliated epithelial shape. Sub-epidermally, a bilater-ally symmetric brain primordium can be seen at the ante-rior pole. Laterally and posteriorly, myoblasts form a

thin layer underneath the epidermis. In late embryos and juveniles of Neochildia, the brain is formed by a 3–4 cell-diameter-thick layer of neurons forming a cor-tex surrounding a neuropile that is relatively free of cellbodies. A highly regular “orthogonal” array of muscle fi-bers penetrates the brain. We have isolated and partiallysequenced homologs of the vertebrate Brn-1 and Brn-3genes, which we call NeocBrn-1 and NeocBrn-3, respec-tively. These sequences contain and span portions of thePOU-specific domain and a homeodomain, and are se-quence similar to their respective homologs in verte-brates and Drosophila. RT-PCR reveals that NeocBrn-1and NeocBrn-3 are expressed from mid-embryonic toadult stages. Whole-mount in situ hybridization showsexpression of both genes in distinct subsets of nerve cellsin juvenile and adult worms. NeocBrn-1 also appears in asubset of intra-epidermal gland cells. These observationsare an initial step towards reconstructing the neural de-velopment of a key group of bilaterians, the Acoela.These flatworms, by virtue of their distinct morphology,development and phylogenetically basal placement, arelikely to provide key insights into the interpretation ofthe evolution of metazoan neural architecture.

Keywords Acoel · Neochildia · Nervous system · Development · POU gene

Introduction

Genes with developmental functions are remarkably con-served across a wide range of animals, a feature that hasproven broadly useful in comparing developmental sys-tems in distantly related taxa. This, in turn, has renewedinterest in comparative embryology as a basis for discus-sion of homologies between cells, tissues, and organs,and as a means of reconstructing the evolution of devel-opmental systems. Here, we focus on generating a betterunderstanding of neural development in acoel flatworms,a group that appears to have a primitive form of neuraldevelopment relative to other Bilateria.

Edited by J. Campos-Ortega

N.B. Ramachandra · V. Hartenstein (✉ )Department of Molecular, Cell, and Developmental Biology, University of California at Los Angeles, 621 Charles E. Young Drive South, Los Angeles, CA 90095–1606, USAe-mail: volkerh@mcdb.ucla.eduTel.: +1-310-2067523, Fax: +1-310-2063987

R.D. Gates · D.K. JacobsDepartment of Organismic Biology, Ecology and Evolution, University of California at Los Angeles, 621 Charles E. Young Drive South, Los Angeles, CA 90095–1606, USA

P. LadurnerInstitut of Zoology and Limnology, University of Innsbruck, Technikerstrasse 25, A-6020, Innsbruck, Austria

Present address:N.B. Ramachandra, Department of Zoology, University of Mysore, Manasagangotri, Mysore 570006, India

Dev Genes Evol (2002) 212:55–69DOI 10.1007/s00427-001-0207-y

O R I G I N A L A RT I C L E

Nallur B. Ramachandra · Ruth D. GatesPeter Ladurner · David K. JacobsVolker Hartenstein

Embryonic development in the primitive bilaterian Neochildia fusca:normal morphogenesis and isolation of POU genes Brn-1 and Brn-3

Received: 16 October 2001 / Accepted: 26 November 2001 / Published online: 21 February 2002© Springer-Verlag 2002

To reconstruct the evolution of specific molecular net-works and morphogenetic events in neurogenesis re-quires a detailed knowledge of neural development intaxa that span the phylogenetic tree. In this context,members of the phylum Platyhelminthes (flatworms) areof special interest, as flatworms have often been recon-structed as basal in bilaterian phylogeny and may haveretained primitive features relative to their more evolvedbilaterian relatives (Hyman 1940; Ehlers 1985; Ax1996). Potentially primitive or ancestral features includea simple gut with a single opening, an epidermal layer ofmulti-ciliated cells that form the main organ of locomo-tion, and a compact anterior brain. In addition, platyhel-minths represent the simplest group of animals possess-ing a central nervous system.

Recent analyses of 18 s rDNA have placed the flat-worms in alliance with spiralian phyla such as molluscsand annelids, as well as lophophorates, in the clade Lophotrochozoa within the protostome Bilateria (Adoutteet al. 1999; Ruiz-Trillo et al. 1999). In addition Acoela, ataxon classically assigned to the platyhelminths, mayhave branched off the metazoan tree earlier than otherflatworms; thus Acoela potentially represent a sistergroup to higher bilaterians (“eubilaterians”). Both theplacements of non-acoel flatworms within the Loph-otrochozoa (Giribet et al. 2000) and the concept that theAcoela are the basal branch of living Bilateria (Berney etal. 2000; Littlewood et al. 2001) continue to be contro-versial. However, structural differences between theacoels and the other platyhelminths are consistent withtheir basal placement (review in Smith and Tyler 1985).Acoels have a gut syncytium, rather than a multicellularepithelial gut, and they lack a paired protonephridial ex-cretory system. The organization of the nervous system inadult acoels lacks a compact brain and neuropile. Nervefibers leaving the brain form an irregular nerve net-likeplexus with no straight connectives or commissures (Kotikova 1986; Reuter and Gustafsson 1995; Reuter etal. 1998), an organization that is reminiscent of the nervenet typical of cnidarians and ctenophores.

Given the basal phylogenetic position of acoels, ananalysis of the function of developmentally relevantgenes in these animals will likely reveal important infor-mation regarding the evolution of fundamental aspects ofbilaterian neural organization. To initiate such an investi-gation, we have begun to screen for homologs of highlyconserved genes involved in acoel neurogenesis. In thispaper we report the partial sequence and expression oftwo Pit-Oct-Unc (POU) genes, Brn-1 and Brn-3.

The POU family of transcription factors was originallydefined on the sequence homology of approximately150–160 amino acids that was identified in the mammaliantranscription factors, Pit-1, Oct-1 and Oct-2 and the nema-tode factor UNC-86 (Herr et al. 1988; Ryan and Rosenfeld1997; Latchman 1999). Since then, a large family of POU-domain containing transcription factors, which form aunique subfamily of homeodomain proteins, has beencharacterized in a variety of species. The POU domain isbipartite in nature, consisting of a highly conserved amino-

terminal region of 75–82 amino acids (POU-specific do-main) and a similarly conserved carboxyl-terminal regionof 60 amino acids (POU homeodomain). These conserveddomains are separated by a less conserved linker region.Based on the sequences of the linker region and the mainbasic cluster at the amino terminus of the POU homeodo-main, the POU-domain family can be divided into sixclasses, I–VI (Wegner et al. 1993).

The role of the POU genes in ontogenetic develop-ment and cellular differentiation has been investigated ina variety of organisms, including vertebrates and Dro-sophila. Many members of the group, in particular thosebelonging to classes III and IV, are expressed in specificsubsets of neurons within embryonic and adult neuraltissues and have been shown to play important roles inthe specification of these neuronal cell types (reviewedin Ryan and Rosenfeld 1997). In this paper, we report thecloning and expression of two POU domain-containinggenes, Brn-1 and Brn-3, in the acoel Neochildia. Bothgenes show a high degree of sequence similarity withtheir vertebrate and Drosophila counterparts, and are ex-pressed in a subset of nerve cells in the brain. In order tointerpret the expression pattern of the POU genes, andother developmentally relevant genes, we have analyzedthe normal development of Neochildia embryos using acombination of live observations, whole-mount prepara-tions, and histology. We introduce a series of stages, de-fined by easily recognizable morphological criteria,which will facilitate comparing the development ofacoels with that of other flatworm taxa. Knowledge ofthe Neochildia POU genes will help clarifying the ances-try and evolution of the POU family of transcription reg-ulators; in addition, it will add a tool with which to ana-lyze neural development in this primitive Bilaterian.

Materials and methods

Animals

Live adult Neochildia were obtained from the Marine BiologicalLaboratory, Woods Hole, Massachusetts, and cultured for 2–6 weeks at 20°C. Eggs produced during this time were collectedand aged.

Fixation

To remove mucus prior to fixation, Neochildia embryos and adultswere treated with 2% HCl in PBS for 5 min, then fixed in a 1:1 so-lution of Carnoy’s solution and 4% formaldehyde for 3 h and de-hydrated through a graded ethanol series. To remove the dark pig-mentation of Neochildia, fixed specimens were incubated in 6%H2O in methanol for 12 h under a bright light at room temperature(Agata et al. 1998). Bleached, fixed and dehydrated embryos werestored in methanol at –20°C until used.

Electron microscopy and histology

Following the fixation described above, embryos were post fixedin a mixture of 1% osmium tetroxide and 2% glutaraldehyde in0.15 M cacodylate buffer for 10 min at 4°C, washed several timesin PBS and dehydrated through a graded ethanol series and ace-tone (all steps at 4°C). Preparations were embedded by infiltrating

56

the specimens in a 1:1 mixture of Epon and acetone overnight andin Epon alone for 5–10 h. The infiltrated specimens were trans-ferred to molds, oriented, and placed at 60°C for 24 h to polymer-ize the Epon. Alternating 1-µm semi-thin sections and sets of80 nm (silver) ultra-thin sections were taken with an LKB ultra-tome. Ultra-thin sections were mounted on net grids (Ted Pella)and treated with uranyl acetate and lead citrate.

Fuchsin labeling of whole-mounts

The whole-mount technique employed here was adapted fromAshburner (1989) and has been used extensively by us, and others,to label nucleic acids in whole embryos of insects and other inver-tebrates. Briefly, following a fixation in 4% PBS-buffered formal-dehyde, 20–50 embryos were placed in small wire mesh baskets,washed in three changes of 70% ethanol and once in distilled water for 5 min each. The embryos were incubated in 2 N HCl for10 min at 60°C to denature the DNA. Following one wash in dis-tilled water and two washes in 5% acetic acid for 5 min each, em-bryos were stained in 2% filtered basic fuchsin in 5% acetic acidfor 15 min. Embryos were destained in 5% acetic acid to removecytoplasmic fuchsin labeling, dehydrated through a graded ethanolseries, transferred to Epon, and individually mounted on slides.

Immunohistochemistry

Neurons were visualized in whole-mount embryos using an anti-acetylated tubulin monoclonal antibody (acTub, 1:1,000; Sigma)and cells in metaphase traced with anti-phosphorylated H3 histone(anti-H3, 1:300; Upstate Biotechnology). Mixed age embryosfixed in 4% PBS-buffered formaldehyde were washed in fromthree to five changes of PBT (PBS plus 0.3% Triton X-100) over a10 min period and incubated overnight in PBT containing the pri-mary antibody at the appropriate dilution. Following a wash inPBT the embryos were incubated in PBT containing the secondaryantibody – peroxidase-conjugated rabbit anti-mouse immunoglob-ulin (Jackson Labs), at a dilution of 1:800 for 4 h. To initiate thecolor reaction embryos were briefly rinsed in, then incubated with,0.1% diamino-benzidine (DAB, Sigma) in 0.1 M phosphate buffer(pH 7.3) containing 0.006% hydrogen peroxide. The reaction wasstopped after 5–10 min by diluting the substrate with 0.1 M phos-phate buffer. Preparations were dehydrated through a graded etha-nol series and acetone, and infiltrated with 1:1 Epon:acetone over-night. Embryos were individually mounted in a drop of freshEpon, covered with a coverslip and analyzed and photographedwith a Zeiss Axiophot photomicroscope.

Cloning of POU genes

DNA extraction

Neochildia genomic DNA was isolated from 10-day-starved indi-viduals using a CTAB protocol (Clark 1992). In brief, flatwormswere incubated in 500 µl 2× CTAB containing 5 µl proteinase K,overnight at 45°C. DNA was sequentially extracted and purifiedusing phenol and chlorofom, then ethanol precipitated and re-sus-pended in TE buffer.

RNA extraction

RNA was extracted from whole embryos or the dissected anteriorends of Neochildia worms using TRIZOL reagent (GibcoBRL) oran RNeasy kit (Qiagen) according to the manufacturers instructions.

Generating cDNA

RNA was reverse transcribed into a cDNA using the poly-T primer SH-1 T17 and SuperScript II according to the manufac-turers instructions (GibcoBRL).

Primer design

To design degenerate oligonucleotide primers that would recog-nize the POU family of genes, representatives of published verte-brate and invertebrate POU genes were identified using GenBank.The amino acid translations of these genes were manually alignedin PAUP, highly conserved areas of the gene identified, and thenucleic acid alignments for these conserved regions were used as atemplate with which to design the forward and reverse primersBrn-F1 and Brn-R2 (Table 1).

PCR amplification of genomic DNA

PCR was carried out using 1 µl genomic DNA in a total volume of50 µl containing 1.5 mM MgCl2, 10% DMSO, 0.5 µM each primerand 1 U AmpliTaq Gold DNA polymerase (Applied Biosystems).Gene fragments were amplified in two rounds of PCR. The prima-ry amplification consisted of an initial 12-min incubation requiredto activate the polymerase, followed by 35 cycles of 94°C for 30 s,45°C for 45 s and 72°C for 1 min and a final elongation at 72°Cfor 7 min. Secondary amplifications were carried out using 1 µl ofa 1:20 dilution of the primary amplification product (including thenegative control) employing the cycling parameters describedabove, but increasing the annealing temperature from 45°C to53°C.

Cloning and sequencing of PCR products

Amplified PCR products were cloned using the TOPO TA clon-ing kit (Qiagen) and the cloned plasmid DNAs isolated using aFlexiprep kit (Pharmacia). Both procedures were carried out ac-cording to the manufacturers’ instructions. All plasmid DNAswere manually sequenced using M13F(–20) to identify clonescontaining gene fragments belonging to the POU family. Theseclones were fully sequenced in both directions using M13F(–20)and M13R at the DNA Sequencing Facility at California StateUniversity, Northridge. The resulting sequences were alignedagainst other POU gene family members and the relationship be-tween the Neochildea gene fragments and the different clades ofPOU genes determined using parsimony analysis and bootstrap-ping.

Elongating the gene fragments

The Neochildia Brn-3-specific forward ACDOB3F2 and SH-1primer (Table 1) were used to PCR amplify the 3′ region of thegene from an adult cDNA in a classic 3′ RACE protocol (RapidAmplification of cDNA Ends) using the cycling parameters described above with an annealing temperature of 56°C.

57

Table 1 Oligonucleotide primer sequences and melting tempera-tures

Primer Primer sequence 5′ – 3′ Tmname

Brn-F1 CAA GCA GMG RMG VAT MAA RYT RGG 57.8Brn-R2 RTT RCA RAA CCA SAC BCK MAC MAC 56.0SH-1 T17 CTC ATT CCT GTT AAG CTT ACC T17 62.0SH-1 CTC ATT CCT GTT AAG CTT ACC 58.6ACDOB3F2 CAT AAC AAC ATG GTG GCA C 58.1AC-B1-F1 TTC ACT CAG GCT GAC GTT G 60.2AC-B1-R1 CCT TGA GCC GCA ATT TTA TC 58.4AC-B3-F1 AAG CTG ACG TTG GAA ACT C 58.0AC-B3-R1 TCA ACA CCC GAC AAA CAG 57.6

58

Fig. 1A–F Embryonic development of Neochildia. Fuchsin-labeledwhole-mounts (top of each panel; embryo shown in ventral view;anterior to the left) and schematic drawings of sagittal sections(bottom of panels; anterior left, dorsal top) of Neochildia embryosat six different stages. A Stage 2 (late cleavage; approx. 16–24 h).Outer layer of cells (descendants of micromeres 1–3a/b) form theprimordium of the epidermis and nervous system. Inner cells (de-scendants of micromeres 4a/b) give rise to muscle cells, digestivesyncytium and neoblasts. B Stage 3 (approx. 24–36 h). Embryoforms solid mesenchyme of homogeneously sized cells with mito-ses (arrowheads) spread over all levels. C Stage 4 (approx.36–48 h). Three concentric layers are distinguished by cell sizeand shape. The outer layer forms the epidermal primordium (ep;blue), middle layer the muscle/brain primordium (mbp), and innerlayer the digestive syncytium primordium (dsp). Cells are still undifferentiated and show mitoses in all layers (arrowheads). D Stage 5 (48–72 h); early organogenesis. Epidermal primordium(ep) has differentiated into a cuboidal ciliated epithelium. Bilater-ally symmetric brain primordium (br; violet) forms at the anteriorpole. Muscle primordium (mp; green) forms a thin layer under-neath the epidermis. E Stage 6 (approx. 72–96 h). Brain primordi-um has condensed. Since a basement membrane is absent inacoels, differentiating muscle cells and cell bodies of epidermalcells intermingle in the body wall. F Stage 7/8 (96–120 h). Em-bryo moves by muscle contraction. Body-wall is differentiated in-to an outer, nuclear free cortex and an inner layer of denselypacked epidermal and muscle nuclei (epn, msn). Brain has innerneuropile (np) surrounded by cortex (co). The statocyst (stc) is acharacteristic landmark located in the posterior cortex. Centralcells have merged into the digestive syncytium (ds). The circularopening in the body wall at a ventral-posterior location representsthe simple pharynx (ph). Neoblasts (nb; turquoise) form a popula-tion of large cells clustered posterior to the brain and around thedigestive syncytium. Scale bar 20 µm

Temporal expression of genes

RT-PCR

The specific primers AC-B1-F1, AC-B1-R1, AC-B3-F1 and AC-B3-R1 (Table 1) were designed for the NeocBrn-1 and NeocBrn-3 gene fragments respectively. These primers were usedto amplify a 198-bp fragment of NeocBrn-1 and a 209-bp frag-ment of NeocBrn-3 from cDNAs isolated from 3-day, 4-day, juvenile and adult Neochildia. The PCR protocol was identical tothe primary round of amplification described above but with anannealing temperature of 56°C.

In situ hybridization

Whole-mount in situ hybridization was performed using a modi-fied Drosophila protocol (Tautz and Pfeifle 1989). DIG-labeledantisense and sense RNA probes were generated using linearizedplasmid DNAs for NeocBrn-1 and NeocBrn-3 and a directionallyappropriate RNA polymerase. The embryos and probes were hy-bridized at a final concentration of 1 ng/µl for 100 µl at 55°C in50% formamide, 5× SSC, 0.1% Triton X-100, 0.1 mg/ml heparin,1 mg/ml yeast tRNA and 10 mg/ml salmon sperm DNA. Follow-ing a series of washes, the embryos were treated with RNase andincubated overnight with an alkaline phosphatase-conjugated anti-DIG-antibody at 4°C. The embryos were repeatedly washed andthe color reaction initiated by adding the enzyme substrateBCIP/NBT (Boehringer Mannheim). The color reaction wasstopped by rinsing in PBS and the embryos examined and photo-graphed using a Zeiss Axiophot photomicroscope.

Results

Normal development and staging

In a recent investigation of embryogenesis of the rhabdocoels Mesostoma lingua, Gieysztoria superba andCraspedella pedum, and the polyclad Imogine mcgrathiwe have introduced a system of stages, based upon mor-phological criteria that can be easily distinguished inwhole-mounted material and that represent major devel-opmental steps (Younossi-Hartenstein and Hartenstein2000a, b; Younossi-Hartenstein et al. 2000, 2001). In thefollowing survey of development in Neochildia we haveattempted to employ this system because it is possible torecognize large-scale morphogenetic events that are sim-ilar to those in the above named flatworms.

Embryogenesis of Neochildia takes 4–5 days (18°C).Early cleavage of Neochildia and the fate of individualblastomeres have been described by Bresslau (1909) andmore recently by Henry et al. (2000). Cleavage follows a pattern called duet spiral cleavage where the pair (“duet”) of blastomeres produced by the first, meridionaldivision undergoes a series of four oblique divisions thatgive rise to four duets of micromeres (1–4a, b) and themacromere duet (4A, B; Henry et al. 2000). Our whole-mounts labeled with fuchsin, anti-phosphohistone andanti-acetylated tubulin antibodies allowed us to followlater stages of cleavage, and describe further advancedmorphogenetic events resulting in the different tissues ofthe juvenile worm.

During late cleavage (12–50 cells; second half of day 1; stage 2 in our nomenclature) the descendants of

the first three duets form an outer layer of blastomeresthat surrounds an inner cluster formed by the macromereduet 4A, B (near vegetal pole) and the micromere duet4a, b (Fig. 1A). These inner cells are the only cells thatproduce the mesoderm and endoderm (Henry et al.2000). Similar to the situation noted for other flatworms(reviewed in Hartenstein and Ehlers 2000), we see here amechanism where gastrulation in the classical sense ofseparating germ layers does not take place. Instead, cellsthat act as stem cells for interior tissues, including diges-tive parenchyma, musculature, and nervous system, aredelivered into the interior of the embryo at an early stageby the orientation of the mitotic spindle.

During the first half of the second day the Neochildiaembryo (stage 3) can be characterized as a solid spheri-cal mesenchyme of rather homogeneously sized cells(Fig. 1B). Mitoses can be seen superficially and deep inthe embryo. Towards the end of the second day (stage 4),three tissue layers crystallize. The surface layer consistsof regularly spaced, mid-sized cells that form the epider-mal primordium (Fig. 1C). These cells have not yet dif-ferentiated, as evidenced by the presence of mitoses andabsence of ciliation. Subepidermally, is a 2–3 cell-diam-eter-thick layer of small, densely packed cells that com-prise the progenitors of muscles and brain. A solid cluster of large cells, the progenitors of the digestivesyncytium, fills the core of the embryo.

Organ formation commences during day 3 (stage 5).The epidermal primordium differentiates into a cuboidal,

▲

densely ciliated epithelium (Fig. 1D). Cilium formationcan be followed with an antibody against acetylated tu-bulin (Fig. 2A, B) and begins simultaneously at all posi-tions within the epidermis. Subepidermally a bilaterallysymmetric brain primordium can be seen at the anteriorpole. Neural cells are small and ovoid and amount to

approximately 250 cells on each side. Laterally and pos-teriorly, myoblasts form a thin layer underneath the epi-dermis.

The “merging” of epidermis and muscle layer, as well as the condensation of the brain primordium charac-terizes stage 6 (day 4). Acoels lack a basement mem-

59

brane in between epidermis and musculature; as a resultcell bodies of both tissues are intermingled (Fig. 1E, F).Thus, whereas nuclei of epidermal cells were spaced ex-tremely regularly and lay all in the same plane duringstage 5, epidermal and muscle cell nuclei form an irregu-lar melange at the surface of the embryo from stage 6onward. The brain becomes more compact, and a central,cell-poor neuropile can be distinguished from an externalcortex. Numerous acTub-positive processes of cells located within the epidermal layer project towards the brain primordium (Fig. 2F). These processes are interpreted as sensory axons. Beside the sensory axons,acTub does not label any other cells in the nervoussystem, indicating that no central neurons with long axons differentiate in the brain during the embryonic period.

Cell differentiation is completed during the fifth dayof embryogenesis (stages 7 and 8; Fig. 1F). Characteris-tic morphological criteria are the spherical statocyst thatdevelops in the midline of the brain, the appearance inthe epidermis of an apical web formed by the rootlets ofepidermal cilia, and the appearance of numerous glandu-lar processes penetrating the epidermal layer. Cells of the

core of the embryo fuse into the digestive syncytiumcharacteristic of the acoel clade. Circularly arrangedmuscle fibers that appear ventrally define the pharynx.The size of the brain in juvenile Neochildia is consider-able. Other flatworm taxa possess a ganglion-like brainconsisting of a central neuropile surrounded by a cortexof cell bodies. In adults of Neochildia and other acoelspecies such a brain has been reported absent. Instead,neuronal somata and neurites, together with muscle fi-bers and gland cell processes form a loose reticulatedmass at the anterior pole of the animal (Bedini and Lafranchi 1991; review in Reuter and Gustafsson 1995).However, we find that in late embryos and juveniles ofNeochildia, the brain is more compact. A 3–4 cell-diam-eter-thick layer of neurons forms a cortex surrounding acentral neuropile that is relatively free of cell bodies. Inaddition, an “orthogonal” array of muscle fibers pene-trates the brain (Fig. 3).

An antibody against phosphorylated H3 histone (anti-H3; Hendzel et al. 1997) allowed us to follow cell divi-sion throughout embryogenesis (Fig. 4). Mitoses wereabsent from all organ primordia from stage 5 onward. Inflatworms, differentiated cells have lost the ability to

60

Fig. 2A–F Neochildia embryoslabeled with acTub antibodythat recognizes cilia, gland cellsand axons. A, B, D, F Parts of whole-mounted embryos, C, E show sections counter-stained with toluidine blue/methylene blue that labels nu-clei. A, B Stage 5. Epidermalcells (ep) at outer surface havestarted to form cilia (ci). C, D Stage 6. Cilia have length-ened and increased in density.Necks of gland cells (gl) scat-tered all over the epidermal layer can be recognized. D, F Stage 7/8. AcTub-positivemicrotubules form the so calledapical web (aw) underneath thecilia. Labeled cell processesemanating from cells within theepidermal layer are interpretedas sensory axons (sn). (epn Epi-dermal cell nuclei, mn nuclei ofmesenchymal cells)

divide. A population of totipotent stem cells called neo-blasts adds cells during postembryonic growth. The dis-tribution of neoblasts in several flatworm species was recently described (Ladurner et al. 2000; Newmark andSanchez-Alvarado 2000; Gschwentner et al. 2001). In juvenile and adult specimens, neoblasts typically form a lateral band posterior to the brain. In Neochildia em-bryos, we see a bilaterally symmetric population of largeanti-H3-positive cells located posterior to the brain pri-mordium. We interpret these cells as the embryonic“stem neoblasts”. On average, they number approximate-ly 20–30 per individual, a number that does not increasefrom mid to late stages. This implies that the cells pro-duced by embryonic neoblast division, rather than in-creasing the number of neoblasts, add to the organ pri-mordia.

Cloning of POU domain genes in Neochildia

PCR amplification between the degenerate primers locat-ed near the 5′ end of the POU-specific domain (Brn-F1)and 3′ end of the homeodomain (Brn-R2 primer) resultedin the isolation of a PCR fragment coding for a two POUsequence (Fig. 5). The first of these, named NeocBrn-1,is 333 bp in length (excluding the primer sequences) and includes 51 amino acids of the POU-specific do-main, 16 linker amino acids and 44 amino acids of thehomeodomain. Four independent PCR clones yieldedidentical sequences. This gene fragment is closest in per-cent amino acid identity to Drosophila drifter ventral-veinless (83%), followed by vertebrate Brn-1 genes(80–81%), Girardia tigrina GtPOU-1 (a planarian flat-worm; 72%), and C. elegans ceh-6 (70%). These search-

61

Fig. 3A–D Cross-sections ofthe body wall and brain of Neo-childia juvenile. Sections takenat different antero-posteriorlevels (A most anterior, D mostposterior). A Ciliated epider-mal cells with apical web (aw)and orthogonal system of so-matic muscle fibers (msf). Anterior brain cortex (br) withloose neuropile crossed by pre-dominantly vertical muscle fibers (msf). C Section takenslightly more posteriorly thanB. Brain cortex, neuropilecrossed by horizontal musclefibers. D Posterior region of thebrain. Layered arrangement ofbody wall [with cilia, apicalweb (aw), nuclei of epidermiscells (epn) and muscle cells(mn)] and brain [cortex (co),neuropile (np)]

62

Fig. 4A–D Labeling of Neochildia with anti-phosphohistone H3to visualize mitotic cells (brown). During early stages (A, B) labeled cells can be seen scattered at the surface, as well as deepinside the embryo. From around stage 5/6 onward (C, D), labeling

is confined to a population of large, deeply located cells in theposterior part of the animal, excluding the brain primordium (br)and epidermis (ep). We interpret these cells as the stem neoblasts(nb)

Fig. 5 Published amino acid translations of human, fly, nematodeand flatworm homologs of Brn-3 and Brn-1, and NeocBrn-1aligned against NeocBrn-3. Dots indicate identity, dashes repres-

ent gaps and where relevant, the asterisks denote the end of the 3′translated region of the gene

es combined with parsimony analyses (see Jacobs et al.1998 for analytical approach) provide strong support forthe placement of this gene in a Brain 1/drifter clade,class III of Wegner (1993; Fig. 6).

The second POU sequence generated using the degen-erate PCR is 348 bp in length excluding the primer se-quence and belongs to the Brn-3 clade (class IV POUgenes); thus, we named it NeocBrn-3. The 3′ region ofthis gene was isolated using a gene-specific primer and3′ RACE. The isolated fragment encodes 99 amino acidsof which 40 show overlap with the NeocBrn-3 clone and59 are new. In all, we have a fragment of the gene en-compassing 572 bases, including the sequences encoding

the POU domain, linker, homeodomain and 158 nucleo-tides of the 3′ end non-coding region. In terms of aminoacid identity the NeocBrn-3 gene is closest to human and mouse Brn-3 genes (80–78%), followed by the Drosophila IPOU/acj6 gene (78%), Girardia tigrina(75%), and C. elegans, unc-86 (73%). These searchescombined with parsimony analyses (e.g. Jacobs et al.1998) provide strong support for the placement of Neo-cBrn-3 in a Brain 3/IPOU clade, class IV of Wegner(1993; Fig. 6).

Temporal expression of POU genes

RT-PCR analysis indicates that NeocBrn-1 and NeocBrn-3are expressed at a low level in embryos and juveniles,and substantially increase in adults (Fig. 7). In situ hy-bridization of early to mid stage embryos with NeocBrn-1probes gives negative results, although RT-PCR showsexpression to be present (Fig. 7). However, in a fractionof embryos 3 days or older we detected signal in a groupof large internal cells which by size and position corre-spond to neoblast progenitors (Fig. 8A). In late embryosand adults, NeocBrn-1 is expressed at low levels in thebrain (Fig. 8B, C) and visualizes an interesting pattern ofepidermal glandular cells, which, along the anterior-posterior axis, roughly coincide with the extension of the brain (Fig. 8D–F). Given that in vertebrates and Drosophila, several POU genes may play a specific rolein the development of ciliated sensory cells (Artinger et al. 1998), expression in these specialized epidermalcells is particularly interesting. NeocBrn-3 labels a small and discrete population of neurons in the brain.The cell group forms a transversally oriented “slice” of densely packed cells (Fig. 8G, H). No expressioncould be detected by in situ hybridization at embryonicstages.

63

Fig. 6 Relationship of the Neochildia sequences with the Brain 1and Brain 3 clades of POU/homeodomain genes. Taxon and genenames are followed by GenBank accession numbers. Sequenceswere selected for the analysis on the basis of completeness andtaxon distribution. Sequence from POU domain, homeodomain,intervening linker and conserved 5′ and 3′ sequence were all in-cluded in the analysis. In many cases this involves the completetranslated product of the gene. The tree was reconstructed using aparsimony analysis implemented in PAUP. In the analysis 432 bas-es of DNA with third positions removed were combined with 260amino acids (see Jacobs et al. 1998; Schubert et al. 2000 for meth-odology). Bootstrap support is shown adjacent to each branch.Note that the major groups of genes are well supported, as aresome of the topologies within the vertebrates. Relationships be-tween genes in different phyla within the gene classes are less wellsupported. However due to the strong support for the major cladesof POU domains including Brain 1 and Brain 3 we can confidentlyplace one Neochildia gene in the Brain 1 group and the other inBrain 3

Fig. 7 An agarose gel of NeocBrn-1 and NeocBrn-3 RT-PCRproducts representing the temporal expression of these genes inNeochildia [3 3-day-old embryos (stage 5), 4 4-day-old embryos(stage 6), J juveniles, A adult, L ladder]

Discussion

Morphological aspects of acoel development

Acoels have attracted the attention of phylogeneticistsfor a long time. Although they share many characteris-tics with other flatworm taxa, they differ in several im-portant ways. For example, they lack both a cellular di-

gestive cavity, and a protonephridial system, and thebody wall is atypical in that a basement membrane is ab-sent. Further, axons of the trunk are not organized in aregular orthogonal system of connectives and commissu-res, but rather form a variable, nerve net-like reticulumof fibers (reviewed in Smith and Tyler 1985; Reuter andGustafsson 1995). The latter has been attributed to thelack of a basement membrane, which in other animals is

64

Fig. 8 Expression pattern ofNeocBrn-1 and Brn-3 in em-bryos (A, B) and adult (C–H)tissues. A, C, G Whole-mountsin ventral view; B, D–H sec-tions. A Stage-5 embryo. Ex-pression of NeocBrn-1 appearsat a low level, and only in asubset of specimens analyzed,in large, deeply situated cellsthat might represent stem neo-blasts. B, C In late embryo andadult, NeocBrn-1 is visible insubset of neurons in the brain(br). The pharynx, located atmid-ventral level, is marked ph in C. D–F Besides the ex-pression in subset of nervecells, Brn-1 also labels a ventral subpopulation of cellsin the epidermal layer whichwe interpret as glandular cells(gl). G, H Expression patternof NeocBrn-3 in a subset ofbrain neurons (br) of adultNeochildia

instrumental in creating a smooth cleft between epider-mis and muscle layer in which axons can form straightfascicles. These properties have either been viewed as primitive characters, inherited from an ancestral cili-ate protozoan (Hadzi 1963; Hanson 1977) or cnidarian(Hyman 1951; von Salvini-Plawen 1978), or as second-ary “regressive” changes (Ax 1963; Remane 1963; Rieger1985; Smith and Tyler 1985). Based on these differ-ences, there have been repeated suggestions that theacoels be taken out of the platyhelminth phylum andplaced in a separate group representing taxa that splitfrom the bilaterian line earlier than the “true” platyhel-minths. Recent evidence, both structural and molecular,supports this view. The cell lineage analysis by Henry et al. (2000) shows that the pattern of cleavages and celllineage in the acoel Neochildia differs significantly fromthose in other flatworms. Perhaps most notable is the pe-culiar type of spiral cleavage, called duet cleavage,found in the acoel, and the developmental origin of themesendoderm. In duet spiral cleavage, the pair of equalsized blastomeres generated by the first cleavage (mac-romere duet 1A, B) sequentially buds off three pairs ofmicromeres (1a, b–3a, b). The 3A, B macromere duetthen sinks into the mass of surrounding micromeres anddivides one more time into 4A, B and the 4a, b micro-mere quartet. These four cells give rise to the entire en-doderm (i.e. the digestive syncytium) and mesoderm (i.e.musculature, neoblasts); the progeny of the 1a, b–3a, bmicromeres produce epidermis and nervous system. Thislineage relationship between tissues and blastomeres issignificantly different from that found in other flatwormsand spiralians in general. In a typical spiralian, meso-derm is derived from two different sources, the 4d mi-cromere (mesoendoderm) and the micromeres (variablythe first, second or third quartets; for review see Henry et al. 2000).

Apart from the different origin of mesoderm inacoels, the time and mechanism by which the dorso-ven-tral axis is established in these animals may differ funda-mentally from what is known in other spiralians. In bothacoels and typical spiralians alike, the first asymmetrythat becomes manifest is the animal-vegetal axis. Thisaxis appears at the time point when the first micromeresare formed. Micromeres take up a position at the animalpole, macromeres at the vegetal pole. This animal-vege-tal axis, in a somewhat oversimplified manner of speak-ing, is translated into the antero-posterior axis of the la-ter embryo (Boyer et al. 1998; Henry et al. 2000). Thus,descendants of the micromeres at the animal pole formthe epidermis, nerve cells and sensory receptors of thehead. Micromere descendants located more vegetally(e.g. descendants of the third quartet/duet) give rise toposterior epidermis. Superimposed upon this a-p axis isthe dorso-ventral axis that, in typical spiralians, is firstmanifested in the asymmetry among the fourth micro-mere quartet (Henry et al. 2000). Thus, the mesodermaldescendants of the 4d micromere move to one side of theearly embryo that is thereby defined as the ventral side.Further dorso-ventral asymmetries soon follow. For ex-

ample, in polyclad embryos, the primordium of the brainand epidermal lobes surrounding the mouth opening de-velop at, or shift towards, the ventral side.

In the acoel Neochildia, the overt “ventralization”originating in the mesoderm primordium does not occur.No asymmetries are visible among the progeny of the 4a,b micromere or the 4A, B macromere duet. Likewise, inthe developing epidermis or brain primordium, it is im-possible to make out any dorso-ventral asymmetries onthe basis of morphological criteria. These asymmetriesonly become apparent at late stages when a mouth open-ing, surrounded by circular muscle fibers, forms at the ventral side of the embryo. It is likely that in acoels,latent dorso-ventral asymmetries exist from early stagesonward, given that there is a fixed relationship betweenmicromere identity and dorso-ventral position of their(epidermal) progeny (Henry et al. 2000). However, themechanism by which the dorso-ventral axis is specified,and in particular, the role of the mesodermal lineage dur-ing this process, may be quite different when comparingacoels with typical spiralians.

Acoels like other flatworm taxa, have neoblasts, aspecial cell type that serves as a totipotent stem cell pop-ulation. In several flatworm species, it has been shownthat once differentiated, cells of the epidermis, muscula-ture, nervous system, or other organs, no longer divide(Baguna 1981; for review of classical literature on neo-blasts, see Ehlers 1985). This in itself is similar to thesituation in other systems where many differentiated celltypes also fail to divide. However, in these other sys-tems, “dedicated” (i.e. tissue-specific) stem cell popula-tions exist that add new cells or replace sequesteredones. Examples are the stratum germinativum of the ver-tebrate epidermis, or the stem cells in the intestinal mu-cosa. Such tissue-specific stem cells are not found inflatworms. Instead, proliferating, motile neoblasts per-vade the entire body and are able to differentiate into anycell type needed. Little is known about the embryonic or-igin and development of neoblasts. Using BrdU or an an-tibody against phosphorylated histone, both of which label proliferating cell populations, neoblasts were visu-alized in several recent papers for postembryonic stagesof planarians (Newmark and Sanchez-Alvarado 2000),Macrostomum sp. (Ladurner et al. 2000), and the acoelConvolutriloba (Gschwentner et al. 2001). In these ani-mals they form a bilaterally symmetric band of quitedensely packed cells flanking the digestive system. Thehead region is devoid of neoblasts.

Using anti-phospho histone, we have visualized inthis study a population of large proliferating cells in midto late stage embryos of Neochildia that we interpret asthe embryonic neoblasts. Their distribution in roughlytwo lateral bands in the posterior half of the embryo isalso reminiscent of the findings in other species citedabove. The fact that the embryonic neoblasts dividecould be interpreted in two different ways. First, it ispossible that neoblasts add to their own number. Second-ly, progeny of embryonic neoblasts might differentiateand contribute to developing organs. According to this

65

view, neoblasts would start to add to and/or replace cellpopulations of various organs during embryogenesis. Inour study, the number of neoblasts labeled at a givenstage did not increase over time making the second pos-sibility more likely. As such, cells found in the variousorgans of hatching flatworms might have two quite dif-ferent origins. Either they are formed in the “traditional”way, that is, they derive from blastomeres that divide in amore or less fixed pattern; alternatively, they are pro-duced in an “ad hoc” manner by totipotent neoblasts thatform a motile cell population, circulating and probingwhere in the embryo additional cells are needed, andthen differentiating into that cell type.

The lack of acTub-positive cells in the central ner-vous system of Neochildia embryos is in contrast to find-ings in other flatworm species where acTub labels subsets of brain neurons and their axons (Hartenstein and Ehlers 2000; Younossi-Hartenstein and Hartenstein2000b; Younossi-Hartenstein et al. 2000, 2001). How-ever, the lack of acTub labeling of CNS structures inNeochildia is not the first instance in which acTub doesnot recognize any epitope in the CNS. We have had thesame experience with various representatives of mol-luscs and annelids (V. Hartenstein, unpublished data).The lack of staining could be an artifact due to insuffi-cient penetration of the antibody. Alternately, acetylationof tubulin in axons does not take place in the embryonicperiod of species, such as Neochildia, where acTub picksup no central neurons. One should note that in all specieswe investigated, only a subset of axons which were rec-ognizable by histology labeled with acTub. In some sys-tems, such as Drosophila melanogaster, only axons thathave reached a certain length and/or diameter are visual-ized with acTub (V. Hartenstein, unpublished).

POU genes in neural development

POU domain genes encode a large family of transcrip-tion factors that have been identified in a basal metazoantaxa such as sponges, cnidarians and ctenophores (D.K.Jacobs, unpublished) as well as in bilaterians, includingnematodes, arthropods, echinoderms and vertebrates.The expression of many POU domain proteins in thenervous system across evolutionary boundaries suggestsa fundamental requirement for these genes during neuro-nal development. In particular, the class III and IV POUproteins are candidate regulators of CNS-specific genes(Josephson et al. 1998). Brn-1 is widely expressed in theembryonic neuroepithelium in rat. In the cortex, Brn-1 isstrongly expressed in the ventricular zone and repressedwhen neural progenitor cells first differentiate into neurons (Alvarez-Bolado et al. 1995). In zebrafish, Brn-1 appears in the neural plate where the brain forms(Hauptmann and Gerster 2000).

The Drosophila Brn-1 homolog, drifter/ventral vein-less, is expressed in several subsets of glial and nervecells, notably the dopaminergic and serotonergic neu-rons, where it is believed to directly upregulate dopa-

decarboxylase, an enzyme in the metabolic pathway ofserotonin and dopamine (Johnson and Hirsh 1990; Yeoet al. 1995). Beside the CNS, drifter/ventral veinless ap-

pears in numerous non-neural tissues, including the tra-cheal system, the epidermis, hind- and foregut, and ringgland (Anderson et al. 1995; Billin and Poole 1995; DeCelis et al. 1995) and, using a genetic approach, a rolefor drifter/ventralveinless in tracheal and epidermal pat-terning has been established. In the Crustacean, Artemiafranciscana a homolog of drifter/ventralveinless is ex-pressed in the larval salt gland, an organ which is in-volved in osmoregulation and disappears in the adult(Chavez et al. 1999). In the adults of the triclad flat-worm, Girardia tigrina, the Brn-1 homolog, GtPOU-1 isexpressed in neurons of the central and peripheral ner-vous system (Munoz-Marmol et al. 1998).

In the present study, the Neochildia Brn-1 homologNeocBrn-1 is expressed in a ventral population of pre-sumed gland cells, as well as a small subset of brain neu-rons. The expression during embryonic stages is alsoconfirmed by isolation of RNA and amplification of Brn-1 gene by RT-PCR from the 3- and 4-day embryos.The interpretation of the spatial expression pattern bywhole-mount in situ techniques is ambiguous. We wereunable to find a signal in most embryos. In a subset ofembryos from day 3 onward we found the gene ex-pressed in a subset of large, internal cells which mostlikely represent neoblast progenitors, given their largesize and internal position. This would point at the possi-bility that the Brn-1-positive neurons we see in the adultbrain are derived from neoblasts in the late embryo. Alternatively, neoblasts could only transiently expressthe gene in the embryo, and neurons turn it on postem-bryonically.

The POU-IV class proteins have critical functions inthe vertebrate CNS and sense organs. The original Brn-3factor was isolated as a novel POU protein by He et al.(1989) by using the degenerate RT/PCR approach. Sub-sequently, three distinct Brn-3 factors were discoveredwhich are encoded by different genes (Theil et al. 1994).These are known as Brn-3a (Brn-3.0), Brn-3b (Brn3.2)and Brn-3c (Brn-3; for review, see Latchmann 1999).Brn-3a is expressed in multiple distinct nuclei of themouse brainstem and spinal cord, as well as sensory neu-rons of the cranial and dorsal root ganglia. Knockoutmice show widespread loss of sensory and motor neurons and die shortly after birth (Xiang et al. 1996).Brn-3b is found in the brainstem and at high levels inretinal ganglion cells; mutant mice lacking Brn-3b areviable but show specific loss of retinal neurons leadingto blindness (Gan et al. 1996, 1999). Brn-3c shows ex-pression in and is absolutely required by the hair cells ofthe inner ear (Xiang et al. 1997a, b). Similarly, POU4F3causes X-linked mixed deafness in humans with the deletion of eight bases at the POU homeodomain in anIsraeli family (Vahava et al. 1998). In spite of these di-verse roles of Brn-3 genes in neural development a pre-ponderance of function appears to relate to the develop-ment of the sensory cells involved in vision, hearing and

66

olfaction, all of which are placoidally derived (Artingeret al. 1998). The Drosophila Brn-3- homolog, IPOU/acj6, encodes a 367-amino-acid protein that acts as atranscriptional repressor in the central and peripheralnervous system from late embryonic to adult stages (Certel et al. 2000). Acj6 may play an important role inregulating synaptic target selection by central neurons. Itis also noteworthy that, as in the vertebrate Brn-3, thereis an association with sensory innervation including ex-pression in olfactory, optic and antennal neurons. C. ele-gans unc-86 POU protein is required for the commitmentof several sensory neuroblast lineages as well as for thespecification and maintenance of particular neural phe-notypes. Mutants of Unc-86 have shown defects in che-motaxis, mechanosensation and egg laying (Finney andRuvkin 1990).

In Neochildia adult brain, NeocBrn-3 is expressed ina fairly small cluster of seemingly contiguous cells of thebrain. Given the paucity of specific markers for subsetsof neurons in flatworm brains, nothing can be stated con-cerning the identity of the NeocBrn-3-positive cells. Inembryos, although NeocBrn-3 RNA was detected from3 days onwards, a signal could not be detected by in situhybridization. We were therefore unable to monitor theorigin of the Brn-3-positive cells. The relatively late on-set of NeocBrn-1 and 3 parallels the expression of homo-logs of these genes in Drosophila. The expression pat-tern supports the hypothesis that POU genes act as tran-scription factors for neuron type-specific proteins, suchas enzymes involved in transmitter synthesis (as for Ddc;Johnson and Hirsh 1990), rather than as selectors forneural lineages.

Although the sequence data recovered for NeocBrn-3and NeocBrn-1 were sufficiently resolved to classifythese genes relative to the nomenclature of POU do-mains, they were not sufficient to resolve the placementof the Acoela relative to other bilaterian taxa. Thus theplacement of the Acoela remains controversial (e.g. Littlewood et al. 2001). It should also be noted that Nemertodermatida, which are thought to be close rela-tives of the Acoela, and are often grouped with them inthe basal clade Acoelomorpha (Littlewood et al. 1999),also lack a true ganglionic brain or neuropile (Raikova etal. 2000). However despite this similar simplicity, thereappear a number of differences in organization of theacoel and nemertodermatid nervous systems. Thus, ad-vances in understanding of the topology of the Metazoantree, as well as numerous detailed studies of neural de-velopment in a variety of metazoans, will be required if adetailed understanding of the evolution of the bilateriannervous system is to unfold.

Conclusion

Recent studies stress the high degree of conservation ofgenetic mechanisms controlling early neural develop-ment in bilaterian animals. Evidence exists that funda-mental aspects of the topology in which precursor cells

of visual organs, neuroendocrine centers and variousbrain structures are laid out in the fate map of the headare conserved in phyla separated as far as chordates andarthropods (Hartenstein and Reh, in press). This indi-cates that the bilaterian ancestor might have possessed ahead in which photoreceptors, various brain structuresand neuroendocrine cells were arranged in a manner thatmay have been similar to the one found in present daytaxa. To further address this hypothesis comparativeanalyses of invertebrate neural development, using mo-

lecular markers for defined cell types, are indispensable.In this paper we have initiated the developmental analy-sis of the nervous system in the acoel Neochildiaand provided two molecular markers, NeocBrn-1 andNeocBrn-3. We plan to expand our effort to other con-served genes expressed in the developing nervoussystem of flatworms, and thereby hope to contribute tothe elucidation of nervous systems in simple inverte-brates.

Acknowledgements We would like to acknowledge VyacheslavPalchevskiy for technical assistance, Amelia Younossi-Hartensteinfor expert help and advice, and support from NSF grant IBN-01110718 to V.H., and NASA Exobiology (NAG5–7207) andAstrobiology to D.J.

References

Adoutte A, Balavoine G, Lartillot N, de Rosa R (1999) Animalevolution: the end of the intermediate taxa? Trends Genet15:104–108

Agata K, Soejima Y, Kato K, Kobayashi C, Umesono Y, WatanabeK (1998) Structure of the planarian central nervous system(CNS) revealed by neuronal cell markers. Zool Sci 15:433–440

Alvarez-Bolado G, Rosenfeld MG, Swanson LW (1995) Model offorebrain regionalization based on spatiotemporal patterns ofPOU-III homeobox gene expression, birthdates, and morpho-logical features. J Comp Neurol 355:237–295

Anderson MG, Perkins GL, Chittick P, Shrigley RJ, Johnson WA(1995) drifter, a Drosophila POU-domain transcription factor,is required for correct differentiation and migration of trachealcells and midline glia. Genes Dev 9:123–137

Artinger KB, Fedtsova N, Rhee JM, Bronner-Fraser M, Turner E(1998) Placodal origin of Brn-3-expressing cranial sensoryneurons. J Neurobiol 36:572–585

Ashburner M (1989) Drosophila. A laboratory manual. ColdSpring Harbor Laboratory Press, Cold Spring Harbor, pp 214–217

Ax P (1963) Relationships and phylogeny of the turbellaria. In:Dougherty EC (ed) The lower Metazoa. University of Califor-nia Press, Berkeley, pp 191–224

Ax P (1996) Multicellular animals, vol I. Fischer, StuttgartBaguna J (1981) Planarian neoblasts. Nature 290:14–15Bedini F, Lanfranchi A (1991) The central and peripheral nervous

system of Acoela (Plathelminthes). An electron microscopicstudy. Acta Zool (Stockh) 72:101–106

Berney C, Pawlowski J, Zaninetti L (2000) Elongation factor 1-alpha sequences do not support an early divergence of theAcoela. Mol Biol Evol 17:1032–1039

Billin AN, Poole SJ (1995) Expression domains of the Cf1a POUdomain protein during Drosophila development. Roux’s ArchDev Biol 204:502–508

Boyer BC, Henry JJ, Martindale MQ (1998) The cell lineage of apolyclad turbellarian embryo reveals close similarity to coelo-mate spiralians. Dev Biol 204:111–123

67

Bresslau E (1909) Die Entwicklung der Acoelen. Verh Dtsch ZoolGes 48:414–324

Certel SJ, Clyne PJ, Carlson JR, Johnson WA (2000) Regulationof central neuron synaptic targeting by the Drosophila POUprotein, Acj6. Development. 127:2395–2405

Chavez M, Landry C, Loret S, Muller M, Figueroa J, Peers B, Rentier-Delrue F, Rousseau GG, Krauskopf M, Martial JA(1999) APH-1, a POU homeobox gene expressed in the salt glandof the crustacean Artemia franciscana. Mech Dev 87:207–312

Clark CG (1992) In: Lee JJ, Soldo A (eds) Protocols in protozool-ogy. Allen Press, Lawrence, Kan. pp D3.1-D3.2

DeCelis JF, Llimargas M, Casanova J (1995) Ventral veinless, thegene encoding the Cf1a transcription factor, links positionalinformation and cell differentiation during embryonic andimaginal development in Drosophila melanogaster. Develop-ment 121:3405–3416

Ehlers U (1985) Das phylogenetische System der Plathelminthes.Fischer, Stuttgart

Finney M, Ruvkin G (1990) The unc-86 gene product couples celllineage and cell identity in Caenorhabditis elegans. Cell63:895–906

Gan L, Xiang M, Zhou L, Wagner DS, Klein WH, Nathans J(1996) POU domain factor Brn-3b is required for the develop-ment of a large set of retinal ganglion cells. Proc Natl AcadSci USA 93:3920–3925

Gan L, Wang SW, Huang Z, Klein WH (1999) POU domain factorBrn-3b is essential for retinal ganglion cell differentiation andsurvival but not for initial cell fate specification. Dev Biol210:469–480

Giribet G, Distel DL, Polz M, Sterrer W, Wheeler WC (2000)Triploblastic relationships with emphasis on the acoelomatesand the position of Gnathostomulida, Cycliophora, Plathel-minthes, and Chaetognatha: a combined approach of 18SrDNA sequences and morphology. Syst Biol 49:539–562

Gschwentner R, Ladurner P, Nimeth K, Rieger R (2001) Stemcells in a basal bilaterian: S-phase and mitotic cells in Con-volutriloba longifissura (Acoela, Platyhelminthes). Cell TissueRes 304:401–408

Hadzi J (1963) The evolution of the Metazoa. Macmillan, NewYork

Hanson ED (1977) The origin and early evolution of animals.Wesleyan University Press, Middletown, Conn.

Hartenstein V, Ehlers U (2000) The embryonic development of therhabdocoel flatworm Mesostoma lingua. Dev Genes Evol210:399–415

Hartenstein V, Reh T (in press) Homologies between vertebrateand invertebrate eyes. In: Moses K (ed) Drosophila eye devel-opment. Springer, Berlin Heidelberg New York

Hauptmann G, Gerster T (2000) Combinatorial expression of Zebrafish Brn-1- and Brn-2-related POU genes in the embry-onic brain, pronephric primordium, and pharyngeal arches.Dev Dyn 218:345–358

He X, Treacy MN, Simmons DM, Ingraham HA, Swanson LW,Rosenfeld MG (1989) Expression of a large family of POU-domain regulatory genes in mammalian brain develop-ment. Nature 340:35–41

Hendzel MJ, Wei Y, Mancini MA, VanHooser A, Ranalli T, Brinkley BR, Bazett-Jones DP, Allis CD (1997) Mitosis-spe-cific phosphorylation of histone H3 initiates primarily withinpericentromeric heterochromatin during G2 and spreads in anordered fashion coincident with mitotic chromosome conden-sation. Chromosoma 106:348–360

Henry JQ, Martindale MQ, Boyer BC (2000) The unique develop-mental program of the acoel flatworm, Neochildia fusca. DevBiol 220:285–295

Herr W, Sturm RA, Clerc RG, Corcoran LM, Baltimore D, SharpPA, Ingraham HA, Rosenfeld MG, Finney M (1988) The POUdomain, a large conserved region in the mammalian Pit-1 Oct-1 Oct-2 and Caenorhabditis elegans unc-86 gene prod-ucts. Genes Dev 12A:1513–1516

Hyman LH (1940) The invertebrates: the Protozoa through Cteno-phora. McGraw Hill, New York

Hyman LH (1951) The invertebrates: Platyhelminthes and Rhyn-chocoela. McGraw Hill, New York

Jacobs DK, Lee SE, Dawson MN, Staton JL, Raskoff KA (1998)The history of development through the evolution of mole-cules: gene trees, hearts, eyes, and dorsoventral inversion. In:DeSalle R, Schierwater B (eds) Molecular approaches to ecol-ogy and evolution. Birkhauser, Basel, pp 323–357

Johnson WA, Hirsh J (1990) Binding of a Drosophila POU-domain protein to a sequence element regulating gene expres-sion in specific dopaminergic neurons. Nature 343:467–470

Josephson R, Muller T, Pickel J, Okabe S, Reynolds K, Turner PA,Zimmer A, McKay RD (1998) POU transcription factors con-trol expression of CNS stem cell-specific genes. Development125:3087–3100

Kotikova EA (1986) Comparative characterization of the nervoussystem of the Turbellaria. Hydrobiologia 132:89–92

Ladurner P, Rieger R, Baguna J (2000) Spatial distribution anddifferentiation potential of stem cells in hatchlings and adultsin the marine platyhelminth Macrostomum sp.: a bromodeoxy-uridine analysis. Dev Biol 226:231–224

Latchman DS (1999) POU family transcription factors in the ner-vous system. J Cell Physiol 179:126–133

Littlewood DTJ, Rohde K, Clough KA (1999) The interrelationshipsof all major groups of Platyhelminthes: phylogenetic evidencefrom morphology and molecules. Biol J Linn Soc 66:75–114

Littlewood DTJ, Olson PD, Telford MJ, Herniou EA, Riutort M(2001) Elongation factor 1-alpha sequences alone do not assistin resolving the position of the Acoela within the Metazoa.Mol Biol Evol 18:437–442

Munoz-Marmol AM, Casali A, Miralles A, Bueno D, BayascasJR, Romero R, Salo E (1998) Characterization of platyhel-minth POU domain genes: ubiquitous and specific anteriornerve cell expression of different epitopes of GtPOU-1. MechDev 76:127–140

Newmark PA, Sanchez-Alvarado A (2000) Bromodeoxyuridinespecifically labels the regenerative stem cells of planarians.Dev Biol 220:142–153

Raikova OI, Reuter M, Jondelius U, Gustafsson MKS (2000) Thebrain of the Nemertodermatida (Platyhelminthes) as revealedby anti-5HT and anti-FMRFamide immunostainings. TissueCell 32:358–365

Remane A (1963) The enterocoelic origin of the coelom. In:Dougherty EC (ed)The lower Metazoa. University of Califor-nia Press, Berkeley

Reuter M, Gustafsson MKS (1995) The flatworm nervous system:pattern and phylogeny. In: Breidbach O, Kutsch W (eds) Thenervous system of invertebrates. Birkhäuser, Basel

Reuter M, Raikova OI, Gustafsson MKS (1998) An endocrinebrain? The pattern of FMRF-amide immunoreactivity in Acoela (Plathelminthes). Tissue Cell 30:57–63

Rieger R (1985) The phylogenetic status of the acoelomate organi-zation within the bilateria. In: Morris SC, George JD, GibsonR, Platt HM (eds) The origins and relationships of lower in-vertebrates. (The Systematics Association special vol 28)Clarendon Press, Oxford, pp 101–122

Ruiz-Trillo I, Riutort M, Littlewood DTJ, Herniou EA, Baguna J(1999) Acoel flatworms: earliest extant bilaterian metazoans,not members of Platyhelminthes. Science 283:1919–1923

Ryan AK, Rosenfeld MG (1997) POU domain family values: flex-ibility, partnerships, and developmental codes. Genes Dev11:1207–1225

Salvini-Plawen L von (1978) On the origin and evolution of thelower Metazoa. Z Zool Syst Evol Forsch 16:40–88

Schubert M, Holland LZ, Holland ND, Jacobs DK (2000) A phy-logenetic tree of the Wnt genes based on all available full-length sequences, including five from the cephalochordateAmphioxus. Mol Biol Evol 17:1896–1903

Smith JPS, Tyler S (1985) The acoel turbellarians: kingpins ofmetazoan evolution or a specialized offshoot? In: Morris SC,George JD, Gibson R, Platt HM (eds) The origins and relation-ships of lower invertebrates. (The Systematics Associationspecial vol 28) Clarendon Press, Oxford, pp 123–145

68

Tautz D, Pfeifle C (1989) A non-radioactive in situ hybridizationmethod for the localization of specific RNA in Drosophila em-bryos reveals translational control of the segmentation genehunchback. Chromosoma 98:81–85

Theil T, Zechner U, Klett C, Adolph S, Möröy T (1994) Chromo-somal localization and sequences of the murine Brn-3 familyof developmental control genes. Cytogenet Cell Genet 66:267–271

Vahava O, Morell R, Lynch ED, Weiss S, Kagan ME, Ahituv N,Morrow JE, Lee MK, Skvorak AB, Morton CC, Blumenfeld A,Frydman M, Friedman TB, King MC, Avraham KB (1998) Mu-tation in transcription factor POU4F3 associated with inheritedprogressive hearing loss in humans. Science 279:1950–1954

Wegner M, Drolet DW, Rosenfeld MG (1993) POU-domain proteins: structure and function of developmental regulators.Curr Opin Cell Biol 5:488–498

Xiang M, Gan L, Zhou L, Klein WH, Nathans J (1996) Targeteddeletion of the mouse POU domain gene Brn-3a causes selec-tive loss of neurons in the brainstem and trigeminal ganglion,uncoordinated limb movement, and impaired suckling. ProcNatl Acad Sci USA 93:1950–1955

Xiang M, Gan L, Li D, Chen ZY, Zhou L, O’Malley BW Jr, KleinW, Nathans J (1997a) Essential role of POU-domain factorBrn-3c in auditory and vestibular hair cell development. ProcNatl Acad Sci USA 94:9445–9450

Xiang M, Gan L, Li D, Zhou L, Chen ZY, Wagner D, O’MalleyBW Jr, Klein W, Nathans J (1997b) Role of the Brn-3 familyof POU-domain genes in the development of the auditory/vestibular, somatosensory, and visual systems. Cold SpringHarb Symp 62:325–336

Yeo SL, Lloyd A, Kozak K, Dinh A, Dick T, Yang X, Sakonju S,Chia W (1995) On the functional overlap between two Drosophila POU homeo domain genes and the cell fate specification of a CNS neural precursor. Genes Dev 9:1223–1236

Younossi-Hartenstein A, Hartenstein V (2000a) Comparative approach to developmental analysis: the case of the dalyel-lid flatworm, Gieysztoria superba. Int J Dev Biol 44:499–506

Younossi-Hartenstein A, Hartenstein V (2000b) The embryonicdevelopment of the polyclad flatworm Imgogine mcgrathi.Dev Genes Evol 210:383–398

Younossi-Hartenstein A, Ehlers U, Hartenstein V (2000) Embry-onic development of the nervous system of the rhabdocoelflatworm Mesostoma lingua (Abildgaard, 1789). J Comp Neu-rol 416:461–476

Younossi-Hartenstein A, Jones M, Hartenstein V (2001) The embryonic development of the nervous system of the tem-nocephalid flatworm Craspedella pedum. J Comp Neurol434:56–68

69