Ordered progression of nematogenesis from stem cells through differentiation stages in the tentacle bulb of Clytia hemisphaerica (Hydrozoa, Cnidaria).

Abstract

Nematogenesis, the production of stinging cells (nematocytes) in Cnidaria, can be considered as a model neurogenic process. Most molecular data concern the freshwater polyp Hydra, in which nematocyte production is scattered throughout the body column ectoderm, the mature cells then migrating to the tentacles. We have characterized tentacular nematogenesis in the Clytia hemisphaerica hydromedusa and found it to be confined to the ectoderm of the tentacle bulb, a specialized swelling at the tentacle base. Analysis by a variety of light and electron microscope techniques revealed that while cellular aspects of nematogenesis are similar to Hydra, the spatio-temporal characteristics are markedly more ordered. The tentacle bulb nematogenic ectoderm (TBE) was found to be polarized, with a clear progression of successive nematoblast stages from a proximal zone (comprising a majority of undifferentiated cells) to the distal end where the tentacle starts. Pulse-chase labelling experiments demonstrated a continuous displacement of differentiating nematoblasts towards the tentacle tip, and that nematogenesis proceeds more rapidly in Clytia than in Hydra. Compact expression domains of orthologues of known nematogenesis-associated genes (Piwi, dickkopf-3, minicollagens and NOWA) were correspondingly staggered along the TBE. These distinct characteristics make the Clytia TBE a promising experimental system for understanding the mechanisms regulating nematogenesis.

Full text

Page 1

Ordered progression of nematogenesis from stem cells through differentiation
stages in the tentacle bulb of Clytia hemisphaerica (Hydrozoa, Cnidaria)
Elsa Denker⁎, Michaël Manuel, Lucas Leclère, Hervé Le Guyader, Nicolas Rabet
Université Pierre et Marie Curie-Paris 6, UMR 7138 CNRS UPMC MNHN IRD, Case 05, 7 quai St. Bernard, 75005 Paris, France
Received for publication 9 November 2007; revised 11 December 2007; accepted 11 December 2007
Available online 27 December 2007
Abstract
Nematogenesis, the production of stinging cells (nematocytes) in Cnidaria, can be considered as a model neurogenic process. Most molecular
data concern the freshwater polyp Hydra, in which nematocyte production is scattered throughout the body column ectoderm, the mature cells
then migrating to the tentacles. We have characterized tentacular nematogenesis in the Clytia hemisphaerica hydromedusa and found it to be
confined to the ectoderm of the tentacle bulb, a specialized swelling at the tentacle base. Analysis by a variety of light and electron microscope
techniques revealed that while cellular aspects of nematogenesis are similar to Hydra, the spatio-temporal characteristics are markedly more
ordered. The tentacle bulb nematogenic ectoderm (TBE) was found to be polarized, with a clear progression of successive nematoblast stages from
a proximal zone (comprising a majority of undifferentiated cells) to the distal end where the tentacle starts. Pulse-chase labelling experiments
demonstrated a continuous displacement of differentiating nematoblasts towards the tentacle tip, and that nematogenesis proceeds more rapidly in
Clytia than in Hydra. Compact expression domains of orthologues of known nematogenesis-associated genes (Piwi, dickkopf-3, minicollagens
and NOWA) were correspondingly staggered along the TBE. These distinct characteristics make the Clytia TBE a promising experimental system
for understanding the mechanisms regulating nematogenesis.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Cnidaria; Clytia hemisphaerica; Nematocyte; Nematogenesis; Medusa; Tentacle bulb; Minicollagens; Piwi; NOWA; Dickkopf-3
Introduction
The Cnidaria (jellyfishes, corals, anemones, hydras, etc.) is a
highly diversified and successful animal phylum, exhibiting a
large variety of habitats, morphologies and life cycles (Brusca
and Brusca, 2003; Collins et al., 2006); however, all cnidarians
share a common highly specialized and unique cell type, the
nematocyte. These single use stinging cells fulfill a predatory,
defensive and sometimes locomotive role (Kass-Simon and
Scappaticci, 2002). They occur at a particularly high concen-
tration on the tentacles and mouth. Nematocyte discharge
results from stimulation of the cnidocil ciliary cone apparatus by
external mechanical and chemosensory stimuli, in some cases
facilitated by signals coming from neighbouring cells through
afferent synapses. The discharge process involves exocytosis of
a complex organelle, the nematocyst, and explosive devagina-
tion of a poison-injecting thread, which prior to discharge is
kept coiled under pressure within the nematocyst wall (Holstein
and Tardent, 1984; Nüchter et al., 2006; Thurm et al., 2004).
Nematogenesis, the multistep process of nematocyte produc-
tion, is a particularly interesting process from developmental
and evolutionary points of view because nematocytes are
thought to be highly derived neuro-sensory cells (Hausmann
and Holstein, 1985; Brinkmann et al., 1996; Westfall, 1996;
Miljkovic-Licina et al., 2004). Nematocytes and typical
neurosensory cells share the ability to detect and integrate
sensory and neuronal stimuli and to transmit efferent action
potentials (Thurm et al., 2004). In hydrozoans, they are also
produced from the same stem cell lineage (the interstitial stem
cell system) (Lentz, 1966; Burnett, 1968; David and Gierer,
1974; David and Murphy, 1977). Finally, the production,
integration and innervation of fully functional nematocytes
from early nematoblasts committed among stem cells is likely to
involve numerous regulator and effector genes, and among
them bilaterian neural gene homologues were found to be
Available online at www.sciencedirect.com
Developmental Biology 315 (2008) 99–113
www.elsevier.com/developmentalbiology
⁎Corresponding author. Fax: +33 1 44 27 58 01.
E-mail address: elsa.denker@snv.jussieu.fr (E. Denker).
0012-1606/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ydbio.2007.12.023

Page 2

expressed during nematogenesis (such as the cnidarian achaete-
scute homologue CnASH (Grens et al., 1995; Lindgens et al.,
2004; Müller et al., 2003), the cnidarian Zic gene (Lindgens et
al., 2004), the paired-like gene prdl-b and COUP-TF genes
(Gauchat et al., 2004; Miljkovic-Licina et al., 2004). Under-
standing nematogenesis can thus provide insights about
fundamental neurogenesis processes in the common Cnidaria/
Bilateria ancestor.
In Hydra, interstitial stem cells are continuously committed
to a nematoblast fate at multiple sites scattered throughout the
ectoderm of the body column, excluding the pedal and oral parts
of the polyp (David and Challoner, 1974; David and Gierer,
1974). The interstitial cell population, which occupy a position
between the myoepithelial cells, comprises multipotent stem
cells as well as their derivatives: the committed precursors,
differentiating intermediates and mature forms of neurons,
nematocytes, glandular cells and germ cells (Bode and David,
1978; David and Murphy, 1977; Bode 1996). Recently, the
existence of an intermediate bipotent stem cell giving rise to
both nematocytes and neurons (ganglion cells), expressing the
Cnox2 homeobox gene, has been reported (Miljkovic-Licina
et al., 2007).
The various steps of nematocyte differentiation in Hydra
have been described by Slautterback and Fawcett (1959) and
Holstein (1981) (see also Bode, 1988). Early committed
nematoblasts undergo a change in nuclear morphology, and
after an initial division, they remain interstitial and divide
synchronously as a syncitial cell cluster of 4 to 32 cells.
Terminal differentiation can occur after any of these division
steps. Cells undergo a first growth phase, during which the
nematocyst primordium forms as a Golgi-derived organelle.
This occurs just before the terminal division that marks the
beginning of the second growth phase. Then the tubule, a
distinct structure that gives rise to the harpoon thread, develops
within the cytoplasm through the fusion of vesicles. When this
stage is completed, the tubule invaginates into the capsule. The
capsule maturates and the tubule differentiates into a shaft and a
thread covered with spines. The cnidocil apparatus forms, and
the capsule wall becomes thinner and acquires its high tensile
strength. The final maturation step involves the synthesis of
poly-γ-glutamate within the nematocyst wall, responsible for
generating the huge internal pressure that will allow discharge
(Szczepanek et al., 2002). At this point, the cluster splits into
individual mature nematocytes that migrate towards the tentacle
to be inserted into the nematocyte batteries in association with
support cells.
Our current knowledge of the nematogenesis process derives
mainly from extensive studies in the freshwater hydrozoan
Hydra. Hydra has proved to be a powerful experimental model
to investigate morphogenetic processes involved in adult tissue
homeostasis and regeneration (Galliot and Schmid, 2002), but it
lacks the medusa stage present in the ancestral life cycle of
Hydrozoa. The medusae of hydrozoan species such as Clytia
and Podocoryne have a distinct nervous and muscular
organization compared to the polyp stages, reflecting different
lifestyles and behaviors (Brusca and Brusca, 2003). The study
of a variety of species and life-cycle stages will aid the
identification of conserved and derived developmental mechan-
isms. In particular, the evolutionary significance of nematogen-
esis as a model of non-bilaterian neurogenesis will emerge only
when cnidarian ancestral features can be sorted from lineage
specific apomorphic features.
We have studied nematogenesis in medusae of the
leptothecate hydrozoan Clytia hemisphaerica (Linnaeus,
1767), which has recently been developed as a model species
for developmental studies (Chevalier et al., 2006; Momose and
Houliston, 2007). The Clytia medusa is relatively large (about
1 cm diameter), transparent and lives for 2 to 3 months.
Histological studies have shown that nematogenesis in hydro-
zoan medusae occurs in specialized zones separate from the
final destination of the mature nematocytes. Oral nematocytes
are produced at the base of the manubrium (the structure that
bears the mouth) and tentacle nematocytes in the tentacle bulbs:
spherical outgrowths on the bell margin from which tentacles
grow (Bouillon, 1994). Newly hatched Clytia medusae have 4
tentacle bulbs and adults about 24, new bulbs being inserted
during medusa growth.
Here we present the tentacle bulb of the C. hemisphaerica
(Hydrozoa) medusa as a novel model to study nematogen-
esis. We have characterized the sequence of nematogenesis
by a variety of techniques of light and electron microscopy,
including histological staining and fluorescent labelling; we
then monitored cell division and nematocyst migration and
performed in situ hybridization of selected markers of
nematogenesis stages. We show that the general cytological
and molecular aspects of nematocyte differentiation are
similar to Hydra, but with a very distinct, highly ordered
spatio-temporal organization. Not only is nematogenesis
confined to a specialized structure, but it involves a spatio-
temporal progression of nematogenesis stages along the
proximo-distal axis of the bulb. Such properties open
perspectives for future experimental studies of nematogenesis
using Clytia.
Materials and methods
Animal cultures
For all experiments, we used C. hemisphaerica medusae cultured in our lab
in Paris. Colonies were established from polyps provided by Evelyn Houliston
(Villefranche-sur-mer). Animals were cultured in 5 glass beakers as described in
Chevalier et al (2006), except that artificial seawater was used (36 g/l Reef
Crystals®, Aquarium systems).
Scanning electron microscopy
Adult medusae were fixed 1 h at room temperature in 4% paraformaldehyde,
washed three times 5 min in PBT and gradually dehydrated in ethanol (20 min
washes in 20%, 40%, 60% and 80% EtOH and two final 20 min 100% EtOH
washes), then transferred into amyl acetate before critical-point drying.
Specimens were mounted on copper stubs with double-sided adhesive tape,
coated with 300 A of gold in a Polaron sputtering apparatus and examined on a
JEOL JSM 6100 scanning electron microscope at 15 kV. For examination of
dissected samples, critical-point dried animals were dissected directly after
mounting on double-sided adhesive coating, either by cutting with a
microscalpel or by rolling them, tearing and spreading out the external
structures and revealing the internal structures.
100 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 3

Transmission electron microscopy and semi-thin histological sections
Clytia tentacle bulbs were isolated with a scalpel blade and forceps, fixed 2 h
at room temperature in 6% glutaraldehyde in cacodylate buffer (CB) (0.1 M
cacodylate and 1.75% NaCl to correct the osmolarity) and washed three times in
CB. Samples were then post-fixed for 45 min in 1% OsO4in CB and washed
three times in CB. They were progressively dehydrated (EtOH 50%, EtOH 70%,
EtOH 95% and twice EtOH 100%, 20 min bathes) and transferred to propylene
oxide (20 min in 50% EtOH and 50% propylene oxide, 20 min in 100%
propylene oxide). Prior to inclusion in Agar 100 epoxy resin mix (Agar 100
resin kit, Agar scientific: 6 g Agar 100 resin, 4 g DDSA, 2.5 g MNA, 0.38 g
BDMA), several incubations were performed in mixtures of Epon resin medium
and propylene oxide with increasing resin proportion (2 h 15 in 25% resin, 3 h
45 in 50% resin, one night in 75% resin, 1 h in 75% resin) until incubation in
pure resin (1 h in 100% resin, 1 h under vacuum at 50 °C in a small glass cupule
to eliminate gas traces). Samples were included and orientated in polyethylene
BEEM®mouldspreviouslyfilled witha thinresinlayerandleft 48h at60 °Cfor
polymerization. Semi-thin and ultra-thin sections were performed with a
Diatomediamondhisto knife.Ultrathinsectionswerestainedwithuranylacetate
and lead citrate before mounting and observation. Semi-thin sections were
stained with toluidine blue and mounted in Eukitt medium (Kindler GmbH and
Cie, Freiburg, Germany).
TRITC pulse nematocyst staining
The protocol was adapted from Weber (1995). Animals were cultured for
30 min in the dark in 1 μM TRITC (tetramethylrhodamine isothiocyanate,
Invitrogen Molecular Probes) in filtered artificial sea water (FASW) under
agitation, washed several times in FASW, quickly rinsed in PBT (PBS
1×+0.1% Tween 20) to eliminate as much salt as possible but avoiding an
osmotic shock and fixed 10 min in EtOH100. Samples were rehydrated
(10 min in EtOH60, 10 min in 30% EtOH, 10 min in PBT), nuclei were
stained by a 10-min incubation in 1 μg/ml DAPI (4′,6-diamidino-2-
phenylindole) in PBT, rinsed quickly in PBT and medusae were mounted
in VectaShield medium. For pulse-chase experiments, TRITC was chased by
several washes in FASW, and animals were cultured in FASW in the dark
under agitation for 24 h, rinsed in PBT and fixed, rehydrated and mounted as
described above.
DAPI poly-γ-glutamate staining
The protocol was adapted from Szczepanek et al. (2002). Samples were
fixed for 45 min in 4% PFA in PBS and washed several times in PBS. To stain
poly-γ-glutamate (PGA), we incubated samples for 30 min in 140 μM DAPI in
PBS. Samples were rinsed several times and mounted in VectaShield. Nuclei
bluestaining andPGAyellow-orange stainingwereobserved by epifluorescence
under DAPI excitation conditions (360 nm).
BrdU pulse labelling assays
The protocol was adapted from Lindgens et al. (2004). Medusae were
incubated for 30 min in 5 mM BrdU in FASW, quickly rinsed twice and
extensively washed twice for 10 min in FASW. Animals were fixed overnight
at 4 °C in 4% paraformaldehyde in FASW and rinsed four times in PBT.
DNA was denatured for 30 min in 2N HCl, and samples were rinsed four
times for 10 min in PBT. Incorporated BrdU was revealed by immunohis-
tochemistry after an overnight incubation at 4 °C with a 1°/100° diluted
mouse anti-BrdU primary antibody (Roche) in PBT+1% BSA, followed by a
short rinse and three 10 min PBT washes. The samples were then incubated
for 3 h at room temperature in a 1/1000° diluted Alexa 568 anti-mouse
antibody (Invitrogen Molecular Probes) in PBT+1% BSA, followed by a
short rinse and three 10 min PBT washes. The last wash contained DAPI at
1 μg/ml. Samples were mounted in VectaShield medium. For pulse-chase
experiments, animals were cultured 30 min in 5 mM BrdU in FASW, then
chased by two rinses and three 5 min washes in FASW, cultured under
agitation in FASW containing 100 U/ml penicillin and 0.1 mg/ml
streptomycin for 2 to 80 h) and then fixed and treated as described above.
Anti-phospho-histone H3 immunostaining
An anti-phospho-histone H3 (Ser10) antibody (rabbit polyclonal, Upstate
(Millipore) 06-570) was used to detect mitotic cells. The corresponding
secondary antibody was anti-rabbit Alexa 568 (goat monoclonal, Invitrogen
molecular probes A11011). Four percent of paraformaldehyde fixed samples
was permeabilized during 10 min in PBS 1% BSA and washed several times for
10 min in PBS 0.1% Triton X-100,and twice for 10 min in PBS 0.01% Triton X-
100.Sampleswere incubatedin the primary antibody at a 1/200°dilution in PBS
0.01% Triton X-100 overnight at 4 °C or 2 h at room temperature. Excess
antibody was removed by four 15 min washes in PBS 0.01% Triton X-100.
Animals were incubated in the corresponding secondary antibody at a 1/1000°
dilution in PBS 0.01% Triton X-100, overnight at 4 °C at room temperature in
the dark and washed four times in PBS 0.01% Triton X-100. The third wash
contained DAPI at 1 μg/ml. Mounting was performed using VectaShield
medium and preparations were observed under epifluorescence using a DAPI
filter and a Cy3.5 filter (for Alexa 568 detection) on an Olympus BX61
microscope with a Q-imaging Camera.
Gene identification in EST collection
The Clytia minicollagen 1 and 3-4a, NOWA, Dkk3 and Piwi genes (see
Supplementary data) were retrieved by BLAST searches on an EST collection
sequenced by the Genoscope (Evry, France) from a C. hemisphaerica
normalized cDNA library (Chevalier et al., 2006). The library was constructed
by Express Genomics (Frederick, Maryland, USA) in pExpress1 plasmids from
total mRNA extracted from a mixture of medusa, embryonic and larval stages.
Sequences of the 5 Clytia genes were deposited in GenBank under the
accession numbers EU195079 (Chemcol1), EU024529 (Chemcol3-4a),
EU195078 (CheNowa), EU195080 (CheDkk-3) and EU199802 (ChePiwi).
Sequence alignment and phylogenetic analyses
For all genes identified by BLAST on Clytia ESTs, the orthology group
assignation was established by phylogenetic analyses. Other cnidarian and
bilaterian sequences were found in GenBank or on www.compagen.org
(Hemmrich and Bosch, 2007). Sequences were automatically aligned using
CLUSTALW in BioEdit (Tom Hall, Ibis Therapeutics, Carlsbad, CA), and the
alignment was corrected manually. Conserved blocks were extracted to perform
phylogenetic analyses.
Phylogenetic analyses were carried out by the Maximum-Likelihood (ML)
method using the PhyML program (Guindon and Gascuel, 2003), with the JTT
model of amino acid substitutions (Whelan and Goldman, 2001). A BioNJ tree
was used as the input tree to generate the ML tree. Among site variation was
estimated using a discrete approximation of the gamma distribution with 4 rate
categories. The gamma shape parameter and the proportion of invariant sites
were optimized during the ML search. Branch support was tested with
bootstrapping (100 replicates).
Single whole-mount in situ hybridization (ISH)
Protocol was adapted from Chevalier et al. (2006). Digoxigenin antisense
probes were synthesized from linearized plasmids using the T7 polymerase as
follows: 2 μl Dig labelling mix, 2 μl 10X Reaction buffer, 1 μl Rnasin 40 U and
2 μl T7 polymerase (60–80 U) were added to 13 µl purified linearized DNA
template. The RNA probe was precipitated by adding 1/10° volume 3 M sodium
acetate and 2.5 volumes ethanol one night at −80 °C, then washed with 70%
ethanol and resuspended in 30 μl DEPC water.
Clytia adult and juvenile medusae were let unfed during one night and fixed
1 h in ice cold 3.7% formaldehyde/0.2% glutaraldehyde/PBS at 4 °C. After five
10 min PBT washes, samples were progressively dehydrated and stored at
−20 °C in absolute methanol. After several rehydration steps, medusae were
incubatedin 0.01 mg/ml proteinase K in PBT for 15 min at 37 °C. Digestion was
stopped with two washes in PBS containing 2 mg/ml glycine. Samples were
gradually transferred to hybridization buffer (HB=5×SSC, 50% formamide,
0.1% Tween) and pre-hybridized for 1 h 30 at 60 °C in HB+0.1% DMSO,
50 μg/ml heparin and 50 μg/ml tRNA. The probe (100 ng/ml) was then added
101 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 4

and hybridization was carried out one night at 60 °C. Excess probe was removed
by extensive washes in HB at 60 °C, in 50% HB in PBTat roomtemperature and
finally in PBT. Samples were saturated using blocking buffer in 1X maleic acid
(Roche) and incubated in 1/2000° anti-DIG Alkaline phosphatase-coupled
antibody (Roche) and washed thoroughly in PBT. Animals were twice
incubated 10 min in TMN (100 mM Tris–Hcl pH 8+50 mM MgCl2+100 mM
NaCl+10 mM levamisole and 0.1% Tween 20), and alkaline phosphatase
activity was revealed by using BM purple reagent (Roche) or Fast Red
TR-naphthol reagent (Sigma). Nuclei were counterstained with 1 μg/ml DAPI.
Samples were post-fixed (1 h in 4% PFA) and mounted in VectaShield (Vector
laboratories). They were observed using a microscope (Olympus BX61) with a
Q-imaging Camera using Image Pro plus software ® (Mediacybernetics). Fast
Red-revealed samples were observed under epifluorescence (Cy 3.5 filter) and
with OptiGrid® Structured-Light System (Qioptiq imaging).
Double whole-mount in situ hybridization
We adapted our simple ISH protocol to double ISH according to Hansen et
al. (2000). The protocol was as described above, except that we hybridized
samples simultaneously with two different probes: one labelled with
digoxigenin, the other with fluorescein. We also performed two rounds of
immunostaining and substrate revelation steps: one for each labelling detection.
After hybridization and washes, we saturated the samples using Blocking
reagent, and we first detected the fluorescein-labelled probe by incubating 2 h in
1/2000anti-fluorescein Alkalinephosphatase-coupled antibody(Roche) at room
temperature. Samples were washed, incubated in TMN and submitted to Fast
Red detection. When the signal was sufficiently intense, the reaction was
stopped by several PBT washes and by a 100 mM glycine–HCl pH 2.2 10 min
incubation at room temperature. Sampleswere thenwashes severaltimes in PBT
and a new Blocking reagent saturation step was undertaken. We then detected
the digoxigenin-labelled probe by incubating 2 h in 1/2000 anti-digoxigenin
alkaline phosphatase-coupled antibody (Roche) at room temperature. Medusae
were washed, incubated in TMN and submitted to BM purple revelation. The
reaction was stopped by several PBT washes, and nuclei were stained with
DAPI. Samples were mounted in VectaShield and analyzed by transmission and
epifluorescence as described above.
Results
Minicollagen gene expression identifies tentacle bulbs as the
major nematogenesis sites in Clytia medusae
To determine the sites of nematogenesis in Clytia medusae,
we examined the expression of minicollagens, small collagen-
like proteins identified in Hydra as the major component of the
nematocyst wall (Kurz et al., 1991). We performed whole-
mount in situ hybridization to detect mRNAs from two of three
minicollagens genes identified in Clytia, mcol1 and mcol3-4a.
Both revealed the same two sites of nematogenesis: the very
proximal part of the manubrium (the structure bearing the
mouth) and the tentacle bulbs (Fig. 1A). The most intense
Fig. 1. Minicollagen transcript localization identifies the tentacle bulbs as major nematogenesis sites of Clytia medusa. (A–D) Whole-mount in situ hybridizations
(FastRed development). (A) minicollagen 1 mRNA localization on adult medusa, subumbrella side. Staining on mature tentacle bulbs (black arrowhead for side view
and red arrowhead for apical view), developing bulb buds (green arrowhead) and manubrium basis (blue arrowhead). (B) Close-up of minicollagen 3-4a RNA
localization on the tentacle bulb (external side). Green arrowhead: isolated cells in the tentacle. (C–D) Close-up of minicollagen 3-4a-expressing-cells in the tentacle
bulb (external side), Z projection of OptiGrid optical sections (Cy 3.5 filter). (C) The staining shape corresponds to an externally interrupted crescent, composed of
round to ovoid cells. Nuclei are stained with DAPI (in blue). (D) Nematoblasts are associated in clusters. A cluster is encircled in yellow. Scale bars: A, B: 250 μm; C:
25 μm; D: 2,5 μm.
102E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 5

staining was found on the latter (Fig. 1B), indicating that the
tentacle bulbs are the major sites of nematogenesis in Clytia.
Minicollagen expression was also detected in isolated cells in
the tentacle (Fig. 1B), but not in mature nematocytes, either in
the tentacles or the manubrium lips.
A crescent-shaped nematogenesis zone in the Clytia tentacle
bulb ectoderm
Close analysis of the tentacle bulb revealed the cells
expressing mcol3-4a as numerous round cells in the ectodermal
layer of the bulb (Fig. 1C) and a few isolated cells in the tentacle
(Fig. 1B). The tentacle bulb cells were grouped in clusters of up
to around 8 cells (Fig. 1D). The minicollagen expressing zone
formed a crescent around the bulb rather than a complete ring,
the in situ staining being interrupted on the external side of the
bulb (visible on the apical view of a stained bulb on Fig. 1A, red
arrow).
Scanning electron microscopy (SEM), phalloidin staining
and toluidine blue staining of semi-thin sections were
performed to clarify the localization of the nematogenic area
(Fig. 2). The bulbs exhibited clear bilateral symmetry associated
with a contrasted external morphology of the inner (adaxial) and
outer (abaxial) faces of the bulb with respect to the symmetry
axis of the animal (Figs. 2A and B). The outer side of the bulb is
flattened whereas the internal side has a thickened morphology.
Moreover, the tentacle insertion site is slightly off center
towards the external side. A group of naturally fluorescent cells
was detected in a triangular area on external side of the bulb,
providing a convenient landmark (Fig. 2B). Optical sections of
phalloidin-stained bulbs demonstrated the marked thickening of
the ectoderm of the tentacle bulb (TBE) compared to that of the
umbrella and tentacle regions (Fig. 2C). DAPI staining of nuclei
revealed that this thickening reflected an increased number of
cells rather than larger cell size (Fig. 2C). In the distal-most part
of the bulb and in the most proximal part of the tentacle, the
ectodermal phalloidin staining was more intense, suggesting
that cell in these regions have thickened actin cortices (Figs. 2C
and D). On transverse semi-thin toluidine blue sections of the
bulb, nematoblasts are easily distinguishable from the other cell
Fig. 2. Morpho-anatomy of the tentacle bulb. (A) Scanning electron microscopy (SEM) pictures of a tentacle bulb (frontal view, external side) showing its insertion on
exumbrellar tissues (eu) on the bell margin (bm), and the relative positions of the velum (v) and the subumbrella (su). This insertion surface defines the most proximal
part of the bulb, the most distal point being the tentacle tip. (B) Side confocal view of a live isolated tentacle bulb, superposition of a DIC transmission picture and of
the natural fluorescence (in blue). en: endoderm, ec: ectoderm. (C–D) Confocal optical section of phalloidin stained (cyan) bulb with DAPI stained nuclei (red), false
colors. Frontal view. (C) The tentacle bulb ectoderm (tbe) appears very thick compared to the umbrella ectoderm (ue) and the tentacle ectoderm (te). tben: tentacle bulb
endoderm. The arrowheads highlight the frontier of the more intensely stained area. (D) The ectoderm of the distal area of the bulb and the proximal area of the tentacle
presents a more intense phalloidin staining (close up). (E) Toluidine blue-stained transverse semi-thin section of a bulb (at the median level). tbe: tentacle bulb
ectoderm, tben: tentacle bulb endoderm (revealing the tentacle axis). (F) Dissected tentacle bulb ectoderm on SEM (internal view) revealing almost only nematoblasts
(one of them is marked by the arrowhead). Scale bars: A, B, C, F: 25 μm; D: 10 μm; F: 5 μm.
103E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 6

types present (notably myoepithelial cells), from the early
nematoblast stage, by the distinct coloration of the cytoplasm.
Nematocysts at various stages of differentiation appeared as
prominent blue spots (Fig. 2E). From examining transverse
sections, it was clear that although most of the TBE is
nematogenic, there is a narrow, thinner segment at the center of
the external side that contains almost no nematoblast (Fig. 2E),
corresponding to the interruption of mcol3-4a expression (Fig.
1C and red arrow on Fig. 1A). SEM observations of dissected
bulbs showed nematoblasts as spherical cells concentrated in
the ectoderm (Fig. 2F), whereas the later stages appeared more
ovoid or spindle shaped. Taken together, these observations
characterize the tentacle bulb ectoderm as a region specialized
for tentacular nematogenesis.
Proximo-distal regionalization of the nematogenic TBE
TEM and histological sections were used to examine the
distribution of nematogenesis stages within the nematogenic
TBE. The following stages were identified by TEM, according
to previous descriptions (Holstein, 1981) (Figs. 3A–F): (i)
undifferentiated, interstitial cells (either stem cells or early
committed nematoblasts) characterized by a high nucleocyto-
plasmic ratio and a large nucleus with a developed nucleolus
(Chapman, 1974) (Fig. 3A); (ii) Early differentiating nemato-
blasts, with the nascent capsule detectable as a highly electron
lucent cytoplasmic spherical element and cytoplasm filled with
rough endoplasmic reticulum (Fig. 3B); (iii) later differentia-
tion stages containing a larger developing capsule and a thread
forming in the cytoplasm visible as cytoplasmic dense
elements (Fig. 3C); and (iv) nematoblasts with a full-sized
capsule containing an internal coiled thread differentiated into
a large shaft area and a thinner tubule (Figs. 3D–F). Within
this category, we could distinguish capsules with a more
electron-lucent matrix (Fig. 3F compared to 3E), correspond-
ing to mature nematocytes as found in the tentacles (data not
shown).
The distribution of nematogenesis stages was determined on
serial toluidine blue stained semi-thin transverse sections (Fig.
3G, see also Supplementary data). We considered only those
stages that could be unambiguously recognized on the basis of
the TEM analysis: early forming capsules (corresponding to the
stage on Fig. 3B), differentiating capsules (corresponding to the
stages on Figs. 3C–E) and mature nematocytes (corresponding
to the stage on Fig. 3F). Early nematogenesis stages were
mainly found in the proximal-most part of the TBE,
progressively disappearing from medial and distal sections.
Very few late stages were found in proximal regions, but they
were found in majority in the distal part of the TBE.
Differentiating nematoblasts were thus mostly found in the
median zone of the TBE. There is thus a clear regionalized
organization of nematoblast stages along the proximo-distal
axis of the bulb. There is not a strict segregation of stages but
rather a graded distribution of each stage. For example, mature
nematocytes are not exclusively located at the distal extremity
of the bulb and in the tentacle, a few mature nematocytes being
detected even in the proximal half of the bulb (Fig. 3G). On
distal transverse sections, a differential distribution of nemato-
blast stages between the center and the periphery of TBE was
also observed, the more mature stages being closer to the
mesoglea (Fig. 3H). Undifferentiated cells were localized on the
most proximal transverse sections. Here, we did not find
nematoblasts with a developing capsule, but we found a
majority of undifferentiated cells with a dark blue cytoplasm
(Fig. 3I) (as well as a large nucleus and a large nucleoli, data not
shown). Moreover, on both sides of the bulb basis, we observed
two symmetric groups of undifferentiated cells with lighter
cytoplasm (Fig. 3I) (and a large nucleus, data not shown).
Fluorescent labelling techniques were used to determine the
distribution of particular stages of late nematogenesis: TRITC
that selectively labels nematoblasts during a short time frame at
the end of nematoblast differentiation (later than minicollagen
expression; Weber, 1995), and DAPI staining to detect poly-γ-
glutamate, present in late stages of differentiation and fully
mature nematocysts (Fig. 3K). TRITC labelling localized late
differentiating nematoblasts to a restricted area of the most
distal part of the bulb and the most proximal part of the tentacle
(Fig. 3J). Orange DAPI staining overlapped with the area of
TRITC staining and also decorated the mature tentacular
nematocytes.
Taken together, these various observations allow us to define
four areas in the TBE with distinct proportions of nematoblast
stages (Fig. 3L). The very thin (about 10 μm) proximal-most
region of the TBE (“α” region) contains only undifferentiated
cells and almost no maturing or mature nematocytes, with
lighter-staining cells being grouped on each side. The “β”
region comprises most of the TBE (excluding the basal-most
and distal-most areas) and contains cells at various stages of
differentiation. The “γ” region, highlighted by TRITC staining
all around the distal bulb and the proximal tentacle, contains
mainly the latest stages of capsule maturation. Finally, the “δ”
region is rest of the tentacle (except its proximal-most area), and
contains only mature nematocytes (stained with DAPI but not
with TRITC).
Active cell proliferation in the Clytia TBE
Cell proliferation in the TBE was examined by BrdU
incorporation (to detect DNA synthesis) and immunochemistry
with an anti-phospho-H3 antibody (to detect mitotic chromatin).
The tentacle bulbs were found to be the major areas of mitosis
and BrdU incorporation within the medusa (not shown). A high
density of mitotic cells was found in a crescent-shaped region of
the TBE, corresponding to the nematogenesis area, indicating
intense division activity among nematoblasts (Fig. 4A). Within
this area, anti-phospho-H3 staining intensity was uniform
suggesting homogeneous cell division between the various
zones. After a 30-min pulse incubation, BrdU incorporation was
also detected in the nematogenesis area (Fig. 4B). DNA
replication appears very active in this region since a 5-min
incubation in BrdU was sufficient to give strongly detectable
incorporation (data not shown). A closer analysis showed that
incorporation was slightly greater in the α and the proximal β
regions of the bulb (Fig. 4B). Altogether, these results suggest
104 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 7

Fig. 3. Thenematogenesisstagesareproximo-distallydistributedinthetentaclebulbectoderm.(A–F)Transmissionelectronmicroscopypicturesofnematoblaststages
inthe TBE.Whitearrowheads:capsules;blackarrowheads:thread.(A)Interstitialcells.(B)Early nematoblast withasmalldevelopingcapsule.(C)Nematoblastwitha
developing capsule and an external thread differentiating in the cytoplasm. (D) Nematocyte with a mature capsule. (E–F) Transition from a nearly mature capsule with
an electron dense matrix (E) to a mature capsule with an electron lucent matrix (F). (G) Graph showing the percentages of three stages of capsule maturation along the
proximo-distal axis. Two different ordinates for the percentage of early forming nematoblasts on the left and for the percentage of differentiating and late maturating
nematoblastsontheright.(H–I)Toluidineblue-stainedtransversesemi-thinsectionsofabulb.tbe:tentaclebulbectoderm;tben:tentaclebulbendoderm.(H)Sectionin
theintermediateregionbetweenthebulbandthetentacleshowingmaturecapsules(inpurple,redarrow)andimmaturecapsules(inblue,bluearrow).(I)Mostproximal
sectionwithundifferentiatedcells,showingtwolateralareaswithcellsdisplayinganinterstitialcellphenotypewithalightcytoplasm(reddottedlines)andaninnerarea
with cells withthe same phenotypebut with a bluecytoplasm(black dotted lines). Asterisksdelimit an area where a part of the section is not present (the most proximal
end of the isolated bulb was reached). (J) TRITC staining revealing nearly matures capsules after 30 min of incubation. (K) DAPI staining (nuclei stained in blue and
poly-γ-glutamate containing capsules in orange). Small box: view of the tentacle. On J and K, arrowhead: tentacle insertion site on the bulb. tb: tentacle bulb; te:
tentacle. (L) Synthetic diagram illustrating the proximo-distal distribution of nematogenesis stages in the tentacle bulb ectoderm. Scale bars: A–F: 1 μm; H–L: 25 μm.
105E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 8

106 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 9

that mitotic division occurs at many of the nematogenesis stages
previously identified in the TBE crescent-shaped area.
Rapid displacement of TBE cells into the tentacle
To determine the fate of cells dividing in the TBE, we
performed BrdU pulse/chase experiments. After 30 min
incubation in BrdU, medusae were cultured 1 to 72 h in
BrdU-free seawater, fixed, and the BrdU was detected by
immunohistochemistry (Figs. 4C–J). BrdU-labelled nuclei were
still located in the TBE nematogenic area after 2 or 8 h of chase
(Figs. 4C, D). However, after 2 h, some cells had become
displaced from the bulb to the bell margin (Fig. 4C, arrow).
After 10 h, a slight shift of the bulk of the labelled nuclei
towards the distal part of the β region could be detected (Fig.
4E); and after 16 h, labelled nuclei were also detectable in the γ
region (Fig. 4F). Thereafter, they were found in progressively
more distal positions in the tentacle (Figs. 4G–I). After 24 h of
chase, labelled nuclei extensively covered the γ region and the
proximal δ region of the tentacle (Fig. 4G); after 48 h, they had
totally evacuated the bulb and began to populate the δ region
(Figs. 4H, I); and after 72 h labelled nuclei had reached the
tentacle tip (Figs. 4J). Scattered stained nuclei were often found
more distally than the main group of stained nuclei. After
around 8 h, the TBE cells progressively leave the bulb and
massively populate the tentacle. To show directly the presence
of nematocytes among the cells moving from the TBE into the
tentacle, we used the live TRITC staining technique to label the
wall of late, maturing nematoblasts. Animals were incubated in
TRITC for 30 min, then washed and cultured for 24 h in
TRITC-free seawater. During this period, labelled nematocysts
moved towards the δ region, populating almost the entire length
of it, except the γ area that was initially stained after 30 min
TRITC incubation (Fig. 5A). Late differentiating, post-mitotic,
nematoblasts are thus included in a massive cellular flow from
the tentacle bulb to the tentacle.
Fig. 4. Cell division and nuclei movements in the TBE. (A) Anti-phosphoH3 immunolocalization in the TBE. (B) Thirty-minute pulse staining with BrdU revealing
nuclei undergoing DNA replication in the TBE. (C–J) Pulse/chase experiments after 30 min incubation in BrdU. (C) Two-hour chase. Arrow: isolated nuclei near the
bell margin. (D) Eight-hour chase. (E) Ten-hour chase. Stained nuclei are now concentrated in the distal β zone. (F) Sixteen-hour chase. Stained nuclei have reached
the γ zone. (G) Twenty-four-hour chase. Stained nuclei cover one third of the tentacle. (H) Forty-eight-hour chase. Stained nuclei populate the middle region of the
tentacle. (I) Forty-eight-hour chase. High magnification of the bulb showing the absence of stained nuclei. (J) Seventy-two-hour chase. Stained nuclei cover the most
distal part of the tentacle. On all pictures, the junctions between the four areas of the bulb are marked with a dotted line. Scale bars: 25 μm.
Fig. 5. Capsule movements in the TBE. (A) TRITC staining after 30 min incubation and 24 h chase. The background staining is due to the weak staining of all
membranes by TRITC. (B) Scheme comparing TRITC and BrdU pulse/chase results. Scale bar: 25 μm.
107E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 10

Overlapping expression domains of nematogenesis genes in the
TBE
Studies in Hydra and other cnidarians have characterized a
number of genes whose expression is associated with different
steps of nematogenesis. In addition to minicollagens, we
identified three of these genes from a Clytia EST collection
and analyzed their expression: the stem cell marker gene Piwi
(previously studied in Podocoryne, Seipel et al., 2004), the late
regulator gene Dkk3 (Fedders et al., 2004) and the structural
Fig. 6. Sequential gene expression territories in the TBE. Bulbs are viewed from the outer side. (A–C) Single in situ hybridizations (BM purple development). (A)
Piwi. The bulb is slightly turned towards the proximal end, so that the continuity of the staining on the inner face is visible. (B) Dkk3. (C) NOWA. (D–F) Double in situ
hybridizations, minicollagen 3-4a developed with Fast Red in each case and the second probe with BM purple. (G–I) Details of double in situ hybridizations. (D, G)
Piwi. (E, H) Dkk3. (F, I) NOWA. Scale bars: A–F: 25 μm; G: 5 μm; H, I: 10 μm. Arrowheads: tentacle bulb insertion site on the umbrella. Green arrows: isolated cells
in the tentacle. White arrows: cells expressing only the probe developed with BMpurple. Black arrows: cells expressing only minicollagen 3-4a (Fast red). Grey
arrows: cells expressing both markers.
108 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 11

gene NOWA (Engel et al., 2002) (see Materials and methods and
Supplementary data for gene identification procedure and
phylogenetic analyses). All three genes were expressed in the
TBE crescent-shaped nematogenic area, but in a sequential
manner along the proximo-distal axis, in accordance with the
cellular characterization of nematogenesis described above.
Piwi was strongly expressed in the proximal TBE, in the α
region and in two lateral areas in the proximal part of the β
region. This gene was also expressed in a more diffuse manner
in the center of the proximal part of the β area (Fig. 6A). Dkk3
was expressed through the entire β region (Fig. 6B). Finally,
NOWA was expressed weakly in the proximal β region and
strongly in the distal part of the β as well as in some isolated
cells in the tentacle (Fig. 6C). NOWA expressing cells were
distributed in clusters and sometimes aligned along the
proximo-distal axis (data not shown).
Double in situ hybridizations positioned the expression
domains with respect to mcol3-4a expression. We found that
Piwi coexpressed with mcol3-4a in two symmetric lateral
patches of cells in the proximal part of the β region, as well as a
narrow stripe of cells in the central area (Figs. 6D and G).
However, in the α region, there were a majority of cells
expressing Piwi alone with few cells coexpressing the two
genes. Dkk3 coexpressed with mcol3-4a in all cells except for a
few isolated mcol3-4a-positive cells in the tentacle and some
cells near the lateral extremities of the crescent (Figs. 6E, H).
The Dkk3/mcol3-4a domain corresponds to the β region.
NOWA also coexpressed with mcol3-4a in part of the β region,
the majority of NOWA-expressing cells occupying the distal
part of the β area, and a small proportion of NOWA-expressing
cells also occurring more proximally and in the tentacle (Figs.
6F, I). NOWA-positive cells were found to contain a more
developed capsule than cells expressing only mcol3-4a, but the
capsules were probably immature in these cells as they
deformed upon fixation (not shown).
Discussion
The Clytia tentacle bulb is a specialized structure dedicated to
the production of tentacle nematocytes
The ectoderm of the tentacle bulb is a specialized territory of
the medusa that functions as a nematocyte production center, in
contrast with diffuse nematogenesis in cnidarian polyps. In
Hydra, interstitial stem cells are distributed throughout the
ectoderm and endoderm, and clusters of nematoblasts at all
stages of differentiation are scattered along the length of the
body column ectoderm, except its basal-most and apical-most
extremities (David and Challoner, 1974; David and Gierer,
1974; Fujisawa et al., 1986; Bode, 1996). Given that the tentacle
ectoderm is the final destination of most mature nematocytes,
differentiated cells must undergo migration along the column
from their differentiation sites to the tentacles (Bouillon, 1994).
In colonial hydrozoans like Hydractinia, nematogenesis is
thought to occur in the basal stolons before nematocytes migrate
towards individual polyps (Teo et al., 2006). Our in situ
hybridization analyses for molecular markers of nematogenesis
demonstrated that in the C. hemisphaerica medusa nematogen-
esis is essentially concentrated into two areas: the inner TBE
and a distinct region at the base of the manubrium ectoderm.
The latter site produces nematocytes of the mouth and
manubrium, whereas the TBE generates tentacle nematocytes.
Mature nematocytes are also present in others locations in the
medusa, the bell margin and the exumbrella (Östman, 1979);
however, the absence of detectable nematogenesis marker
expression and TRITC labelling in these area indicates that they
are not produced locally. The BdrU pulse-chase analyses show
that there are nuclei movements from the tentacle bulb to the
bell margin. Nematoblasts or nematocytes could be part of these
moving cells, and the bell margin nematocytes could thus
possibly be also produced in the TBE. Further experiments will
be needed to investigate the origin of exumbrellar nematocytes.
The existence of a graded progression of nematoblast stage
populations along the TBE, from the base of the bulb to the site
of tentacle insertion, is remarkable. No such spatio-temporal
ordering of nematogenesis exists in the previously studied
models, particularly Hydra. We hypothesize that this ordered
progression is a consequence of (i) the spatial restriction of stem
cells at the bulb base, as revealed by histological sections (Fig.
3H) and localization of Piwi+/mcol3-4a−transcripts (Figs. 6A,
D, G) and (ii) the continuous translocation of differentiated
stages towards the tentacle. The degree of spatial restriction of
nematogenesis stages along the bulb axis is sufficient for genes
involved in distinct steps of the process to be expressed in
staggered crescents, located more basally for earlier genes and
more distally for later genes (Fig. 6). Stage segregation along
the bulb axis was not, however, perfect (Figs. 3G, L). Likely
explanations include a certain amount of mixture between cells
within the TBE, some degree of heterogeneity between clusters
in the timing of commitment and nematogenesis steps, and
variability in the rate of cell translocation towards the tentacle.
The presence of a few mature nematocytes within the β region
could represent a subpopulation dedicated to a defensive role on
the TBE surface.
From the evolutionary point of view, tentacle bulbs represent
an innovation (synapomorphy) of the Hydroidolina clade
(Collins et al., 2006), a large group that comprises the vast
majority of hydrozoan species. Similar expression pattern of the
Piwi homologue Cniwi in the tentacle bulb of Podocoryne
(Seipel et al., 2004) compared to Clytia suggests that the basic
features of TBE nematogenesis are conserved among Hydro-
idolina despite evolutionary divergence between these two
species. Non-Hydroidolina hydromedusae (Trachylina) are
devoid of tentacle bulbs and their nematogenesis area extends
around the whole bell margin (in the so-called “Nesselring”)
(Bouillon, 1994). Thus, the acquisition of tentacle bulbs in an
ancestor of Hydroidolina resulted in a higher spatial restriction
of tentacle nematogenesis.
Dynamics of nematogenesis in the TBE
We were able to recognize all nematogenesis stages
previously described in Hydra and other Cnidaria (Westfall,
1966; Holstein, 1981) in the Clytia TBE. As in other
109 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 12

hydrozoans like Hydra (Slautterback and Fawcett, 1959),
Obelia (Westfall, 1966) or Podocoryne (Boelsterli, 1977), the
nematoblasts were found to be grouped in clusters. In Hydra, it
has been shown that nematogenesis requires several mitosis
steps during early stages and an additional terminal division
seems to occur after capsule formation (Campbell and David,
1974; Engel et al., 2002). In Clytia, we observed from our anti-
H3 and BrdU experiments a broad proliferation zone covering
the whole TBE crescent, with slightly more BrdU incorporation
detected in the proximal area, suggesting that nematoblasts
divide throughout all early stages as in Hydra, but possibly also
in later stages of differentiation with a developing nematocyst.
In Podocoryne, tentacle bulbs have also been shown to be
intensive proliferation sites, as well as the manubrium and the
bell margin (Spring et al., 2000).
The total duration of nematogenesis in Clytia is clearly
shorter than in Hydra. We showed that a dividing cell in the
proximal-most area of the bulb, where cells express the stem
cell marker Piwi will reach the tentacle after 1 to 2 days. In
comparison, in Hydra it was shown by [3H] incubation that
nematoblasts differentiate within 5 to 8 days depending on the
nematocyst type (5–7 days for desmonemes and isorhizas; 7–
8 days for stenoteles). The timing difference between
nematocyte types is thought to be due to a “lag phase” during
which stenoteles undergo a special maturation step before
migration (Weber, 1995). Clytia medusae are known to possess
a majority of small oblong nematocytes, known as microbasic-
b-mastigophores, in addition to a minority of larger and more
spherical nematocytes, termed atrichous isorhiza (Östman,
1979). Within the TBE, we followed the nematogenesis of the
first category, which is mostly used for prey capture, as are
Hydra stenoteles. As for stenoteles in Hydra, colonization of
the tentacles by nematocytes in Clytia following BrdU
incorporation involved a latent phase, the cells moving into
the tentacle after 8 to 12 h. This phase of latency could involve a
tissue reorganization step necessary to elaborate the final
tentacle tissue structure. Within the TBE, nematoblasts are
numerous in the interstices between adjacent myoepithelial
cells, whereas in the tentacle there is only one layer of interstitial
nematocytes. The intense phalloidin staining in the γ region
(Fig. 2D) could reflect tissue remodelling in the “bottle neck”
zone.
Stem cells and Piwi expression in TBE
The narrow band, or “α” region, at the base of the TBE, in
which a majority of cells express the Clytia Piwi homologue but
not the differentiation marker mcol3-4a, is likely to comprise a
population of stem cells. Histologically this region was found to
contain undifferentiated cells with a high nucleocytoplasmic
ratio. The Piwi gene is a widely conserved stem cell marker
throughout multicellular eukaryotes, including Drosophila (Lin
and Spradling,1997; Cox et al., 1998), annelids (Rebscher et al.,
2007), platyhelminths (Rossi et al., 2006; Reddien et al., 2005),
vertebrates (Cox et al., 1998; Cikaluk et al., 1999; Tan et al.
2002), sea urchins (Qiao et al., 2002) and the cnidarian Podo-
coryne (Seipel et al., 2004). A homologue has even been
identified in the ciliates Tetrahymena (Mochizuki et al., 2002)
and Stylonychia (Fetzer, 2002). In all studied Bilateria, the Piwi
protein is crucial for germ stem cell maintenance and division as
well as somatic stem cell maintenance because it silences gene
expression by RNA interference (Lin and Spradling, 1997; Cox
et al., 1998; Lingel and Sattler, 2005).
In the Clytia TBE, Piwi was expressed not only in stem cells
but during the first steps of nematogenesis, as show by
coexpression with mcol3-4a in the β region and in a few
isolated cells in the α region. Coexpression of Piwi and mcol3-
4a was extended in two symmetrical patches located laterally in
the proximal part of the β region. The Piwi staining was more
diffuse in the central part of the proximal β area, where it also
partly overlapped with mcol3-4a pattern. Coexpression of Piwi
and mcol3-4a suggests that after commitment, early nemato-
blasts initially maintain Piwi expression but then progressively
lose it. This could reflect a progressive loss of stem cell
potential. The wider staining in the lateral areas compared to the
centre could be explained by the number of Piwi+/mcol3-4a−
cells that is more important in the lateral areas, consistent with
the two patches observed on semi-thin sections (Fig. 3I, see also
Supplementary data). In consequence, the number of proliferat-
ing and differentiating cells is higher in the lateral areas.
An antagonistic Frizzled family receptor CheFz3 is also
expressed in lateral parts of the tentacle bulb base, while the
classical Frizzled CheFz1 is expressed more widely (Momose
and Houliston, 2007). The Wnt pathway may thus be involved
in regulating early steps of Clytia nematogenesis, for instance
the balance between proliferation and commitment of nemato-
blast precursors, since the Wnt pathway has been implicated in
the control of interstitial cell proliferation in Hydractinia (Teo et
al., 2006)
Nematocyte differentiation markers in the TBE
We showed that the expression of four homologues of Hydra
nematocyte differentiation markers was conserved in Clytia
tentacle bulb nematogenesis, and that their relative expression
patterns were consistent with our spatio-temporal model. mcol1
and mcol3-4a were mostly expressed in the β crescent-shaped
differentiation area. Dkk3 almost totally overlapped with
mcol3-4a, whereas NOWA was expressed in a subset of
mcol3-4a-positive cells, mainly in the distal part of the β
area, comprising a majority of late differentiating nematoblasts
with a well-formed but still immature capsule.
This mRNA distribution in differentiating nematoblasts was
coherent with published data on NOWA and Dkk3 expression
timing and on NOWA function, even if the absence of published
double in situ hybridizations with minicollagens precludes
estimation of the exact overlap between the expression of these
genes in Hydra. In Hydra, the NOWA protein was shown to be
involved in minicollagen condensation and cross-linking (Engel
et al., 2002), leading to capsule wall maturation, at the end of
the nematogenic process. Consistently, NOWA and mcol3-4a
mRNAs are mostly coexpressed in the distal differentiation
area, and more proximally than the TRITC-stained nematocysts
(revealing nematoblasts undergoing the wall maturation step).
110 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 13

The Dkk3 gene belongs to the Dkk3 Dickkopf subfamily.
Contrary to Dkk 1, 2 and 4, Dkk3 genes have not been shown to
be implied in the Wnt pathway, and their function remains to be
investigated (Fedders et al., 2004). The Hydra Dkk3 gene is
expressed in differentiating nematoblasts in the body column, a
pattern that parallels the Clytia staining in the β zone, which
contains a majority of differentiating stages. In Hydra, Dkk3 is
also expressed in isolated, mature migrating nematocytes near
the tentacle. In Clytia, these cells would be predicted to lie in
the γ zone, but no Dkk3-positive cells were detected in this
region, indicating that Dkk3 expression in Clytia stops earlier
than in Hydra.
The transcription factor-coding FoxB gene, implicated in
neurogenesis in vertebrates, has also been reported to be
expressed in the proximal β area in a crescent-shaped pattern
(Chevalier et al., 2006). As Zic or CnASH, the expression of
FoxB in nematoblasts (if it is confirmed) would thus be
reminiscent of the expression of FoxB during neurogenesis in
Bilateria (Odenthal and Nusslein-Volhard, 1998; Gamse and
Sive, 2001; Mazet and Shimeld, 2002). Moreover, FoxB was
also expressed in several Clytia medusa sense organs (Chevalier
et al., 2006): the statocysts, gravity-sensing organs linked to the
neural coordination system of hydromedusae as well as in the
larval endoderm, and possibly in the gonad photoreceptors. In
addition, it is expressed in the endoderm of the planula, in
particular at the posterior pole, where the nematocytes and
neurons are committed. Therefore, the expression of FoxB in
cnidarian neuron-related cells could represent an ancestral
legacy; alternatively, the gene could have been secondarily
recruited.
No differentiation markers have yet been found to be
specifically expressed in the γ region. In contrast the maturation
markers TRITC and DAPI were able to stain nematoblasts and
nematocytes in this area. Weber (1995) observed that in Hydra
nematocysts lose their TRITC staining ability precisely when
the final osmotic pressure is reached in the matrix, i.e. when
PGA accumulation has reached its maximum. Our results are
consistent with this observation because we observed (i) an area
of late maturating nematoblasts stained with both TRITC and
DAPI, and containing a maturating cyst wall and a poly-γ-
glutamate-filling matrix (the γ area), and (ii) more distally in the
rest of the tentacle (the δ area), only mature nematocytes whose
cyst wall had lost the ability to fix TRITC but whose cyst matrix
was brightly stained with DAPI.
minicollagen and NOWA in situ hybridizations detected a
few isolated positive cells within some of the tentacles,
suggesting that a cell migration occurs before the end of
differentiation, whereas in Hydra nematocytes migrate only
once differentiated. We propose two hypotheses to explain this
difference. First, these cells may be microbasic-b-mastigophore
nematoblasts that are carried into the tentacle with the general
massive cell movement, together with the fully differentiated
cells. Alternatively, these isolated undifferentiated cells and the
differentiated nematocytes that continuously colonize the
tentacle may represent two distinct cell populations correspond-
ing to two nematocyte types. Supporting this hypothesis, the
Fig. 7. General model of nematogenesis in the Clytia TBE. The nematogenesis stages are distributed in a proximo-distal succession. Gene expression patterns as well
as cellular characteristics are shown. Dashes represent minority staining of scattered cells.
111 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 14

isolated cells were slightly bigger and rounder than other
nematoblasts and could thus belong to the atrichous isorhiza
category. Although in Hydra different nematoblasts populations
cannot be distinguished by differential minicollagen mRNA
expression (Kurz et al., 1991), differential migration pathways
were described among nematocyte categories. Stenoteles start to
migrate after a “lag-phase” not observed for atrichous isorhiza
(Weber, 1995), which is reminiscent of what we observed in
Clytia.
Advantages of the Clytia TBE as an experimental model for
nematogenesis
The characteristics of nematogenesis in the tentacle bulb
ectoderm of Clytia open unique experimental perspective for
future studies of gene expression and regulation. In our model,
the existence of a spatial progression of nematoblast stages
along the bulb axis should greatly facilitate the identification of
new genes involved in the various phases of nematogenesis
(Fig. 7).
Because nematogenesis progresses from basis to tip of the
bulb, the time frame in which a gene is expressed during the
process is spatially integrated in the form of a crescent-shaped
expression zone on the inner TBE. The lower and upper limits
of this band along the bulb axis can in turn be taken as a rough
indication about the time frame of the gene expression during
nematogenesis (e.g. early, middle, late). Such property should
be particularly helpful to develop systematic approaches, e.g. in
situ screenings for identification of genes involved in the
various phases of nematogenesis, or microarrays to compare
transcriptomes at various levels of the TBE.
An additional useful property of the bulb is its capacity to
survive and to regenerate tentacles and nematoblasts in culture
following excision from the medusa in only sea water for
several days (personal observation). This ability is likely to rely
on the food digestion and storage capacity of the underlying
endoderm (Bouillon, 1994). This feature would offer the
opportunity to work on an individual, autonomous organ
isolated from the influence of the rest of the organism and to
restrict the effect of functional experiments to this area. Future
attempts to investigate functional aspects of molecular regula-
tion during nematogenesis could take advantage of the unique
experimental opportunities offered by the tentacle bulb model.
Indeed, the function of studied genes during nematogenesis
could be tested and targeted by direct injection of dsRNA in the
endodermal cavity or siRNA lipofection directly to the
ectodermal surface.
Acknowledgments
We thank Pierrette Lamarre for technical help. We also
specially thank Michel Vervoort for kindly providing the anti-
phosphoH3 antibody; Aldine Amiel and Evelyn Houliston for
the phalloidin staining; David Montero for help with SEM
preparations and observation; Ghislaine Frébourg for Epon
inclusion protocol; Claire Fayet for help with semi-thin and
ultra-thin sections, coloration and TEM observation.
We would like to specially thank Evelyn Houliston for lab
facilities, improving the English and, together with Eric
Quéinnec, for their useful comments and critical reading of
the manuscript. We also thank Muriel Jager and Roxane Chiori
for fruitful discussions and advice on technical and theoretical
aspects of our work. This work was supported by grants from
the French Ministry of Research (ACI jeunes chercheurs), and
a grant from the GIS “Institut de la Génomique Marine”—
ANR blanche NT_NV_52. We thank the Consortium National
de Recherche en Génomique and the Genoscope (Evry,
France) for sequencing ESTs from C. hemisphaerica.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.ydbio.2007.12.023.
References
Bode, H.R., 1988. Control of nematocyte differentiation in Hydra. In:
Hessinger, D.A., Lenhoff, H.M. (Eds.), Biology of Nematocysts. Academic
Press, Inc. Dan Diego, pp. 209–232.
Bode, H.R., 1996. The interstitial cell lineage of hydra: a stem cell system that
arose early in evolution. J. Cell Sci. 109, 1155–1164.
Bode, H.R., David, C.N., 1978. Regulation of a multipotent stem cell, the
interstitial cell of hydra. Prog. Biophys. Mol. Biol. 33, 189–206.
Boelsterli, U., 1977. An electron microscopic study of early developmental
stages, myogenesis, oogenesis and cnidogenesis in the anthomedusa,
Podocoryne carnea M. Sars. J. Morphol. 154, 259–289.
Bouillon, J., 1994. Classe des Hydrozoaires. In: Grassé, P.-P. (Ed), Traité de
Zoologie, Cnidaires, Cténaires, 3(2), Masson, Paris, pp. 174, 202, 205, 206.
Brinkmann, M., Oliver, D., Thurm, U., 1996. Mechanoelectric transduction in
nematocytes of a hydropolyp (Corynidae). J. Comp. Phys. 178, 125–138.
Brusca, R.C., Brusca, G.J., 2003. Invertebrates. Sinauer Associates, Sunderland.
Burnett, A.L., 1968. The acquisition, maintenance, and liability of the
differentiated state in hydra. In: Hagens, H.-W. (Ed.), Results and Problems
in Cell Differentiation. Springer-Verlag, Berlin, pp. 109–127.
Campbell, R.D., David, C.N., 1974. Cell cycle kinetics and development of
Hydra attenuata. II. Interstitial cells. J. Cell Sci. 16, 349–358.
Chapman, D.M., 1974. Cnidarian histology. In: Muscatine, L., Lenhoff, H.M.
(Eds.), Coelenterate Biology: Reviews and Perspectives. Academic Press,
New York, pp. 1–92.
Chevalier, S., Martin, A., Leclère, L., Amiel, A., Houliston, E., 2006. Polarised
expression of FoxB and FoxQ2 genes during development of the hydrozoan
Clytia hemisphaerica. Dev. Genes Evol. 216, 709–720.
Cikaluk, D.E., Tahbaz, N., Hendricks, L.C., DiMattia, G.E., Hansen, D.,
Pilgrim, D., Hobman, T.C., 1999. GERp95, a membrane-associated protein
that belongstoafamilyof proteinsinvolvedinstemcell differentiation.Mol.
Biol. Cell. 10, 3357–3372.
Collins, A.G., Schuchert, P., Marques, A.C., Jankowski, T., Medina, M.,
Schierwater, B., 2006. Medusozoan phylogeny and character evolution
clarified by new large and small subunit rDNA data and an assessment of the
utility of phylogenetic mixture models. Syst. Biol. 55, 97–115.
Cox, D.N., Chao, A., Baker, J., Chang,L., Qiao, D., Lin, H., 1998. A novel class
of evolutionarily conserved genes defined by piwi are essential for stem cell
self-renewal. Genes Dev. 12, 3715–3727.
David, C.N., Challoner, D., 1974. Distribution of interstitial cells and
differentiating nematocytes in nests of Hydra attenuata. Am. Zool. 14,
537–542.
David, C.N., Gierer, A., 1974. Cell cycle kinetics and development. of Hydra
attenuata. III. Nerve and nematocyte differentiation. J. Cell Sci. 16,
359–375.
David, C.N., Murphy, S., 1977. Characterization of interstitial stem cells in
hydra by cloning. Dev. Biol. 58, 372–383.
112 E. Denker et al. / Developmental Biology 315 (2008) 99–113

Page 15

Engel, U., Özbek, S., Streitwolf-Engel, R., Petri, B., Lottspeich, F., Holstein,
T.W., 2002. Nowa, a novel protein with minicollagen Cys-rich domains, is
involved in nematocyst formation in Hydra. J. Cell. Sci. 115, 3923–3934.
Fedders, H., Augustin, R., Bosch, T.C., 2004. A Dickkopf-3-related gene is
expressed in differentiating nematocytes in the basal metazoan Hydra. Dev.
Genes Evol. 214, 72–80.
Fetzer, C.P., 2002. A PIWI homolog is one of the proteins expressed exclusively
during macronuclear development in the ciliate Stylonychia lemnae. Nucleic
Acids Res. 30, 4380–4386.
Fujisawa, T., Nishimiya, C., Sugiyama, T., 1986. Nematocyte differentiation in
hydra. Curr. Top. Dev. Biol. 20, 281–290.
Galliot, B., Schmid, V., 2002. Cnidarians as a model system for understanding
evolution and regeneration. Int. J. Dev. Biol. 46, 39–48.
Gamse, J.T., Sive, H., 2001. Early anteroposterior division of the presumptive
neurectoderm in Xenopus. Mech. Dev. 104, 21–36.
Gauchat, D., Escriva, H., Miljkovic-Licina, M., Chera, S., Langlois, M.C.,
Begue, A., Laudet, V., Galliot, B., 2004. The orphan COUP-TF nuclear
receptors are markers for neurogenesis from cnidarians to vertebrates. Dev.
Biol. 275, 104–123.
Grens, A., Mason, E., Marsh, J.L., Bode, H.R., 1995. Evolutionary conservation
of a cell fate specification gene: the Hydra achaete-scute homolog has
proneural activity in Drosophila. Development 121, 4027–4035.
Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to
estimate large phylogenies by maximum likelihood. Syst. Biol. 52,
696–704.
Hansen, G.N., Williamson, M., Grimmelikhuijzen, C.J., 2000. Two-color
double-labeling in situ hybridization of whole-mount Hydra using RNA
probes for five different Hydra neuropeptide preprohormones: evidence for
colocalization. Cell Tissue Res. 301, 245–253.
Hausmann, K., Holstein, T.W., 1985. Sensory receptor with bilateral
symmetrical polarity. Naturwissenschaften 72, 145–146.
Hemmrich, G., Bosch, T.C.G., 2007. Compagen—a comparative genomics
platform for basal Metazoa. Manuscript in preparation.
Holstein, T., 1981. The morphogenesis of nematocytes in Hydra and Forskålia:
an ultrastructural study. J. Ultrastruct. Res. 75, 276–290.
Holstein, T., Tardent, P., 1984. An ultrahigh-speed analysis of exocytosis:
nematocyst discharge. Science 223, 830–833.
Kass-Simon, G., Scappaticci, A.A., 2002. The behavioral and developmental
physiology of nematocysts. Can. J. Zool. 80, 1772–1794.
Kurz, E.M., Holstein, T.W., Petri, B.M., Engel, J., David, C.N., 1991. Mini-
collagens in hydra nematocytes. J. Cell. Biol. 115, 1159–1169.
Lentz, T.L., 1966. The Cell Biology of Hydra. North-Holland Publishing Co,
Amsterdam.
Lin, H., Spradling, A.C., 1997. A novel group of pumilio mutations affects the
asymmetric division of germline stem cells in the Drosophila ovary.
Development 124, 2463–2476.
Lindgens, D., Holstein, T.W., Technau, U., 2004. Hyzic, the Hydra homolog of
the zic/odd-paired gene, is involved in the early specification of the sensory
nematocytes. Development 131, 191–201.
Lingel, A., Sattler, M., 2005. Novel modes of protein–RNA recognition in the
RNAi pathway. Curr. Opin. Struct. Biol. 15, 107–115.
Mazet, F., Shimeld, S.M., 2002. The evolution of chordate neural segmentation.
Dev. Biol. 251, 258–270.
Miljkovic-Licina, M., Gauchat, D., Galliot, B., 2004. Neuronal evolution:
analysis of regulatory genes in a first-evolved nervous system, the hydra
nervous system. Biosystems 76, 75–87.
Miljkovic-Licina, M., Chera, S., Ghila, L., Galliot, B., 2007. Head regeneration
in wild-type hydra requires de novo neurogenesis. Development 134,
1191–1201.
Mochizuki, K., Fine, N.A., Fujisawa, T., Gorovsky, M.A., 2002. Analysis of a
piwi-related gene implicates small RNAs in genome rearrangement in
Tetrahymena. Cell 110, 689–699.
Momose, T., Houliston, E., 2007. Two oppositely localised frizzled RNAs as
axis determinants in a cnidarian embryo. PLoS Biol. 4 (e70), 889–899.
Müller, P., Seipel, K., Yanze, N., Reber-Müller, S., Streitwolf-Engel, R.,
Stierwald, M., Spring, J., Schmid, V., 2003. Evolutionary aspects of
developmentally regulated helix–loop–helix transcription factors in striated
muscle of jellyfish. Dev. Biol. 255, 216–229.
Nüchter, T., Benoit, M., Engel, U., Özbek, S., Holstein, T.W., 2006.
Nanosecond-scale kinetics of nematocyst discharge. Curr. Biol. 16,
R316–R318.
Odenthal, J., Nusslein-Volhard, C., 1998. Fork head domain genes in zebrafish.
Dev. Genes Evol. 208, 245–258.
Östman, C., 1979. Nematocysts of the Phialidium medusae of Clytia
hemisphaerica (Hydrozoa, Campnulariidae) studied by light and electron
microscopy. Zoon. 7, 125–142.
Qiao, D., Zeeman, A.M., Deng, W., Looijenga, L.H., Lin, H., 2002. Molecular
characterization of hiwi, a human member of the piwi gene family whose
overexpression is correlated to seminomas. Oncogene 21, 3988–3999.
Rebscher, N., Zelada-Gonzalez, F., Banisch, T.U., Raible, F., Arendt, D., 2007.
Vasa unveils a common origin of germ cells and of somatic stem cells from
the posterior growth zone in the polychaete Platynereis dumerilii. Dev. Biol.
306, 599–611.
Reddien, P.W., Oviedo, N.J., Jennings, J.R., Jenkin, J.C., Sanchez Alvarado, A.,
2005. SMEDWI-2 is a PIWI-like protein that regulates planarian stem cells.
Science 310, 1327–1330.
Rossi, L., Salvetti, A., Lena, A., Batistoni, R., Deri, P., Pugliesi, C., Loreti, E.,
Gremigni, V., 2006. DjPiwi-1, a member of the PAZ-Piwi gene family,
defines a subpopulation of planarian stem cells. Dev. Genes Evol. 216,
335–346.
Seipel, K., Yanze, N., Schmid, V., 2004. The germ line and somatic stem cell
gene Cniwi in the jellyfish Podocoryne carnea. Int. J. Dev. Biol. 48, 1–7.
Slautterback, D.B., Fawcett, D.W., 1959. The development of the cnidoblasts of
hydra. J. Biophys. Biochem. Cytol. 5, 441–452.
Spring, J., Yanze, N., Middel, A.M., Stierwald, M., Groger, H., Schmid, V.,
2000. The mesoderm specification factor twist in the life cycle of jellyfish.
Dev. Biol. 228, 363–375.
Szczepanek,S.,Cikala,M.,David,C.N.,2002.Poly-gamma-glutamatesynthesis
duringformationofnematocystcapsulesinHydra.J.CellSci.115,745–751.
Tan, C.H., Lee, T.C., Weeraratne, S.D., Korzh, V., Lim, T.M., Gong, Z., 2002.
Ziwi, the zebrafish homologue of the Drosophila piwi: co-localization with
vasa at the embryonic genital ridge and gonad-specific expression in the
adults. Mech. Dev. 119 (Suppl 1), S221–S224.
Teo, R., Möhrlen, F., Plickert, G., Müller, W.A., Frank, U., 2006. An
evolutionary conserved role of Wnt signaling in stem cell fate decision.
Dev. Biol. 289, 91–99.
Thurm, U., Brinkmann, M., Golz, R., Holtmann, M., Oliver, D., Sieger, T.,
2004. Mechanoreception and synaptic transmission of hydrozoan nemato-
cytes. Hydrobiologia 530–531, 97–105.
Weber, J., 1995. Novel tools for the study of development, migration and
turnoverofnematocytes(cnidarianstingingcells).J.Cell.Sci.108,403–412.
Westfall, J.A., 1966. The differentiation of nematocytes and associated
structures in Cnidaria. Cell Tissue Res. 75, 381–403.
Westfall, J.A., 1996. Ultrastructure of synapses in the first-evolved nervous
systems. J. Neurocytol. 25, 735–746.
Whelan, S., Goldman, N., 2001. A general empirical model of protein evolution
derived from multiple protein families using a maximum-likelihood
approach. Mol. Biol. Evol. 18, 691–699.
113 E. Denker et al. / Developmental Biology 315 (2008) 99–113