R E S E A R CH A R T I C L E
Developmental architecture of the nervous system in Themiste
lageniformis (Sipuncula): New evidence from confocal laser
scanning microscopy and gene expression
Allan M. Carrillo-Baltodano1 | Michael J. Boyle2 | Mary E. Rice2 | Néva P. Meyer1
1Biology Department, Clark University,
Worcester, Massachusetts
2Smithsonian Institution, Smithsonian Marine
Station at Fort Pierce, Fort Pierce, Florida
Correspondence
Néva P. Meyer, Biology Department, Clark
University, Worcester, MA 01610.
Email: nmeyer@clarku.edu
Funding information
American Museum of Natural History, Grant/
Award Number: Lerner-Gray Fund; National
Museum of Natural History, Grant/Award
Number: Short-Term Visitor Award
Abstract
Sipuncula is a clade of unsegmented marine worms that are currently placed among
the basal radiation of conspicuously segmented Annelida. Their new location provides a
unique opportunity to reinvestigate the evolution and development of segmented body
plans. Neural segmentation is clearly evident during ganglionic ventral nerve cord (VNC) for-
mation across Sedentaria and Errantia, which includes the majority of annelids. However,
recent studies show that some annelid taxa outside of Sedentaria and Errantia have a med-
ullary cord, without ganglia, as adults. Importantly, neural development in these taxa is
understudied and interpretation can vary widely. For example, reports in sipunculans range
from no evidence of segmentation to vestigial segmentation as inferred from a few pairs of
serially repeated neuronal cell bodies along the VNC.We investigated patterns of pan-neu-
ronal, neuronal subtype, and axonalmarkers using immunohistochemistry andwholemount
in situ hybridization (WMISH) during neural development in an indirect-developing
sipunculan, Themiste lageniformis. Confocal imaging revealed two clusters of 5HT+ neurons,
two pairs of FMRF+ neurons, and Tubulin+ peripheral neurites that appear to be serially
positioned along the VNC, similar to other sipunculans, to other annelids, and to spiralian
taxa outside of Annelida. WMISH of a synaptotagmin1 ortholog in T. lageniformis (Tl-syt1)
showed expression throughout the centralized nervous system (CNS), including the VNC
where it appears to correlate with mature 5HT+ and FMRF+ neurons. An ortholog of elav1
(Tl-elav1) showed expression in differentiated neurons of the CNS with continuous expres-
sion in the VNC, supporting evidence of a medullary cord, and refuting evidence of ontoge-
netic segmentation during formation of the nervous system. Thus, we conclude that
sipunculans do not exhibit any signs ofmorphological segmentation during development.
K E YWORD S
Annelida, elav, nervous system, Sipuncula, synaptotagmin
1 | INTRODUCTION
Across prominent spiralian groups (annelids, mollusks, flatworms,
nemerteans), there is notable body-plan diversity. Despite this
diversity, there are also some shared features in centralized nervous
system (CNS) architecture, including an anterior brain, pairs of
circumesophageal connectives, and one or more longitudinal nerve
cords within developmental and adult stages (Hay-Schmidt, 2000;
Hejnol & Lowe, 2015; Martín-Durán et al., 2018; Schmidt-Rhaesa,
2007). In order to determine whether such similarities reflectAllan M. Carrillo-Baltodano and Michael J. Boyle contributed equally to this study.
Received: 6 June 2019 Revised: 19 July 2019 Accepted: 1 August 2019
DOI: 10.1002/jmor.21054
Journal of Morphology. 2019;1–23. wileyonlinelibrary.com/journal/jmor © 2019 Wiley Periodicals, Inc. 1
common ancestry (homology), convergence, or another evolution-
ary path, it is essential that we investigate the developmental
formation of the CNS with a combination of modern genetic and
morphological techniques in different spiralians, and at multiple life
history stages. These analyses enable us to visualize characteristics
during development that are otherwise undetectable through
exclusive comparisons of adult morphology. In this context, the
ventral nerve cord (VNC) as a component of the CNS is proving to
be highly informative.
Annelids (bristle worms, earthworms, and leeches) are a spiralian
taxon with segmented body plans and a predominant VNC with vari-
able composition (Hejnol & Lowe, 2015; Müller, 2006; Purschke,
Bleidorn, & Struck, 2014). Within Annelida, the ancestral nervous sys-
tem was historically inferred to include a rope-ladder-like ventral cord
(Bullock & Horridge, 1965; Orrhage & Müller, 2005; Purschke, 2002),
with bilateral pairs of iterated ganglia joined by longitudinal connec-
tives and transverse commissures (Richter et al., 2010). However,
recent work has highlighted a diversity of neural arrangements among
adult annelids, including species in some taxa with medullary cords
and others with ventral cords containing intraepidermal neurons
(Helm et al., 2018; Purschke, 2016; Purschke et al., 2014). And within
the last decade, a series of phylogenomic studies (Andrade et al., 2015;
Helm et al., 2018; Struck et al., 2011, 2015; Weigert et al., 2014;
Weigert & Bleidorn, 2016) have produced new hypotheses for alterna-
tive clade arrangements, followed by subsequent ancestral character
state reconstructions of annelid VNC architecture suggesting that rope-
ladder-like arrangements may have evolved more than once, through
convergence (Helm et al., 2018). These efforts raise an interesting pos-
sibility that evolution of the VNC within Annelida was more compli-
cated than previously thought, and further highlights the need to
explore the development of neural architecture in additional taxa.
Molecular phylogenies continue to gain support for the placement
of unsegmented sipunculan worms within the overtly segmented
annelids (Dordel, Fisse, Purschke, & Struck, 2010; Rousset, Pleijel,
Rouse, Erséus, & Sidall, 2007; Struck et al., 2007; Zrzavý, Pavel,
Piálek, & Janouškovec, 2009), and more recently, in a sister-group
relationship to Errantia plus Sedentaria (Andrade et al., 2015; Helm
et al., 2018; Struck et al., 2011, 2015; Weigert et al., 2014; but see
Parry, Edgecombe, Eibye-jacobsen, & Vinther 2016; Parry, Tanner, &
Vinther 2014; Sperling et al. 2009). Historically, phylogenetic analyses
based upon comparative morphology suggested sipunculan affinities
among a variety of different animal groups, including holothurians,
echiurans, sternaspids, and priapulids, other echinoderms, and with
annelids and/or molluscs (see Schulze et al., 2005). The Sipuncula
have also been placed as the sister group to Annelida, and as a distinct
metazoan phylum (Brusca & Brusca, 2003; Hyman, 1959; Stephen &
Edmonds, 1972; Young, Sewell, & Rice, 2002). Different phylogenetic
positions of Sipuncula using morphological characters emphasize how
unique their body plan is when compared with other taxa (for reference,
see Boyle & Rice, 2014), including the rest of Annelida. Challenges asso-
ciated with sipunculan affinity reflect notable disagreements about the
origins and definition of “segmentation” (Budd, 2001; Graham, Butts,
Lumsden, & Kiecker, 2014; Hannibal & Patel, 2013; Scholtz, 2002;
Seaver, 2003; Seaver & Kaneshige, 2006). Currently, genetic, develop-
mental, and morphological characteristics of segmentation in Annelida,
Arthropoda, and Chordata are not considered homologous. And repeti-
tive structures alone have not been recognized as “true” segmentation
(Budd, 2001; Scholtz, 2002; Seaver, 2003). Here, we combine definitions
as presented by Scholtz (2002) and Seaver (2003) that consider an ani-
mal's body plan as segmented along the anterior–posterior axis when it
is composed of repeated structural units of tissues and organ systems
that develop from both ectoderm and mesoderm. Accordingly, the most
recent phylogenetic position of Sipuncula makes them a pivotal group
with respect to studies on the evolution and development of segmented
body plans in general, and nervous systems in particular.
Previously, larval and adult sipunculan nervous systems have been
described as a medullary cord, that is, an unpaired VNC lacking dis-
tinct clusters of neuronal somata and segmentally arranged commis-
sures (Åkesson, 1958; Helm et al., 2018; Rice, 1985). However, in two
sipunculan species, Phascolosoma agassizii and Themiste pyroides,
Kristof, Wollesen, Maiorova, and Wanninger (2011) and Kristof,
Wollesen, and Wanninger (2008) reported what they suggested were
serotonergic-positive (5HT+) “metamerically arranged perikarya”
within the VNC. This pattern led the authors to suggest that during
development, sipunculan nervous systems would “recapitulate” a pre-
sumed segmental organization of the last common ancestor of anne-
lids. However, based on the hypothesis that the last common
ancestor of Annelida may not have possessed a rope-ladder-like VNC
(Helm et al., 2018), additional sipunculan taxa should be carefully
examined to characterize the number, arrangement, and relative posi-
tions of neuronal somata, commissures, and related developmental
evidence that may support or refute previous suggestions of vestigial
segmentation within Sipuncula.
Here, we examined the developing nervous system in larvae and
juveniles of the marine sipunculan worm Themiste lageniformis Baird,
1868 using cross-reactive antibodies against serotonin, Tubulin, and
FMRFamide, along with in situ detection of antisense riboprobes
against two neural genes, Tl-elav1 and Tl-syt1. Metazoan elav genes
encode RNA-binding proteins that are initially expressed in neuronal
cells as they begin the process of differentiation, with expression
maintained in mature neurons (Berger, Renner, Luer, & Technau,
2007; Pascale, Amadio, & Quattrone, 2008; Ratti et al., 2006;
Robinow & White, 1991; Wakamatsu & Weston, 1997). Orthologs of
elav have been shown to be expressed in nondividing neural cells in
the annelids Capitella teleta (Meyer & Seaver, 2009; Sur, Magie,
Seaver, & Meyer, 2017) and Platynereis dumerilii (Denes et al., 2007;
Simionato et al., 2008). Orthologs of synaptotagmin1 are Ca2+-binding
proteins that are important for exocytosis of vesicles containing neu-
rotransmitters in the presynaptic terminals of mature neurons (Rizo &
Rosenmund, 2008). Where studied, the expression of synaptotagmin1
is initiated and maintained in mature, differentiated neurons (Denes
et al., 2007; Littleton, Bellen, & Perin, 1993; Meyer, Carrillo-
Baltodano, Moore, & Seaver, 2015; Nomaksteinsky et al., 2009;
Santagata, Resh, Hejnol, Martindale, & Passamaneck, 2012; Simionato
et al., 2008). To date, the expression patterns of neural genes have
not been investigated in a member of Sipuncula. Collectively, all of the
2 CARRILLO-BALTODANO ET AL.
patterns we observed, through detection of immunohistochemical
markers and the expression of neural transcripts by whole-mount in
situ hybridization (WMISH), provide supporting evidence for a non-
ganglionated, medullary nerve cord with some iteration of cell bodies
and peripheral nerves within an unsegmented sipunculan body plan.
This work complements a growing body of comparative studies on ner-
vous system development among spiralian taxa, and enables broader
understanding of both taxonomic variabilities and commonalities within
the developmental architecture of nervous systems across Metazoa.
2 | MATERIALS AND METHODS
2.1 | Animal collection, development, and fixation
Adult specimens of Themiste lageniformis Baird, 1868 were collected from
oyster reefs within the Indian River Lagoon estuary near Fort Pierce,
FL. On the reef, specimens were extracted from cryptic spaces among
clusters of live oysters by separating them with a rock hammer, and
removing the worms by hand into containers of natural seawater. In the
laboratory, worms were placed in glass finger bowls of 0.35 μl filtered
seawater (FSW) at room temperature (22–23 
C). Adult worms were
then dissected to remove coelomic fluid containing oocytes, which were
placed into gelatin-coated plastic petri dishes and cleaned with multiple
exchanges of FSW, following protocols of Boyle and Seaver (2010).
Embryonic development was artificially activated by the treatment
of oocytes in FSW with a final concentration of 1.0 mmol/L N,20-O-
dibutyryladenosine-30:5-cyclic monophosphate (cAMP; Millipore-
Sigma, St. Louis, MI, cat #D-0627). cAMP stock solution was prepared
at 0.5 mol/L in dimethylformamide. cAMP activation was terminated
after 1.5 hr by FSW exchanges. Embryonic and larval development
was cultured in FSW with antibiotics (0.6 mg/ml penicillin; 0.5 mg/ml
streptomycin) at room temperature. At designated stages of develop-
ment, specimens were transferred from culture dishes to new, coated
plastic petri dishes for anesthetization and fixation.
Embryonic and larval stages were anesthetized following Boyle and
Seaver (2010). In brief, embryos were fixed directly in a solution of 4%
paraformaldehyde (pfa) in FSW. Postgastrula and prelarval stages were
relaxed with a combination of isotonic magnesium chloride (MgCl2) and
ethanol (EtOH, 2.5% final concentration), followed by 4% pfa fixation.
Stages with active muscular contraction were relaxed with a combina-
tion of MgCl2, EtOH, and 0.25% bupivacaine hydrochloride (BH;
Millipore-Sigma, cat# B-5274) followed by 4% pfa. For confocal laser
scanning microscopy (CLSM), fixation occurred at room temperature
and was terminated within 1.0 hr by exchanges of FSW, followed by
multiple exchanges and storage in phosphate-buffered saline (PBS). For
WMISH (see below), fixation occurred overnight at 4.0 
C followed by
termination with exchanges of FSW and PBS.
2.2 | Cloning of Tl-elav1 and Tl-synaptotagmin1
Total RNA was extracted using TriReagent (Molecular Research
Center, Inc., Cincinnati, OH) or an RNeasy Mini Kit (Qiagen, Hilden,
Germany). cDNA was generated using the SMARTer RACE kit
(Clontech, Mountain View, CA) or the Quantitect Reverse Transcrip-
tion Kit (Qiagen). A synaptotagmin 1 homolog was recovered from
T. lageniformis by degenerate PCR using the following nested
primers: forward 50-TYAAYCCNGTNTTYAAYG and reverse 50-
TCRTTRTARTANGGRTT; second forward 50-TAYGAYTTYGAYMG-
ideoxyl-TT and second reverse 50-SWRAARCADATRTC-ideoxyl-CC.
Multiple sequences were recovered from the degenerate PCR, likely
representing multiple splice variants. One 663-bp fragment
(MN193560) spanning all of the C2A domain and part of the C2B
domain was used to generate an antisense riboprobe. Homologs of
Elav were identified from assembled transcriptomes (MJBoyle;
unpublished) with Sequenceserver (Priyam et al., 2015; https://www.
sequenceserver.com) from Stages 2–9 of T. lageniformis, blastula—
pelagosphera stages in Nephasoma pellucidum, and 1–8 days of
development in Phascolion cryptum, and gene orthology was assigned
as detailed below. A 750-bp fragment for Tl-elav1 (MN207127)
was amplified with gene-specific primers: forward primer 50-
CGAGCAGTGAAAGCATAAAGG and reverse primer 50-
ACTAGCGGCAACAAATGAAGA. PCR fragments were cloned using
either a pGEM-T easy (Promega, Madison, WI) or pCRII vector
(Invitrogen) and sequenced. DIG-labeled RNA probes were generated
using the MegaScript T7 transcription kit (ThermoFisher Scientific,
Waltham, MA).
2.3 | Gene orthology analysis
To determine gene orthology of the elav genes in sipunculans, amino
acid sequences for the RRM1–3 regions minus the hinge region of
multiple bilaterian Elav orthologs, as well as multiple Bruno/CELF3–5
orthologs (outgroup) (Table S1), were aligned using muscle and edited
by hand in MEGA7 (Figure S5). Nexus files were generated using
AliView and Notepad (Microsoft, Redmond, WA). Bayesian inference
was performed with the amino acid model RtRev (ProtTest3) (Darriba,
Taboada, Doallo, & Posada, 2011; Guindon & Gascuel, 2003) using Mr
Bayes 3.2.6 (Huelsenbeck & Ronquist, 2001). Convergence was
reached after 7 × 105 generations between two simultaneous runs
with sampling every 100 generations and 4 chains. A consensus tree
with posterior probabilities representing 7.5 × 103 trees and a burn-in
of 25% was generated in FigTree v1.4.4 (Figure S4).
2.4 | Immunostaining
Relaxation and fixation were performed as above. Depending on the
stage, larvae were relaxed and fixed in either FSW or artificial sea
water at room temperature, then washed several times with PBS and
transferred into PBS + 0.2% Triton X-100 (PBT). Larvae were
incubated in blocking solution (5% heat-inactivated goat serum [Mil-
lipore-Sigma] + PBT) for 1 hr at room temperature while rocking.
Specimens were then incubated in blocking solution with the follow-
ing primary antibodies: 1:500 rabbit antiserotonin (5HT; Millipore-
Sigma, cat #S5545), 1:500 rabbit anti-FMRFamide (Immunostar,
Hudson, WI, cat# 20091), 1:600 mouse antityrosinated Tubulin (1A-2;
Millipore-Sigma, cat# T9028), and 1:10 mouse antiβ-Tubulin (E7;
CARRILLO-BALTODANO ET AL. 3
Developmental Studies Hybridoma Bank, Iowa City, IA). Serotonin-
like, FMRFamide-like and β-Tubulin-like immunoreactivity are desig-
nated as 5HT+, FMRF+, and Tubulin+ (Tub+) elements throughout this
manuscript. Secondary antibodies were administered in blocking solu-
tion at 1:300 goat anti-mouse F(ab0)2 FITC (Millipore-Sigma, cat
#F8521) and 1:600 sheep anti-rabbit F(ab0)2 Cy3 (Millipore-Sigma, cat
#C2306). DNA staining were performed by incubating the larvae in
1:1000 TO-PRO-3 (Life Technologies, Carisbad, CA, cat #T3605) in
PBS according to Meyer et al. (2015) and Meyer, Boyle, Martindale,
and Seaver (2010). Larvae labeled with propidium iodide (Millipore-
Sigma, cat# P4864) were pretreated with RNase A (Millipore-Sigma,
cat# R6513) at 1.0 mg/ml PBT for 1 hr at 37 
C, and washed with
PBT before addition of primary antibodies.
2.5 | Whole-mount in situ hybridization
After fixation, specimens were dehydrated stepwise into 100% meth-
anol and stored at −20 
C. WMISH was performed as per details in Boyle
and Seaver (2010). Specimens were hybridized for 72 hr at 65 
C with
1 ng/μL of either a DIG-labeled antisense Tl-elav1 or Tl-synaptotagmin1
RNA probe. Probes were detected by overnight treatments of 1:5,000
anti-DIG-AP (Roche, Basel, Switzerland) at 4 
C, followed by NBT/BCIP
substrate color reactions that were periodically monitored for 2–5 hr at
RT. Color reactions were stopped with PBS + 0.1% Tween-20.
2.6 | Image handling and imaging
WMISH specimens were cleared and mounted in 40% glycerol (1×
PBS), and imaged using DIC-optics on an AxioImager M2 compound
microscope (Zeiss, Oberkochen, Germany) coupled with an 18.0
megapixel EOS Rebel T2i digital camera (Canon, Tokyo, Japan). Con-
trast and brightness adjustments of immunohistochemistry and
WMISH images were performed in Photoshop CC and figure panels
were constructed using Illustrator CC (Adobe Systems, Inc.). Z-
projections were obtained from the CLSM scans using Fiji (Schindelin
et al., 2012). Multiple DIC image focal planes (.TIFF) were rendered in
Helicon Focus (HeliconSoft). Autofluorescent dust and precipitates
near, but not touching animals, were digitally removed using Photo-
shop CC version 14.0 (Adobe Systems, Inc.).
Animals imaged by CLSM were dehydrated stepwise in iso-
propanol, and cleared and mounted in 2:1 benzyl benzoate:benzyl
alcohol; or cleared and mounted in SlowFade® Gold (Life Technolo-
gies, cat# S36936). Visualization and imaging were performed with
Zeiss Axioplan 2 LSM510 or Leica TCS SP5-X technology.
3 | RESULTS
3.1 | Serotonin-like (5HT+) and Tubulin-like (Tub+)
immunoreactivity in Themiste lageniformis
A developmental staging system for Themiste lageniformis was previ-
ously developed by Boyle and Seaver (2010). Briefly, embryos develop
via unequal spiral cleavage (Stage 1) and then gastrulate by epiboly
(Stage 2) to form lecithotrophic, nonfeeding trochophore-like larvae
(Stage 4). The trochophore-like larvae elongate to form nonfeeding
pelagosphera larvae (Stage 8), which become competent to metamor-
phose into juveniles by Stage 10 (Figure 1). At Stage 5, we observed
serotonergic neurons and neurites in the developing CNS (Figure 2a–d),
with four pairs of 5HT+ neurons near the neuropil in the brain. One pair
of 5HT+ neurons is positioned dorsal to the neuropil (dS, Figure 2d),
and the other three pairs are ventral to the neuropil (vS, two pairs visi-
ble in Figure 2e). In addition to the four pairs of 5HT+ neurons in the
brain, there is one bilateral pair of flask-shaped 5HT+ cells in the dorsal-
most head region (fsS, Figure 2c–e). These cells each extend a single
neurite to the dorsal-anterior surface of the head (Figure 2d,e). The
brain connects to the VNC via paired circumesophageal connectives,
which extend in a ventral-posterior direction from the brain. One pair
of circumesophageal connectives is more ventral and one more dorsal
(double closed arrowheads, Figure 2f). In the trunk at Stage 5, the
developing VNC is visible as a concentration of cell nuclei, ranging in
thickness from one to three cells, along the ventral midline (Figure 2a,c).
Within the developing VNC, there are four 5HT+ longitudinal bundles
of neurites that extend posteriorly (Figure 2a–d) toward the developing
terminal organ. In the anterior VNC, there are two iterated clusters of
bilaterally symmetric 5HT+ cells from anterior to posterior (Figure 2b,d,
brackets in g). Each cluster consists of a dorsal and a ventral pair of
neurons (Figure 2g), and the neurons within a cluster extend a neurite
to approximately the same anterior–posterior position along the longi-
tudinal tracts in the VNC. In addition to these two clusters of 5HT+
cells, there is a more posteriorly localized pair of 5HT+ cells (asterisks,
Figure 2g) in the VNC, each of which extends a single neurite to the
respective left and right longitudinal tracts of the VNC. The neurites of
these posterior 5HT+ single neurons join the longitudinal tracts of the
VNC anterior to the insertion point of the anterior-most 5HT+ cell clus-
ter (Movie S1 and Figure S1). Because of the anterior insertion point
for these single cells, we do not think that they are part of either of the
two 5HT+ clusters in the VNC or are part of a presumptive third cluster
that is just beginning to form. In the vicinity of the first cluster of 5HT+
neurons, there are peripheral neurites on each side of the trunk that
extend dorsal-laterally from the VNC, and then turn anteriorly toward
the head (open arrowheads, Figure 2b,d,g).
At later stages of development, we found increased 5HT+ immu-
noreactivity in the head, making it difficult to determine whether
there are changes in the number of 5HT+ cells in the head. At Stage
6, there are still paired 5HT+ circumesophageal connectives and the
same number and pattern of 5HT+ neurons in the VNC as at Stage
5 (Figure 2h–j). Changes from Stages 5 to 6 are that the 5HT+ longitudi-
nal tracts of the VNC now form a “loop” just anterior to the developing
terminal organ (closed arrow, Figure 2i,n, inset 2i00). Furthermore, each
of the peripheral neurites associated with the first cluster of 5HT+ neu-
rons extend dorsal-laterally from the VNC and now branch, with one
neurite extending anteriorly toward the head, and the other neurite
extending posteriorly (open arrowhead, Figure 2i,j, inset 2i0).
We also examined Tub+ elements of the nervous system at Stage
6 (Figure 2h,j–n). Within the head, the neuropil in the brain is Tub+.
Additional Tub+ elements include the paired circumesophageal
4 CARRILLO-BALTODANO ET AL.
connectives and at least seven longitudinal tracts in the VNC, includ-
ing the four 5HT+ tracts (Figure 2h,j). There at least five Tub+ trans-
verse tracts that cross the midline in the VNC (Figure S2a–d). The
transverse neurites appear to connect the left and right longitudinal
tracts and are at the same dorsal-ventral position as the longitudinal
tracts. The number of neurites contained within each transverse tract
is unclear, but each one crosses multiple longitudinal tracts. We also
detected a total of six pairs of Tub+ peripheral nerves in the trunk (ante-
rior-most pair not shown; Figure 2h,j). Interestingly, the insertion point in
the VNC for four pairs of the peripheral nerves is at the same anterior-
posterior position as the Tub+ transverse neurites (Figure S2a–d). Five of
the six pairs of peripheral nerves connect to corresponding pairs of Tub+
epidermal organs or papillae in the trunk (Figure 2h, closed arrowheads
in j), and these are discussed in more detail below. The second pair of
peripheral nerves does not connect with papillae, but colocalizes with
the 5HT+ branching peripheral neurites that extend from the anterior-
most 5HT+ cell cluster in the VNC (compare open arrowheads in
Figure 2i,j). Additional Tub+ elements in the trunk include a dorsal pair of
Tub+ cells that are positioned between the second and third pairs of
peripheral nerves near the endoderm (open arrows, Figure 2k,n). Finally,
at Stage 6, the developing terminal organ at the most posterior end of
the larva contains six flask-shaped Tub+ cells, with three ventral and
three dorsal cells (vT and dT, respectively, Figure 2l,m). Unlike the flask-
shaped cells in the head, which are similar in morphology to sensory cells
found in apical organs of other animals, the bulbous region of the Tub+
cells in the terminal organ is apical while the narrower region is basal.
During metamorphosis from the trochophore-like stage to the
pelagosphera larva (Stages 8 and 9), there are at least 10 5HT+ neurons
in the brain (Figure 3a–d), similar to the number observed at Stage
5. The 5HT+ dorsal flask-shaped cells (fsS) are still visible and extend
neurites to the dorsal-anterior surface of the head (Figure 3d). A ring of
5HT+ neurites near the metatroch, which comprises cells bearing
motile, compound cilia, is now visible (mtr, Figure 3b). The paired
circumesophageal connectives are still 5HT+ and connect to longitudi-
nal 5HT+ tracts in the VNC that are positioned closer to the ventral
midline when compared to Stages 5 and 6 (Figure 3b vs. Figure 2b,i)
and extend toward the terminal organ (Figure 3b,h, arrow in j). The
VNC is now clearly visible by nuclear staining as a continuous band of
cells (~3–5 cells deep) that run along the ventral midline (Figure 3a,f).
Within the VNC, two clusters of 5HT+ neurons are still present
(brackets, Figure 3e); however, the number of 5HT+ neurons per cluster
is different than the four per cluster present at Stages 5 and 6. The first
5HT+ cluster has six neurons while the second 5HT+ cluster now has
only two neurons (Figure 3b,e). We did not observe formation of any
additional 5HT+ clusters in the VNC through Day 6 of development
(Figure 3b,g,i). The anterior pair of peripheral 5HT+ neurites associated
with the first cluster of 5HT+ neurons is still present at Stages 8 and
9 and there is now a new pair of 5HT+ peripheral neurites associated
with the second 5HT+ cell cluster (open arrowheads, Figure 3e,k).
Within the first 5HT+ cluster, the pair of 5HT+ peripheral neurites is
positioned with two neurons anterior and four neurons posterior.
Finally, in the pelagosphera, there are additional 5HT+ longitudinal
tracts on the lateral sides of the trunk (double open arrowheads,
Figure 3b,k). In contrast to the pattern of 5HT+ cells in the VNC of
larva, in 37-day-old juveniles, we observed an increased number of
5HT+ cells scattered throughout the VNC, with no apparent clusters
of 5HT+ neurons (Figure S3a–c).
Within pelagosphera larvae, the larval cuticle covers the trunk,
making antibody labeling difficult. In order to obtain a more distinct
staining of Tub+ elements in pelagosphera, we microdissected either
part or all of the head to enable primary and secondary antibodies to
reach the posterior end of the trunk. Using this method, we were able
to visualize innervation of the epidermal papillae by the Tub+ periph-
eral nerves (closed arrowheads, Figure 3h,i) discussed in more detail
F IGURE 1 Themiste lageniformis,
diagram of developmental stages used
during this study. Development
staging chart modified from Boyle and
Seaver (2010), showing the stages
and their corresponding ages in hours
postactivation (hpa). Stages 4 and
5 are ventral views, Stage 6 and
pelagosphera are lateral views with
ventral to the left and anterior up. cu,
cuticle; ee, egg envelope; es,
esophagus; ey, eye; in, intestine; mc,
metatroch cells; mg, midgut; mt,
metatroch; pc, prototroch cell; pt,
prototroch; st, stomodeum; to,
terminal organ; VNC, ventral
nerve cord
CARRILLO-BALTODANO ET AL. 5
below. At Stage 9, there are at least five Tub+ longitudinal tracts and
at least eight transverse neurites that cross the midline (arrows
pointing at five of them, Figure S2e). In juveniles, the longitudinal
tracts have condensed, forming two paired longitudinal cords in the
anterior half of the VNC (Figure S2f and prlt in Figure S2g) and a sin-
gle, unpaired longitudinal cord in the posterior half of the VNC
(Figure S2f and unlt in Figure S2h). We were not able to detect in
juveniles the transverse neurites present at earlier larval stages
(Figure S2f–h).
In summary, there are differentiated 5HT+ neurons and neu-
rites in the brain and VNC of T. lageniformis starting in early larval
development through juvenile stages. During larval stages, there
F IGURE 2 Legend on next page.
6 CARRILLO-BALTODANO ET AL.
are two 5HT+ clusters of neurons in the VNC with corresponding
5HT+ neurites. After metamorphosis, juveniles have more 5HT+
cells scattered along the A–P axis of the VNC, but the clusters of
5HT+ neurons in the VNC, including the two seen in larvae, are no
longer detectable.
3.2 | FMRFamidergic-like (FMRF+) immunoreactivity
in Themiste lageniformis
We used a cross-reactive, polyclonal antibody against FMRFamide to
detect FMRF-like immunoreactivity (FMRF+) in combination with anti-
Tubulin antibodies (Tub+) from Stages 5–9 in T. lageniformis (Figure 4).
At Stage 5 in the brain, we detected two flask-shaped FMRF+ cells
positioned just anterior to the neuropil (fsF; one fsF shown in
Figure 4b). We also detected smaller FMRF+ neurons near the neuro-
pil in the brain. At Stage 5, a pair of bilaterally symmetric, Tub+ cells
also are present in the most anterior region of the head (Figure 4e).
From the brain, paired FMRF+ circumesophageal connectives
(Figure 4a, double closed arrowhead in Figure 4c) extend ventrally
into the VNC where there are a pair of lateral FMRF+ longitudinal
tracts and a single median FMRF+ longitudinal tract, none of which
extend to the developing terminal organ (Figure 4a,e). Similar to the
two clusters of 5HT+ cells in the VNC, we found two pairs of FMRF+
cells along lateral sides of the anterior half of the VNC (Figure 4a,
brackets in 4d,e). Each FMRF+ cell sends an axon to the longitudinal
tract on their respective side of the animal, to the median longitudinal
tract, and to the longitudinal tract on the opposite side of the VNC at
this stage (Figure 4d). The lateral pair of FMRF+ longitudinal tracts and
two pairs of FMRF+ cells are more laterally positioned than the 5HT+
longitudinal tracts and cell clusters in the VNC. At least one peripheral
FMRF+ neurite on each side of the body extends laterally from just
anterior to the first pair of FMRF+ neurons in the VNC, and then turns
dorsal-laterally toward the head (open arrowhead, Figure 4e).
At Stage 6, due to an excess of background signal we were not
able to discern the pattern of FMRF+ cells and neurites in the head,
although in some preparations we were able to see two flask-shaped
FMRF+ cells, as well as two Tub+ cells in the head (Figure 4g). Within
the trunk, two pairs of FMRF+ cells continue to be visible in the ante-
rior half of the VNC (brackets, Figure 4f,g), but no additional FMRF+
neurons were observed in the VNC at Stage 6. A lateral pair and a
single median FMRF+ longitudinal tract continue to be visible at Stage
6 (Figure 4f), and these extend into the posterior end of the larva in
the region of the developing terminal organ. As seen at Stage 5, the
peripheral FMRF+ neurites extend from a position anterior to the first
pair of FMRF+ neurons and sweep toward the lateral-posterior before
turning anteriorly (open arrowhead, Figure 4f,g).
In early pelagosphera larvae, we observed approximately five
FMRF+ neurons in the head, but no flask-shaped cells. In the trunk,
the two lateral, two medial, and one median FMRF+ longitudinal tracts
in the VNC continue to be visible (Figure 4h) and now join together in
a loop at the posterior end of the body, where they appear to connect
with the dorsal side of the Tub+ terminal organ (Figure 4h,m,n). Within
the VNC, the two pairs of FMRF+ neurons in the anterior half of the
VNC continue to be visible (Figure 4h,n), and there are now three new
medial FMRF+ neurons in the posterior half of the VNC (arrows in
Figure 4n and inset). In the periphery of the trunk, the anterior (first)
pair of peripheral neurites (anterior pair of open arrowheads,
Figure 4n) is still visible, and there are two new pairs of peripheral
FMRF+ neurites that now extend from each side of the VNC (second
and third pairs of open arrowheads, Figure 4n). The second pair of
peripheral FMRF+ neurites is positioned posterior to the second pair
of FMRF+ cell bodies in the VNC while the third, more posterior pair
of peripheral FMRF+ neurites extends toward a bilaterally symmetric
pair of two dorsal-lateral FMRF+ neurons (open arrows, Figure 4h,n).
Interestingly, these peripheral FMRF+ neurons (open arrows,
Figure 4i–l) end near dorsal-lateral papillae in the epidermis, but do
not colocalize with papillar cells bearing external cilia (closed arrow-
heads, Figure 4k,l).
In summary, during early larval stages, there are two pairs of
FMRF+ cells and lateral longitudinal tracts in the VNC, in addition to
two pairs of FMRF+ peripheral neurites. Later, one more pair of
FMRF+ peripheral neurites and three FMRF+ neurons in the posterior
VNC are present.
3.3 | Peripheral nerves and the epidermal papillae in
Themiste lageniformis
Beginning at Stage 5 through Stage 6, there are six pairs of Tub+
peripheral nerves that enter/exit the VNC in the trunk, which we refer
to as pairs one through six from anterior to posterior (anterior-most
F IGURE 2 Themiste lageniformis, 5HT+ and Tub+ neural elements in prelarval stages. Images are z-stack projections from confocal laser
scanning micrographs of larvae labeled with anti-serotonin (red), anti-tyrosinated Tubulin (yellow), and nuclear stain (propidium iodide; grayscale)
or TO-PRO3 (cyan). (a,b) Ventral and (c,d) lateral views of Stage 5 specimens showing the developing br and VNC. (e–g) Cropped and magnified
images of the same animal as in (a,b), showing in (e), the fsS in the head and vS in the brain, in (f), the circumesophageal connectives (pairs of
double closed arrowheads), and in (g), the two 5HT+ clusters of cells (brackets) in the developing VNC and a posterior pair of 5HT+ cells
(asterisks). dS present in the brain are visible in (d), but not visible in (e). Transverse 5HT+ peripheral neurites in the trunk are labeled with open
arrowheads in (b,d) and (g). (h–n) Ventral views of a Stage 6 specimen showing the composition of 5HT+ and tub+ neural elements, and tub+
ciliated cells. (i0) Inset showing the left transverse 5HT+ neurite branching toward the anterior and posterior (open arrowhead), and (i00) a view of
the 5HT+ loop just anterior to the terminal organ (arrow). (j–n) Cropped magnifications of the same animal as in (h,i). Closed arrowheads in
(j) point to the five pairs of Tub+ papillae and brackets show the position of the two 5HT+ clusters. (k,n) Deeper projections showing the two Tub+
cells (open arrows). (l,m) The position of the vT and dTof the terminal organ. br, brain; dS, dorsal 5HT+ cells; dT, dorsal Tub+ cells of the terminal
organ; fsS, flask-shaped 5HT+ cells; mt, metatrochal cells; np, neuropil; st, stomodeum; to, terminal organ; VNC, ventral nerve cord; vS, ventral
5HT+ cells; vT, ventral Tub+ cells of the terminal organ. Scale bar is 50 μm in (a,h,j) and 25 μm in (e,k,l)
CARRILLO-BALTODANO ET AL. 7
F IGURE 3 Themiste lageniformis, 5HT+ and Tub+ neural elements in the pelagosphera larva. Images are z-stack projections from confocal laser
scanningmicrographs of larvae labeledwith anti-serotonin (red), anti-tyrosinated tubulin (yellow), and nuclear stain TO-PRO3 (grayscale). (a–c,e) Ventral
and (d) lateral views of Stage 9 pelagosphera larvae showing the developing brain (br) and ventral nerve cord (VNC). (b) Themetatroch (mt) andmetatrochal
5HT+ ring nerve (mtr) and one of the lateral 5HT+ longitudinal tracts (double open arrowhead). (c)Magnified views of the brain and (e) the anterior region of
the trunk from the same specimen as in (a,b). (d) Lateral magnified view of the head showing one of the pair of dorsal 5HT+ flask-shaped cells. (f,g) Lateral
views of a Stage 8 pelagosphera larva. (h,j) Ventral and (i,k) lateral views of a Stage 9 pelagospherawith the head removed. Brackets in (h,i) mark two 5HT+
clusters, and closed arrowheads point to Tub+ papillae on the surface of the trunk. These papillae are innervated by Tub+ peripheral nerves. Arrow in
(j) points to the posterior end of the 5HT+ VNC near the terminal organ. (k) Single channel from (i) showing transverse 5HT+ neurites (open arrowheads)
and a lateral 5HT+ longitudinal tract (double open arrowhead). br, brain; dS, dorsal 5HT+ cells; fg, foregut; fsS, flask-shaped 5HT+ cells; mt, metatroch; mtr,
metatrochal 5HT+ ring nerve; st, stomodeum; to, terminal organ; VNC, ventral nerve cord. Scale bar is 50 μm in (a,e,f,h,j,k) and 25 μm in (c)
8 CARRILLO-BALTODANO ET AL.
F IGURE 4 Legend on next page.
CARRILLO-BALTODANO ET AL. 9
or first pair not shown; Figure 2h,j). At Stage 6, Pairs 1 and 3–6 appear
to innervate five Tub+ papillae at ventral, lateral, and dorsal locations
throughout the epidermis in the trunk. Each papilla has one Tub+
cilium that projects externally from the body (Figure 2h, closed arrow-
heads in j). The second pair of Tub+ peripheral nerves enters/exits the
VNC near the first 5HT+ cluster and colabel with antibodies against
5HT and FMRF (discussed in more detail previously), but they do not
appear to extend to ciliated papillae. This suggests that the second
pair of Tub+ peripheral nerves is different than the other pairs of Tub+
peripheral nerves. The third and fourth pairs of Tub+ peripheral nerves
connect to the VNC at the position of the second 5HT+ cluster, while
the fifth and sixth pairs of Tub+ peripheral nerves enter/exit the VNC
at more posterior locations, and extend toward the two most poste-
rior pairs of papillae (Figure 2h,j).
The pattern of peripheral nerves is slightly different at pelagosphera
stages. There are now five instead of six pairs of Tub+ peripheral nerves;
the sixth pair is missing in the pelagosphera. The second pair of Tub+
peripheral nerves continues to colocalize with the first pair 5HT+ and
FMRF+ peripheral neurites; however, while the 5HT+ neurites end at the
5HT+ peripheral longitudinal tracts in the mid-body (Figure 3k), the sec-
ond pair of Tub+ neurites extends dorsally past this point and appears to
innervate the second pair of papillae. Likewise, the third pair of Tub+
peripheral nerves in the pelagosphera colocalizes with the second pair of
peripheral 5HT+ (Figure 3e) and FMRF+ neurites (Figure 4h). However,
neither 5HT+ nor FMRF+ neurites directly contact the third pair of papil-
lae. The fourth pair of Tub+ peripheral nerves colocalizes with the third
pair of peripheral FMRF+ neurites, which branch laterally from the VNC
toward two dorsal FMRF+ cells located near the fourth pair of papillae.
The fourth pair of Tub+ peripheral nerves, however, continues toward,
and appears to innervate the ciliated papillae; whereas, the peripheral
FMRF+ neurites appear to be associated with papillary support cells
(Figure 4i–l). Finally the fifth pair of Tub+ peripheral nerves extends a little
anterior to the terminal organ. Overall, from early to late larval develop-
ment, there are pairs of iterated Tub+ peripheral neurites that extend, in
their majority, to ciliated papillae located in the epidermis. Two of these
pairs of Tub+ peripheral neurites colocalize with the 5HT+ and FMRF+
neurites and are in phase with the 5HT+ clusters.
3.4 | Tl-elav1 expression
In order to have a more comprehensive view of nervous system devel-
opment in a sipunculan, we cloned orthologs of the pan-neuronal
markers elav and synaptotagmin1 from T. lageniformis (Tl-elav1 and Tl-
syt1) and examined their expression patterns. Previous phylogenetic
analyses of Elav family members recovered two clades, one neural
clade (Elav1) and one clade (Elav2), that appeared to be unique to
annelids and mollusks (Meyer & Seaver, 2009; Shigeno, Parnaik,
Albertin, & Ragsdale, 2015). To investigate gene orthology of these
Elav proteins in sipunculans, we performed a Bayesian analysis on
Elav proteins recovered from three sipunculan species, T. lageniformis,
N. pellucidum, and P. cryptum. Similar to previous analyses, we recov-
ered two Elav clades: one containing bilaterian neural Elav proteins
(Elav1) and a separate clade (Elav2) containing Elav proteins from
annelids and mollusks. All three sipunculan species had one gene that
clustered in the Elav1 clade and one gene that clustered in the Elav2
clade (Figure S4). We examined expression of Tl-elav1, the “neural”
elav gene.
In Stage 4, trochophore-like larvae, Tl-elav1 is expressed in the left
and right lobes of the brain (long black arrow, Figure 5a,b), including in
two flask-shaped cells in the dorsal-anterior region of each lobe (fs in
inset, Figure 5b). At this stage in the trunk, there also are two localized,
bilaterally symmetric regions of Tl-elav1 expression on either side of
the ventral midline (long white arrow, Figure 5a,b), likely the anlage of
the developing VNC. Within each of these trunk subdomains, there are
at least five Tl-elav1+ cells, all of which are subsurface (Figure 5a,b).
There are also single Tl-elav1+ cells in the posterior and dorsal-posterior
regions of Stage 4 animals (closed arrowheads, Figure 5b), which may
be neurons in the developing peripheral nervous system.
At Stage 5, Tl-elav1 continues to be expressed throughout the
developing brain (Figure 5c,d) and in subsurface cells in a bilaterally
symmetric pattern in the trunk (long white arrow, Figure 5c,d), which
likely corresponds to the developing VNC. Within each of these sub-
domains (left and right VNC), there are at least 10 subsurface Tl-elav1+
cells. There are also two new regions of Tl-elav1 expression in the
trunk, one associated with the stomodeum (double arrowhead,
Figure 5c,d), and one in the dorsal-posterior region (open arrowhead,
F IGURE 4 Themiste lageniformis, FMRF+ and Tub+ neural elements in Stages 5 and 6 and the pelagosphera. Images are z-stack projections
from confocal laser scanning micrographs of larvae labeled with anti-FMRFamide (magenta or grayscale), anti-β-Tubulin or anti-tyrosinated
Tubulin (yellow), and nuclear stain TO-PRO3 (cyan). (a) Ventral and (e) lateral views of Stage 5. (b–d) Magnified views of the same animal in (a).
(b) Magnified view of the brain showing a flask-shaped FMRF+ cell. (c) Double FMRF+ circumesophageal connectives (double closed arrowheads).
(d) Two pairs of FMRF+ cells in the VNC (brackets; one pair visible). (e) Tub+ cells in the head, and a FMRF+ peripheral neurite extending anteriorly
(open arrowhead), and a second Tub+ papilla on the right side of the trunk (closed arrowhead). Brackets label the positions of the two pairs of
FMRF+ cells in the VNC. (f) Ventral and (g) lateral views of a Stage 6 animal showing the two pairs of FMRF+ cells (brackets) in the VNC and
peripheral FMRF+ neurites (open arrowheads). (h,n) Ventral views of a Stage 8 pelagosphera larva showing two posterior FMRF+ cells (open
arrows), the second, third, and fifth pairs of Tub+ papillae (closed arrowheads), and peripheral FMRF+ neurites (open arrowheads).
(m) Magnification of the posterior end of the same animal in (h) and (n) showing the fifth pair of Tub+ papillae and the posterior FMRF+ loop near
the terminal organ. (o) Lateral view of a Stage 9 pelagosphera larva with two pairs of FMRF+ cells in the VNC (brackets). (i–l) Close-up of the inset
from (o). Open arrows in (j–l) point to one of the posterior FMRF+ cells that is not colocalized with the adjacent Tub+ neurite innervating the
papillae (closed arrowhead). fsF, flask-shaped FMRF+ cells; mtr, metatrochal FMRF+ ring nerve; st, stomodeum. Scale bar is 50 μm in (a,e,f,h,m–o)
and 25 μm in (b,i)
10 CARRILLO-BALTODANO ET AL.
F IGURE 5 Themiste lageniformis, expression of Tl-elav1 in the developing nervous system. All images are differential interference contrast
(DIC) micrographs of whole-mount in situ hybridization results. (a) Ventral and (b) lateral views showing expression in the brain (black arrows) and
the developing trunk (white arrows) of a trochophore-like Stage 4. Closed arrowheads in (b) point to expression in discrete cells on the dorsal-
posterior side. Inset in (b) shows Tl-elav1 expression in flask-shaped cells. (c) Ventral and (d) lateral views of Stage 5 larvae with expression in the
brain and developing VNC (white arrows), and around the stomodeum (double closed arrowheads). There is at least one pair of Tl-elav1+ cells on
the dorsal side, opposite to the VNC (open arrowhead). (e) Ventral and (f) lateral views of Stage 6 showing bilateral expression within the
developing VNC (short white arrows), and a pair of Tl-elav1+ cells on the posterior end (closed arrowhead). Additionally, there is expression on the
dorsal-lateral side (short white arrow), and at the expected positions of Tub+ papillae (open white arrowhead). (g) Ventral and (h) lateral views of a
Stage 8 pelagosphera larva with Tl-elav1 expression in a continuous sheet of cells in the region of the VNC (white bracket). Double closed
arrowheads point to expression associated with the stomodeum. The open white arrowhead in (h) points at the position of another presumptive
Tub+ papillae. Scale bar is 50 μm in (a,c,e,g). Asterisks mark the stomodeum
CARRILLO-BALTODANO ET AL. 11
Figure 5d). The single Tl-elav1+ cells in the posterior region are still
present at this stage (closed arrowheads, Figure 5d).
In comparison with previous stages, at Stage 6, the two well-
formed brain lobes are increased in both size and cell number and
show increased levels of Tl-elav1 expression, while the Tl-elav1+ cells
around the stomodeum were less detectable (Figure 5e,f). The devel-
oping VNC at Stage 6 is morphologically visible, and Tl-elav1 is
expressed throughout most of the VNC (Figure 5e,f). In the most
anterior part of the VNC, Tl-elav1 is expressed in bilaterally symmetric
clusters of cells that extend laterally (short white arrows, Figure 5e).
Within the anterior two-thirds of the VNC, expression is continuous,
with no sign of “breaks” that would indicate ganglionic structures or
remnants of ganglia (Figure 5e,f). In the posterior third of the VNC, Tl-
elav1 expression was less detectable; a few Tl-elav1+ cells are indicated
with a closed arrowhead in Figure 5e,f. Outside of the CNS, Tl-elav1 is
expressed in a few dorsal-lateral cells posterior to the metatroch (short
white arrow, Figure 5f). At this stage, Tl-elav1 is also expressed in two
dorsal, bilaterally symmetric single cells on either side of the animal,
near one of the epidermal papillae (white open arrowhead points to
one Tl-elav1+ cell near a papilla, Figure 5f).
In early pelagosphera larvae (Stage 8), Tl-elav1 continues to be
expressed in the brain (Figure 5g,h), in a few cells near the stomodeum
(double arrowhead, Figure 5g), and in the VNC (brackets, Figure 5g,h).
Notably, expression in the VNC appears largely continuous from ante-
rior to posterior. Finally, Tl-elav1 is expressed in a bilaterally symmet-
ric pattern in a few cells on the dorsal-posterior side of the trunk,
which may correspond to positions of some epidermal papillae at this
stage (open arrowhead, Figure 5h). Similar to larval stages, after meta-
morphosis, there is continuous expression of Tl-elav1 in the VNC
(Figure S3d–f).
3.5 | Tl-synaptotagmin1 expression
Across all stages observed, Tl-syt1 is expressed in the developing brain
and the VNC (Figure 6). This gene is also expressed near the stomo-
deum, and in cells throughout the ectoderm of the trunk. As observed
for Tl-elav1, the expression of Tl-syt1 is consistent with an expression
in neurons of the central and peripheral nervous systems, including
cells that may innervate the epidermal papillae and the developing ter-
minal organ.
At Stage 4¸ Tl-syt1 is expressed in both brain lobes (long black
arrow, Figure 6a). There are also two patches of subsurface expres-
sion on the ventral side of the trunk (white arrows, Figure 6a,b). These
two patches of expression extend transversely in “rows” across the
ventral side of the embryo and correlate with a subset of the Tl-elav1
pattern in the VNC anlage (compare Figure 6a with Figure 5a). At
Stage 5, Tl-syt1 is expressed throughout the brain and in a patch of
cells posterior to the stomodeum (double arrowhead, Figure 6c,d). In
the region of the developing VNC, Tl-syt1 continues to be expressed
in two transverse “rows”, each of which now contains up to 10 cells
(white arrows, Figure 6c,d). More posteriorly, there are Tl-syt1+ cells
along the ventral midline (open arrowhead, Figure 6c,d).
Tl-syt1 continues to be expressed in the developing brain and near
the stomodeum at Stage 6 (Figure 6e,f; double-arrowhead in 6f points
to expression near the stomodeum). In the VNC at Stage 6, the two
transverse “rows” of Tl-syt1 now appear as “finger-like” projections
that extend laterally from a central domain of expression at the ven-
tral midline (white arrows, Figure 6e,f). However, Tl-syt1 is not
expressed in the most posterior region of the forming VNC. Tl-syt1
continues to be expressed in cells at the most posterior end of the
trunk (open arrowhead, Figure 6e,f), near the developing terminal
organ. In addition, pairs of Tl-syt1+ cells were detected dorsally and
laterally along the larval body (Figure 6e,f), and at least one of the two
posterior pairs of cells is associated with a pair of epidermal papillae
(closed arrowheads, Figure 6e).
At the pelagosphera stage, Tl-syt1 continues to be expressed in
the brain (Figure 6g,h), and in cells posterior to the stomodeum (dou-
ble arrowhead, Figure 6g,h). As seen in earlier stages, Tl-syt1 does not
label the entire VNC, but is instead localized to the anterior half of the
VNC. The two “finger-like” projections of expression in the VNC are
still distinguishable (white arrows, Figure 6g,h), along with Tl-syt1+
cells near the ventral midline (Figure 6g,h). We did not detect Tl-syt1+
peripheral cells at this stage (Figure 6g,h). One reason for this could
be that the cuticle in the trunk prevented efficient probe penetration,
making visualization of single Tl-syt+ cells in the trunk difficult.
4 | DISCUSSION
In this study, we combined immunohistochemistry and confocal micros-
copy with whole-mount gene expression profiles of neuronal markers
to build upon and extend existing knowledge about the development
of sipunculan nervous system architecture. This is the first use of all
three markers (5HT+, FMRF+, and Tub+) and neuronal genes (Tl-elav1
and Tl-syt1) in any species of sipunculan marine worm. The data pres-
ented here are novel, both contrasting and complementing previous
studies of sipunculan nervous systems (Åkesson, 1958; Kristof et al.,
2008; Wanninger, Koop, Bromham, Noonan, & Degnan, 2005).
4.1 | Interpretation of the CNS architecture in
Themiste lageniformis
We took a step forward from previous studies of the nervous system
in sipunculans, which relied upon classical morphology and immuno-
histochemistry (Åkesson, 1958; Kristof et al., 2008, 2011; Wanninger
et al., 2005), to study expression of pan-neuronal genes that have
been shown to play a role in neuronal differentiation and maturation
across Metazoa (Craxton, 2010; Pascale et al., 2008; Samson, 2008).
Based on the expression of Tl-elav1 and Tl-syt1 in the head and the
ventral trunk at Stage 4, which is shortly after gastrulation, there
appear to be many differentiating and possibly mature neurons in the
developing brain and VNC. We were not able to obtain information
about the presence of 5HT+, FMRF+, and Tub+ cells and neurites at
Stage 4, but it is likely that these elements are already present. For
example, Tl-elav1 is expressed in two flask-shaped cells on the dorsal-
12 CARRILLO-BALTODANO ET AL.
F IGURE 6 Themiste lageniformis, expression of Tl-syt1 in the developing nervous system. All images are differential interference contrast
(DIC) micrographs of whole-mount in situ hybridization results. (a) Ventral and (b) lateral views of Stage 4 with expression in the brain (black
arrow) and as two transverse “fingers” on the ventral side within the developing trunk (short white arrows). (c) Ventral and (d) lateral views of
Stage 5 showing broader expression in the brain and on the ventral side (short white arrows), in a pair of cells on the ventral-posterior end (open
arrowheads), and around the stomodeum (double closed arrowheads). (e) Ventral and (f) lateral views of Stage 6 showing a more defined pattern
of Tl-syt1 expression in the brain, in transverse “fingers” across the ventral midline (short white arrows), within additional cells along the posterior
region of the ventral nerve cord, in a group of cells at the posterior end (open arrowheads), and in patches of cells on left and right lateral as well
as mid-dorsal sides of the trunk (closed arrowheads). (g) Ventral and (h) lateral views of a Stage 8 pelagosphera larva with transverse expression
on the anterior portion of the trunk (short white arrows), but no expression on the posterior half of the ventral trunk. There is also expression in
the brain and around the stomodeum at this stage (double black arrowheads). Scale bar is 50 μm in (a,c,e,g). Asterisks mark the stomodeum
CARRILLO-BALTODANO ET AL. 13
anterior side of the head at Stage 4, and these are in a similar position
to the flask-shaped 5HT+ cells we see later at Stage 5. Moreover, the
expression of Tl-syt1 in the developing VNC matches the relative
positions of the 5HT+ clusters and two posterior 5HT+ cells that we
see later in Stage 5, as well as the pairs of FMRF+ cells, providing us again
with a possibility that the trochophore-like Stage 4 already has 5HT+ and
FMRF+ cells in the trunk. Previous observations of T. lageniformis col-
lected in Hawaii found at least two pairs of 5HT+ cells in the trunk (NPM,
unpublished data) at Stage 4, but more detailed observations at Stage
4 and earlier are needed to understand when and where the first
neurons form.
The brain continues to add neurons from Stages 4–6, possibly
later. There are many Tl-elav1+ and Tl-syt1+ cells by Stages 5 and 6 ver-
sus the ~10 5HT+ and four FMRF+ cells we detected (Figure 7),
suggesting the presence of additional neural subtypes. The 5HT+ and
FMRF+ cells in the brain are arranged in a bilaterally symmetric man-
ner around a central neuropil and are connected to the VNC by paired
circumesophageal connectives, which were labeled with antibodies
against 5HT, FMRF, and Tubulin (Figure 7). Additional elements we
detected in the brain included two dorsal flask-shaped 5HT+ cells and
two dorsal flask-shaped FMRF+ cells, which were in close proximity to
each other and may contribute to an apically positioned larval organ,
similar to other apical organs present in the larvae of many other
marine invertebrates (Marlow et al., 2014; Wanninger, 2008).
In the trunk, we found evidence of a nerve cord that runs along
the ventral midline. At Stage 4, there are two bilaterally symmetric Tl-
elav1+ domains on either side of the ventral midline. As the larvae
develop, these domains appear to converge toward the ventral
midline, either by the addition of cells or by cell movement. By Stage
6, Tl-elav1 is expressed continuously in the developing VNC from
anterior to posterior and left to right. The VNC appears to be
narrower from left to right at Stage 6 versus Stages 4 and 5, and this
lengthening of the VNC occurs at the same time as the lengthening of
the overall larval body. Tl-elav1 continues to be expressed in the VNC
from anterior to posterior in Stage 8 pelagospherae and juveniles.
This, in combination with an absence of any visible “breaks” (somata-
free connectives) in the nuclear stain along the VNC (e.g., Figure 3a,
suggests that T. lageniformis does not have ganglia (concentrations of
neuronal cell bodies sensu; Richter et al., 2010) in the VNC. Further-
more, there are several Tub+ longitudinal tracts in the VNC by Stage
6, some of which are 5HT+ and some of which are FMRF+ (Figure 7).
By Stages 8 and 9, the longitudinal tracts in the VNC are closer to the
ventral midline than at earlier stages (Figure 7), and after metamor-
phosis, the tracts have condensed into two paired longitudinal cords
in the anterior half of the VNC and a single, unpaired cord in the pos-
terior half of the VNC (Figure S2f–h).
Although we did not find evidence of ganglia in the VNC during
development, we did find evidence of two 5HT+/FMRF+ cell clusters
within the VNC (Figure 7). As 5HT+ and FMRF+ neurons only repre-
sent a small fraction of cells in the nervous system, it would be inter-
esting to examine a variety of neurotransmitters in T. lageniformis to
see how other neural subtypes are arranged within the continuous tis-
sue of the VNC. Within the developing VNC, Tl-syt1 expression
overlaps spatially with Tl-elav1 expression. However, while Tl-elav1
forms a continuous expression domain from Stages 5 to 8, Tl-syt1 is
expressed in two distinct, transverse “rows” at Stages 4 and 5, and in
“fingers” that extend laterally from a medial expression domain at
Stages 6 and 8 (e.g., Figure 6e,g). Based on the relative positions of
cells, we infer that the 5HT+ clusters in the VNC are within the
domains of Tl-elav1 and Tl-syt1 expression near the ventral midline,
while the FMRF+ cells in the VNC are positioned within the more lat-
eral Tl-syt1+ “fingers.” Expression of Tl-syt1 in two transverse bands
suggests that neurons may mature in two clusters in T. lageniformis,
which is supported by the presence of two distinct groups of 5HT+
and FMRF+ cells in the VNC. In addition to the presence of these two
cell clusters, we also found evidence of at least five transverse bun-
dles of neurites in the VNC, the first of which corresponds in position
to the first cluster of neurons while the other bundles of transverse
neurites are positioned anterior and posterior to the second cluster of
neurons (Figure 7). By juvenile stages, once the longitudinal tracts have
condensed, transverse neurites are no longer visible (Figure S2f–h). Fur-
thermore, we did not observe the addition of any more clusters of
5HT+ cells or additional pairs of FMRF+ cells that appeared iterated
along the VNC in pelagosphera larvae or juveniles. Within juveniles, dis-
tinct clusters of 5HT+ cells are no longer visible, with 5HT+ neurons
scattered throughout the VNC. To further complicate this pattern, at
Stage 9, the number of cells in the first cluster of 5HT+ cells changes
from four to six, and in the second cluster of 5HT+ cells, from four to
two, suggesting the maturation of more neurons on the anterior cluster,
and possible cell death or a change in the type of neurotransmitter
expressed in the second cluster. In conclusion, the developing VNC in
T. lageniformis larvae does not exhibit a rope-ladder-like arrangement,
but instead reveals a pattern of continuous neuronal somata surround-
ing multiple, central longitudinal neurite bundles with multiple trans-
verse neurites. In juveniles, there are paired cords in the anterior half of
the VNC and a single, unpaired cord present in the posterior half of the
VNC with no detectable commissures. We find evidence of two ante-
rior clusters of neural cell types within the continuous VNC at larval
stages, but these clusters were not detectable in juveniles. Taken
together, this pattern is most similar to a medullary cord (Purschke
et al., 2014; Richter et al., 2010).
4.2 | Interpretation of the Peripheral Nervous
System (PNS) architecture in Themiste lageniformis
We here provide a detailed description of peripheral nervous system
formation in a sipunculan, including 5HT+ neurites and FMRF+ and
Tub+ cells and neurites. At the anterior end, we found a bilateral pair
of Tub+ cells located outside of the brain. These Tub+ cells are similar
to the bilateral Tub+ cells found in larvae of C. teleta (scac+) (Amiel,
Henry, & Seaver, 2013; Meyer et al., 2015) and P. dumerilii (cPRC)
(Arendt, Tessmar-raible, Snyman, Dorresteijn, & Wittbrodt, 2004;
Gühmann et al., 2015), which were proved to be photoreceptor cells
(Verasztó et al., 2018) in those annelids. We also found 5HT+ and
FMRF+ neurite rings in the pelagosphera larva correlating with the
14 CARRILLO-BALTODANO ET AL.
metatroch, but not the prototroch, and these may be involved in con-
trolling the beating of the cilia of the metatroch, which is the primary
organ of locomotion in sipunculan pelagospherae. In P. dumerilii, ciliary
beating of the prototroch was shown to be influenced by the
presence of serotonin (Verasztó et al., 2017), which may also influ-
ence cilia in the metatroch of T. lageniformis. As T. lageniformis does
not develop a ciliated prototroch, we were not able to assess whether
5HT+ and FMRF+ neurite rings correlate with that band in sipunculan
F IGURE 7 Themiste lageniformis, summary diagrams of the developing nervous system. 5HT+ (red, left), FMRF+ (magenta, center), and Tub+ (yellow,
right) elements in the developing nervous system of Stages 5 and 6 and pelagosphera larvae. cc, circumesophageal connectives; dS, dorsal 5HT+ cells;
dTub, dorsal Tub+ cells in the head; ep1–ep5, Tub+ epidermal papillae one through five; fsF, flask-shaped FMRF+ cells; fsS, flask-shaped 5HT+ cells; mt,
metatroch; mtr, metatrochal ring nerve; pn1–pn6, Tub+ peripheral neurites; pt, prototroch; to, terminal organ; vS, ventral 5HT+ cells
CARRILLO-BALTODANO ET AL. 15
trochophore larvae, a question that should be pursued in sipunculans
with trochophore larval stages.
Within the trunk by Stage 5, there are pairs of bilaterally symmet-
ric 5HT+, FMRF+, and Tub+ peripheral nerves that enter/exit the VNC
at the position of the first 5HT+ cluster and first pair of FMRF+ cells.
In the pelagosphera, we detected a second pair of bilaterally symmet-
ric 5HT+, FMRF+, and Tub+ peripheral nerves that enter/exit the VNC
from the position of the second 5HT+ cluster. In addition to these
peripheral nerves, we also observed Tub+ nerves that enter/exit the
VNC at multiple anterior–posterior locations and innervate ciliated
papillae in the epidermis of the trunk, most likely providing a connec-
tion between external cues outside the body and the CNS. The Tub+
neurites are arranged in an iterated pattern from anterior to posterior
across all the larval stages studied. It is interesting that these neurites
zig–zag laterally and reach their presumptive target (e.g., papillae),
without any mesodermal septa or mesenteries that could restrict them
to stereotypic branching patterns. An annelid-wide analysis with an
emphasis on oligochaetes proposed that the position of ganglia is
more useful than mesodermal septa to identify and describe the pat-
tern of peripheral nerves (Zattara & Bely, 2015). Our data in
T. lageniformis do not show any evidence of ganglia, but they do show
that some of the Tub+ peripheral neurites are associated with two
clusters of neurons, which leads us to consider whether this pattern
may suggest a putative morphological precursor for the formation of
ganglia, instead of a sign of vestigial segmentation as proposed by
others (Kristof et al., 2008).
4.3 | Comparison of 5HT+ and FMRF+ neural
elements among sipunculans
Within Sipuncula, descriptions of nervous system formation using
immunohistochemistry and CLSM have been published on four spe-
cies: Themiste pyroides and Thysanocardia nigra, indirect lecithotrophic
development with a pelagosphera larva (Kristof, 2011; Kristof et al.,
2011; Kristof & Maiorova, 2016), Phascolion strombus, indirect
lecithotrophic development without a pelagosphera (Wanninger et al.,
2005), and Phascolosoma agassizii, indirect planktotrophic develop-
ment with a pelagosphera (Kristof et al., 2008). Here, we only com-
pare 5HT+ and FMRF+ neural elements between those four species
and T. lageniformis, as no other reports of nuclear DNA or Tub+ are
available. Previous studies did not present replicative timelines of
species–specific developmental stages. Thus, we are cautious with
direct temporal comparisons of morphology and associated inferences
across contrasting life history patterns.
Within trochophore and pelagosphera larval stages of T. pyroides,
5HT+ elements include pairs of flask-shaped cells with anterior projec-
tions and other neurons associated with the neuropil of the brain, as
well as paired circumesophageal connectives between the brain and
VNC (Kristof, 2011; Kristof et al., 2011; Kristof & Maiorova, 2016). In
“early” pelagospherae, the pattern of 5HT+ cell bodies in the VNC is
similar to the pattern in Stage 6 of T. lageniformis: two clusters of four
cell bodies each, and a pair of posterior cells along paired 5HT+ longitu-
dinal tracts (Kristof et al., 2011). As observed in T. lageniformis, there is
also a 5HT+ ring-like metatrochal neurite in pelagospherae of
T. pyroides, and one or more peripheral 5HT+ neurites that extend from
within the trunk toward the head (Kristof, 2011). FMRF+ elements in
pelagospherae of T. pyroides include flask-shaped cells in the brain, dou-
ble circumesophageal connectives, median and lateral longitudinal
tracts in the VNC, and two pairs of cells in the VNC (Kristof &
Maiorova, 2016). The two pairs of FMRF+ cells and longitudinal tracts
are positioned laterally to 5HT+ cell clusters and longitudinal tracts in
the VNC (Kristof, 2011; Kristof et al., 2011; Kristof & Maiorova, 2016),
similar to the pattern we observed in T. lageniformis, suggesting that
development of the nervous system in both species is similar. Further-
more, 5HT+ and FMRF+ cells of the VNC in both species indicate there
are two distinct groups of cell bodies along the anterior–posterior axis
during VNC development in Themiste. Similarities between FMRF+
neural elements of T. nigra and T. lageniformis include flask-shaped
cells near the brain neuropil, paired circumesophageal connectives, one
lateral pair and one median longitudinal tract in the VNC, and two pairs
of FMRF+ cells in the VNC. Additionally, one pair of FMRF+ transverse
neurites extends from each side of the VNC near the anterior pair of
FMRF+ cells, and bends in a dorsal-anterior direction toward the head
(Kristof, 2011). Both pairs of FMRF+ cells and longitudinal tracts are
also positioned laterally to the midline in T. nigra (Kristof, 2011),
suggesting a similar spatial relationship with corresponding 5HT+ cells
and longitudinal tracts in the VNC of T. lageniformis. No confocal data
showing 5HT+ elements in T. nigra have been published.
Phascolion strombus was the first sipunculan species in which both
5HT+ and FMRF+ elements were described (Wanninger et al., 2005).
From their micrographs, it was not possible to determine the number
or arrangement of 5HT+ cells in the brain of P. strombus. However,
the brain appears to extend single 5HT+ circumesophageal connec-
tives toward the VNC, in contrast to the paired 5HT+ connectives in
Themiste. Notably, only one cluster of 5HT+ cells was detected along
the developing VNC of P. strombus, which included an associated
transverse 5HT+ neurite, leading the authors to infer an absence of
iteration or “cryptic” segmentation in the VNC of P. strombus
(Wanninger et al., 2005). In contrast, they detected two pairs of
FMRF+ cells along the VNC, which become more apparent in early
juveniles. Again, as in Themiste and T. nigra, FMRF+ cell bodies and
FMRF+ longitudinal tracts are positioned more laterally to the midline
than 5HT+ neural elements (Wanninger et al., 2005). Thus, the relative
positions of 5HT+ and FMRF+ elements along the midline appear to
be a common pattern within Sipuncula.
5HT+ and FMRF+ elements have also been investigated in
P. agassizii. In an “early larva,” Kristof et al. (2008) detected four 5HT+
cells without flask-shaped morphology in the “apical organ,” and a pair
of 5HT+ circumesophageal connectives extending from the brain
toward the developing VNC. Within the VNC, there was initially one
pair of 5HT+ cells, followed by development of a second pair in a
“slightly older larva,” With subsequent growth and elongation of the
larva, they detected two more pairs of 5HT+ cells, which the authors
collectively interpreted as a metameric pattern of four pairs of 5HT+
cells along the VNC (Kristof et al., 2008). They found one pair of
FMRF+ cells, and two or more FMRF+ transverse peripheral neurites
16 CARRILLO-BALTODANO ET AL.
branching from the VNC (Kristof et al., 2008). Based on these and
other observations, Kristof et al. (2008) suggested that sequential
pairs of 5HT+ cells and associated “commissures” reflected anterior-
to-posterior development of the rope-ladder-like VNC architecture
found within Annelida. From this, they concluded that formation of
the VNC in P. agassizii, based primarily on 5HT+ elements, represented
the ontogenetic recapitulation of an ancestrally segmented body plan
for Sipuncula. After further comparison with T. lageniformis and similar
studies of three other sipunculans discussed above, we do not follow
that conclusion as discussed by Kristof et al. (2008). Upon closer
examination of their data, we suggest that the VNC commissures may
have been misidentified and that the relative anterior–posterior
arrangements of 5HT+ cells along the VNC may be reinterpreted as
two prominent clusters, not four.
There are also histological descriptions of the developing nervous
system in Phascolion strombi, Gofingia minuta, and P. agassizii (Åkesson,
1958; Rice, 1973). Briefly, the VNC in P. strombi and G. minuta was
described to originate as a singular, unpaired median thickening of
ectoderm that undergoes delamination to produce the ventral neuro-
blasts of an unpaired, unsegmented nerve cord (Åkesson, 1958). A
similar pattern was observed by Rice (1973) during development in
P. agassizii, although in that species, the VNC is initially paired and
subsequently becomes united into a single cord within pelagosphera
larvae after 2 months of growth. Interestingly, our Tl-elav1 expression
data suggest that the VNC in T. lageniformis forms as two bilaterally
symmetric domains on either side of the ventral midline, which later
becomes one continuous neural domain along the ventral midline.
Based on development of the nervous system in T. lageniformis,
and a comparison of all sipunculan species for which similar data are
published, we propose that the last common ancestor of Sipuncula
had a larva possessing a medullary-like VNC with continuous neural
tissue from anterior to posterior and at least two clusters of neural
subtypes (including 5HT+ and FMRF+ neurons) in the anterior VNC.
Several species of sipunculans appear to have at least two neural sub-
type clusters (containing both 5HT+ and FMRF+ neurons) at the stages
examined (Kristof et al., 2008; Kristof & Maiorova, 2016; Wanninger
et al., 2005), which as noted above, has been interpreted as a vestigial
rope-ladder-like VNC. However, based on our data, these clusters form
within a continuous field of neurons in T. lageniformis. We are not aware
of any other reports using “pan-neuronal” markers in sipunculans,
suggesting that any inference of somata-free areas in the VNC is cur-
rently problematic. Additionally, the presence or absence of VNC com-
missures within Sipuncula needs to be investigated in much more detail
before any conclusions can be made. In T. lageniformis, there are Tub+
tracts that are transversely oriented and cross the midline along the VNC
(Figure S2). These neurites appear to connect particular left and right lon-
gitudinal tracts, although in some cases they extend peripherally from
one medial bundle within the VNC toward lateral sides of the trunk
body. In general, several transverse tracts are at a similar dorsal-ventral
position as the longitudinal tracts. The number of neurites contained
within each transverse tract is unclear; however, some do cross the
longitudinal tracts. We remain cautious as to whether these trans-
verse neurites represent definitive commissures.
In adult Sipuncula, decades of histological studies of their VNCs
suggest a consistent neuroanatomical pattern, that is, the VNC com-
prises a ventral aggregation of cells with a dorsal core of tracts (neuro-
pil), all of which is enclosed in a peritoneal covering of connective
tissue (Bullock & Horridge, 1965; Gerould, 1938; Hilton, 1917; Rice,
1973, 1993; von Mack, 1902). Interestingly, Clark (1969) pointed out
that such a distribution of neurons, being positioned ventral to the neu-
ropil mass in Sipuncula, is in contrast to what is observed in segmented
invertebrates, where the ganglionic neurons “invariably form a rind”
around a centralized neuropil along the VNC. Adult VNCs in Sipuncula
are subepidermally located within the coelom along the body wall of
the trunk (Åkesson, 1958; Rice, 1993), with one exception. Rice (1973)
noted that the cord is transiently located within the epidermal layer
during early development in P. agassizii, which may reflect Åkesson's
(1958) observations on “neuroblast delamination” prior to muscle for-
mation in larvae. As described in Phascolosoma sp. by Hilton (1917) (see
Bullock & Horridge, 1965), short nerves extend from the VNC to multi-
ple attachment sites on right and left sides along the body in patterns
that may be opposite, alternate, or irregular in orientation along the
cord, and within the same worm. In all species examined, neuronal cell
bodies appear to be evenly distributed along most of the length of the
cord, with no clear evidence of ganglionic aggregations of neurons
(Rice, 1993). And finally, as the histological studies briefly discussed
above show, and as typically observed with most histological prepara-
tions of animals, there is a complex integration of cells, tissues, and
organ systems within larval, juvenile, and adult body plans of
sipunculans. It will be interesting to determine how the 5HT+ and
FMRF+ neural elements in embryonic and larval sipunculans, especially
the two clusters of neurons in the developing VNC, compare with adult
neural elements, including which tissues (e.g., gut, muscle, nephridia,
epidermis) they innervate.
4.4 | Comparison of VNCs in Annelida
Within the segmentally organized body plans of annelids in Errantia
and Sedentaria (referred to as Pleistoannelida sensu Struck et al.,
2011), the developing and adult VNC typically consists of iterated
ganglia with transverse commissures along the anterior–posterior axis,
somata-free connectives that connect the ganglia, and paired cords
(Bullock & Horridge, 1965; Meyer et al., 2015; Müller, 2006; Orrhage &
Müller, 2005; Purschke, 2002; Richter et al., 2010; Starunov,
Voronezhskaya, & Nezlin, 2017; Zattara & Bely, 2015), although there
are also many exceptions to this rope-ladder-like arrangement (Müller,
2006; Orrhage & Müller, 2005). This pattern is distinctly different from
the unpaired medullary cord of adult sipunculans, which lack ganglia
and somata-free connectives (Åkesson, 1958; Boyle & Rice, 2014;
Clark, 1969; Kristof et al., 2008; Rice, 1973; Wanninger et al., 2005).
However, the VNC in T. lageniformis shows several similarities with the
VNCs of other annelids that branch outside of Errantia and Sedentaria.
Current phylogenies place Oweniidae and Magelonidae as a sister clade
to the rest of annelids (Andrade et al., 2015; Helm et al., 2018; Weigert
et al., 2014), and the architecture of the VNC has been described in a
few species in each clade (Helm, Vöcking, Kourtesis, & Hausen, 2016;
CARRILLO-BALTODANO ET AL. 17
Rimskaya-Korsakova, Kristof, Malakhov, & Wanninger, 2016). Oweniids
are reported to have unpaired medullary VNCs with continuous neural
tissue along the ventral midline throughout the trunk (i.e., no ganglia;
Helm et al., 2016, 2018; Rimskaya-Korsakova et al., 2016). However,
within this unpaired cord, bilaterally symmetric, 5HT+ longitudinal tracts
were visible (Helm et al., 2018; Rimskaya-Korsakova et al., 2016), and
in premetamorphic and postmetamorphic animals of Owenia fusiformis,
multiple clusters of 5HT+ neurons were observed in the developing
VNC (Helm et al., 2016, 2018). Interestingly, in at least one oweniid
species, Myriochele heeri, the 5HT+ neurons in the VNC are uniformly
distributed in the adults (Helm et al., 2018), while in another species,
Galathowenia oculata, they are uniformly distributed in all but the most
posterior segments (Rimskaya-Korsakova et al., 2016). This is similar to
T. lageniformis, where we observed a few clusters of 5HT+ and FMRF+
neurons during larval stages, but a fairly uniform distribution of 5HT+
neurons in juveniles. Based on these similarities, and as suggested
above, we think it is possible that the cellular clusters in the developing
VNC of Sipuncula either represent precursors of ganglia in this group
of worms, or reflect a more general pattern across Spiralia.
In at least two species within Magelonidae, the adult VNC has also
been described as medullary with irregularly arranged 5HT+ and
FMRF+ neurons (i.e., no serial repetition). The cords are paired in the
anterior and unpaired in the posterior VNC (Helm et al., 2018; Jones,
1968). The adult sipunculan VNC is unpaired based on serial cross
sections (Åkesson, 1958; Rice, 1973, 1993; von Mack, 1902). This
final unpaired structure is again similar to what has been described in
adult oweniids, and in the posterior VNC of adult magelonids. How-
ever, in the developing VNC of sipunculan larvae, at least some of
the longitudinal tracts arise as paired structures (this study; Kristof,
2011; Rice, 1973; Wanninger et al., 2005). By juvenile stages in
T. lageniformis, these paired tracts are only visible as two closely juxta-
posed Tub+ cords in the anterior, while in the posterior trunk, only a
single Tub+ cord is visible. The paired nature of the longitudinal tracts
in the VNC in larval and juvenile sipunculans could represent a devel-
opmental transition from two bilaterally symmetric domains of neural
tissue on either side of the midline to the eventual single domain
across the ventral midline (based on Tl-elav1 expression). The longitu-
dinal tracts of T. lageniformis and other sipunculans begin as bilaterally
symmetric, but then appear to undergo fusion as development pro-
ceeds. Similarly, it is possible that different early branching annelid
taxa undergo different degrees of fusion of their longitudinal tracts. It
is interesting to note that all adult sipunculan species described thus
far have an unpaired medullary cord (Åkesson, 1958; Rice, 1973,
1993), suggesting that the unpaired condition may be important bio-
logically or that this feature arose early during evolution of the group,
and has not undergone significant change since.
Other annelid taxa that branch outside of Sedentaria and Errantia
include Chaetopteridae, Apistobranchidae, and Psammodrilidae (Helm
et al., 2018), and paired medullary cords lacking somata-free connec-
tives have been described in adult species within Chaetopteridae and
Apistobranchidae (Helm et al., 2018). Within all three of these clades,
the presence of serially-repeated 5HT+ neuronal clusters as well as
segmentally arranged commissures in the adult VNC have been
described (Helm et al., 2018). Furthermore, serially repeated FMRF+
neuronal clusters in the adult VNC have been described in a few spe-
cies of chaetoperids and apistobranchids (Helm et al., 2018). So, while
a segmented body plan is prevalent throughout Annelida except in a
few groups, notably Sipuncula (Åkesson, 1958; Boyle & Rice, 2014;
Clark, 1969; Wanninger et al., 2005), the presence of serially terated
ganglia is not. Instead, as more taxa are being described, it appears
that the rope-ladder-like VNC that has been traditionally used as a
defining character for annelids (Purschke, 2002) may have evolved
incrementally within different clades, that is, is not a synapomorphy
of Annelida. Indeed, it has even been suggested that the rope-ladder-
like architecture evolved multiple times within Annelida (Helm et al.,
2018; Purschke et al., 2014). Reconstructing evolution of the VNC in
annelids will require sampling across annelid groups with different
morphologies, for example, with or without parapodia, sampling of
groups with different life history strategies, and, importantly, sampling
from early development through adult stages of the same species. If
the rope-ladder-like architecture was not present in the last common
ancestor of Annelida, it is interesting to speculate why clusters of neu-
ral subtypes are visible during developmental stages of at least
sipunculans and oweniids, but not at later stages (i.e., juveniles or
adults, respectively). This hints at the possibility that serial repetition
of neural subtypes may be a developmental feature and/or a ground
pattern found in other spiralian taxa. Within Annelida, what started as
repetition of neural subtypes may have become codified into ganglia
of rope-ladder-like VNCs in selected clades.
4.5 | Arrangement of 5HT+ and FMRF+ cells in the
trunks of different spiralian taxa
When comparing the trunk nervous systems in larvae of other
spiralians, the arrangements of 5HT+ and FMRF+ perikarya exhibit
notable variation. In two aplacophoran larvae, there are two initial clus-
ters of 5HT+ cell bodies on each of two parallel ventral neurite bundles
(Redl, Scherholz, Todt, Wollesen, & Wanninger, 2014). This pattern is
relatively similar to the arrangement of 5HT+ clusters along the single
cord in T. lageniformis, although no obvious pattern of FMRF+ cells in
those solenogastres is detectable along the neurite bundles, unlike
what we observed in T. lageniformis. Along parallel connectives in the
ventral cord of the polyplacophoran Ischnochiton, 5HT+ and FMRF+
perikarya are more or less continuous from anterior to posterior, and
do not appear to form distinct clusters (Voronezhskaya, Tyurin, &
Nezlin, 2002). Within the veliger larva of a freshwater bivalve (Dreissena
polymorpha), a pair of 5HT+ ventral neurites extends between the brain
and a posterior sensory cell, but no 5HT+ perikarya are observed along
either of the two connectives (Pavlicek, Schwaha, & Wanninger, 2018).
In the veliger of the Pacific oyster (Crassostrea gigas), paired 5HT+ and
FMRF+ VNCs have paired sets of perikarya that are considered pedal
and visceral ganglia (Yurchenko, Yurchenko, Skiteva, Voronezhskaya, &
Dyachuk, 2018). During development of a polyclad larvae, 5HT+ peri-
karya show bilaterally symmetric patterns around a pharyngeal nerve
ring, a median nerve, and dorsal-lateral connectives (Rawlinson, 2010).
Compared with 5HT+ cell bodies, FMRF+ perikarya were not detected
18 CARRILLO-BALTODANO ET AL.
along any of those neurites in the same species (Rawlinson, 2010). In
the pilidium larva of Lineus albocinctus, FMRF+ cell bodies of multipolar
interneurons occur throughout a complex “nerve net” within lappet
fields, while 5HT+ perikarya primarily occur along lateral nerve cords,
suboral neurites, and the oral nerve ring (Hindinger, Schwaha, &
Wanninger, 2013). However, in the direct-developing nemertean,
Pantinonemertes californiensis, 5HT+ perikarya are continuous along
each of two lateral nerve cords, and they are also found posterior to
the lateral cords within a single cluster flanking the ventral midline
(Hiebert & Maslakova, 2015).
Therefore, among spiralian larvae there are notable variations in
the distribution and extent of 5HT+ and FMRF+ neural architecture.
However, there is also a common developmental pattern in the way
that serotonergic and FMRFamidergic nervous systems are generally
wired along the anterior–posterior axis. As mentioned above, the pat-
tern observed across larval types superficially shows an aggregation
of anterior ganglia from which ventral longitudinal cords extend pos-
teriorly. Within this pattern, 5HT+ and FMRF+ perikarya are typically
spaced along paired nerve cords in either distinct clusters or in a linear
series of cell bodies, from which neurites extend to connect them with
their respective longitudinal elements (5HT+ or FMRF+) in the VNC. It
is not clear as to what specific roles 5HT+ and FMRF+ perikarya have
in generating different components of the CNS, or whether their particu-
lar arrangements correlate closely with functional anatomy or even body-
plan organization. In segmented spiralians, 5HT+ and FMRF+ perikarya
may align with ganglia, peripheral neurites, or appendages during early
development, but then show considerable variation in later stages (Helm
et al., 2016; Meyer et al., 2015; Starunov et al., 2017; Zattara & Bely,
2015). Yet, in the larvae and juveniles of unsegmented spiralians, there
can also be relatively regular arrangements of 5HT+ and FMRF+ perikarya
along the anterior–posterior axis (Friedrich, Wanninger, Brückner, &
Haszprunar, 2002; Hay-Schmidt, 2000; Hiebert & Maslakova, 2015;
Hindinger et al., 2013; Rawlinson, 2010; Reuter, Wikgren, & Lenthonen,
1986; Wanninger & Haszprunar, 2003). And in segmented animals
outside Spiralia, 5HT+ neurons can exhibit a repeated or even scattered
pattern that is not correlated with segmental anatomy (Mayer & Harzsch,
2007). Thus, while observed developmental patterns of neural architec-
ture (e.g., circumesophageal connectives, nerve cords, perikarya) may
appear to correlate with or perhaps reflect the organization of some, but
not all, body plans in which they are described, we suggest that 5HT+
and FMRF+ neural elements, although integral components of animal ner-
vous systems, are not reliable markers of body-plan organization, espe-
cially in the context of segmented animals within Spiralia.
4.6 | Concluding remarks
Sipuncula are an ancient clade of worms with unambiguous fossil evi-
dence of an unsegmented body plan in the Early Cambrian from ~520
million years ago (Huang, Chen, Vannier, & Saiz Salinas, 2004).
Because of this, they are currently considered to represent the oldest
annelid crown group (Parry et al., 2016; Weigert et al., 2014), or possi-
bly coeval with the earliest stem group polychaetes (Parry et al.,
2014). All of this brings into question the interpretations by Kristof
et al. (2008), who discussed possible 5HT+ evidence for the “establish-
ment and loss” of segmentation during development of the VNC in
the sipunculan, P. agassizii. For almost a decade, that study has been
cited as the primary source of morphological evidence for the segmen-
tal ancestry of Sipuncula (Kristof et al., 2008). Our results, together with
fossils pointing to the ancient history of sipunculan worms (Huang
et al., 2004), suggest that such interpretations may have been prema-
ture. Clark's (1969) hypothesis that sipunculans may have diverged
from an unsegmented “preannelid” lineage with a centralized nervous
system, are more in line with our data on T. lageniformis and Helm
et al. (2018).
The developmental mechanisms responsible for segmenting an
animal body are not as well understood in annelids, when compared
to insects or vertebrates (Chipman, 2010; Seaver, 2003). Some of the
candidate “segmentation genes” such as engrailed and wingless, or
pathways such as Notch signaling, have been shown as inconclusive
across the few species of annelids that have been studied thus far
(Gazave, Lemaı, & Balavoine, 2017; Prud'homme et al., 2003; Seaver &
Kaneshige, 2006; Seaver, Paulson, Irvine, & Martindale, 2001; Seaver,
Thamm, & Hill, 2005; Steinmetz, Kostyuchenko, Fischer, & Arendt,
2011; Thamm & Seaver, 2008). Only recently, teloblast cells have
been found to contribute to the mesoderm of segments in the annelid
P. dumerilii similar to the mechanisms used by clitellates (Balavoine,
2014; Özpolat, Handberg-thorsager, Vervoort, & Balavoine, 2017). It
remains to be studied if this phenomenon is widespread across other
annelids. Although we still are trying to understand how annelids
make segments, sipunculans have been mentioned as one example of
the loss of whole-body segmentation (Struck et al., 2011; Weigert
et al., 2014; Weigert & Bleidorn, 2016). It has been suggested that this
loss of segmentation is a result of the transition to a sedentary or
“sessile” life style (Chipman, 2010); however, many clades of
tubiculous polychaetes have a life style similar to sipunculans but pos-
sess a clearly segmented body plan (Rouse & Pleijel, 2001). It will be
necessary to determine the molecular control of segmentation across
multiple annelid taxa, both within and outside of Errantia plus
Sedentaria, in order to ascertain if segmentation was either lost or
had not yet evolved in the last common ancestor of Sipuncula.
In conclusion, our investigation did not uncover definitive evi-
dence of segmental morphology during development of the
sipunculan nervous system. Although we observed two clusters of
perikarya, a transition from paired to unpaired longitudinal bundles,
and well-supported evidence for the architecture of a medullary VNC,
ancestry of a segmented body plan is not supported by patterns of
common immunohistochemical markers or neuronal gene expression
patterns in T. lageniformis.
ACKNOWLEDGMENTS
The authors wish to thank R. Bellin at the College of the Holy Cross
for access to their confocal microscope, and the Smithsonian
Marine Station Life Histories Program and associated support staff
for assistance with field collections. AMC-B received a grant from
the Lerner-Gray Fund by the American Museum of Natural History,
CARRILLO-BALTODANO ET AL. 19
and NPM received a Smithsonian Short-Term Visitor award to con-
duct work at the Smithsonian Marine Station at Fort Pierce, Florida.
This publication is Smithsonian Marine Station contribution
no. 1114.
CONFLICT OF INTEREST
The authors declare no potential conflict of interest.
AUTHOR CONTRIBUTIONS
MJB, NPM, and MER designed the study. AMC-B, MJB, and NPM col-
lected and analyzed the data. MJB and AMC-B collected the animals
and provided all developmental stages. AMC-B prepared all figures,
except for Figure 1. AMC-B, MJB, and NPM wrote the manuscript. All
authors commented and agreed on the final version of the manuscript.
ORCID
Néva P. Meyer https://orcid.org/0000-0002-2002-3280
REFERENCES
Åkesson, B. (1958). A study of the nervous system of the Sipunculoidae,
with some remarks on the development of the two species Phascolion
strombi (Montagu) and Golfingia minuta (Keferstein). Undersöksinar
Over Öresund, 38, 1–149.
Amiel, A. R., Henry, J. Q., & Seaver, E. C. (2013). An organizing activity is
required for head patterning and cell fate specification in the poly-
chaete annelid Capitella teleta: New insights into cell–cell signaling in
Lophotrochozoa. Developmental Biology, 379(1), 107–122. https://doi.
org/10.1016/j.ydbio.2013.04.011
Andrade, C. S., Novo, M., Kawauchi, G. Y., Worsaae, K., Pleijel, F.,
Giribet, G., & Rouse, G. W. (2015). Articulating “Archiannelids”: Phy-
logenomics and annelid relationships, with emphasis on Meiofaunal
taxa. Molecular Biology and Evolution, 32(11), 2860–2875. https://doi.
org/10.1093/molbev/msv157
Arendt, D., Tessmar-raible, K., Snyman, H., Dorresteijn, A. W., &
Wittbrodt, J. (2004). Ciliary photoreceptors with a vertebrate-type
opsin in an invertebrate brain. Science, 306(5697), 869–871.
Balavoine, G. (2014). Segment formation in annelids: Patterns, processes
and evolution. International Journal of Developmental Biology, 58(6–8),
469–483. https://doi.org/10.1387/ijdb.140148gb
Berger, C., Renner, S., Luer, K., & Technau, G. M. (2007). The commonly
used marker ELAV is transiently expressed in neuroblasts and glial cells
in the Drosophila embryonic CNS. Developmental Dynamics, 236,
3562–3568.
Boyle, M. J., & Rice, M. E. (2014). Sipuncula: An emerging model of spiralian
development and evolution. International Journal of Developmental Biol-
ogy, 58(6–8), 485–499. https://doi.org/10.1387/ijdb.140095mb
Boyle, M. J., & Seaver, E. C. (2010). Expression of FoxA and GATA tran-
scription factors correlates with regionalized gut development in two
lophotrochozoan marine worms: Chaetopterus (Annelida) and Themiste
lageniformis (Sipuncula). EvoDevo, 1(1), 2. https://doi.org/10.1186/
2041-9139-1-2
Brusca, R. C., & Brusca, G. J. (2003). Invertebrates (2nd ed.). Sunderland,
MA: Sinauer Associates, Inc.
Budd, G. E. (2001). Why are arthropods segmented? Evolution & Development,
3(5), 332–342. https://doi.org/10.1046/j.1525-142X.2001.01041.x
Bullock, T. H., & Horridge, G. A. (1965). Structure and function in the
nervous systems of invertebrates. New York, NY: Freeman & Company.
Chipman, A. D. (2010). Parallel evolution of segmentation by co-option of
ancestral gene regulatory networks. BioEssays, 32(1), 60–70. https://
doi.org/10.1002/bies.200900130
Clark, R. B. (1969). Systematics and phylogeny: Annelida, Echiura,
Sipuncula. In M. Florkin & B. T. Scheer (Eds.), Chemical zoology volume
4 (pp. 1–68). New York, NY: Academic Press.
Craxton, M. (2010). A manual collection of Syt, Esyt, Rph3a, Rph3al, Doc2,
and Dblc2 genes from 46 metazoan genomes—An open access
resource for neuroscience and evolutionary biology. BMC Genomics,
11, 37. https://doi.org/10.1186/1471-2164-11-37
Darriba, D., Taboada, G. L., Doallo, R., & Posada, D. (2011). ProtTest 3:
Fast selection of best-fit models of protein evolution. Bioinformatics,
27(8), 1164–1165.
Denes, A. S., Jékely, G., Steinmetz, P. R. H., Raible, F., Snyman, H.,
Prud'homme, B., … Arendt, D. (2007). Molecular architecture of anne-
lid nerve cord supports common origin of nervous system centraliza-
tion in bilateria. Cell, 129(2), 277–288. https://doi.org/10.1016/j.cell.
2007.02.040
Dordel, J., Fisse, F., Purschke, G., & Struck, T. H. (2010). Phylogenetic posi-
tion of Sipuncula derived from multi-gene and phylogenomic data and
its implication for the evolution of segmentation. Journal of Zoological
Systematics and Evolutionary Research, 48, 197–207. https://doi.org/
10.1111/j.1439-0469.2010.00567.x
Friedrich, S., Wanninger, A., Brückner, M., & Haszprunar, G. (2002). Neuro-
genesis in the mossy chiton, Mopalia muscosa (Gould) (Polyplacophora):
Evidence against molluscan metamerism. Journal of Morphology, 253,
109–117. https://doi.org/10.1002/jmor.10010
Gazave, E., Lemaı, Q. I. B., & Balavoine, G. (2017). The notch pathway in
the annelid Platynereis: Insights into chaetogenesis and neurogenesis
processes. Open Biology, 7, 160242.
Gerould, J. H. (1938). Eyes and nervous system of Phascolosoma verillii and
other sipunculids. Travaux de La Station Zoologique de Wimereux, 13,
313–325.
Graham, A., Butts, T., Lumsden, A., & Kiecker, C. (2014). What can verte-
brates tell us about segmentation? EvoDevo, 5(1), 24. https://doi.org/
10.1186/2041-9139-5-24
Gühmann, M., Jia, H., Randel, N., Verasztó, C., Bezares-Calderón, L. A.,
Michiels, N. K., … Jékely, G. (2015). Spectral tuning of phototaxis by a
Go-Opsin in the rhabdomeric eyes of Platynereis. Current Biology, 25
(17), 2265–2271. https://doi.org/10.1016/j.cub.2015.07.017
Guindon, S., & Gascuel, O. (2003). A simple, fast, and accurate algorithm to
estimate large phylogenies by maximum likelihood. Systematic Biology,
52(5), 696–704.
Hannibal, R. L., & Patel, N. H. (2013). What is a segment? EvoDevo, 4(1),
35. https://doi.org/10.1186/2041-9139-4-35
Hay-Schmidt, A. (2000). The evolution of the serotonergic nervous system.
Proceedings of the Royal Society. Biological Sciences, 267(1448),
1071–1079. https://doi.org/10.1098/rspb.2000.1111
Hejnol, A., & Lowe, C. J. (2015). Embracing the comparative approach:
How robust phylogenies and broader developmental sampling impacts
the understanding of nervous system evolution. Philosophical Transac-
tions of the Royal Society of London. Series B, Biological Sciences, 370,
20150045.
Helm, C., Beckers, P., Bartolomaeus, T., Drukewitz, S. H., Kourtesis, I.,
Weigert, A., … Bleidorn, C. (2018). Convergent evolution of the ladder-
like ventral nerve cord in Annelida. Frontiers in Zoology, 15(36), 1–17.
Helm, C., Vöcking, O., Kourtesis, I., & Hausen, H. (2016). Owenia
fusiformis—A basally branching annelid suitable for studying ancestral
features of annelid neural development. BMC Evolutionary Biology, 16
(129), 1–19. https://doi.org/10.1186/s12862-016-0690-4
Hiebert, L. S., & Maslakova, S. A. (2015). Expression of Hox, Cdx, and
Six3/6 genes in the hoplonemertean Pantinonemertes californiensis
offers insight into the evolution of maximally indirect development in
20 CARRILLO-BALTODANO ET AL.
the phylum Nemertea. EvoDevo, 6(26), 1–15. https://doi.org/10.1186/
s13227-015-0021-7
Hilton, W. A. (1917). The central nervous system of a sipunculid. Journal of
Entomology and Zoology, 9(1), 30–33.
Hindinger, S., Schwaha, T., & Wanninger, A. (2013). Immunocytochemical
studies reveal novel neural structures in nemertean pilidium larvae and
provide evidence for incorporation of larval components into the juve-
nile nervous system. Frontiers in Zoology, 10(1), 31. https://doi.org/10.
1186/1742-9994-10-31
Huang, D.-Y., Chen, J.-Y., Vannier, J., & Saiz Salinas, J. I. (2004). Early
Cambrian sipunculan worms from Southwest China early Cambrian
sipunculan worms from Southwest China. Proceedings of the Royal
Society B Biological Sciences, 271, 1671–1676. https://doi.org/10.
1098/rspb.2004.2774
Huelsenbeck, J. P., & Ronquist, F. (2001). MRBAYES: Bayesian inference
of phylogenetic trees. Bioinformatics, 17(8), 754–755.
Hyman, L. H. (1959). Phylum Sipuncula. In The invertebrates. Vol.5: Smaller
coelomate groups (pp. 610–690). New York, NY: McGraw-Hill.
Jones, M. L. (1968). On the morphology, feeding, and behavior of
Magelona sp. Biological Bulletin, 134(272–297), 272–297.
Kristof, A. (2011). The molecular and developmental basis of bodyplan pat-
terning in Sipuncula and the evolution of segmentation, Copenhagen,
Denmark: University of Copenhagen.
Kristof, A., & Maiorova, A. S. (2016). 23. Annelida: Sipuncula. In
A. Schmidt-Rhaesa, S. Harzsch, & G. Purschke (Eds.), Structure and evo-
lution of invertebrate nervous systems (pp. 248–253). Oxford, England:
Oxford University Press.
Kristof, A., Wollesen, T., Maiorova, A. S., & Wanninger, A. (2011). Cellular
and muscular growth patterns during sipunculan development. Journal
of Experimental Zoology. Part B, Molecular and Developmental Evolution,
316B(3), 227–240. https://doi.org/10.1002/jez.b.21394
Kristof, A., Wollesen, T., & Wanninger, A. (2008). Segmental mode of neu-
ral patterning in sipuncula. Current Biology, 18(15), 1129–1132.
https://doi.org/10.1016/j.cub.2008.06.066
Littleton, J. T., Bellen, H. J., & Perin, M. S. (1993). Expression of
synaptotagmin in Drosophila reveals transport and localization of syn-
aptic vesicles to the synapse. Development (Cambridge, England), 118(4),
1077–1088. Retrieved from. http://www.ncbi.nlm.nih.gov/pubmed/
8269841
Marlow, H., Tosches, M. A., Tomer, R., Steinmetz, P. R., Lauri, A.,
Larsson, T., & Arendt, D. (2014). Larval body patterning and apical
organs are conserved in animal evolution. BMC Biology, 12(1), 7.
https://doi.org/10.1186/1741-7007-12-7
Martín-Durán, J. M., Pang, K., Børve, A., Lê, H. S., Furu, A., Cannon, J. T., …
Hejnol, A. (2018). Convergent evolution of Bilaterian nerve cords.
Nature, 553(7686), 45–50. https://doi.org/10.1038/nature25030
Mayer, G., & Harzsch, S. (2007). Immunolocalization of serotonin in
Onychophora argues against segmental ganglia being an ancestral fea-
ture of arthropods. BMC Evolutionary Biology, 7, 118. https://doi.org/
10.1186/1471-2148-7-118
Meyer, N. P., Boyle, M. J., Martindale, M. Q., & Seaver, E. C. (2010). A com-
prehensive fate map by intracellular injection of identified blastomeres
in the marine polychaete Capitella teleta. EvoDevo, 1(1), 8. https://doi.
org/10.1186/2041-9139-1-8
Meyer, N. P., Carrillo-Baltodano, A., Moore, R. E., & Seaver, E. C. (2015).
Nervous system development in lecithotrophic larval and juvenile
stages of the annelid Capitella teleta. Frontiers in Zoology, 12, 15.
https://doi.org/10.1186/s12983-015-0108-y
Meyer, N. P., & Seaver, E. C. (2009). Neurogenesis in an annelid: Charac-
terization of brain neural precursors in the polychaete Capitella sp.
I. Developmental Biology, 335(1), 237–252. https://doi.org/10.1016/j.
ydbio.2009.06.017
Müller, M. C. M. (2006). Polychaete nervous systems: Ground pattern and
variations—cLS microscopy and the importance of novel
characteristics in phylogenetic analysis. Integrative and Comparative
Biology, 46(2), 125–133. https://doi.org/10.1093/icb/icj017
Nomaksteinsky, M., Röttinger, E., Dufour, H., Chettouh, Z., Lowe, C. J.,
Martindale, M. Q., & Brunet, J.-F. (2009). Centralization of the deu-
terostome nervous system predates chordates. Current Biology, 19(5),
1264–1269. https://doi.org/10.1016/j.cub.2009.05.063
Orrhage, L., & Müller, M. C. M. (2005). Morphology of the nervous system
of Polychaeta (Annelida ). Hydrobiologia, 535, 79–111.
Özpolat, B. D., Handberg-thorsager, M., Vervoort, M., & Balavoine, G.
(2017). Cell lineage and cell cycling analyses of the 4d micromere using
live imaging in the marine annelid Platynereis dumerilii. eLife, 6
(e30463), 1–35.
Parry, L., Tanner, A., & Vinther, J. (2014). The origin of annelids. Frontiers in
Paleontology, 57(6), 1091–1103. https://doi.org/10.1111/pala.12129
Parry, L. A., Edgecombe, G. D., Eibye-jacobsen, D., & Vinther, J. (2016).
The impact of fossil data on annelid phylogeny inferred from discrete
morphological characters. Proceedings of the Royal Society. Biological
Sciences, 283(1837), 20161378.
Pascale, A., Amadio, M., & Quattrone, A. (2008). Defining a neuron: Neuro-
nal ELAV proteins. Cellular and Molecular Life Sciences, 65, 128–140.
https://doi.org/10.1007/s00018-007-7017-y
Pavlicek, A., Schwaha, T., & Wanninger, A. (2018). Towards a ground pat-
tern reconstruction of bivalve nervous systems: Neurogenesis in the
zebra mussel Dreissena polymorpha. Organisms Diversity & Evolution, 18
(1), 101–114.
Priyam, A., Woodcroft, B. J., Rai, V., Munagala, A., Moghul, I., & Ter, F.
(2015). Sequence server: A modern graphical user interface for custom
BLAST databases. bioRxiv, 033142, 1–18.
Prud'homme, B., de Rosa, R., Arendt, D., Julien, J.-F., Pajaziti, R.,
Dorresteijn, A. W. C., … Balavoine, G. (2003). Arthropod-like expres-
sion patterns of engrailed and wingless in the annelid Platynereis
dumerilii suggest a role in segment formation. Current Biology, 13(21),
1876–1881. https://doi.org/10.1016/j.cub.2003.10.006
Purschke, G. (2002). On the ground pattern of Annelida. Organisms Diver-
sity & Evolution, 2, 181–196.
Purschke, G. (2016). Annnelida: Basal groups and Pleistoannelida. In
A. Schmidt-Rhaesa, S. Harzsch, & G. Purschke (Eds.), Structure and evo-
lution of invertebrate nervous systems (pp. 254–312). Oxford, England:
Oxford University Press.
Purschke, G., Bleidorn, C., & Struck, T. H. (2014). Systematics, evolution
and phylogeny of Annelida—A morphological perspective. Memoirs of
Museum Victoria, 71, 247–269.
Ratti, A., Fallini, C., Cova, L., Fantozzi, R., Calzarossa, C., Zennaro, E., …
Silani, V. (2006). A role for the ELAV RNA-binding proteins in neural
stem cells: Stabilization of Msi1 mRNA. Journal of Cell Science, 119(7),
1442–1452. https://doi.org/10.1242/jcs.02852
Rawlinson, K. A. (2010). Embryonic and post-embryonic development of
the polyclad flatworm Maritigrella crozieri; implications for the evolu-
tion of spiralian life history traits. Frontiers in Zoology, 7(12), 1–25.
Redl, E., Scherholz, M., Todt, C., Wollesen, T., & Wanninger, A. (2014).
Development of the nervous system in Solenogastres (Mollusca)
reveals putative ancestral spiralian features. EvoDevo, 5(48), 1–17.
Reuter, M., Wikgren, M., & Lenthonen, M. (1986). Immunocytochemical
demonstration of 5HT-like and FMRFamide-like substances in whole
mounts of Microstomum lineare (Turbellaria). Cell and Tissue Research,
246(1), 7–12.
Rice, M. E. (1973). Morphology, behavior, and histogenesis of the
pelagosphera larva of Phascolosoma agassizii (Sipuncula). Smithsoninan
Contribution to the Marine Science, 132, 1–51.
Rice, M. E. (1985). Sipuncula: Developmental evidence for phylogenetic
inference. In S. Morris, J. George, R. Gibson, & H. Platt (Eds.), The ori-
gins and relationships of lower invertebrates (pp. 274–296). Oxford,
England: Oxford University Press.
CARRILLO-BALTODANO ET AL. 21
Rice, M. E. (1993). Sipuncula. In F. W. Harrison & M. E. Rice (Eds.), Micro-
scopic anatomy of invertebrates Onychophora, Chilopoda, and Lesser
Protostomata (Vol. 12, pp. 274–296). New York, NY: Wiley-Liss.
Richter, S., Loesel, R., Purschke, G., Schmidt-Rhaesa, A., Scholtz, G.,
Stach, T., … Stegner, M. E. J., S. H. (2010). Invertebrate neuro-
phylogeny: Suggested terms and definitions for a neuroanatomical
glossary. Frontiers in Zoology, 7(1), 29. https://doi.org/10.1038/nrg2417.
EvoD/Vo
Rimskaya-Korsakova, N., Kristof, A., Malakhov, V. V., & Wanninger, A.
(2016). Neural architecture of Galathowenia oculata Zach, 1923
(Oweniidae, Annelida). Frontiers in Zoology, 13(5), 1–19. https://doi.
org/10.1186/s12983-016-0136-2
Rizo, J., & Rosenmund, C. (2008). Synaptic vesicle fusion. Nature Struc-
tural & Molecular Biology, 15(7), 665–674.
Robinow, S., & White, K. (1991). Characterization and spatial distribution
of the ELAV protein during Drosophila melanogaster development.
Journal of Neurobiology, 22, 443–461.
Rouse, G. W., & Pleijel, F. (2001). Polychaetes (1st ed.). Oxford, England:
Oxford University Press.
Rousset, V., Pleijel, F., Rouse, G. W., Erséus, C., & Sidall, M. E. (2007). A
molecular phylogeny of annelids. Cladistics, 23(1), 41–63.
Samson, M. (2008). Rapid functional diversification in the structurally con-
served ELAV family of neuronal RNA binding proteins. BMC Genomics,
9(392), 1–11. https://doi.org/10.1186/1471-2164-9-392
Santagata, S., Resh, C., Hejnol, A., Martindale, M. Q., & Passamaneck, Y. J.
(2012). Development of the larval anterior neurogenic domains of
Terebratalia transversa (Brachiopoda) provides insights into the diversi-
fication of larval apical organs and the spiralian nervous system. Evo-
Devo, 3, 1–20.
Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., …
Cardona, A. (2012). Fiji: an open-source platform for biological-image
analysis. Nature Methods, 9, 676–682. https://doi.org/10.1038/
nmeth.2019
Schmidt-Rhaesa, A. (2007). The evolution of organ systems. Oxford,
England: Oxford University Press.
Scholtz, G. (2002). The Articulata hypothesis–or what is a segment? Organ-
isms Diversity & Evolution, 2(3), 197–215. https://doi.org/10.1078/
1439-6092-00046
Schulze, A., Cutler, E. B., & Giribet, G. (2005). Reconstructing they phylog-
eny of the Sipuncula. Hydrobiologia, 535/536, 277–296. https://doi.
org/10.1007/s10750-004-4404-3
Seaver, E. C. (2003). Segmentation: Mono- or polyphyletic? International
Journal of Developmental Biology, 47(7–8), 583–595. Retrieved from.
http://www.ncbi.nlm.nih.gov/pubmed/14756334
Seaver, E. C., & Kaneshige, L. M. (2006). Expression of “segmentation”
genes during larval and juvenile development in the polychaetes
Capitella sp. I and H. elegans. Developmental Biology, 289(1), 179–194.
https://doi.org/10.1016/j.ydbio.2005.10.025
Seaver, E. C., Paulson, D. a., Irvine, S. Q., & Martindale, M. Q. (2001). The
spatial and temporal expression of Ch-en, the engrailed gene in the
polychaete Chaetopterus, does not support a role in body axis segmen-
tation. Developmental Biology, 236(1), 195–209. https://doi.org/10.
1006/dbio.2001.0309
Seaver, E. C., Thamm, K., & Hill, S. D. (2005). Growth patterns during seg-
mentation in the two polychaete annelids, Capitella sp. I and Hydroides
elegans: Comparisons at distinct life history stages. Evolution & Devel-
opment, 7(4), 312–326. https://doi.org/10.1111/j.1525-142X.2005.
05037.x
Shigeno, S., Parnaik, R., Albertin, C. B., & Ragsdale, C. W. (2015). Evidence
for a cordal, not ganglionic, pattern of cephalopod brain neurogenesis.
Zoological Letters, 1(26), 1–13. https://doi.org/10.1186/s40851-015-
0026-z
Simionato, E., Kerner, P., Dray, N., Le Gouar, M., Ledent, V., Arendt, D., &
Vervoort, M. (2008). Atonal- and achaete-scute-related genes in the
annelid Platynereis dumerilii: Insights into the evolution of neural basic-
helix-loop-helix genes. BMC Evolutionary Biology, 8, 170. https://doi.
org/10.1186/1471-2148-8-170
Sperling, E. A., Vinther, J., Moy, V. N., Wheeler, B. M., Sémon, M.,
Briggs, D. E. G., & Peterson, K. J. (2009). MicroRNAs resolve an appar-
ent conflict between annelid systematics and their fossil records. Pro-
ceedings of the Royal Society B, 276, 4315–4322. https://doi.
org/10.1098/rspb.2009.1340
Starunov, V. V., Voronezhskaya, E. E., & Nezlin, L. P. (2017). Development
of the nervous system in Platynereis dumerilii (Nereididae, Annelida).
Frontiers in Zoology, 14(27), 1–20. https://doi.org/10.1186/s12983-
017-0211-3
Steinmetz, P. R. H., Kostyuchenko, R. P., Fischer, A., & Arendt, D. (2011).
The segmental pattern of otx, gbx, and Hox genes in the annelid
Platynereis dumerilii. Evolution & Development, 13(1), 72–79. https://
doi.org/10.1111/j.1525-142X.2010.00457.x
Stephen, A. C., & Edmonds, S. J. (1972). They phyla Sipuncula and Echiura.
London, England: British Museum.
Struck, T. H., Golombek, A., Weigert, A., Franke, F. A., Westheide, W.,
Purschke, G., … Halanych, K. M. (2015). The evolution of annelids
reveals two adaptive routes to the interstitial realm. Current Biology,
25(15), 1993–1999. https://doi.org/10.1016/j.cub.2015.06.007
Struck, T. H., Paul, C., Hill, N., Hartmann, S., Hösel, C., Kube, M., …
Bleidorn, C. (2011). Phylogenomic analyses unravel annelid evolution.
Nature, 471(7336), 95–98. https://doi.org/10.1038/nature09864
Struck, T. H., Schult, N., Kusen, T., Hickman, E., Bleidorn, C., McHugh, D., &
Halanych, K. M. (2007). Annelid phylogeny and the status of Sipuncula
and Echiura. BMC Evolutionary Biology, 7, 57. https://doi.org/10.1186/
1471-2148-7-57
Sur, A., Magie, C. R., Seaver, E. C., & Meyer, N. P. (2017). Spatiotemporal
regulation of nervous system development in the annelid Capitella
teleta. EvoDevo, 8, 1–23. https://doi.org/10.1186/s13227-017-0076-8
Thamm, K., & Seaver, E. C. (2008). Notch signaling during larval and juvenile
development in the polychaete annelid Capitella sp. I. Developmental
Biology, 320(1), 304–318. https://doi.org/10.1016/j.ydbio.2008.04.015
Verasztó, C., Gühmann, M., Jia, H., Rajan, V. B. V., Bezares-Calderón, L. A.,
Piñeiro-López, C., … Jékely, G. (2018). Ciliary and rhabdomeric
photoreceptor-cell circuits form a spectral depth gauge in marine zoo-
plankton. eLife, 7(e36440), 1–19.
Verasztó, C., Ueda, N., Bezares-Calderón, L. A., Panzera, A., Williams, E. A.,
Shadidi, R., & Jékely, G. (2017). Ciliomotor circuitry underlying whole-
body coordination of ciliary activity in the Platynereis larva. eLife, 6,
e26000. https://doi.org/10.7554/eLife.26000
von Mack, H. (1902). Das Centralnervensystem Von Sipunculus nudus
L. (Bulchstrang). Mitt Besonderer Berucksightigung Untersuchung.
Arbeiten Aus Dem Zoologischen Instituten Der Universität Wien Und Der
Zoologischen Station in Triest, 13, 237–334.
Voronezhskaya, E. E., Tyurin, S. A., & Nezlin, L. P. (2002). Neuronal devel-
opment in larval chiton Ischnochiton hakodadensis (Mollusca: Poly-
placophora). The Journal of Comparative Neurology, 38, 25–38. https://
doi.org/10.1002/cne.10130
Wakamatsu, Y., & Weston, J. (1997). Sequential expression and role of
Hu RNA-binding proteins during neurogenesis. Development, 124,
3349–3460.
Wanninger, A. (2008). Comparative lophotrochozoan neurogenesis and
larval neuroanatomy: Recent advances from previously neglected taxa.
Acta Biologica Hungarica, 59, 127–136. https://doi.org/10.1556/ABiol.
59.2008.Suppl.21
Wanninger, A., & Haszprunar, G. (2003). The development of the seroto-
nergic and FMRF-amidergic nervous system in Antalis entalis
(Mollusca, Scaphopoda). Zoomorphology, 122, 77–85. https://doi.org/
10.1007/s00435-003-0071-6
Wanninger, A., Koop, D., Bromham, L., Noonan, E., & Degnan, B. M.
(2005). Nervous and muscle system development in Phascolion strombus
(Sipuncula). Development Genes and Evolution, 215(10), 509–518. https://
doi.org/10.1007/s00427-005-0012-0
22 CARRILLO-BALTODANO ET AL.
Weigert, A., & Bleidorn, C. (2016). Current status of annelid phylogeny.
Organisms Diversity & Evolution, 16(2), 345–362. https://doi.org/10.
1007/s13127-016-0265-7
Weigert, A., Helm, C., Meyer, M., Nickel, B., Arendt, D., Hausdorf, B., …
Struck, T. H. (2014). Illuminating the base of the annelid tree using
transcriptomics. Molecular Biology and Evolution, 31(6), 1391–1401.
https://doi.org/10.1093/molbev/msu080
Young, C. M., Sewell, M. A., & Rice, M. E. (2002). Atlas of marine inverte-
brates larvae (1st ed.), Cambridge, MA: Academic Press.
Yurchenko, O. V., Skiteva, O. I., Voronezhskaya, E. E., & Dyachuk, V. A.
(2018). Nervous system development in the Pacific oyster, Crassostrea
gigas (Mollusca: Bivalvia). Frontiers in Zoology, 15(10), 1–21.
Zattara, E. E., & Bely, A. E. (2015). Fine taxonomic sampling of nervous sys-
tems within Naididae (Annelida: Clitellata) reveals evolutionary lability
and revised homologies of annelid neural components. Frontiers in
Zoology, 12(8), 1–21. https://doi.org/10.1186/s12983-015-0100-6
Zrzavý, J., Pavel, Ř., Piálek, L., & Janouškovec, J. (2009). Phylogeny of
Annelida (Lophotrochozoa): Total-evidence analysis of morphology
and six genes. BMC Evolutionary Biology, 9(189), 1–14. https://doi.org/
10.1186/1471-2148-9-189
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of this article.
How to cite this article: Carrillo-Baltodano AM, Boyle MJ,
Rice ME, Meyer NP. Developmental architecture of the
nervous system in Themiste lageniformis (Sipuncula): New
evidence from confocal laser scanning microscopy and gene
expression. Journal of Morphology. 2019;1–23. https://doi.org/
10.1002/jmor.21054
CARRILLO-BALTODANO ET AL. 23