3  P or i f era

Sally P. Leys and Nathan Farrar

Introduction The soft tissues of sponges vary in density from cobweb thin
to massively thick (Fig.  3.1G). In all instances they consist of
Porifera is a taxon of morphologically diverse benthic animals an outer epithelial layer formed by plate-like or in some cases
that inhabit marine and freshwater environments from des- T-shaped cells—both exopinacocytes—that enclose a collagen-
ert ponds to caves, from abyssal to littoral, and from arctic to ous region with wandering cells. Internally, bordering canals and
tropical oceans. About 8500 species are recognized of some feeding chambers, endopinacocytes form the inner epithelium.
11,000 described, but more than twice that number are esti- Sponges attach to substrates by a third type of epithelial ­cell—
mated to exist based on current trends of discovery (Appeltans basopinacocytes—which also secrete a particular adhesive extra-
et  al.  2012, Van Soest et  al.  2012). Four classes are presently cellular matrix. The collagenous middle region of sponge tissues
recognized, with Hexactinellida and Demospongiae (together is partly what gives the sponge body its elasticity. Elasticity (spon-
loosely termed the Silicea) forming a sistergroup to Calcarea and giness) is also provided by a type of collagenous skeleton called
Homoscleromorpha (Fig.  3.1A–D). Although some analyses spongin that is found in the class Demospongiae. Three classes
have suggested sponges may be paraphyletic, the current con- (Demospongiae, Hexactinellida, and Homoscleromorpha) have
sensus is that the phylum Porifera is monophyletic (Nosenko skeletal components, called spicules, made of silica—in contrast
et al. 2013). to the fourth class Calcarea, in which the spicules are made of
In contrast to the great range of gross morphologies of calcium carbonate. There is one unusual group of coralline dem-
Porifera, the taxon is united in sharing a common internal form, osponges which, in addition to producing a siliceous skeleton,
consisting of flagellated cells whose beating action drives water also constructs a massive skeleton of aragonite. Skeletogenesis
from microscopic openings on the animal’s surface, through dis- differs so much in each of the groups that it is considered that the
crete epithelial-lined canals in the body, and out of larger collect- use of silica to form spicules arose independently in at least the
ing vents (Fig. 3.1A–E). Filtration of water is the primary mode Silicea and Homoscleromorpha, as did calcification in Calcarea
of food uptake and excretion, and nutrient and gas exchange, and coralline demosponges.
and this strategy has only been lost in a few members of one Sponges are suspension feeders that filter water to extract bac-
deep-sea group of demosponges, which have turned to carnivory teria and other picoplankton. Some species obtain a large por-
(these are not discussed in this chapter). tion of their nutrition from bacterial or algal symbionts, much
The ‘aquiferous’ or canal system forms the central polariz- as do stony corals, but generally a complement of food sources is
ing structure of any sponge. Incurrent openings, termed ostia, required. The apparent simplicity of filtration as a process belies
litter the surface tissue and are invisible to the naked eye. The the complexity of having the precise dimensions of filtration
excurrent vent or osculum, in contrast, is a millimetre to half system to maintain the correct pressure drop across the filter.
a metre in diameter and usually lies at the upper surface of the The sponge’s principal sensory system is epithelial and where it
animal such that water filtered, as visualized by food colouring is concentrated at particular locations in the aquiferous system
or fluorescent dye, is ejected in a jet above and directly away as ‘organs’ (Ludeman et al. 2014), it allows modifications of the
from the sponge body. The body of a sponge consists of (a) the canal dimensions to either prevent uptake of bad water by clos-
aquiferous system (Fig. 3.1E–F), (b) soft tissues, including a ing ostia, or eject particles taken in inadvertently, by contracting.
compressible collagenous middle layer (mesohyl) and, where In some cases contractions are complicated behaviours involv-
it is present, (c) an elastic and/or brittle skeleton of collagen ing coordinated inflations and contractions of the whole ani-
and minerals. mal, which have been described as ‘sneezing’ (Elliott and Leys

Structure and Evolution of Invertebrate Nervous Systems Edited by Andreas Schmidt-Rhaesa, Steffen Harzsch,
and Günter Purschke © Oxford University Press 2016. Published 2016 by Oxford University Press.

9
S t ru c t ure a n d E volu t i o n o f I n vert e b rat e Nervous S y s t e m s

Fig.  3.1.  General morphology of sponges. A–D: Representatives of the canal, arrows: apopyle exits from chambers to the canals). G: A schematic
four classes of Porifera: A: Aphrocallistses vastus, Hexactinellida, B: Ephydatia of the differences in epithelia and mesohyl cell density in a gradient from
muelleri, Demospongiae, C: Sycon coactum, Calcarea, D: Oscarella n.sp. thin- (1) to thick- (3) walled sponges (from Pavans de Ceccatty 1974a,
Homosleromorpha (C,D, courtesy of J. Chu and S. Nichols). E, F: The aquif- with kind permission by Oxford University Press) (ic: incurrent canal; ec:
erous system with canals (E) and a single choanocyte chamber (F) shown excurrent canal; ch: choanocytes). H: A larva of Amphimedon queenslandica.
by scanning electron microscopy (SEM) (ch: choanocyte chamber, ca:

2007, Meech 2008). These sensory and responsive systems are with a speed indicative of electrical signalling (Bergquist 1978,
described later in ‘Architecture’. Reiswig and Mackie 1983). Modern molecular phylogenies have
Porifera reproduce sexually and asexually. Sexual reproduc- confirmed that hexactinellids and demosponges form a clade
tion is by production of gametes—some species are hermaphro- that is sister to the other two classes, but hexactinellids remain as
ditic and others have separate sexes—and while the majority of the only group with syncytial tissues, and so far they are the only
species studied brood the fertilized egg and release a fully devel- group known to use electrical signalling. Glass sponges, which
oped ciliated swimming larva (Fig.  3.1H), in other instances comprise 8% of described species, are also unusual in being
both males and females release gametes. In the latter case oocyte restricted to deep marine habitats where silica levels are high and
fertilization and development into a swimming larva occur in temperatures low and stable. Demosponges (83% of sponge spe-
the water column. Sperm can be taken up by feeding chambers cies) inhabit the greatest range of ecosystems, from freshwater
of conspecifics or, if released as sperm packets, can be ‘caught’ to marine, and have even adapted to the deep sea by adopting
on the surface of conspecifics allowing sperm to penetrate the carnivory in addition to filtration. Calcarea, which comprise 8%
body wall and fertilize oocytes. There is some indication that of described sponges, and Homoscleromorpha (1% of sponge
like other metazoans sponges have and use pheromones, which species), the newest sponge class, are both marine groups with
sperm use to locate conspecifics and recognize gametes (Riesgo wide distribution but often cryptic morphologies (Van Soest
et al. 2014). et al. 2012).
Sponge larvae are ciliated propagules 20 µm to 2 mm in
length (Fig. 3.1H). They are all non-feeding and can stay in the
water column for a month, but in laboratory settings larvae com- Historical Investigations
monly metamorphose within 1–2 days of release from the par- Of Nervous Systems
ent. Sponge larvae can respond to gravity and light (Warburton
1966); where it has been identified, the sensory organ is a ring- It is unclear exactly when people began to consider sponges
like arrangement of cilia and pigment at one pole (Leys and in the light of the evolution of nervous systems. From Parker
Degnan 2001, Leys et al. 2002, Maldonado et al. 2003, Collin (1919), we know that Aristotle wondered at the mechanism of
et al. 2010). the responsiveness of sponges, but Lendenfeld (1885) was one
Prior to molecular studies, Hexactinellida were proposed to of the first to publish a study of the histology of nerve-like cells
form a separate subphylum (or even phylum) because of their of sponges and to suggest that sponges should no longer be con-
unusual syncytial tissues and ability to arrest the feeding current sidered to be protists. The late 1800s was a time of reflection

10
 P or i f era

on the origins of coordination and while most of those writ- Conduction Pathways
ing had studied cnidarians not sponges, they all contemplated
the early origin of, and connection between, nerves and mus-
And Effectors: Epithelia
cle, receptor, and effector (Moroz 2009, and references therein).
Cellular sponges carry out contractions in response to a broad
Parker tied these thoughts together in a seminal review in 1919,
selection of stimuli; these are not just reflex responses, as
but importantly he also carried out the first systematic experi-
described by Aristotle, in response to being torn off a rock (cited
ments on sponge responsiveness to stimuli (Parker 1910). With
by Parker, 1919); they are also able to propagate contractions in a
that work there began a series of investigations on responsive-
coordinated manner (Weissenfels 1990, Elliott and Leys 2007).
ness to chemicals and histological evidence of nervous tissues in
Most past work has focused on instant responses to stimuli, but
sponges, beginning with McNair (1923) and ending in Jones’
it is clear that there are much longer responses involving contrac-
(1962) conclusion that there is no nervous system in sponges.
tion of both epithelia and mesohyl cells, and that cells crawling
This work included important contributions by Prosser (1967),
in the mesohyl are also affected because they stop crawling dur-
Emson (1966), Lentz (1966), as well as Pavans de Ceccatty
ing contractions (Elliott and Leys 2007). Studies have shown
(1955, 1960, 1971, 1974b), and Pavans de Ceccatty et  al.
these contractions are propagated, and, most importantly, they
(1960). In recent years a very similar approach has been applied
are stereotypical and can be triggered by a threshold of stimuli
(histological studies, as well as effects of chemicals on behaviour)
(Elliott and Leys 2010, Ludeman et al. 2014).
with little further advance in cellular sponges.
The speed of contraction is typically slow—ranging from less
A major breakthrough was made in syncytial sponges (hex-
than 1 to 130µm/s (McNair 1923, Nickel 2004, Elliott and Leys
actinellids), however, with the discovery that they arrest their
2007, Bond 2013)—well below the speed of electrically propa-
feeding current upon touch (Lawn et al. 1981) and that this
gated events. This is what we fully expect for a signalling system
is carried out by an electrically propagated action potential
based on membrane-bound receptors. Propagation requires diva-
(Leys and Mackie 1997). This is not to say that there is no
lent cations and in particular calcium (Elliott and Leys 2007).
electrical signalling in demosponges. Loewenstein (1967)
There are a couple of instances where more rapid propagation
claims to have recorded coupling between cells in Microciona
occurs: McNair (1923) reported that waves of contraction can
prolifera and Haliclona occulata, and although this work has
move down the osculum of the freshwater sponge at 0.35 mm/s;
not been reproduced, Reiswig also reported a rapid (<1 s)
the other is the rapid dropping of flaps over the ostial pore fields,
response of tissues covering the ostial (incurrent) pore fields
previously discussed, as reported briefly by Reiswig (1979).
in Phorbas amaranthas (formerly Hymedesmia) (Reiswig
Except for those two instances we have no expectation of elec-
1979). The recent finding of ionotropic glutamate recep-
trical signalling in cellular sponges. Even in Calcarea the types of
tors in expressed transcripts from several sponges (Riesgo
motility measured are extremely slow (Bond 2013).
et  al.  2014), and functional work which shows that inward
rectifying potassium channels have ionic properties that sug-
gest they may function to rapidly reset membrane potential
in demosponges (Tompkins-MacDonald et  al.  2009), indi- Conduction Pathways
cates there may be a more rapid mechanism of signalling in
cellular sponges than is currently understood.
And Effectors: Syncytia
And Flagella
Architecture Glass sponge syncytia provide uninterrupted conduits for electrical
signalling. While it is quite certain sponges lack nerves, the glass
Sponges lack conventional nerves—but do sense and respond to sponge syncytium functions analogously. About 75% of the glass
stimuli. As Parker noted, the most obvious activity of a sponge sponge body consists of a single syncytial tissue that penetrates all
is its filtration of vast volumes of water—up to 1000 times the parts of the sponge (Leys 1995, 1999). The few cellular regions
body volume each day—which from some sponges pours out are also joined to the syncytium by a continuous ‘cell’ membrane,
like ‘a vigorous spring’ (Parker 1919). The flagellated pump but remain separated only by a protein plug in a narrow junction,
cells form one set of effectors. Another more gradual activity which offers no impedance to electrical impulses (Mackie and
is contraction and relaxation of part of or the whole sponge Singla 1983). A stimulus to any part of the syncytium propagates
body. This behaviour, which has been noted since the time of via calcium channels at a rate of 0.17–0.3 cm/s throughout the
Aristotle, is controlled by a second set of effectors, cells form- syncytium and causes the flagellated pump units to stop beating
ing the epithelia and/or cells of the middle layer (mesohyl) (e.g. (Leys et al. 1999). The 5 s-long action potential is slightly retarded
Pavans de Ceccatty et al. 1970, Pavans de Ceccatty 1976, Nickel by a 25–75% reduction in sodium concentration, but is com-
et al. 2011). It is easiest to address the architecture of the effect- pletely blocked by the divalent ions Co2+ and Mn2+, and by the
ors in cellular and syncytial sponges separately, for although potassium channel blocker TEA (Leys et al. 1999). It is thought
there is evidence that both effectors respond in both types of that calcium is the principal ion involved and that the slow rate of
sponge, we know far more about the epithelial and mesohyl con- propagation is due to slow restoration of the membrane potential
duits in cellular sponges, and far more about flagella as effectors and or low channel abundance. The action potential is also sen-
in hexactinellids. sitive to temperature with a Q10 of 3 (Leys and Meech 2006),

11
S t ru c t ure a n d E volu t i o n o f I n vert e b rat e Nervous S y s t e m s

and this is given as one reason why, globally, hexactinellid sponges that sense changes in tension. This must be the case in
might be restricted to deeper, cooler waters. Hexactinellida, which respond to sharp jabs and to clogging
of the filtration system by arresting the flagella pumps (Lawn
et al. 1981, Leys et al. 1999). In cellular sponges, a fish bite
Sensory Structures or clogging of the sponge filter by a sudden increase in par-
ticulate concentration may affect stretch sensitive channels,
A sensory organ by definition receives stimuli from the envir- triggering the contraction responses noted above. McNair’s
onment and signals to nerves. As sponges do not have nerves, (1923) early work with freshwater sponges showed that the
they are also not usually considered to possess sense organs, but osculum contracted rapidly if pin pricks were applied to the
recently this has come into question (Ludeman et al. 2014). If top or to the bottom of the osculum chimney (Fig. 3.2A).
sponges respond, then they are able to sense, and experimental (2) Sensory cilia. Often cilia and flagella are considered identical
work has shown three ways in which this can happen. organelles, but in sponges the different modes of function of
flagella on choanocytes and of motile and non-motile cilia
(1) The first is simple stretch sensitive channels. The epithelium on epithelia reflect hidden ultrastructural differences (Linck
of any sponge is probably sensitive as a whole, via channels 1973). In contrast to the flagella pumps, cells at the exit

Fig. 3.2.  Sensory systems in adult and larval sponges. A: McNair’s 1923 Tethya wilhelma (from Nickel 2010, with kind permission by John Wiley
drawings of the response of the osculum of Ephydatia to touch, showing and Sons). F: The ciliary response to a rapid increase (I) and decrease (II)
presumed contraction of sphincter-like bands (a,b) equally along the oscu- in light intensity in a larva from Amphimedon queenslandica (from Leys
lum, (c,d) only at the tip of the osculum, and (e,f ) in two places along the et al. 2002, with kind permission by Springer). H: Cross-section through
osculum and at its base (modified from McNair 1923, with kind permission the posterior pole of the larva of Amphimedon queenslandica (TEM) (inset
of the Marine Biological Laboratory, Woods Hole). B, C: Primary cilia in shows an SEM of the posterior pole of the larva). I,J: Cytoplasmic bridges
the osculum of Ephydatia muelleri, shown in (B) thin section transmission connect pigment-filled projections at the posterior pole of the larva of
electron microscopy (TEM), and (C) scanning electron microscopy (SEM). Sigmadocia caerulea (TEM). The presence of microtubules stabilizing the
D: Relaxed and contracted Tethya wilhelma (from Ellwanger et  al.  2007, bridges is visible in panel J (from Maldonado et al. 2003, with kind permis-
with kind permission by Springer). E: Primary cilium from the osculum of sion by Springer).

12
 P or i f era

of choanocyte chambers in many sponges have cilia, which would not be unexpected in light of the behavioural repertoire
beat in a single direction with a single recovery stroke (SPL listed above. A near classic post-synaptic density (PSD) scaffold
personal observation), as do most metazoan cilia. In add- could be present in sponges based on what is present in their gen-
ition, non-motile cilia occur in discrete locations in the omes and transcriptomes (Sakaraya et al. 2007, Alie and Manuel
aquiferous system. Careful study of specimens, preserved 2010)—even rapid acting channel receptors (ionotropic gluta-
and dissected so as to open their canal system to viewing mate receptors or iGluRs) were found in three sponge transcrip-
by scanning electron microscopy, shows that there are non- tomes (Riesgo et  al.  2014). For synaptic signalling one would
motile cilia 4–6 µm long here and there along the aquifer- expect to see clusters of vesicles at the sites of contact of two
ous system (Ludeman 2010) (Fig. 3.2B,C). The inner lining cells. Several good examples of vesicle exchange between cells
of the osculum in all demosponges studied so far possesses were shown by Pavans de Ceccatty et al. (1970) in a survey of
cilia (Nickel 2001, Ludeman et al. 2014) (Fig. 3.2D,E)— cell–cell connections in demosponges, but so far a clear example
even glass sponge oscula have 6µm-long cilia (Ludeman of a PSD structure has yet to be found in a sponge (Sakaraya
et al. 2014). Experimental work now shows that the oscula et al. 2007, Alie and Manuel 2010).
cilia may function as sensory hair cells. These cilia label with There is some evidence that sponges use molecules associated
dyes used to label ion channels in primary cilia, and chem- with synaptic transmission, though many of these studies are
icals used to block sensation by hair cells or lateral line cilia difficult to assess. Early histochemical work by Lentz (1966) on
in fish, and also block propagated contractions that are usu- mesohyl cells in the osculum of Sycon showed serotonin (5HT),
ally easily triggered by mechanical stimuli or by L-glutamate epinephrine (Epi), and norepinephrine (NE) staining in vesicles
(Ludeman et  al.  2014). The defined location and specific situated throughout the cytoplasm and near the cell membrane
function of the ciliary epithelium suggest that it is a region (Fig. 3.3A,B). Acetylcholinesterase activity was also demonstrated
that receives signals from the environment and transmits in cells near the osculum using a thiolacetic acid and lead nitrate
them to the animal, and so in essence it functions as a sen- assay, although those vesicles were located in the perinuclear area
sory organ. rather than at the cell membrane (Fig.  3, Lentz 1966). Recent
(3) ‘Photosensory’ organelles. Almost all sponge larvae are cov- analysis of several sponge transcriptomes has identified enzymes
ered with cilia, which beat unidirectionally and constantly that synthesize these molecules including 5HT (e.g. tryptophan
to move the larva forward or prevent it from sinking. These hydroxylase) (Riesgo et al. 2014), albeit none have been found
‘swimming cilia’ are approximately 20 µm long. In some so far in Sycon, the genus in which Lentz’s work was carried out.
sponge species, towards the posterior swimming pole of the Transmission electron micrographs (EM) showing exchange
larva, cilia vary in length from extremely short to those that of vesicles between cells strongly suggest that sponges do
are ten times the length of ‘swimming cilia’ (Leys and Degnan have vesicle-based transmitter signalling (Pavans de Ceccatty
2001, Collin et  al.  2010). The ‘pole’ cilia are also respon- et al. 1970) (Fig. 3.3C,D). More recent work has used immu-
sive to changes in light. In some larvae, when light intensity nohistochemical methods, which give staining patterns that are
is suddenly increased they straighten without beating, and more diffuse, with labels seen generally throughout many cells
when light intensity decreases they bend over the posterior (e.g. Ramoino et  al.  2007). While antibodies are a promising
pole (Leys and Degnan 2001). In other larvae the pole cilia technique for exploring ultrastructure, because sponge proteins
do the reverse, bending under high light and straightening are highly divergent from even their closest metazoan relatives,
when light dims (Collin et al. 2010) (Fig. 3.2F,G). The cells the use of antibodies raised against non-sponge antigens suffers
from which long cilia arise have unusual bulbous extensions from lack of cross-reactivity, so that in the absence of rigorous
of the cell surface filled with pigment granules. It seems that controls the results are difficult to interpret. Antibodies raised
these are well positioned to block light to the lower portion against sponge-specific epitopes and coupled to EM imaging
of the cilium (Fig. 3.2H). Maldonado et al. (2003) showed (immunogold) in regions of cell–cell contact, such as those
that in some larvae cytoplasmic bridges connected neigh- shown by de Ceccatty (Fig.  3C–E), promise to be informa-
bouring cytoplasmic extensions, raising the possibility that tive, and controls, such as those carried out by Lentz (1966) by
there might exist some need for rapid (electrical) signalling depleting 5HT, Epi, and NE, will be essential.
around the ciliary ring (Fig. 3.2I,J). Sponge larvae respond Reports of other neuroactive molecules in sponges have come
not only to light but also to gravity (Warburton 1966) and from physiological experiments where behaviours such as con-
chemicals (Jackson et al. 2002), so other roles for cilia on the tractions and alteration of pumping rate have been monitored
surface of sponges cannot be ruled out. (e.g. Emson 1966, Ellwanger and Nickel 2006). One question
that emerges from these findings is why sponges would require
signalling components typically associated with fast signalling
at synapses, since few sponge responses are rapid. The simplest
Neuro -Molecules reason would be that molecules are involved in vesicle-based
And Neurotransmitters endo- and exocytosis, such as occurs during feeding, as explored
by Ramoino et al. (2011). But it is also likely that contractions
Sponges may not have nerves but they do possess genes that of the whole sponge body in response to irritants in the water
suggest they either have the potential to carry out synaptic- would require coordination of all canals and therefore a signal
like signalling or carry it out in an unconventional way. This that would travel across all epithelia within seconds to minutes.

13
S t ru c t ure a n d E volu t i o n o f I n vert e b rat e Nervous S y s t e m s

Fig.  3.3.  Evidence for cell–cell signalling sponges. A, B: Stains for (A) D: Vesicle exchange via a potential cathrin-coated vesicle (C,D, from
serotonin and (B) acetylcholinesterase activity label vesicles around the Pavans de Ceccatty et al. 1970, with kind permission by Academic Press).
nucleus (N) and near the plasma membrane in cells of the osculum of E: Particles on the membrane of pinacocytes in Ephydatia muelleri shown
Sycon sp. (from Lentz 1966, with kind permission by John Wiley and by freeze fracture deep etch (from Lethias et al. 1983, with kind permis-
Sons). C: Vesicular exchange between two cells in Hippospongia communis. sion by Elsevier).

In spite of the difficulty in localizing distribution of mol- Experiments to test for immunoreactivity of regions around the
ecules in tissues, glutamatergic signalling has nonetheless been osculum to custom antibodies to these receptors and molecules
shown unequivocally to occur in freshwater sponges—where a are needed.
threshold concentration of 60–80µM L-glutamate (L-Glu) trig-
gers a propagated contraction behaviour of the whole sponge
(Elliott and Leys 2007, Ludeman et al. 2014). Lower concentra- Evolutionary
tions cause smaller, uncoordinated contractions ‘flinches’, and
higher concentrations cause the entire surface of the animal to Considerations
rupture due to the force of the contraction. While many chem- Of Neurogenesis
icals trigger slight contractions of sponges (e.g. Ellwanger and
Nickel 2006), so far L-Glu alone has been shown to cause this The expression of genes involved in neural development in
predictable, stereotypical response to precise concentrations. sponges, which do not have neurons, is intriguing. Probably the
That glutamate signalling is somehow involved in coordinating most interesting of these is bHLH, which has been shown to
whole animal behaviour is also supported by the demonstration convert cells of a non-neuronal fate to a neuronal phenotype;
that incubation in gamma-aminobutyric acid (GABA) inhibits but bHLH also has roles in muscle development (and mesoder-
all behaviour, whether triggered by L-Glu or by mechanical dis- mal cell differentiation), hematopoeisis, and skeleton develop-
turbance (Elliott and Leys 2010). Glutamate was shown by high ment (Simionato et al. 2007). A sponge bHLH, AmqbHLH1,
pressure liquid chromatography to be present in the freshwater was shown to have neurogenin-like activity in Xenopus and
sponge and glutamate decarboxylase is also present, as are a proneural activity in Drosophila, where it was able to induce the
number of glutamate and GABA receptors, and even ionotropic formation of more sensory bristles than in wild types (Richards
glutamate receptors (Elliott and Leys 2010, Riesgo et al. 2014). et al. 2008). In the Amphimedon demosponge embryo, bHLH

14
 P or i f era

genes are expressed in cells populating an unusual ‘middle layer’ system, yet appear to respond to some neurotransmitters and
of cells which differentiate into ciliated swimming cells, flask- encode genes used to build and maintain nervous systems. How
shaped cells, and cells that contain mucous inclusions, all of these genes shape the morphology and physiology of the sponge
which populate the outer epithelium of the swimming larva. remains an active area of research.
Notch Delta expression was also studied in the homosclero-
morph sponge Oscarella lobularis and found to largely occur in R e f ere n c es
choanocytes, the pump cells of the feeding chambers (Gazave
et al. 2008). It is not possible to draw inferences about neuro- Alié, A. and Manuel, M. (2010). The backbone of the post-synaptic
genic cell lineages from these data without additional character- density originated in a unicellular ancestor of choanoflagellates and
ization of the signalling properties of the cells that these genes metazoans. BMC Evolutionary Biology, 10, 34.
go on to define. This sort of work has not yet been done, but is Appeltans, W., Ahyong, S.T., Anderson, G., et  al. (2012). The mag-
likely to yield exciting data, especially if coupled with studies nitude of global marine species diversity. Current Biology, 22,
following the histogenesis of cells that form oscula and sensory 2189–2202.
structures in the sponge. Bergquist, P.R. (1978). Sponges. Hutchinson and Co., London.
In the absence of a link between neurogenic genes and func- Bond, C. (2013). Locomotion and contraction in an asconoid calcar-
eous sponge. Invertebrate Biology, 132, 283–290.
tional sensory systems in sponges, we can imagine that, as with
Carmeliet, P. and Tessier-Lavigne, M. (2005). Common mechanisms of
the PSD scaffold, neurogenic proteins may be involved in devel-
nerve and blood vessel wiring. Nature, 436, 193–200.
oping some aspect of the sponge that is not specifically involved Collin, R., Mobley, A.S., Lopez, L.B., Leys, S.P., Diaz, M.C., and
with coordination (beyond the usual need for each cell to sig- Thacker, R.W. (2010). Phototactic responses of larvae from the
nal to its neighbours), which could have been co-opted to con- marine sponges Neopetrosia proxima and Xestospongia bocatorensis
struct more rigid designs for signalling in later evolving animals. (Haplosclerida: Petrosiidae). Invertebrate Biology, 129, 121–128.
Immunogold-EM studies with custom antibodies to sponge pro- Conaco, C., Neveu, P., Zhou, H., et al. (2012). Transcriptome profiling
teins will help clarify some of the PSD proteins’ roles in sponges. of the demosponge Amphimedon queenslandica reveals genome-wide
There are other genes such as the axon guidance molecules events that accompany major life cycle transitions. BMC Genomics,
(AGMs), however, that must play alternative roles in sponges 13, 1209.
because clearly neural patterning and synapse formation are Elliott, G.R.D. and Leys, S.P. (2007). Coordinated contractions effect-
ively expel water from the aquiferous system of a fresh water sponge.
not required. Homologues of plexinA1, semaphorin3B, and
Journal of Experimental Biology, 210, 3736–3748.
EphB1 are AGMs expressed in A. queenslandica larvae (Conaco
Elliott, G.R.D. and Leys, S.P. (2010). Evidence for glutamate, GABA
et  al.  2012). AGMs create permissive and repulsive zones for and NO in coordinating behaviour in the sponge, Ephydatia muel-
migrating cells and processes during development, and are leri (Demospongiae, Spongillidae). Journal of Experimental Biology,
known to contribute to the formation of the vertebrate vascular 213, 2310–2321.
system, itself an elaborate set of interconnected tubes (Carmeliet Ellwanger, K., Eich, A., and Nickel, M. (2007). GABA and glutamate
and Tessier-Lavigne 2005). It may be that AGMs participate in specifically induce contractions in the sponge Tethya wilhelma.
the patterning and formation of the aquiferous canals in sponges Journal of Comparative Physiology A, 193, 1–11.
by restricting the ways in which the canal system can develop. Ellwanger, K. and Nickel, M. (2006). Neuroactive substances specific-
Why do sponges have a large set of genes typically associated ally modulate rhythmic body contractions in the nerveless metazoon
with nervous systems? If ctenophores are basal even to sponges, Tethya wilhelma (Demospongiae, Porifera). Frontiers in Zoology, 3, 7.
Emson, R.H. (1966). The reactions of the sponge Cliona celata to
as recently suggested (Ryan et al. 2013), then sponges may be
applied stimuli. Comparative Biochemistry and Physiology, 18,
secondarily simplified having lost a nervous system. The residual
805–827.
components may have been used in a simpler, non-neural sig- Gazave, E., Lapébie, P., Renard, E., et al. (2008). NK homeobox genes
nalling system that is well suited to the sponge’s physiological with choanocyte-specific expression in homoscleromorph sponges.
needs, and is less energetically costly. Alternatively, the signalling Development Genes and Evolution, 218, 479–489.
genes possessed by sponges may be used in a more conventional Jackson, D., Leys, S.P., Hinman, V.F., Woods, R., Lavin, M.F., and
way, and these ‘signalling units’ have been co-opted into the later Degnan, B.M. (2002). Ecological regulation of development: induc-
evolving nervous systems. A third framework, not entirely separ- tion of marine invertebrate metamorphosis. International Journal of
ate from the former, is that sponges use ‘neural’ components in Developmental Biology, 46, 679–686.
an unconventional manner. It may be that some sponge cells are Jones, W.C. (1962). Is there a nervous system in sponges? Biological
coupled through gap junctions which allow electrical signalling Reviews and Biological Proceedings of the Cambridge Philosophical
Society, 37, 1–50.
through epithelia, and this system is linked to mesohyl cells via
Lawn, I.D., Mackie, G.O., and Silver, G. (1981). Conduction system
slower vesicle-based signalling, as is suggested by some cell ultra-
in a sponge. Science, 211, 1169–1171.
structure work. This would serve to effectively couple electrical Lendenfeld, R.V. (1885). Das Nervensystem der Spongien. Zoologischer
and chemical signalling, but at speeds far slower than seen in typ- Anzeiger, 8, 47–50.
ical electrically coupled synapses, and more consistent with the Lentz, T.L. (1966). Histochemical localization of neurohumors in a
speed of behaviours observed in sponges. Unconventional use sponge. Journal of Experimental Zoology, 162, 171–179.
of neural molecules may also point to convergent mechanisms, Lethias, C., Garrone, R., and Mazzorana, M. (1983). Fine structure
though without a more complete characterization of the path- of sponge cell membranes: comparative study with freeze-fracture
ways this remains an open question. Sponges have no nervous and conventional thin section methods. Tissue & Cell, 15, 523–535.

15
S t ru c t ure a n d E volu t i o n o f I n vert e b rat e Nervous S y s t e m s

Leys, S.P. (1995). Cytoskeletal architecture and organelle transport Parker, G.H. (1919). The Elementary Nervous System. J.B. Lippincott
in giant syncytia formed by fusion of hexactinellid sponge tissues. Company, Philadelphia.
Biological Bulletin, 188, 241–254. Pavans de Ceccatty, M. (1955). Le systeme nerveux des eponges calcaires
Leys, S.P. (1999). The choanosome of hexactinellid sponges. Invertebrate et siliceuses. Annales des Sciences Naturelles, Zoologie et Biolologie
Biology, 118, 221–235. Animale, 17, 203–288.
Leys, S.P., Cronin, T.W., Degnan, B.M., and Marshall, J.N. (2002). Pavans de Ceccatty, M. (1960). Les structures cellulaires de type
Spectral sensitivity in a sponge larva. Journal of Comparative nerveux et de type musculaire de l´éponge siliceuse Tethya lyn-
Physiology A, 188, 199–202. curium LMK. Comptes Rendus de l´Academie des Sciences, 251,
Leys, S.P. and Degnan, B.M. (2001). The cytological basis of pho- 1818–1819.
toresponsive behaviour in a sponge larva. Biological Bulletin, 201, Pavans de Ceccatty, M. (1971). Effects of drugs and ions on a primitive
323–338. system of spontaneous contractions in a sponge Euspongia officinalis.
Leys, S.P. and Mackie, G.O. (1997). Electrical recording from a glass Experientia, 27, 57–59.
sponge. Nature, 387, 29–30. Pavans de Ceccatty, M. (1974a). Coordination in sponges. The founda-
Leys, S.P., Mackie, G.O., and Meech, R.W. (1999). Impulse con- tions of integration. American Zoologist, 14, 895–903.
duction in a sponge. Journal of Experimental Biology, 202, Pavans de Ceccatty, M. (1974b). The origin of the integrative
1139–1150. systems: a change in view derived from research on coelent­-
Leys, S.P. and Meech, R.W. (2006). Physiology of coordination in erates and sponges. Perspectives in Biology and Medicine, 17,
sponges. Canadian Journal of Zoology, 84, 288–306. 379–390.
Linck, R.W. (1973). Chemical and structural differences between cilia Pavans de Ceccatty, M. (1976). Cellular movements and pathways
and flagella from the lamellibranch mollusc, Aequipecten irradians. of coordination in the sponges. Bulletin de la Societe de Zoologie
Journal of Cell Science, 12, 951–981. Francaise, 100, 154.
Loewenstein, W.R. (1967). On the genesis of cellular communication. Pavans de Ceccatty, M., Cargouil, M., and Coraboeuf, E. (1960).
Developmental Biology, 15, 503–520. Les reactions motrices de l’éponge Tethya lyncurium (Lmk.)
Ludeman, D. (2010). Origin of sensory cells within the metazoa: do sponges a quelques stimulations éxperimentales (I). Vie et Milieu, 11,
have sensory cilia? Hons Thesis, Biological Sciences, University of 594–600.
Alberta, Edmonton. Pavans de Ceccatty, M., Thiney, Y., and Garrone, R. (1970). Les bases
Ludeman, D., Farrar, N., Riesgo, A., Paps, J., and Leys, S. (2014). ultrastructurales des communications intercellulaires dans les oscu-
Evolutionary origins of sensation in metazoans: functional evidence les de quelques éponges. In W. Fry, ed. The Biology of the Porifera,
for a new sensory organ in sponges. BMC Evolutionary Biology, 14, 3. pp. 449–466. Academic Press, London.
Mackie, G.O. and Singla, C.L. (1983). Studies on hexactinellid Prosser, C.L. (1967). Ionic analysis and effects of ions on contrac-
sponges. I. Histology of Rhabdocalyptus dawsoni (Lambe, 1873). tions of sponge tissues. Zeitschrift für Vergleichende Physiologie, 54,
Philosophical Transactions of the Royal Society of London B Biological 109–120.
Sciences, 301, 365–400. Ramoino, P., Gallus, L., Paluzzi, L., et al. (2007). The GABAergic-like
Maldonado, M., Durfort, M., McCarthy, D.A., and Young, C.M. system in the marine demosponge Chondrilla nucula. Microscopy
(2003). The cellular basis of photobehavior in the tufted parenchy- Research and Technique, 70, 944–951.
mella larva of demosponges. Marine Biology, 143, 427–441. Ramoino, P., Ledda, F.D., Ferrando, S., et al. (2011). Metabotropic
McNair, G.T. (1923). Motor reactions of the fresh-water sponge γ-aminobutyric acid (GABAB) receptors modulate feeding behavior
Ephydatia fluviatilis. Biological Bulletin, 44, 153–166. in the calcisponge Leucandra aspera. Journal of Experimental Zoology
Meech, R.W. (2008). Non-neural reflexes: sponges and the origins of A, 315A, 132–140.
behaviour. Current Biology, 18, R70–R72. Reiswig, H.M. (1979). A new sponge with rapid contraction systems.
Moroz, L.L. (2009). On the independent origins of complex brains and Annual Meeting of the Canadian Society of Zoology. Laval University,
neurons. Brain, Behavior and Evolution, 74, 177–190. Quebec, 83a.
Nickel, M. (2001). Cell biology and biotechnology of marine invertebrates. Reiswig, H.M. and Mackie, G.O. (1983). Studies on hexactinellid
Sponges (Porifera) as model organisms. Thesis, University of Stuttgart. sponges. III. The taxonomic status of Hexactinellida within the
Nickel, M. (2004). Kinetics and rhythm of body contractions in the Porifera. Philosophical Transactions of the Royal Society of London B
sponge Tethya wilhelma (Porifera: Demospongiae). Journal of Biological Sciences, 301, 419–428.
Experimental Biology, 207, 4515–4524. Richards, G.S., Simionata, E., Perron, M., Adamska, M., Vervoort, M.,
Nickel, M. (2010). Evolutionary emergence of synaptic nervous sys- and Degnan, B.M. (2008). Sponge genes provide new insight into
tems: what can we learn from the non-synaptic, nerveless Porifera? the evolutionary origin of the neurogenic circuit. Current Biology,
Invertebrate Biology, 129, 11–16. 18, 1156–1161.
Nickel, M., Scheer, C., Hammel, J.U., Herzen, J., and Beckmann, F. Riesgo, A., Farrar, N., Windsor, P.J., Giribet, G., and Leys, S.P. (2014).
(2011). The contractile sponge epithelium sensu lato—body con- The analysis of eight transcriptomes from all poriferan classes reveals
traction of the demosponge Tethya wilhelma is mediated by the surprising genetic complexity in sponges. Molecular Biology and
pinacoderm. Journal of Experimental Biology, 214, 1692–1698. Evolution, 31(5), 1102–1120.
Nosenko, T., Schreiber, F., Adamska, M., et  al. (2013). Deep meta- Ryan, J.F., Pang, K., Schnitzler, C.E., et al. (2013). The genome of the
zoan phylogeny: when different genes tell different stories. Molecular ctenophore Mnemiopsis leidyi and its implications for cell type evo-
Phylogenetics and Evolution, 67, 223–233. lution. Science, 342, 1336.
Parker, G.H. (1910). The reactions of sponges with a consideration of Sakaraya, O., Armstrong, K.A., Adamska, M., et al. (2007). A post-
the origin of the nervous system. Journal of Experimental Zoology, synaptic scaffold at the origin of the animal kingdom. PLoS One,
8, 765–805. 2, e506.

16
 P or i f era

Simionato, E., Ledent, V., Richards, G., et al. (2007). Origin and Van Soest, R.W.M., Boury-Esnault, N., Vacelet, J., et al. (2012). Global
diversification of the basic helix-loop-helix gene family in meta- diversity of sponges (Porifera). PLoS One, 7, 4e35105.
zoans: insights from comparative genomics. BMC Evolutionary Warburton, F.E. (1966). The behavior of sponge larvae. Ecology, 47,
Biology, 7, 133. 672–674.
Tompkins-MacDonald, G.J., Gallin, W.J., Sakarya, O., Degnan, B., Weissenfels, N. (1990). Condensation rhythm of fresh-water sponges
Leys, S.P., and Boland, L.M. (2009). Expression of a poriferan (Spongillidae, Porifera). European Journal of Cell Biology, 53,
potassium channel: insights into the evolution of ion channels in 373–383.
metazoans. Journal of Experimental Biology, 212, 761–767.

17