вход по аккаунту


Cell-substrate interactions in Cnidaria

код для вставкиСкачать
Towards an Understanding of the Molecular Basis
of Immune Responses in Sponges: The Marine
Demosponge Geodia cydonium as a Model
Institut für Physiologische Chemie, Abteilung Angewandte Molekularbiologie, Universität, Duesbergweg 6, 55099 Mainz, Germany
Geodia cydonium; sponges; immune response; evolution; allorecognition; aggregation factor; aggregation receptor; SRCR module; SCR module; immunoglobulinlike molecules; receptor tyrosine kinase; phenylalanine hydroxylase
The phylogenetic position of the phylum Porifera (sponges) is near the base of the
kingdom Metazoa. During the last few years, not only rRNA sequences but, more importantly,
cDNA/genes that code for proteins have been isolated and characterized from sponges, in particular
from the marine demosponge Geodia cydonium. The analysis of the deduced amino acid sequences of
these proteins allowed a molecular biological approach to the question of the monophyly of the
Metazoa. Molecules of the extracellular matrix/basal lamina, with the integrin receptor, fibronectin,
and galectin as prominent examples, and of cell-surface receptors (tyrosine kinase receptor),
elements of sensory systems (crystallin, metabotropic glutamate receptor) as well as homologs/
modules of an immune system (immunoglobulin-like molecules, scavenger receptor cysteine-rich
[SRCR]- and short consensus repeats [SCR]-repeats), classify the Porifera as true Metazoa. As living
fossils, provided with simple, primordial molecules allowing cell-cell and cell-matrix adhesion as
well as processes of signal transduction as known in a more complex manner from higher Metazoa,
sponges also show pecularities not known in later phyla. In this paper, the adhesion molecules
presumably involved in the sponge immune system are reviewed; these are the basic adhesion
molecules (galectin, integrin, fibronectin, and collagen) and especially the highly polymorphic
adhesion molecules, the receptor tyrosine kinase as well as the polypeptides comprising scavenger
receptor cysteine-rich (SRCR) and short consensus repeats (SCR) modules. In addition, it is reported
that in the model sponge system of G. cydonium, allogeneic rejection involves an upregulation of
phenylalanine hydroxylase, an enzyme initiating the pathway to melanin synthesis. Microsc. Res.
Tech. 44:218–236, 1999. r 1999 Wiley-Liss, Inc.
The study of self/non-self recognition in invertebrates
offers a significant insight into the origin of the sophisticated immune responses of vertebrates. Invertebrates
comprise ⬎95% of all known species in the animal
kingdom. It seems, therefore, likely that a study of
invertebrate self/non-self recognition will lead also to
the discovery of novel defense responses, as yet undetected in the more complex immune systems of vertebrates. An understanding of the nature of more ‘‘simple’’
allorecognition phenomena may result in a unifying
concept for immunology.
Sponges (Porifera) possess remarkable regeneration
capabilities as demonstrated by classical reaggregation
studies (Galtsoff, 1925; Humphreys, 1963; Wilson, 1907).
The capacity of regeneration in combination with inflammatory responses to injury (Jones, 1957) is an essential
component of their ability to survive. In addition,
sponges have developed mechanisms to distinguish
between self/non-self. The little that is known about
natural challenges to self-integrity in sponges, comes
from experimental transplantation studies. Smith and
Hildemann (1986) in their extensive review have
grouped sponge alloimmune responses in experimental
transplantations into two major rejection processes.
Some species, such as the marine sponge Axinella
verrucosa (Buscema and van de Vyver, 1983) or the
freshwater sponge Ephydatia muelleri (Mukai, 1992)
may form barriers to separate self from non-self tissue,
while others, such as the marine sponges Callyspongia
diffusa (Hildemann et al., 1979) or Geodia cydonium
(Pfeifer et al., 1992) may react by producing cytotoxic
factors that destroy the transplant.
Recently, progress in understanding allorecognition
in sponges has been made on the cellular level. In
grafted branches from two individuals of Microciona
prolifera, a specific cell type, the gray cells, accumulate
in the contact areas (Humphreys, 1994). Furthermore,
Abbreviations used: AF, aggregation factor; AR, aggregation receptor; aa,
amino acids; nt, nucleotides; SRCR, scavenger receptor cysteine-rich; SCR, short
consensus repeats.
Contract grant sponsor: Deutsche Forschungsgemeinschaft; Contract grant
number: Müu 348/12–1; Contract grant sponsor: Minerva-Foundation; Contract
grant sponsor: International Human Frontier Science Program; Contract grant
number: RG-333/96-M.
*Correspondence to: Prof. Dr. W.E.G. Müller, Institut für Physiologische
Chemie, Abteilung Angewandte Molekularbiologie, Universität, Duesbergweg 6,
55099 Mainz, Germany. E-mail: WMUELLER@mail.UNI-Mainz.DE
Received 13 March 1998; accepted in revised form 14 September 1998
Fig. 1. Scheme of AF-mediated cell recognition in G. cydonium. A
29-kDa AR is inserted into the plasma membrane to which one
galectin molecule binds. Evidence is reported in this review that
polypeptide(s) featuring SRCR and SCR domains, might function as
an AR. In the presence of Ca2⫹, a second galectin molecule binds to the
first one. Then these two molecules form a bridge between the AR and
the 140 kDa polypeptide, associated with the AF. Following this
scheme, the interactions between the AF and the AR involve the
galectin, which might bind to carbohydrate both at the 140 kDa
polypeptide and at the AR. In addition, the sponge contains an integrin
receptor, which is assumed to interact with fibronectin and collagen.
some combinations of dissociated cells originating from
unrelated sponges react in vitro by immediate cytotoxic
reactions (Humphreys, 1994). Based upon these data,
the authors suggested that the gray cells may be the
functional immunocytes of sponges.
Until now none of the classical vertebrate self- or
non-self receptors, such as MHC- or T-cell receptors and
antibodies, have been reported from invertebrates
(Humphreys, 1994). In our approach, we search for
putative structures of invertebrate immune systems,
primarily adhesion molcules, by cloning genes that
display sequence homologies to those receptors involved in self/non-self recognition in vertebrates. For
most of the studies we use the marine sponge Geodia
cydonium (Porifera, Demospongiae, Tetractinomorpha,
Astrophorida, Geodiidae) as a model animal.
Two classes of adhesion molecules have been isolated,
which can be grouped into (1) basic adhesion molecules
and (2) polymorphic adhesion molecules. In addition,
an enzyme was studied, which is included in the
phenoloxidase(s) pathway, activating the melanization
process, occurring during allorecognition tissue responses. For the transplantation studies, two different
techniques for allo- and autografting, (1) the parabiotic
attachment and (2) the insertion technique, were used.
As a result of the simplicity in their cellular organization, sponges were used for the first studies of the
mechanisms underlying specific cell adhesion in multicellular eukaryotic organisms (Wilson, 1907). The following characteristics of the sponge system make it attractive as a suitable model to investigate basic mechanisms
of cell-cell and cell-matrix interaction: (1) the loose and
flexible embedding of the cells in the mesohyl compartment (Garrone, 1978); (2) the high motility of the cells
in the organism (Simpson, 1984); (3) the relatively low
specialization and high differentiation and dedifferentiation potency of the cells; and (4) the availability of
large amounts of relatively homogeneous starting cell
material for biochemical investigations.
Cell-Cell Adhesion Molecules
Porifera possess the two adhesion systems typical for
Metazoa: the cell-cell and the cell-matrix adhesion
system. The cell-cell adhesion system in sponges consists of two major elements, the intercellular aggregation factor (AF) and the cell surface-associated aggregation receptor (AR) to which the AF binds (Fig. 1). After
the discovery of the first AFs, found simultaneously in
the marine sponges Microciona prolifera (Henkart et
Fig. 2. Adhesion molecules of the sponge Geodia cydonium; electron micrographs. A: Core structure of the aggregation factor with the
circular center and approximately 25 radiating arms. B: Sheet-like
structure formed by the sponge galectin and glycoconjugates onto
which sponge cells adhere. C: Lateral alignment of collagen fibrils in
the presence of the collagen assembly factor; formation of collagen
bundles in vitro. Preparations are shadowed with platinum. Magnification: bars in A ⫽ 0.1 µm, B ⫽ 1 µm, and C ⫽ 0.2 µm.
al., 1973; Weinbaum and Burger, 1973) and G. cydonium (Müller and Zahn, 1973), a molecular explanation
of the operation of cell adhesion molecules became
possible. The AF is a multiprotein complex with a
sedimentation coefficient of 90S (Müller, 1982). In its
‘‘native form,’’ the AF appears as a sphere with a
diameter of 1,000 Å, which displays a concave cup
structure. After treatment of the AF with detergents, a
core structure is obtained that appears as ‘‘sunbursts’’
with a circular center (diameter 1,000 Å) and approximately 25 radiating arms (Fig. 2A). Incubation of
isolated single cells with the AF, which functions speciesspecifically, for 1–3 days causes a rearrangement of the
cells, first into aggregates and finally into a functional
sponge (Müller and Zahn, 1973).
The AF interacts with a membrane component, the
AR, that has been identified, but not yet purified; recent
studies suggest that protein(s) composed of scavenger
receptor cysteine-rich (SRCR) modules (module: proteincoding domains that are flanked by introns; see Müller
and Müller, 1998) participate in the binding of the AF to
the plasma membrane. Monoclonal antibodies have
been used as tools to identify the binding domains of the
AF (Wagner-Hülsmann et al., 1996). A 140 kDa polypeptide was found to be involved in the AF-mediated
reaggregation process. This polypeptide interacts with
a galectin that links individual AF molecules to the AR
at the plasma membrane and consequently bridges two
cells together (Müller et al., 1997) (Fig. 1). Confocal
laser scanning microscopical analysis demonstrated
that both the galectin and the AF are present at the rim
of the cells (Wagner-Hülsmann et al., 1996).
Galectin. The sponge AF interacts with the AR via
galectin and the 140 kDa polypeptide (Fig. 1). The
galectin, which occurs in isoforms, was studied in
detail. The purified molecules reveal forms of Mr 13 to
18 kDa (Bretting et al., 1981), which bind specifically to the
sugars DGalNAc, DGal␤1=4DGlcNAc, DGal␤1=3DGlcNAc,
and DGalNAc. In the presence of Ca2⫹ or glycoconjugates,
the sponge galectins undergo conformational changes and
‘‘polymerize’’ to large three-dimensional clumps (DiehlSeifert et al., 1985b) (Fig. 2B). The cDNAs of two
isoforms of the galectins from G. cydonium were cloned
(Pfeifer et al., 1993; Wagner-Hülsmann et al., 1996).
The predicted proteins deduced from the complete
sequences display high similarity with the corresponding molecules from vertebrates and Caenorhabditis
elegans (Hirabayashi and Kasai, 1993; Müller et al.,
1997). The deduced aa sequences of the two isoforms
feature the characteristic carbohydrate-recognition domain at LHFNPR-G-V-N-W-E-R[H]-PF (the amino acids [aa] given in bold are those directly involved in
binding to the carbohydrate); this domain is conserved
from sponge to human (Pfeifer et al., 1993) (Fig. 3).
Based on the extent of aa substitutions the two sponge
galectins were calculated to have diverged from the
galectin, isolated from the nematode C. elegans, 800
million years (Myrs) ago (for details see: Hirabayashi
and Kasai, 1993; Müller et al., 1994) (Müller, 1997)
(Fig. 3).
Cell-Substrate Adhesion Molecules
Integrin. One major class of extracellular matrix
(ECM) receptors are the integrin receptors. Integrins
Fig. 3. Phylogenetic tree computed from the deduced aa sequences
of galectins from (1) Vertebrates: Human (HUMAN), rat (RAT), chicken
(GALLUS), frog (XENOPUS), conger eel (EEL) and (2) Invertebrates:
Nematode (CAEEL) and G. cydonium (GEODIA; isoform I and II).
Time scale indicates the time of divergence, based on aa substitution
analysis. The consensus sequence of the carbohydrate-recognition
domain is given.
are membrane-anchored heterodimer receptors composed of ␣- and ␤ subunits. At least 16 different ␣- and 8
␤-subunits have been identified which yield more than
20 heterodimeric integrin receptors. Upon binding of
their respective ligands, the integrins activate intracellular signalling pathways. The first integrin molecule
(cDNA) from sponges was isolated from G. cydonium
(Pancer et al., 1997a). The cDNA clones encoding the
subunit contain an open reading frame that encodes a
putative 118,628 Da polypeptide (Fig. 4A). Most ␣
subunits of the integrins, including the one from the
sponge, possess seven to eight repeating domains. Like
other ␣ subunits of integrins, the sponge molecule has
putative divalent cationic-binding sites. A dendrogram
was computed from the deduced amino acid [aa] sequences of integrin ␣ subunits (Fig. 4B). The sponge
integrin ␣ sequence was found to be most homologous to
the corresponding sequences from the invertebrate
species Drosophila melanogaster and C. elegans and
also to vertebrate species, e.g., mouse integrin ␣V, chick
integrin ␣VIII, and human fibronectin receptor ␣ subunit. In addition, the phylogenetic analysis revealed
that the sponge sequence branches off first from the two
invertebrate sequences; later, the three vertebrate sequences were derived.
The integrin receptor is known to bind molecules
primarily in the fibronectin and collagen families.
Fibronectin. Fibronectins are high molecular weight
glycoproteins present in most ECM and also in blood
plasma. This class of adhesive matrix proteins has been
subdivided into several types, each of them composed of
repeating modules of types I, II, and III (survey: Kreis
and Vale, 1993). A typical fibronectin consists of ⬎ 10
type I-, ⬇ 2 type II-, and ⬎ 15 type III (FN3) modules. In
contrast to the types I and II modules, the type III (FN3)
contains no disulphide bond.
Protein(s) that immunologically cross-react with human anti-fibronectin antiserum from G. cydonium have
been isolated (Pahler et al., 1998). The main bands have
sizes of 230 and 210 kDa. In addition, the cDNA was
cloned that encodes a putative ‘‘multiadhesive protein,’’
which comprises three interesting modules: (1) a fibronectin-, (2) a scavenger receptor cysteine-rich (SRCR)-,
and (3) a short consensus repeat [SCR]-module (Fig. 5).
The fibronectin module of the deduced sponge protein
comprises the characteristic topology and aa found in
fibronectin type-III (FN3) elements (Bork et al., 1996).
Even though it remains to be proven if this FN3 module
functions as an adhesion molecule in the Porifera, the
finding supports earlier immunochemical data on the
presence of fibronectin-like molecules in the freshwater
sponge Ephydatia muelleri (Labat-Robert et al., 1981).
The sponge FN3 module is phylogenetically the oldest
such module in the Metazoa.
Collagen. Collagens constitute a superfamily of extracellular matrix proteins. All collagens have triple
helical domains, with subunits (␣-chains) containing
the (Gly-Xaa-Yaa)n repetitive sequence motif. Based on
the configuration of the sequence domains, collagens
are classified into (1) fibrillar collagens, which contain
only one triple-helical domain that encompasses the
entire molecule, (2) basement membrane collagens
(type IV), (3) Facit collagens, which both comprise
several shorter triple-helical domains that are separated by nontriple-helical segments, (4) short chain
collagens, and (5) some specialized types of collagen
(Bork et al., 1996).
Until recently it was accepted that collagens are
present only in Metazoa. However, a new class of
collagens, which is assumed to have formed by convergent evolution, has been identified in fungi (Celerin et
al., 1996). In sponges, primitive fibrillar collagens have
been visualized in a number of species, e.g., Chondrosia
reniformis (Garrone et al., 1975), and G. cydonium
(Diehl-Seifert et al., 1985a). The 20- to 25-nm diameter
thin fibrils from G. cydonium display a 19.5-nm periodicity with one intraperiod band. A collagen assembly
factor that caused a reconstitution of collagen bundles
in an ordered sequential event was identified from the
extracellular matrix material (Fig. 2C) (Diehl-Seifert et
al., 1985a). In 1990, Exposito and Garrone (1990)
succeeded for the first time, in identifying a sponge
collagen of that type at the cDNA level.
After having elucidated the structural elements in
the extracellular matrix of sponges, it was tempting to
search also for peptidases involved in the metabolism of
them, e.g., of collagen, in sponges. The aa Pro dominant
in collagen is known to impose strong constraints on the
conformation of collagen molecules. Two enzymes are
involved in the final hydrolysis of Pro-containing dipeptides: the prolinase splitting dipeptides with Pro in both
Fig. 4. Sponge integrin receptor. A: Schematic presentation of the
structural features in G. cydonium integrin ␣ subunit. The heavy and
light chains are indicated (the light chain is shaded). Positions of the 8
characteristic repeats of integrins are marked 1 to 8. Three putative
Ca2⫹-binding sites as well as the C-terminal transmembrane region
are indicated. B: Dendrogram computed from the deduced aa se-
quences of integrin ␣ subunits found to be most homologous to sponge
integrin ␣ (INTG_GEODIA) with the corresponding sequences from (1)
invertebrate species, Caenorhabditis elegans (YMA1_CAEEL), and Drosophila melanogaster (ITAP_DROME); and (2) vertebrate species, mouse
integrin ␣V (ITAV_MOUSE), chicken integrin ␣VIII (ITA8_CHICK) and
human fibronectin receptor ␣ subunit (ITA5_HUMAN).
Fig. 5. Scheme of the putative ‘‘multiadhesive protein,’’ cloned
from G. cydonium. Three modules could be identified in this protein;
the fibronectin- (FN3), the scavenger receptor cysteine-rich (SRCR)-,
and the short consensus repeat (SCR or also termed Sushi) module. (1)
The deduced aa sequences of FN3 modules displayed highest similarity to those found in deuterostomes and protostomes FN3 as well as in
the pseudocoelomate C. elegans. (2) The SRCR scavenger module is
characterized by high similarity to those modules found in polypeptides of lymphocytes and macrophages, e.g., human CD6 antigen,
human CD5 surface glycoprotein, human M130 antigen, and bovine
WC1 surface antigen. (3) The SCR modules is highly related to
modules in proteins from human beta-2-glycoprotein I precursor,
mouse CR2 complement receptor type 2 precursor, D. melanogaster
locomotion-related protein, Limulus clotting factor C precursor from
Tachypleus tridentatus. While fibronectin is known to interact with
integrin, both the SRCR-rich polypeptides and the SCR-rich proteins
are involved in immune reactions.
positions and the prolidase cleaving prolyl residues in
the carboxyl-terminal position (Vanhoof et al., 1995).
The latter enzyme was selected and the corresponding
cDNA was cloned from the marine demosponge Suberites domuncula (Wiens et al., unpublished data).
Protein tyrosine kinases (PTKs) play important roles
in the response of cells to different extracellular stimuli.
PTKs are divided into two major groups, the receptor
tyrosine kinases (RTKs), which are membrane spanning molecules with similar overall structural topologies, and the non-receptor TKs, also composed of structurally similar molecules (Hunter et al., 1992). The first
RTK from lower Metazoa was identified and cloned
from G. cydonium (Müller and Schäcke, 1996). The
putative aa sequence comprises (1) the extracellular
part with a Pro/Ser/Thr-rich region, and two complete
immunoglobulin [Ig]-like domains, (2) the transmembrane domain, (3) the juxtamembrane region, and (4)
the catalytic tyrosine [TK]-domain (Fig. 8A). A similarity search with the G. cydonium TK-domain aa sequence showed that all metazoan RTKs lie on one
branch of the tree while the non-receptor TKs are
grouped in a second one; the sponge RTK is placed in a
cDNA encoding two cell-surface receptors, the receptor tyrosine kinase (RTK) and the SRCR-rich polypeptides have been isolated from G. cydonium, which have
been shown to be highly polymorphic.
Receptor Tyrosine Kinase (RTK)
Focusing on the RTK, first we provide some experimental evidence, suggesting that this molecule is involved in allorecognition events in G. cydonium. Second, polymorphism, both on aa and on nt basis, was
analyzed within the extracellular region of the RTK.
Fig. 6. Map of area around Rovinj (Croatia; Northern Adriatic Sea) with the locations at which the G.
cydonium specimens were sampled. The specimens were
collected either by dredging from a depth of 20–25 m [䊏]
or by skin diving from 3–7 m [䊐]. A total number of 14
specimens came from the deeper area and 22 from
shallow places. The animals from locations indicated by
numbers are those which fused: nos. 1 with 2 and 3
with 4.
separate branch, which splits off first from the common
tree of metazoan PTKs. It is estimated that the sponge
RTK diverged from RTKs of other metazoans 650–665
MYA (Gamulin et al., 1997).
Live specimens of a size ranging between 14– 30 cm
in diameter of G. cydonium were collected near Rovinj
(Croatia) (Fig. 6). In total, 36 specimens were used for
the experiments; 14 were collected from the deeper
sand and mud area and 22 from the shallow hard
bottom at the shore of the Limski Kanal. Two techniques for grafting were applied. First, slices of approximately 2– 5 cm thickness were cut from the surface of
animals. Size matching pieces were attached to each
other and loosely fixed by rubber bands: ‘‘parabiotic
attachment.’’ Second, applying the ‘‘insertion technique,’’ tissue pieces removed with a cork drill from one
specimen (diameter of 1 cm; approximate length
of 4 cm) were inserted into holes in the recipients,
which had a slightly narrower diameter (0.9 cm) as
described by Müller et al. (1983) and Pancer et al.
(1996) (Fig. 7).
A total of 36 specimens of G. cydonium were assayed.
Autografting experiments revealed that all paired grafts
(n ⫽ 36) fused, irrespective of the grafting method used.
In contrast, 34 out of 36 allografts showed rejection and
only two of them fused. One pair of sponges that fused
was collected in the Limski Kanal area (specimen 1 and
2) within a distance of 5.3 km, the second pair came
from the deeper region (3 and 4) from a distance of 6.5
km (Fig. 6). The contact zone of rejecting or fusing
parabiotic allografts was sliced and analyzed. Rejections were noted by the identification of a typical
demarcation line, which is absent in fusions. With the
insertion technique, the rejected graft tissue formed a
pronounced demarcation boundary and underwent necrotic degeneration as previously described (Müller et
al., 1981, 1983). No boundary line is seen in fusions
(Pancer et al., 1996).
Microscopical analysis of rejected and fused allografts revealed that the boundary zone between rejections was 250 µm thick and almost devoid of cells. At
most places the boundaries in the demarcation zones
were separated from each other (Pancer et al., 1996).
Inspection of the attachment zones at higher magnification showed that the nearly cell-free demarcation boundary in allografts was filled with collagen fibers. In
contrast, the attachment zones of fused pairs were
extremely dense with cells. The data also show that the
previously described distinction between barrier formation and cytotoxic reactions at the attachment zones
between allografts (Smith and Hildemann, 1986) should
be ascribed to the mode of grafting, i.e., ‘‘insertion’’ or
‘‘parabiosis,’’ and is not characteristic for a specific
sponge species.
Polymorphism in the Ig-Like Domains of RTK.
To test the possibility of polymorphism in the extracel-
Fig. 7. Transplantation of allogeneic tissue from the sponge G. cydonium applying the ‘‘insertion
technique’’. The grafts (arrows) were analyzed by sectioning immediately after grafting (A), or 2 days (B)
and 5 days later (C). Magnification x 2.
lular part, the Ig-like domains of RTK, the corresponding nt sequences of the gene (GCTK_2) were analysed
from six different specimens of G. cydonium by application of the technique of reverse transcriptase/polymerase chain reaction (RT-PCR). We chose for analysis the
region of the two Ig-like domains (nt 400–1047; Fig. 8),
which includes ␤-strands B1 to G1 of Ig-like domain 1,
and ␤-strands A2 to D2 of Ig-like domain 2 (nomenclature according to Williams and Barclay, 1988). For this
analysis, RNA was isolated from both fused and rejected graft pairs. Alignment of the amplified region is
shown in Figure 8B, which includes sequences from two
fused pairs, F-1 (animals 1 and 2; the sequences are
abbreviated TK1 and TK2) and F-2 (3 and 4; TK3 and
TK4) as well as from a pair that showed rejection R-1 (5
and 6; TK5 and TK6). In sponges 1, 2 , 3, and 6, the
clones that were analyzed from each sample were
identical while in sponges 4 and 5 two different sequences were found, and named TK4.1, TK 4.2, TK5.1,
and TK5.2, respectively (Pancer et al., 1996).
Assuming that the genome of G. cydonium contains
only a single locus coding for RTK, then animals 4 and 5
are heterozygous at the RTK locus. Sequence data of
the complete sponge RTK gene revealed two different
loci (Pancer et al., 1998); the primers used for PCR
analysis in the present study were designed in a way
that amplification of only one form can be expected.
Among the six animals, 18 nt substitutions were recorded (Fig. 8B/I), 12 of them leading to aa substitutions (Fig. 8 B/II). One gap of three nt is present in
sequence TK5.2, which results in loss of the aa Thr.
Substitutions in 9 of the 13 polymorphic aa positions
resulted in residues of similar physico-chemical properties; the others showed no alteration in the predicted
secondary structure of the Ig-like domains (Fig. 8 B/II).
Hence, it is reasonable to assume that the overall
structure of the extracellular polymorphic domains
remains stable.
We tested the possibility of linkage disequilibrium
between the polymorphic nt sites that were recorded
(Fig. 8 B/I) (Pancer et al., 1996). Among 9 informative
polymorphic sites (sites that segregate for only two nts
that are present at least twice) that were analyzed, 11
pairwise comparisons were found to be significant for
linkage disequilibrium by the chi-square test, indicating that some of the nt substitutions are conserved
among these alleles.
The fact that only 6 individual sponges were sampled
in this study, two pairs of fusions (1 ⫹ 2; 3 ⫹ 4) and one
pair of rejection (5 ⫹ 6), precludes at present conclusions concerning possible linkage between the phenomenon of fusion/rejection in G. cydonium and the allelic
polymorphism recorded in the Ig-like domains of the
GCTK_2 locus. Nevertheless, the data strongly indicate
distinct polymorphic alleles in this locus.
Analysis of Exon-Intron-Exon and Intron Sequences in the Ig-Like Domains in RTK. In the
previous section, we summarized the intra-species polymorphism in RT-PCR products of amplification across
most of the two Ig-like domains in the RTK gene
(475–478 bp) of G. cydonium (GCRTK_2; Pancer et al.,
1968). Subsequently, we characterized polymorphism
at the genomic level (Pancer et al., 1998).
Analysis of two clones from each of the six sponges
revealed seven distinct sequences, differing in 18 nt
positions and a tri-nucleotide deletion, which translated into 13 divergent aa positions. In order to characterize this polymorphism at the genomic level, SacIdigested G. cydonium DNA was probed either with the
two Ig-like domains, or the TK domain of GCRTK
cDNA. The Southern blot revealed the presence of
multiple hybridization-bands with the Ig-like probe
(1.8–5.3 kb), while only two bands hybridized with the
TK probe (4.8, 12 kb; Pancer et al., 1998). Northern
blotting with the same probes revealed two bands each
for the Ig-like probe (2.1, 3.3 kb) and the TK probe (1.6,
3.3 kb). Hence, we conclude that in addition to the
GCRTK_2 gene (transcript of 3.3 kb; Pancer et al.,
1998) there is at least one additional gene with a TK
domain and numerous genes with Ig-like-features.
Fig. 8.
Searching for additional Ig-like featuring genes of
G. cydonium, we performed genomic-PCR across the
two Ig-like domains that are split by an intron (Intron I
in Fig. 8A; Gamulin et al., 1997). The amplification
products displayed several DNA bands, ranging in
length from 0.65 to 1.5 kb, which appeared heterogeneous among the individuals (not shown). Fifteen independent clones were randomly selected for sequencing:
four from sponge 3 (termed TKIg3.1-TKIg3.4), three
from 4 (TKIg4.1-TKIg4.3), five from 5 (TKIg5.1TKIg5.5), and three from 6 (TKIg6.1-TKIg6.3). Fourteen of the sequences ranged in length 636–676 bp, and
one from sponge 3 was 927 bp long (TKIg3.3). Sequence
analysis revealed only two identical clones from sponge
4 (TKIg4.2, and TKIg4.3) that were discarded.
Multiple alignment of these 14 distinct genomic-PCR
clones and the corresponding region from another
long-intron clone (933 bp) from a genomic library
(GCTKGe2; Gamulin et al., 1997) appears in Figure 9.
Evidently, most of the heterogeneity (up to 83% of
length and sequence divergence) is confined to the
region within the intron boundaries (161–198 and 458
bp for the short and long introns, respectively). Independent analysis of the introns and of exon 1 and 2 regions
revealed in most cases a different pattern of clusters for
regions from the same clones, and close relatedness of
all sequences of exon 1 regions (up to 6% divergence),
and of exon 2 regions (up to 6% with the exception of
TKIg3.3 that is 7–10% divergent) (data not shown).
This is probably evidence for an independent diversification in the history of each of the three regions.
Thus, among the Ig-like featuring genes of G. cydonium we have recorded 19 distinct putative coding
regions (15 genomic clones and 4 cDNA clones) and 14
different intron sequences. Furthermore, five unique
Ig-like coding regions were cloned from sponges 3 and 5,
and four from individual 4, while only two sequences
were identical in two of seven sponges (which were not
analyzed here).
In an attempt to estimate the number of Ig-like
featuring genes of G. cydonium, we performed PCR
amplification across an intron that splits the Ig-like
domains. Amplification products from animals 1–6 were
resolved on a 7% polyacrylamide sequencing gel, revealing 6–7 or more clear bands within the range of the
short introns (Fig. 10), as well as several bands corre-
Fig. 8. Receptor tyrosine kinase (RTK) gene. A: Schematic presentation of the RTK gene from G. cydonium (termed GCRTK). The coding
region of the receptor consists of several domains: Pro-Ser-Thr rich
domain (P/S/T), Ig-like 1 separated by an intron (Intron I) from Ig-like
2, then a second intron (Intron II), a transmembrane domain (TM), the
juxtamembrane region (JM), and the TK-domain (TK). Positions of the
primers used for PCR amplification are indicated by arrows.
B: Polymorphism in the Ig-like domain of G. cydonium RTK. Multiple
alignment of RT-PCR products of amplification across the Ig-like
domains in RTK, from six individuals. Sequences were cloned from (1)
each fused pair F-1 (animals 1 and 2; the sequences are abbreviated
TK1 and TK2) and F-2 (3 and 4; TK3 and TK4) and from (2) each pair
that showed rejection R-1 (animal 5 and 6; TK5 and TK6). Five
independent clones were analyzed. Two independent clones were
sequenced from each individual; only those that were different are
shown here (TK4.1, TK4.2, TK5.1 and TK5.2). (I) nucleotide sequence
(a stretch of nt 457 to 937; Pancer et al., 1996) and (II) deduced
protein. Substitutions with respect to sequence TK1 are shown;
identities (●) and gaps are marked (-). Those nt substitutions or gaps
that caused aa changes are indicated (ⴱ). Substitutions of residues
with similar physico-chemical properties are marked (⫹).
sponding to longer introns (not shown). Strikingly, most
of the band-patterns are polymorphic among five of the
sponges while only individuals 3 and 4 share all the
intron-bands. The actual number of length-varying
introns may be larger than what appears in the gel
because the annealing-sites of the primers are in several cases divergent. Hence, some of the intron loci
might have been less efficiently amplified (perhaps the
faint bands in Fig. 10). Furthermore, this data represents only length-heterogeneity, but sequenced introns
of equal length were often diverse (some differed by as
much as 7 nt; Fig. 9). However, even in the case of
identically sized introns (TKIg3.1 and TKIg4.1) we still
recorded an aa substitution (Fig. 9).
In summary, our data on a sponge multigene family
of Ig-like genes that are divergent within and among
individuals, is supported by cloning, Southern and
Northern analyses, and specific PCR-typing. Presently,
this phenomenon only appears to be paralleled by the
human multigene family named killer cell inhibitory
receptors [KIR]. These MHC class I-specific receptors of
primate natural killer [NK] cells are transmembrane
glycoproteins consisting of either two or three extracellular C2 Ig-like domains, of which 29 distinct human
transcripts have been reported (reviewed in Valiante et
al., 1997). The number of KIR genes in the human
genome and their allelic variation are unknown, but
eight moderately different KIR sequences have already
been identified from a cDNA library from resting NK
cells of a single individual. Despite sequence identity
being greater than 75% among the extracellular domains of KIR molecules, receptors differing by only 3–5
residues appear to have very different specificities for
alleles of class I receptors (reviewed in Long et al., 1997;
Salter et al., 1997). The similarity between the sponge
and human multigene families of C2 Ig-like featuring
genes raises the intriguing question of whether the
moderately divergent receptors (89–99% aa identity
among 19 members) from the sponge also bind to
divergent ligands, and if so, what is their biological
Whatever evolutionary forces drove the extensive
multiplication of the sponge Ig-like featuring genes,
and the diversification of their coding regions and
introns, it is a striking fact that multicopies of these
genes exist in the genome of one of the lowest contemporary metazoans. Hypothetically, such divergent Ig-like
genes could have served as the evolutionary templates
for the vertebrate loci of multicopy germline V-type Ig
genes, which through V(D)J recombination rearrange
into antibodies and T-cell receptors (reviewed in Rothenfluh et al., 1995; Thompson, 1995).
Future studies will be performed to search for a possible
correlation between these polymorphic haplotype-like patterns and the results of histocompatibility tests by fusion/
rejection experiments (Pancer et al., 1996). Additionally,
we intend to look for the ligand of these polymorphic
SRCR-Rich Polypeptides: The Putative
Aggregation Receptor
Until recently the cell surface ligand for the AF of
M. prolifera or G. cydonium remained unknown. Our
initial attempts to identify this ligand, termed aggregation receptor (AR) (Müller et al., 1976), employed
screening of G. cydonium expression cDNA libraries
Fig. 9. Multiple alignment of 15 genomic sequences of the Ig-like
domains and the splitting intron; the delimiation within the exons are
the same as those represented in Figure 8. Fourteen PCR clones are
from sponges 3 (TKIg3.1–TKIg3.4), 4 (TKIg4.1, TKIg4.2), 5 (TKIg5.1–
TKIg5.5), and 6 (TKIg6.1–TKIg6.3) and the corresponding region of a
clone obtained from a genomic library (GCTKGe2). The sequences are
arranged by order of similarity; nt positions conserved among all the
sequences are shown in inverted type, while those conserved in at
least 11 of the sequences are shaded; dashes represent gaps in the
alignment. The position of the intron is superimposed above of the
alignment. The arrows mark the position in the two longer introns at
which 223 nts were omitted; the numbers at the end of the columns
indicate the numbers of the nt shown.
with polyclonal and monoclonal antibodies against its
membrane proteins. This approach was unsuccessful
due to binding of the various antibodies almost exclusively to clones expressing galectin, even in the presence of different galactans or bird nest glycoprotein as
blockers. Therefore, cDNA and genomic libraries of
G. cydonium were screened with heterologous probes
encoding adhesion receptors of higher Metazoa, as well
as PCR with degenerate primers designed to match
conserved domains.
One molecule turned out to be an important candidate for the AR in G. cydonium, a protein featuring
SRCR domains (Pancer et al., 1997b). Proteins featuring SRCR domains comprise a superfamily that includes two invertebrate and several vertebrate protein
members (Pancer et al., 1997b; Resnick et al., 1994). In
vertebrates, they function during endocytosis, e.g., the
macrophage scavenger receptor, or are involved in
binding to lectins. In sea urchins, the SRCR-containing
receptor binds speract, the egg sperm-activating pep-
Fig. 10. Genomic PCR amplification across the intron that splits
the Ig-like domains in GCRTK gene from sponges 1–6. Amplification
was performed with two primers that flank the splice site. Amplification products were resolved through a 7% sequencing gel, and
visualized by the laser imager of an automatic DNA sequencer (Li-Cor
4200). Only the region of the gel corresponding to the short intron is
shown. A sequencing reaction served to determine the length of the
amplification products (not shown), which are depicted in bp on the
left side (because of the primers these lengths are 43 bp longer than
the actual introns).
tide (Dangott et al., 1989). Previously we reported two
out of at least three different forms of the G. cydonium
SRCR proteins (6.5, 4.9, and 3.9 kb), most likely
alternatively-spliced transcripts of the same gene. One
form (4.9 kb, termed SRCR-Re) features twelve SRCR
repeats, a C-terminal transmembrane domain, and a
cytoplasmic tail, while the second (3.9 kb, SRCR-Mo)
consists only of the first ten SRCR domains and has no
transmembrane domain (Pancer et al., 1997b) (Fig.
PCR-Cloning of the SRCR-Car of G. cydonium
A third form of G. cydonium SRCR-containing protein (6.5 kb) that consists of fourteen SRCR domains,
six short consensus repeats (SCRs), a transmembrane
region, and a cytoplasmic tail is described here. Interestingly, the membrane proximal SRCR repeat contains
the fibronectin receptor binding site Arg-Gly-Asp (RGD).
A recombinant fragment of the protein, from a region
present in all three forms (of alternatively spliced
messages) of the sponge SRCR-containing proteins,
was found to inhibit AF-mediated cell-cell reaggregation, in a similar way as the antibodies raised against
this recombinant fragment. Based on these findings we
propose that the new SRCR/SCR-containing protein is
the putative cell-adhesion receptor and it was therefore designated SRCR-Car (Fig. 11).
A schematic presentation of the three sponge SRCRcontaining proteins is shown in Figure 11.
In comparison to SRCR-Car, SRCR-Re lacks SRCR
repeats 13 and 14 and SCR repeats 1–5, while SRCR-Mo
is missing SRCR repeats 11–14 and all of the SCR
repeats as well as the transmembrane region. Based
on the assumption that the three molecules represent products of alternatively spliced mRNA, it is
proposed that the gene encoding SRCR-SCR-Car consists of at least 4 introns, which are indicated in Figure
11A. A schematic presentation of the domains in the three
forms is drawn in Figure 11B (Blumbach et al., 1998).
The SRCR repeats region consists of 14 SRCR repeats. Their lengths range from 124 aa (SRCR 3) to 96
aa (SRCR 13), and aa sequence identities are between
26 and 53%. Based on the arrangements of the Cys
residues, the SRCRs were subdivided into Group A and
Group B (Resnick et al., 1994).
There are six SCR repeats in the extracellular membrane proximal domain of the SRCR-Car molecule.
These repeats range in length from 59 to 66 aa and
share aa sequence identities of 19–61%. SCR modules
have been grouped into four classes based the position
of Cys and other conserved residues (Nonaka et al.,
1993). The derived sponge consensus pattern of Cys-X29Cys-X8-Gly-X5-Cys-X3-Gly-X-Trp-X4-Pro-X-Cys is
closely related to the type III SCR consensus (Cys-X7Gly-X5-Cys-X2-Gly-X-Trp-X3-Pro-X-Cys) and only distantly related to the other three common consensus
sequences (Nonaka et al., 1993).
The indicated sequence for the predicted transmembrane region, aa1967 to aa1984 is given in Figure
Identification of Sponge Polypeptides
Cross-Reacting With McAb-rSRCR
To identify the sponge SRCR-containing proteins, a
170-aa-long polypeptide that is found in all three forms
of the SRCR-containing proteins (aa824 to aa993 in
SRCR-Mo, Fig. 11A) was expressed in Escherichia coli.
The recombinant SRCR-polypeptide (rSRCR) was affinity purified, and subsequently used to immunize mice
for the production of monoclonal antibodies, of which
the most specific was termed McAb-rSRCR. This antibody was then used to identify immuno-reacting polypeptides in a lysate from sponge cells, that was prepared in a buffer containing Triton X114 in order to
solubilize also ‘‘group 1’’ receptors, such as the putative
SRCR-containing receptor(s). The antibody reacted
strongly with two polypeptides of 220 and 117 kD,
presumably representing the 220 kD SRCR-Car. No
bands corresponding to the 166 kD SRCR-Re or the 130
kD SRCR-Mo molecules were identified. In addition,
two smaller polypeptides of 36 and 32 kD weakly
reacted with McAb-rSRCR. In the absence of the detergent Triton X114 in the lysis buffer, only the 36 and 32 kD
polypeptides could be visualized. When McAb-rSRCR was
pre-adsorbed with rSRCR, no cross-reacting polypeptides
could be seen in the cell extracts, indicating the specificity
of this monoclonal antibody (Blumbach et al., 1998).
Fig. 11. Sponge proteins featuring SRCR and SCR domains;
scheme. A: The putative gene whose alternatively-spliced transcripts
most likely encode the SRCR-Car, SRCR-Re, and SRCR-Mo molecules.
It is proposed that the gene contains at least four introns (I1– I4). In the
5’-putative exon, ten complete SRCR domains (R1– R10) and one
partial domain are encoded; the second exon contains two other SRCR
domains (R11– R12); exon three includes SRCR domains R13 and R14
together with five SCR segments (1– 5); the fourth exon contains the
sixth SCR and the transmembrane domain (TM) and the fifth exon,
the intracellular segment. SRCR domain R13 features the cell attachment signature RGD (aa 1441–1443). Dots mark the aa bordering the
exons. The portion that was expressed in E. coli as recombinant
SRCR-polypeptide is delimited by arrows [> aa824 to aa993]. The total
length of the putative coding region is 2,043 aa. B: Models of the three
deduced sponge proteins featuring either SRCRs only (SRCR-Mo) or
SRCRs and SCR domains (SRCR-Re and SRCR-Car) are shown. The
models illustrate (1) the SRCR-Car molecule, a membrane-bound form
that contains 14 SRCR repeats with an RGD signature in SRCR 13, 6
SCR repeats, a C-terminal transmembrane domain and a cytoplasmic
tail; (2) SRCR-Re that is also a membrane-bound form, features the
first 12 SRCRs and one SCR and (3) SRCR-Mo, a soluble molecule
featuring the first 10 SRCRs.
Immunohistochemical Localization
of G. cydonium SRCR-Containing Proteins
To determine the localization of the SRCR-containing
protein(s) of G. cydonium, immunohistochemistry was
performed on formaldehyde-fixed sponge sections. As
shown in Figure 12, McAb-rSRCR recognizes structures in the plasma membrane of cells; these are
especially prominent at the surface of the spherulous
cells (Fig. 12A). These cells, 20 µm in size, are characterized by their large intracellular inclusions (Simpson,
1984). This cell type, which has a mulberry-like shape,
is highlighted by inspection with Nomarsky interference contrast optics (Fig. 12B).
Effect of rSRCR-Polypeptide and McAb-rSRCR
on Reaggregation of Sponge Cells
When a single-cell suspension of G. cydonium cells is
incubated in the absence of enriched AF-fraction, the
cells form only small aggregates of approximately 100
µm in diameter during the course of one hour. Addition
of increasing concentrations of AF enhances aggregation (Müller and Zahn, 1973; Wagner-Hülsmann et al.,
1996), generating aggregates up to 1,350 ⫾ 210 µm in
the presence of 100 µg of AF-fraction.
Two types of experiments are reported here. First,
cells were pre-incubated with McAb-rSRCR Fab’ fragments, washed and subsequently incubated with in-
Fig. 12. Localization of SRCR-containing proteins of G. cydonium by immunohistochemistry. Sections
through formaldehyde-fixed tissue were probed with McAb-rSRCR and the immunocomplexes were
visualized with FITC-labeled secondary antibody. The slices were either inspected by immunofluorescence microscopy (A) or by light microscopy (Nomarsky interference contrast optics) (B). Magnifications:
(A and B) x400.
creasing concentrations of AF-fraction. As shown in
Figure 13A, the size of the aggregates did not increase
significantly. Thus, this McAb specifically blocks reaggregation of sponge cells, perhaps by interference with
the recognition sites for the AF. In the second set of
experiments, the AF-fraction was pre-incubated with
different concentrations of rSRCR-polypeptide, and was
subsequently added to the single-cell suspension. Under these conditions, the size of the aggregates formed
from single cells gradually decreased in parallel with
the increase in the amount of rSRCR-polypeptide added
(Fig. 13B). This finding is taken as an indication that
the soluble recombinant protein competes with the AF
in binding to the membrane-associated receptor(s),
thus preventing reaggregation.
A defense system, apparently common to both protostomes and deuterostomes, is the prophenoloxidase
activating system (for reviews see Johansson and Söderhäll, 1996; Smith, 1996). The final product of the
phenoloxidase activity is melanin, which is ubiquitously present throughout the metazoan kingdom. The
enzyme is activated by serine protease(s) from its
inactive form, the prophenoloxidase(s) (Aspán and
Söderhäll, 1991). Melanin and/or its reactive intermediates have been shown to be fungistatic and bacteriostatic, and to display also antiviral activity (reviewed in
Johansson and Söderhäll, 1996); it has been suggested
that melanization occurs also during allorecognition
tissue responses (Sabbadin, 1982). Melanin is a polymer formed via the phenoloxidase(s) from tyrosine, an
aa that is converted from phenylalanine by the phenylalanine hydroxylase (PAH) (see Hufton et al., 1995).
PAH is the rate-limiting enzyme in the pathway to
catabolize phenylalanine (Hufton et al., 1995). This
enzyme, together with tyrosine hydroxylase and tryptophan hydroxylase, belongs to the family of pterindependent aromatic L-amino acid hydroxylases (Grenett et al., 1987). In Figure 14, the metabolic pathway of
phenylalanine to the hormone epinephrine on one side
and to melanin at the second is outlined.
In sponges, melanization has been observed in a
number of species, e.g., Halichondria, Hymeniacidon,
Microciona, and Verongia (Kennedy, 1997). Until now,
no specific cells producing melanin, such as melanocytes known in higher invertebrates, have been identified in sponges (Simpson, 1984). Some evidence has
been presented that indicates that melanin is synthesized in those cells, which are destined to die (Pavans
de Ceccatty, 1958).
Recently, we cloned the cDNA encoding the PAH from
the sponge G. cydonium and we showed that this
enzyme is induced in those tissues that undergo death
in response to allorecognition (unpublished data).
After transplantation of allogeneic tissue from
G. cydonium by the ‘‘insertion technique,’’ the graft
underwent necrotic degeneration (Fig. 7). Already after
2 days the colour of the graft turned from yellow (Fig.
7A) to brownish (Fig. 7B). After a transplantation
period of 5 days, the graft tissue was black (Fig. 7C).
Tissue samples from the graft were analyzed for PAH
activity. The initial PAH activity, determined immediately after transplantation, was 2.6 U/mg of protein
(Fig. 15). The activity of PAH increased after already 2
days to 8.2 U/mg and reached a maximum with 9.7
U/mg after 3 days. The decrease in enzyme activity that
occurred thereafter was attributed to the necrotic process occurring in the grafts. The upregulation of the
PAH was also demonstrated on the level of PAH transcripts. mRNA isolated from the grafts after 2 days
contained 7-fold more PAH-transcripts than controls.
After 3 days the level of mRNA for PAH decreased but
was still 5.2-fold higher than at time zero.
The potential role of the G. cydonium PAH as a
regulatory protein is also supported by the calculation
that the estimated half-life of PAH is only 3.5 hours.
Based on this observation, it can be expected that this
type of immune reaction is not passive but involves the
induction of several genes. Recently, it was found that
during the apoptotic process in sponges, which is
induced by defined extracellular stimuli, the expression
of genes characteristic for apoptosis in higher vertebrates, e.g., the MA-3 gene (Wagner et al., 1998), is
Fig. 13. Effects of the monoclonal antibody McAbrSRCR, or recombinant rSRCR-polypeptide on AFmediated cell reaggregation of G. cydonium single-cell
suspension. A: Cells were pre-incubated with 100 µg of
McAb-rSRCR Fab’ fragments in 1 ml assays for 30
minutes at 20°C in Ca2⫹- and Mg2⫹ containing sea water,
washed and subsequently incubated with 3 to 100 µg of
AF-fraction. Sixty minutes later the size of the aggregates was determined optically. B: AF-fraction (100 µg/ml
final concentration) as well as different concentrations (3
to 100 µg/ml) of rSRCR-polypeptide were pre-incubated
(30 minutes at 20°C) and subsequently added to the
single cell suspension for a period of 60 minutes. The
results are from ten experiments; means ⫾ SD are
upregulated. This result may also support the assumption that allogeneic incompatibility reaction in sponges
is under genetic control.
A main protein for the AF has also been cloned from a
second sponge, Microciona prolifera (Fernandez-Busquets et al., 1996). Recently sequence analysis revealed
that this protein is polymorphic (Fernandez-Busquets
and Burger, 1997), and it was concluded that the
deduced aa sequence showed similarity to the Na⫹Ca2⫹ exchanger protein. The cDNA libraries from
G. cydonium were successfully screened for a related
sequence (unpublished data), and sequence comparison
of the G. cydonium cDNA revealed highest similarity to
P- and E-selectins that are known to be transmembrane glycoproteins with the characteristic motifs of an
N-terminal Ca2⫹-type (C-type) lectin, an EGF motif,
series of contiguous complement regulatory domains, a
transmembrane domain, and a short cytosolic tail
(Kreis and Vale, 1993).
An alignment study of the G. cydonium deduced aa
sequence, preliminarily also termed AF, GeoCy_AF,
with those polymorphic sequences from the presumed
AF from M. prolifera, MiPro_AF, is shown in Figure 16.
Only the segment displaying highest similarity is
shown here. The high degree of similarity between the
G. cydonium and M. prolifera sequences is striking.
However, a definite conclusion about a potential relationship can only be given after further cell biological
studies using antibodies raised against recombinant
Sponges may be considered as a key group of Metazoa
because—based on sequence data—they are the direct
descendants of the first multicellular animal from
which all other Metazoa originated (Müller, 1995, 1997,
1998a,b). Like other invertebrates, sponges possess a
natural immunity. As reported earlier for other marine
and freshwater sponges, the sponge species studied
here, G. cydonium, forms a physico-chemical barrier,
which separates two incompatible allografts from each
other. In G. cydonium grafting of incompatible tissue by
the ‘‘insertion technique’’ resulted in cellular death and
necrosis of the allograft, possibly as a result of cytotoxic
interactions and perhaps also due to humoral toxic
factor(s). It is interesting to recall that these so-called
simple sponges may already possess an alloimmune
memory, as suggested for C. diffusa (Bigger et al.,
The theories concerning a possible genetic system
that underlies allorecognition in sponges are still under
debate. While Hildemann and his colleagues favored
the hypothesis of identical histocompatibility antigens
as a prerequisite for fusion (Hildemann et al., 1980),
Curtis and coworkers (Curtis, 1979; Curtis et al., 1982)
concluded from their experiments that allograft acceptance could also occur between genetically different
specimens. In the present study, we attempted to
approach the question of genetic control of allorecogni-
Fig. 14. The metabolic pathway of phenylalanine to the hormone
epinephrine as well as to melanin, a polymer synthesized during
allorecognition tissue responses. The first enzyme is the phenylala-
nine hydroxylase (PAH), which requires tetrahydrobiopterin [BH4] as
coenzyme, which undergoes reduction to dihydrobiopterin [BH2] during the monooxygenase reaction.
tion in sponges, at the molecular level for the first time.
Molecules that might be involved in host defense have
recently been cloned from the sponge G. cydonium: a
galectin (Pfeifer et al., 1993) and the RTK with its two
Ig-like domains (Schäcke et al., 1994). The experimental data showed that the Ig-like domains of the RTK
were highly polymorphic.
To the best of our knowledge, this is the first report
that demonstrates extensive polymorphism within Iglike domains of an invertebrate receptor, which may be
a part of the mechanism of allorecognition. Further-
more, these data demonstrate in an unequivocal manner that fusing sponge grafts need not be genetically
identical. We suggest that a monomorphic receptor,
such as the galectin of G. cydonium, is a candidate for
mediating species-specific recognition, and a polymorphic receptor such as the Ig-like domains of the RTK
are mediators of individual-specific recognition. However, due to the fact that there is no further molecular
data than what is presented here, which might explain
the basis of invertebrate immunity (especially with
respect to sponges), it is inappropriate to speculate
Fig. 15. Increase of phenylalanine hydroxylase (PAH) activity in
allografts from G. cydonium. Tissue was removed immediately after
grafting (time zero) or 1 to 10 days after transplantation. The activity
of PAH is given in units/mg of protein.
further about a potential involvement of the Ig-like
domains of G. cydonium RTK in innate or acquired
immunity at the present time.
In animals, many of the soluble and cell-surface
receptors involved in immune and host defense contain
SRCR or SCR repeats. In vertebrates, SRCR repeats
are found predominantly in receptors of the immune
system, expressed on the surface of B and T cells and
macrophages (Resnick et al., 1994). SCRs are protein
motifs of 60–70 aa found in several vertebrate immune
and nonimmune proteins (Reid and Day, 1989). SCRs
are common in receptors and regulatory proteins of the
vertebrate complement system (Reid and Day, 1989), as
well as selectins and mucins (Migaki et al., 1995).
Invertebrate SCR containing proteins include the Unc-5
protein of the Nematode Caenorhabditis elegans, the
cell adhesion molecule gp70 from a slime mold (see
Lipke, 1996), Drosophila hig-protein of the central
nervous system, Limulus (horseshoe crab) coagulation
factor C, two sea urchin (Strongylocentrotus purpuratus) complement-like components, and a tunicate (Botryllus schlosseri) protein (see Pancer et al., 1995).
We characterized the first SRCR-containing proteins
of the lowest contemporary Metazoans, from the sponge
G. cydonium (Pancer et al., 1997b). Three sponge
proteins were reported: a membrane-bound form with
12 SRCR repeats and a single SCR (SRCR-Re), a
soluble form that contained only the first ten SRCR
repeats but no transmembrane domain (SRCR-Mo),
apparently alternatively spliced forms of the same
gene, as well as the a third form, SRCR-Car.
The sponge SRCR repeats belong to Group A SRCRs,
together with MARCO, macrophage scavenger receptor
I, cyclophilin-C-binding protein, MAC2-binding protein, speract, and CFI. Group A SRCRs contain six Cys
residues per domain (termed C2, C3, C5, C6, C7, and C8),
while group B SRCRs contain eight Cys residues,
including those at positions C1 and C4 (located approximately 15 residues before C2 and 4 residues after C3,
respectively), a conserved Gly adjacent to C4 and often
also a Trp two residues before C6 (Resnick et al., 1994).
An interesting feature of the G. cydonium SRCR-Car
molecule is the presence of the fibronectin receptor
binding site Arg-Gly-Asp (reviewed in Ruoslahti, 1996)
in the thirteenth SRCR domain. It is possible that the
putative AR, SRCR-Car, interacts also with cell surface
integrin(s) that have been cloned from G. cydonium
(Pancer et al., 1997a) and recently also from another
sponge, Ophlitaspongia tenuis (Brower et al., 1997).
As known since 1973 (Weinbaum and Burger, 1973),
cell-cell interactions in the sponge M. prolifera are
mediated by heterophilic interactions of the third order
(AR-AF-AR), and a similar mechanism has been proposed for G. cydonium (Müller, 1982). The 220 kD new
polypeptide of G. cydonium closely corresponds in size
to that of the proposed AR protein-complex of
M. prolifera, for which two proteins of 210 and 68 kD
have been described (Varner, 1995). Therefore, based on
sequence data, immunohistochemistry and in vitro
reaggregation assays, we propose that the 220 kD
polypeptide might represent the sponge AR. Thus, it is
possible that some of the G. cydonium SRCR-containing proteins are part of the AR that mediates the
heterophilic cellular interactions via the AF to adjacent
cells. Future studies will be conducted to determine the
detailed interaction between the AF and the proposed
AR, the SRCR-containing protein(s), and to determine
whether this molecule interacts as a monomer or in a
multimeric form with its ligand(s).
In conclusion, the demosponge G. cydonium is a
surviving species of a phylogenetic lineage that, according to paleontological data, was established more than
550 MYA (Müller and Müller, 1998; Reitner and Mehl,
1995). The spicules, one morphological characteristic of
the taxon Geodiidae, changed very little since that
time. According to the determinations of the rates of aa
substitutions of the deduced sequences from the
G. cydonium galectin(s) and the receptor tyrosine kinase, this taxon diverged 800 to 665 MYA from other
Until 1993, when the first cDNA, coding for a sponge
adhesion molecule, was isolated and characterized
(Pfeifer et al., 1993), Porifera were considered as merely
colonies of unspecialized cells— ‘‘individual flagellates’’– which require only cells that secrete adhesive
glycoprotein and bind to it (e.g., Loomis, 1988). Now
that several cDNAs coding for ‘‘metazoan’’ adhesion
molecules and receptors have been cloned, it has become clear that these proteins show high homology to
the corresponding molecules from higher metazoan
One major step in the evolution of the Metazoa from
protists was the production of molecules forming the
ECM. These molecules provide the scaffold and the
support for tissues and organs. Until now, only constituents of the ECM have been detected in sponges. However, these elements alone are not sufficient to allow the
development of tissues and organs; receptors are required for a tuned and controlled interaction between
cells themselves and between cells and extracellular
molecules. In sponges, two main signalling receptors
have been detected, (1) a receptor tyrosine kinase and
(2) elements of seven-spanning G protein-linked receptors (Müller, 1998a,b). Hence, the animals, belonging to
the lowest metazoan phylum, the Porifera, are provided
with an elaborated intercellular communication network, allowing a coordination of growth, differentiation, tissue formation, and allorecognition.
Fig. 16. Alignment of the G. cydonium deduced aa sequence,
GeoCy_AF, with the sequence from the M. prolifera, MiPro_AF. The
following polymorphic sections of the M. prolifera sequence are
compared: MiPro_AFco, core protein; MiPro_AFb, AF form B;
MiPro_AFc, AF form C; MiPro_AFd, AF form D and MiPro_AFe, AF
form E. Residues conserved in all sequences are shown in inverted
type; those present in at least four sequences are shaded.
gene in sponges reveals the early evolutionary appearance of two
collagen families. Proc Natl Acad Sci USA 87:6669–6673.
Fernandez-Busquets X, Burger MM. 1997. The main protein of the
aggregation factor responsible for species-specific cell adhesion in
the marine sponge Microciona prolifera is highly polymorphic. J Biol
Chem 272:27839–27847.
Fernandez-Busquets X, Kammerer RA, Burger MM. 1996. A 25 kDa
protein is the basic unit of the core from the 2x104-kDa aggregation
factor responsible for species-specific cell adhesion in the marine
sponge Microciona prolifera. J Biol Chem 271:23558–23565.
Galtsoff PS. 1925. Regeneration after dissociation (an experimental
study on sponges); II. J Exp Zool 42:223–251.
Gamulin V, Skorokhod A, Kavsan V, Müller IM, Müller WEG. 1997.
Experimental indication against blockwise evolution of metazoan
protein molecules: example receptor tyrosine kinase gene from the
sponge Geodia cydonium. J Mol Evol 44:242–252.
Garrone R, Huc A, Junqua S. 1975. Fine structure and physicochemical studies on the collagen of the marine sponge Chondrosia
reniformis Nardo. J Ultrastruct Res 52:261–275.
Garrone R. 1978. Phylogenesis of connective tissue. Basel: Karger
Grenett HE, Ledley FD, Reed LL,Woo SLC. 1987. Full-length cDNA
for rabbit tryptophan hydroxylase: functional domains and evolution of aromatic amino acid hydroxylases. Proc Natl Acad Sci USA
Henkart P, Humphreys S, Humphreys T. 1973. Characterization of
sponge aggregation factor. A unique proteoglycan complex. Biochemistry 12:3045–3050.
Hildemann WH, Johnston IS, Jokiel PL. 1979. Immunocompetence in
the lowest metazoan phylum: transplantation immunity in sponges.
Science 204:420–422.
Hildemann WH, Bigger CH, Johnston IS, Jokiel PL. 1980. Characteristics of transplantation immunity in the sponge Callyspongia
diffusa. Transplantation 30:362–367.
Hirabayashi J, Kasai K. 1993. The family of metazoan metalindependent ␤-galactoside-binding lectins: structure, function and
molecular evolution. Glycobiology 3:297–304.
Hufton SE, Jennings GI, Cotton RGH. 1995. Structure and function of
the aromatic amino acid hydroxylases. Biochem J 311:353–366.
Humphreys T. 1963. Chemical and in vitro reconstitution of sponge
cell adhesion. I. Isolation and functional demonstration of the
components involved. Dev Biol 8:27–47.
Aspán A, Söderhäll K. 1991. Purification of prophenoloxidase from
crayfish blood cells and its activation by an endogenous serine
proteinase. Insect Biochem 21:363–373.
Bigger CH, Jokiel PL, Hildemann WH, Johnston IS. 1982. Characterization of alloimmune memory in a sponge. J Immunol 129:1570–1572.
Blumbach B, Pancer Z, Diehl-Seifert B, Steffen R, Münkner J, Müller I
Müller WEG. 1998. The putative sponge aggregation receptor:
Isolation and characterization of a molecule composed of scavenger
receptor cysteine-rich domains and short consensus repeats. J Cell
Sci 111:2635–2644.
Bork P, Downing AK, Kieffer B, Campbell ID. 1996. Structure and
distribution of modules in extracellular proteins. Q Rev Biophys
Bretting H, Phillips SG, Klumpart HJ, Kabat EA. 1981. A mitogenic
lactose-binding lectin from the sponge Geodia cydonium. J Immunol
Brower DL, Brower SM, Hayward DC, Ball EE. 1997. Molecular
evolution of integrins: genes encoding integrin ␤ subunit from a
coral and a sponge. Proc Natl Acad Sci USA 94:9182–9187.
Buscema M, Van de Vyver G. 1983. Variability of allograft rejection
processes in Axinella verrucosa. Dev Comp Immunol 7:613–616.
Celerin M, Ray JM, Schisler NJ, Day AW, Stetter-Stevenson WG,
Laudenbach DE. 1996. Fungal fimbriae are composed of collagen.
EMBO J 15:4445–4453.
Curtis ASG. 1979. Individuality and graft rejection in sponges or a
cellular basis for individuality in sponges. In: Larwood G, Rosen BR,
editors. Biology and systematics of colonial organisms. New York:
Academic Press, p 39–48.
Curtis ASG, Kerr J, Knowlton N. 1982. Graft rejection in sponges.
Transplantation 33:127–133.
Dangott LJ, Jordan JE, Bellet RA, Galbers DL. 1989. Cloning of the
mRNA for the protein that crosslinks to the egg peptide speract.
Proc Natl Acad Sci USA 86:2128–2132.
Diehl-Seifert B, Kurelec B, Zahn RK, Dorn A, Jerecevic B, Uhlenbruck
G, Müller WEG. 1985a. Attachment of sponge cells to collagen
substrata: effect of a collagen assembly factor. J Cell Sci 79:271–285.
Diehl-Seifert B, Uhlenbruck G, Geisert M, Zahn RK, Müller WEG.
1985b. Physicochemical and functional characterization of the polymerization process of the Geodia cydonium lectin. Eur J Biochem
Exposito JY, Garrone R. 1990. Characterization of a fibrillar collagen
Humphreys T. 1994. Rapid allogeneic recognition in the marine
sponge Microciona prolifera. Ann NY Acad Sci 712:342–345.
Hunter T, Lindberg RA, Middlemas DS, Tracy S, Geer P. 1992.
Receptor protein kinases and phosphatases. Cold Spring Harbor
Symp Quant Biol 58:25–41.
Johansson MW, Söderhäll K. 1996. The prophenoloxidase activating
system and associated proteins in invertebrates. Prog Mol Subcell
Biol 15:46–66.
Jones WC. 1957. The contractility and healing behavior of pieces of
Leucosolenia complicata. Q J Microsc Sci 71:203–217.
Kennedy GY. 1997. Pigments of marine invertebrates. In: Russell FS,
Young M, editors. Advances in marine biology. New York: Academic
Press, 16:314–381.
Kreis T, Vale R. 1993. Guidebook to the extracellular matrix and
adhesion proteins. Oxford: Oxford University Press.
Labat-Robert J, Robert L, Auger C, Lethias C, Garrone R. 1981.
Fibronectin-like proteins in Porifera. Its role in cell aggregation.
Proc Natl Acad Sci USA 78:6261–6265.
Lipke PN. 1996. Cell adhesion proteins in the non-vertebrate eukaryotes. Prog Mol Subcell Biol 17:119–157.
Long EO, Burshtyn DN, Clark WP, Peruzzi M, Rajagopalan S, Rojo S,
Wagtmann N, Winter CC. 1997. Killer cell inhibitory receptors:
diversity, specificity and function. Immunol Rev 155:135–144.
Loomis WF. 1988. Four billion years: an assay on the evolution of
genes and organisms. Sunderland, MA: Sinnauer.
Migaki GI, Kahn J, Kishimoto TK. 1995. Mutational analysis of the
membrane-proximal cleavage site of L-selectin: relaxed sequence
specificity surrounding the cleavage site. J Exp Med 182:549–447.
Mukai H. 1992. Allogeneic recognition and sex differentiation in
chimeras of the freshwater sponge Ephydatia muelleri. J Exp Zool
Müller WEG. 1982. Cell membranes in sponges. Intern Rev Cytol
Müller WEG. 1995. Molecular phylogeny of metazoa [animals]: monophyletic origin. Naturwissenschaften 82:321–329.
Müller WEG. 1997. Origin of metazoan adhesion molecules and
adhesion receptors as deduced from their cDNA analyses from the
marine sponge Geodia cydonium. Cell Tissue Res 289:383–395.
Müller WEG. 1998a. Molecular phylogeny of Eumetazoa: experimental evidence for monophyly of animals based on genes in sponges
[Porifera]. Prog Mol Subcell Biol 19:89–132.
Müller WEG. 1998b. Origin of Metazoa: sponges as living fossils.
Naturwissenschaften 85:11–25.
Müller WEG, Zahn RK. 1973. Purification and characterization of a
species-specific aggregation factor in sponges. Exp Cell Res 80:95–104.
Müller WEG, Müller IM. 1998. Transition from Protozoa to Metazoa:
an experimental approach. Prog Mol Subcell Biol 19:1–22.
Müller WEG, Schäcke H. 1996. Characterization of the receptor
protein-tyrosine kinase gene from the marine sponge Geodia cydonium. Prog Mol Subcell Biol 17:183–208.
Müller WEG, Müller I, Zahn RK, Kurelec B. 1976. Species-specific
aggregation factor in sponges. VI. Aggregation receptor from the cell
surface. J Cell Sci 21:227–241.
Müller WEG, Bernd A, Zahn RK, Kurelec B, Dawes K, Müller I,
Uhlenbruck G. 1981. Xenograft rejection in marine sponges. Eur J
Biochem 116:573–579.
Müller WEG, Müller I, Zahn RK, Kurelec B, Batel R, Uhlenbruck G.
1983. Systematic value of histocompatibility reactions in the establishment of closely related sponge species; description of Geodia
rovinjensis n. sp. Senckenbergiana Biol 63:303–311.
Müller WEG, Schröder HC, Müller IM, Gamulin V. 1994. Phylogenetic
relationship of ubiquitin repeats of the polyubiquitin gene from the
marine sponge Geodia cydonium. J Mol Evol 39:369–377.
Müller WEG, Blumbach B, Wagner-Hülsmann C, Lessel U. 1997.
Galectins in the phylogenetically oldest metazoa, the sponges
[Porifera]. Trends Glycosci Glycotechnol 9:123–130.
Nonaka M, Matsuda Y, Shiroishi T, Moriwaki K, Nonaka M, NatsuumeSakai S. 1993. Molecular cloning of the ␤ subunit of mouse coagulation
factor XIII and assignment of the gene to chromosome 1: close evolutionary relationship to complement factor H. Genomics 15:535–542.
Pahler S, Blumbach B, Müller I, Müller WEG. 1998. A putative
multiadhesive basal lamina protein from the marine sponge Geodia
cydonium: cloning of the cDNA encoding a fibronectin-, an SRCR- as
well as a complement control protein module. J Exp Zool 282:332–
Pancer Z, Gershon H, Rinkevich B. 1995. Cloning of a protochordate
cDNA featuring mammalian short consensus repeats (SCR) of
complement-control protein superfamily. Comp Biochem Physiol
Pancer Z, Kruse M, Schäcke H, Scheffer U, Steffen R, Kovács P, Müller
WEG. 1996. Polymorphism in the immunoglobulin-like domains of
the receptor tyrosine kinase from the sponge Geodia cydonium. Cell
Adhesion Commun 4:327–339.
Pancer Z, Kruse M, Müller I, Müller WEG. 1997a. On the origin of
adhesion receptors of metazoa: cloning of the integrin ␣ subunit cDNA
from the sponge Geodia cydonium. Molec Biol Evol 14:391–398.
Pancer Z, Münkner J, Müller I, Müller WEG. 1997b. A novel member
of an ancient superfamily: sponge Geodia cydonium, Porifera)
putative protein that features scavenger receptor cysteine-rich
repeats. Gene 193:211–218.
Pancer Z, Skorokhod A, Blumbach B, Müller WEG. 1998. Multiple
Ig-like featuring genes divergent within and among individuals of
the marine sponge Geodia cydonium. Gene 207:227–233.
Pavans de Ceccatty M. 1958. La melanisation chez quelques éponges
calcaires et siliceuses: ses rapports avec le systéme reticulohistocytaires. Arch Zool Exp Gén 96:1–51.
Pfeifer K, Schröder HC, Rinkevich B, Uhlenbruck G, Hanisch F-G,
Kurelec B, Scholz P, Müller WEG. 1992. Immunological and biological identification of tumor necrosis factor in sponges: role of this
factor in the formation of necrosis in xenografts. Cytokine 4:161–
Pfeifer, K, Haasemann M, Gamulin V, Bretting H, Fahrenholz F,
Müller WEG. 1993. S-type lectins occur also in invertebrates: high
conservation of the carbohydrate recognition domain in the lectin
genes from the marine sponge Geodia cydonium. Glycobiol 3:179–
Reid KBM, Day AJ. 1989. Structure-function relationship of the
complement components. Immunol Today 10:177–180.
Reitner J, Mehl D. 1995. Early Paleozoic diversification of sponges:
new data and evidences. Geol Paläont Mitt Innsbruck 20:335–347.
Resnick D, Pearson A, Krieger M. 1994. The SRCR superfamily: a
family reminiscent of the Ig superfamily. Trends Biochem Sci
Rothenfluh H, Blanden S, Steele EJ. 1995. Evolution of V genes: DNA
sequence structure of functional germline genes and pseudogenes.
Immunogenetics 42:159–171.
Ruoslahti E. 1996. RGD and other recognition sequences for integrins.
Annu Rev Cell Dev Biol 12:697–715.
Sabbadin A. 1982. Formal genetics of ascidians. Am Zool 22:765–773.
Salter RD, Chan H-W, Tadikamalla R, Lawlor DA. 1997. Domains
organization and sequence relationship of killer cell inhibitory
receptors. Immunol Rev 155:175–182.
Schäcke H, Rinkevich B, Gamulin V, Müller I M, Müller WEG. 1994.
Immunoglobulin-like domain is present in the extracellular part of
the receptor tyrosine kinase from the marine sponge Geodia cydonium. J Mol Recog 7:27–276.
Simpson TL. 1984. The cell biology of sponges. New York: SpringerVerlag.
Smith LC, Hildemann WH. 1986. Allogeneic cell interactions during
graft rejection in Callyspongia diffusa (Porifera; Demospongia); a
study with monoclonal antibodies. Proc R Soc Lond B 226:465–377.
Smith VJ. 1996. The prophenoloxidase activating system: a common
defense pathway for deuterostomes and protostomes. Adv Comp
Environ Physiol 23:75–114.
Thompson CB. 1995. New insights into V(D)J recombination and its
role in the evolution of the immune system. Immunity 3:531–539.
Valiante NM, Lienert K, Shiling HG, Smits BJ, Parham P. 1997. Killer
cell receptors: keeping pace with MHC class I evolution. Immunol
Rev 155:155–164.
Vanhoof G, Goossens F, De Meester I, Hendriks S, Sharpe S. 1995.
Proline motifs in peptides and their biological processing. FASEB J
Varner J A. 1995. Cell adhesion in sponges: potentiation by a cell surface 68
kDa proteoglycan-binding protein. J Cell Sci 108:311–3126.
Wagner C, Steffen R, Koziol C, Batel R, Lacorn M, Steinhart H, Simat
T, Müller WEG. 1998. Apoptosis in marine sponges: a biomarker for
environmental stress (cadmium and bacteria). Marine Biol 131:411–
Wagner-Hülsmann C, Bachinski N, Diehl-Seifert B, Blumbach B,
Steffen R, Pancer Z, Müller WEG. 1996. A galectin links the
aggregation factor to cells in the sponge [Geodia cydonium] system.
Glycobiology 6:785–793.
Weinbaum G, Burger MM. 1973. A two-component system for surface
guided reassociation of animal cells. Nature 244:510–512.
Williams AF, Barclay AN. 1988. The immunoglobulin superfamily:
domains for cell surface recognition. Ann Rev Immunol 6:381–405.
Wilson HV. 1907. On some phenomena of coalescence and regeneration
in sponges. J Exptl Zool 5:245–258.
Без категории
Размер файла
1 184 Кб
interactions, cnidaria, substrate, cells
Пожаловаться на содержимое документа