Báo cáo khoa học: Selenium affects biosilica formation in the demosponge Suberites domuncula Effect on gene expression and spicule formation pot

Selenium affects biosilica formation in the demosponge
Suberites domuncula
Effect on gene expression and spicule formation
Werner E. G. Mu
¨
ller
1
, Alexandra Borejko
1
, David Brandt
1
, Ronald Osinga
2
, Hiroshi Ushijima
3
,
Bojan Hamer
4
, Anatoli Krasko
1
, Cao Xupeng
1
, Isabel M. Mu
¨
ller
1
and Heinz C. Schro
¨
der
1
1 Institut fu

¨
r Physiologische Chemie, Abteilung Angewandte Molekularbiologie, Universita
¨
t, Mainz, Germany
2 Wageningen University, Fish Culture and Fisheries Group, the Netherlands
3 Department of Developmental Medical Sciences, Institute of International Health, Graduate School of Medicine, University of Tokyo, Japan
4 Center for Marine Research, ‘Ruder Boskovic’ Institute, Rovinj, Croatia
The synthesis of siliceous spicules in sponges (phylum
Porifera) is unique in the metazoan kingdom. This
form of biomineralization, which results in the forma-
tion of polymerized amorphous silica, leads to the
production of filigree and highly structured skeletal
elements with morphologies specific to sponge species.
The process of silica ⁄ spicule formation in sponges can
be designated biologically controlled mineralization [1],
as the reactions are driven by cellular activities that
govern (a) nucleation, (b) growth, (c) morphology, and
(d) location of the spicules within the specimen.
A breakthrough in our understanding of spicule for-
mation in siliceous sponges came from the studies of
Shimizu et al. [2] and Cha et al. [3], who showed that
the formation of the skeletal framework and silica is
enzymatically controlled. The major enzyme that initi-
ates nucleation of spicule formation was termed silica-
tein [3]. In the last few years, several silicatein enzymes
(which catalyze biosilicification) from different demo-
sponges have been identified; according to their protein
sequences they belong to the family of cathepsin L pro-
teolytic enzymes [3,4].
Keywords

selenium; silica; silicatein; spicules; sponges
Correspondence
W. E. G. Mu
¨
ller, Institut fu
¨
r Physiologische
Chemie, Abteilung Angewandte
Molekularbiologie, Universita
¨
t,
Duesbergweg 6, 55099 Mainz, Germany
Fax: +49 6131 3925243
Tel: +49 6131 3925910
E-mail:
(Received 24 March 2005, revised 18 May
2005, accepted 26 May 2005)
Note
The cDNA sequences for selenoprotein M
(AJ875186) and spicule-associated protein
(AJ872182) have been deposited at
EMBL ⁄ GenBank.
doi:10.1111/j.1742-4658.2005.04795.x
Selenium is a trace element found in freshwater and the marine environ-
ment. We show that it plays a major role in spicule formation in the demo-
sponge Suberites domuncula. If added to primmorphs, an in vitro sponge
cell culture system, it stimulates the formation of siliceous spicules. Using
differential display of transcripts, we demonstrate that, after a 72-h expo-
sure of primmorphs to selenium, two genes are up-regulated; one codes for
selenoprotein M and the other for a novel spicule-associated protein.

The deduced protein sequence of selenoprotein M (14 kDa) shows charac-
teristic features of metazoan selenoproteins. The spicule-associated protein
(26 kDa) comprises six characteristic repeats of 20 amino acids, composed
of 10 distinct hydrophobic regions ( 9 amino acids in length). Recombin-
ant proteins were prepared, and antibodies were raised against these two
proteins. Both were found to stain the central axial filament, which compri-
ses the silicatein, as well as the surface of the spicules. In the presence of
selenium, only the genes for selenoprotein M and spicule-associated protein
are up-regulated, whereas the expression of the silicatein gene remains
unchanged. Finally we show that, in the presence of selenium, larger silica
aggregates are formed. We conclude that selenium has a stimulatory effect
on the formation of siliceous spicules in sponges, and it may be involved in
the enzymatic synthesis of biosilica components.
Abbreviations
DMEM, Dulbecco’s modified Eagle’s medium; PoAb, polyclonal antibody.
3838 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
The growth of spicules probably starts intracellularly
[5]. After reaching a crucial size, e.g. in Suberites
domuncula 1 lm in diameter and 8 lm in length, the
spicules are extruded from the cells and their growth
proceeds in the extracellular space. In S. domuncula
only one type of spicule is formed, megascleres (styles ⁄
oxea), with lengths of up to 450 lm and diameters of
5–7 lm. It is still not known which morphogenetic
processes control the morphology of the spicules. It
has recently been reported that extra-organismic, mor-
phogenetic inorganic elements, e.g. silicate and ferric
iron, and also homeodomain transcription factors, e.g.
Iroquois, are major factors that control the organiza-
tion of the skeletal architecture of spicules [5,6]. The

last step in the biologically controlled mineralization
process, i.e. the final location of the spicules within the
specimen, involves active transport by specialized cells
[7]. The spicules are finally embedded ⁄ cemented into
an organic matrix which contains collagen [8].
The spicules harbor in their center an organic axial fil-
ament in an  1 lm wide canal. In this study we show
that selenium ⁄ selenite induces genes which lead to the
synthesis of proteins associated with silicatein fibers in
the axial filament. These studies with selenium were trig-
gered by observations that this element is required for
the growth of sponge cells [9]. Serine undergoes chemical
conversion into selenocysteine [10]. In eukaryotes as well
as prokaryotes [11], selenium is biochemically incorpor-
ated into proteins through selenocysteine, a process dur-
ing which a tRNA–EF complex is delivered to a codon
that would normally be read as a stop.
We used differential display to further identify
genes ⁄ proteins in the siliceous demosponge S. domun-
cula that may be involved in biosilicification in sponges.
The experiments were performed in cultures of sponge
cells, the primmorph system, which represent 3D cell
aggregates; they contain proliferating and differentiating
cells [12,13]. We found that, in response to selenium,
the expression of two genes, selenoprotein M and the
(sponge-specific) spicule-associated protein, is up-regu-
lated. Cell biological data revealed that the increase in
spicule formation in primmorphs caused by selenium
is paralleled by increased expression of these two
genes ⁄ proteins. Finally, we found that, in the presence

of selenium, larger silica condensation products are
formed in the in vitro assay with recombinant silicatein.
Results
Spicule formation in primmorphs
Light microscopy showed that, after a total incubation
period of 14 days in the presence of 30 lm silicic acid
and 10 lm Fe(III), many spicules were embedded in
the thin rim region that surrounds the body of the
primmorphs (Fig. 1C,D) and inside the 3D cell aggre-
gates. In contrast, almost no spicules were found in
primmorphs that had been cultured without additional
silicic acid and 10 lm Fe(III) (Fig. 1A,B).
The formation of spicules (monactinal tylostyles) in
primmorphs was also demonstrated by transmission
electron microscopy. Cuts through primmorphs that
had been incubated in the absence of silicic acid and
Fe(III) did not show any spicules or cells forming
spicules. This is in contrast with primmorphs that
had been incubated for longer than 7 days in the
presence of 30 lm silicic acid and 10 lm Fe(III), under
conditions described in Experimental procedures.
AB
DC
Fig. 1. Light microscopic images of prim-
morphs, grown for 14 days (7 days in RPMI
medium ⁄ seawater and an additional 7 days
in RPMI ⁄ DMEM ⁄ seawater) in the absence
(A and B) or presence of 30 l
M silicate and
10 l

M ferric citrate (C and D). Spicules can
be seen in the rim surrounding the body of
the primmorphs (>).
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3839
Cross-sections showed either cells (sclerocytes) in the
process of forming the axial filaments of the spicules
or regions with extracellularly located spicules. In the
primordial stage, the axial filaments are synthesized
intracellularly in vesicles (Fig. 2A,B); the filaments
vary in size between 0.3 and 1.5 lm in diameter. It
should be highlighted that during formation of the
axial filaments, characterized by rods filled with
highly electron-dense material, these filaments are clo-
sely associated with  15-nm round fibrils (Fig. 2B).
These fibrils may be involved in the extrusion of
the spicules into the extracellular space. During
maturation, the diameters of the axial filaments
decrease to 0.4 lm (Fig. 2D); in parallel the siliceous
spicules are formed around the filaments (Fig. 2D).
Before the spicules are extruded they grow to lengths
of 8 lm; Fig. 2C shows an intracellular spicule
1.5 lm in length.
Effect of selenium on glutathione peroxidase
activity
First 14-day-old primmorphs were exposed to 0.1 lm,
1 lm and 10 lm sodium selenite for 0 h (control) or
72 h. Then extracts were prepared and glutathione per-

oxidase activity was determined as described in Experi-
mental procedures. In the absence of selenium (time
zero), the enzyme activity was found to be 8.2 ±
1.2 UÆ(mg protein)
)1
. There was no significant change
in activity even when selenium was added to the
primmorphs for 72 h at the indicated concentration
ranges: 7.1 ± 1.1 UÆmg
)1
at 0.1 lm selenium; 7.9 ±
1.0 UÆmg
)1
at 1 lm selenium; 9.1 ± 1.8 UÆ mg
)1
at
10 lm selenium.
Effect of selenium on spicule formation
in primmorphs
A semiquantitative determination revealed that the
amount of polymerized silica in primmorphs cultured
in the absence of additional silicon was low, amount-
ing to <0.4 mgÆ (g wet weight of primmorphs)
)1
. When
30 lm silicic acid [and 10 lm Fe(III)] was added for
7 days to RPMI medium, the concentration increased
to 13 ± 5 mgÆg
)1
. When it was added for 7 days in

RPMI medium and then 7 days in RPMI ⁄ Dulbecco’s
modified Eagle’s medium (DMEM), the concentration
increased to 15 ± 7 mgÆg
)1
. In comparison, the
amount of silica in the tissue of adult sponge speci-
mens is 74 ± 18 mgÆ(g wet tissue)
)1
.
If 10 lm sodium selenite was added together with
silicic acid and Fe(III) to the primmorphs, a doubling
of the silica concentration was seen after 7 days
(32 ± 12 mgÆg
)1
), but no further increase was meas-
ured after 14 days (34 ± 12 mgÆg
)1
). If sodium selenite
was added alone, without additional silicic acid and
Fe(III), no significant change in the amount of silica
was found in the primmorphs.
AB
DC
Fig. 2. Formation of spicules in primmorphs
(transmission electron microscopic images).
Primmorphs cultured for 14 days were ana-
lyzed. (A, B) Section through a primmorph,
showing the formation of an axial filament
(af), a process that proceeds intracellularly.
At higher magnification (B), the 15-nm round

fibrils (fi) adjacent to the axial filament (af)
become visible. (C) A small spicule that is
still intracellularly located is shown in a
sclerocyte. (D) A more mature spicule sp,
now present extracellularly, is shown which
surrounds the axial filament with its
siliceous material.
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3840 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
Selenium in spicules
To clarify whether selenium is incorporated into spi-
cules (axial filaments and the surrounding silica) of
S. domuncula, primmorphs were incubated with
75
Se,
as described in Experimental procedures. After incu-
bation for up to 3 days, spicules were isolated,
mechanically disintegrated, and the radioactivity was
assessed. At time zero, no activity was detected; after
a 1-day incubation, 2500 c.p.m.Æ(5 mg solid mater-
ial)
)1
was measured. The amount increased further
to 5300 c.p.m.Æ5mg
)1
after the extended incubation
period of 3 days. The extract from
75

Se metabolically
labeled spicules was subjected to SDS ⁄ PAGE. As
shown in Fig. 3, the predominant band identified
corresponds to a 14-kDa protein (lanes a and b).
Weaker bands corresponding to sizes of 35 kDa and
23 kDa were also detected if larger amounts of pro-
teins were analyzed (Fig. 3, lane a).
Cloning of the cDNA encoding selenoprotein M
from S. domuncula by differential display
Among the differentially expressed transcripts, one was
found to encode selenoprotein M. The 652-nucleotide
cDNA (accession number AJ875186) contained one
ORF, spanning nucleotides 94–96 to 463–465(stop)
(Fig. 4). One TGA stop codon exists at nucleotides
187–189, which can also function as a codon for seleno-
cysteine [14]. As outlined below, the TGA stop codon
may be suppressed and (very likely) used for the inser-
tion of selenocysteine. The complete protein, with a
calculated molecular mass of 13 918 Da (comprising
the 123-amino-acid ORF) shares the highest sequence
similarity with the 15-kDa selenoprotein M from
humans (accession number NP_536355M) [15]. There-
fore, the sponge molecule was termed selenoprotein M
(SelM_SUBDO) and its cDNA SDSelM. A sequence
comparison revealed that the sponge selenoprotein M
has the highest similarity to human selenoprotein M
(‘expect value’ E of e
)33
); comparatively low is the rela-
tionship to the Drosophila melanogaster putative pro-

tein CG7484-PB (E ¼ 2e
)11
). There are only very
distant – if at all – relationships to the Caenorhabditis
elegans protein with a coiled coil domain (E ¼ 0.13),
the Saccharomyces cerevisiae deduced polypeptide
Yjl049wp (E ¼ 0.56), and the Arabidopsis thaliana
putative protein At1g08340 (E ¼ 0.23).
Cloning of the cDNA for the spicule-associated
protein
A further differentially expressed transcript was char-
acterized, the cDNA encoding a sponge-specific pro-
tein which was termed spicule-associated protein
(SPIaP_SUBDO). The cDNA, SDSPIaP, is 860-bp long
and comprises one ORF at nucleotides 4–6 to nucleo-
tides 757–759 (stop) (accession number AJ872182).
The 251-amino-acid polypeptide (Fig. 5A) has a calcu-
lated molecular mass of 25 602 Da. This protein dis-
played no striking homology to any protein reported
in the database.
One selection criterion used to study this protein
was the existence of repeats; the spicule-associated
protein has high structural regularity (Fig. 5A). Sec-
ondary-structure analysis revealed a-helical regions
at the C-terminus and N-terminus of the protein, and
the central part of the molecule has extended stret-
ches regularly interspersed with predicted turns and
coil conformations. Very interesting with respect to
the amino-acid sequence is the existence of six highly
similar segments of 20 amino acids (60–79; 80–99;

100–119; 120–139; 140–159; 160–179; Fig. 5A). A
closer analysis of the distribution of polar ⁄ nonpolar
amino acids and calculation of the hydropathicity
within the molecule revealed that the borders of the
spicule-associated protein display highly hydrophobic
(potential transmembrane) regions (Fig. 5B). Further-
more, in the central part of the deduced protein, the
hydropathicity plot indicates 10 distinct regular
hydrophobic regions of approximately nine amino
acids. The dominant domain consensus sequence (as
in region number 4) reads (Q)TVNVTATPS, with
Ma b
14 -
20 -
30 -
45 -
62 -
90 -
Fig. 3. SDS ⁄ PAGE analysis of
75
Se-labeled protein(s) in the spi-
cules. Labeled protein(s) were isolated from ground spicules as
described in Experimental procedures. After electrophoresis, the
dried gel was exposed to the film. Positions of the molecular mass
markers are shown on the left; the arrowhead points to the 14-kDa
protein. Extracts from 10 mg (lane a) and 1 mg (lane b) solid mater-
ial were applied to the gel.
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation

FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3841
the amino acids glutamine, lysine and histidine at
the N-terminus and the nonpolar amino acid proline
and the hydroxy amino acids serine and threonine at
the C-terminal border.
Recombinant proteins and antibodies
To identify selenoprotein M and the spicule-associated
protein in sponge extract, the respective recombinant
A
B
13
579
TM TM
1 50 100 200150 250
0
-30
-10
-20
10
20
30
40
50
Fig. 5. Spicule-associated protein SPIaP_
SUBDO. (A) The deduced protein was ana-
lyzed for the predicted secondary structure
as described by Garnier et al. [50]; the
helical conformation (X), the extended
conformation (–), the turn (>) and the coil
conformation (w) are indicated. The six

highly similar segments of 20 amino acids
are marked in white on black or are
underlined. In addition, the 10 hydrophobic
regions, present in the six 20-amino-acid
blocks, are indicated and numbered (#).
(B) Hydropathicity plot of the S. domuncula
SPIaP_SUBDO; the calculation was per-
formed by the method of Kyte and Doolittle
[51]. The horizontal axes show the amino
acid numbers along the protein vs. the cor-
responding hydropathicity. The dotted lines
at the )5 value divide hydrophobic regions
(above) from hydrophilic regions (below).
The 10 distinct hydrophobic regions are
consecutively labeled. The two highly
hydrophobic (potential transmembrane; TM)
regions were identified [52]; they range
from amino acids 32–55 and 204–237.
Fig. 4. S. domuncula selenoprotein M. From the S. domuncula nucleotides sequence (SDSelM), selenoprotein M (SelM_SUBDO) is predic-
ted and aligned with the human selenoprotein M precursor (SelM_HUMAN, NP_536355 [2]); the human sequence was shortened between
amino acids 20–35 and 119–124, indicated by square brackets []). Residues conserved (similar or related with respect to their physicochemi-
cal properties) in the two sequences are shown in white on black. The TGA triplet that encodes selenocysteine (U) is underlined.
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3842 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
proteins were prepared. In parallel, the distribution of
silicatein was analyzed with the tools available [16].
The cDNAs, SDSelM and SDSPIaP, were cloned
into the expression vector pBAD ⁄ gIII as described in

Experimental procedures. After induction with arabi-
nose, the proteins could be identified in the bacterial
lysate (Fig. 6A, lane a vs. b). The recombinant protein
was purified (Fig. 6A, lane c) and used to raise anti-
bodies in rabbits. The polyclonal antibody against
SelM (PoAb-SelM) was found to react with the puri-
fied recombinant fusion selenoprotein M (Fig. 6A, lane
d). The size of this fusion protein was 17 kDa, which
is in accordance with the size of the selenoprotein M
fragment of 84 amino acids expressed (10 kDa)
together with the protein stretch covering the myc
epitope and polyhistidine.
Similarly, the antibodies against the spicule-associ-
ated protein (PoAb-SPaP) were tested. Antibodies were
prepared against the 31-kDa recombinant fusion pro-
tein (Fig. 6B, lane a); they reacted in the western blot
assay with the recombinant protein (Fig. 6B, lane b).
In parallel, the antibodies against the recombinant
silicatein (Fig. 6C, lane a) were subjected to western
blotting and found to react with the 35-kDa protein
(Fig. 6C, lane b). In control assays, it was established
that the adsorbed polyclonal antibodies PoAb-SelM
and PoAb-SPaP did not react with any protein on the
filter (data not shown).
Protein expression of selenoprotein M after
exposure to selenium
In a first series of experiments, 12-day-old primmorphs
(7-day-old primmorphs were incubated for additional
days in RPMI ⁄ DMEM⁄ seawater) were exposed to
10 lm sodium selenite for 24 or 72 h (Fig. 7, lanes a

and b). Primmorphs incubated in seawater ⁄ medium in
the absence of selenium were used as a control.
Extracts were prepared and subjected to western blot
analysis. The blotting results revealed that, in the
absence of selenium, the 14-kDa selenoprotein M is
missing from the extract (not shown). After incubation
for 24 h (lane a) and 72 h (lane b) with sodium selen-
ite, the 14-kDa band, reflecting selenoprotein M, is
clearly present on the immunoblot.
Gene expression studies
The effect of selenium on the expression of the genes
encoding selenoprotein M, spicule-associated protein,
and silicatein was studied (Fig. 8). In the absence of
sodium selenite, almost no transcripts were detected
during the 72-h incubation period for selenoprotein M
(the probe SDSelM was used) and spicule-associated
protein (SDSPIaP), whereas a large number of silicatein
(SDSILIC) transcripts could be detected by northern
blotting. However, during the 72-h incubation with
sodium selenite, within the concentration range 10–
SelM SAP SILIC
Mabc d ab a b
14 -
20 -
30 -
45 -
62 -
90 -
17
31

35
-
++
arab
CBA
Fig. 6. Recombinant selenoprotein M, spicule-associated protein
and silicatein. Sponge SDSelM (selenoprotein M), SDSPIaP (spi-
cule-associated protein) and SDSILIC (silicatein) cDNA was
expressed in E. coli. (A) Expression of selenoprotein M (SelM): lane
a, PAGE analysis (10% gels) of bacterial lysate obtained from
E. coli grown in the absence of arabinose (– arab); lane b, lysate
from bacteria that had been induced with arabinose (+ arab); lane
c, affinity-purified fusion protein; molecular mass 17 kDa; lane d,
western blot analysis of purified fusion protein using PoAb-SelM.
(B) Spicule-associated protein (SAP). The purified fusion protein
(lane a) was used to raise antibodies. The resulting PoAb-SPaP
were found to react with the 31-kDa protein in the western blot
assay (lane b). (C) The recombinant silicatein (SILIC) was used (lane
a) to prepare antibodies (PoAb-SILIC); they recognized the 35-kDa
recombinant protein (lane b). The size markers (M) are given.
a bM
10
15
25
35
24 72 hrs
Fig. 7. Formation of selenoprotein M in primmorphs after exposure
to selenium. Twelve-day-old primmorphs, which had been incuba-
ted for 7 days in RPMI ⁄ seawater and 5 days in RPMI ⁄ DMEM ⁄ sea-
water were exposed to 10 l

M sodium selenite for 24 h (lane a) or
72 h (lane b). Then the 3D cell aggregates were extracted and the
cleared extract (30 lg per lane) was size separated by SDS ⁄ PAGE
(15% gel). Blotting was performed; the blots were incubated with
PoAb-SelM as described in Experimental procedures. In lane M,
size markers were separated.
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3843
1000 lm, considerable up-regulation of the expression
of the genes for selenoprotein M and spicule-associated
protein was detected (Fig. 8). On the basis of the data
obtained, the steady-state expression was almost identi-
cal within the selenium concentration range chosen. The
expression level of silicatein remained almost unchanged
during the exposure to selenium (Fig. 8).
From these data, we conclude that the steady con-
centration of selenoprotein M and spicule-associated
protein in primmorphs is controlled by selenium at the
transcriptional level, whereas the expression of silica-
tein is independent of the presence of sodium selenite.
Identification of selenoprotein M, spicule-
associated protein and silicatein in total extract
or in axial filaments by western blotting
To clarify whether selenoprotein M and spicule-associ-
ated protein are associated with the axial filaments of
the spicules from S. domuncula, western blot ⁄ antibody
studies were performed. Extracts from sodium selenite-
treated primmorphs and axial filaments from spicules

were separated by size and subjected to western blot
experiments (Fig. 9). The blots were incubated with
antibodies against selenoprotein M (PoAb-SelM), spi-
cule-associated protein (PoAb-SPaP), or silicatein
(PoAb-SILIC).
It was shown that selenoprotein M (the 14-kDa
band) exists in large amounts in the soluble extracts
and to a small extent also in the axial filaments. Like-
wise, the spicule-associated protein was identified in
the total extracts and also in lower amounts in the
axial filament (26-kDa band). Silicatein exists predom-
inantly in the axial filaments (24-kDa protein) (Fig. 9).
Localization of selenoprotein M and spicule-
associated protein by immunofluorescence
analysis
The proteins were localized in tissue from S. domuncula
using the antibodies PoAb-SILIC, PoAb-SelM and
PoAb-SPaP. Sections were cut through tissue from
which the spicules had not been removed, and incuba-
ted with the antibodies.
The fluorescence images obtained with antibodies
against silicatein show that primarily ⁄ exclusively the
surfaces of the 5–7-lm thick and up to 450-lm long
monactinal tylostyles and to a smaller extent the diac-
tinal oxeas as well as the axial filaments are recognized
by PoAb-SILIC (Fig. 10A); the corresponding Nomar-
sky interference image is shown in parallel (Fig. 10B).
This finding is interesting, as it indicates that silicatein
is not restricted to the axial filament but also exists
around the spicules.

The images obtained with the antibodies raised
against selenoprotein M (Fig. 10C) and spicule-associ-
ated protein (Fig. 10E) surprisingly revealed strong
immunoreaction not only on the surfaces of the spicules
but also in their canals which harbor the axial filaments.
Parallel Nomarsky interference images (Fig. 10D,F)
show that, in addition to these structures, areas in the
mesohyl of the tissue are decorated by the antibodies.
0 10 100 1000 µM Se
SDSILIC
SDSelM
SDSAP
0.9
0.7
1.4
Fig. 8. Effect of selenium on the expression of selenoprotein M,
spicule-associated protein and silicatein. Fourteen-day-old prim-
morphs were exposed to 0–1000 l
M sodium selenite for 72 h. Sub-
sequently, RNA was isolated and equal amounts (5 lg) were
size-separated, transferred, and probed with labeled selenopro-
tein M (SDSelM), spicule-associated protein (SDSPIaP) or silicatein
(SDSILIC) cDNA.
SILIC
45
14
10
20
30
75

37
25
20
26
150
100
75
50
37
25
20
15
24
Ext AF
SelM SAP
M Ext AF
M Ext AF AF
WB WB WBPAGE
Fig. 9. Identification of selenoprotein M and spicule-associated pro-
tein in the axial filaments by western blotting. Primmorphs (12 days
old) were incubated with 10 l
M sodium selenite for 72 h and then
extracted. The cleared extract was analyzed by SDS ⁄ PAGE (PAGE)
and then by western blotting (WB), using PoAb-SelM, PoAb-SPaP
and PoAb-SILIC. In parallel, axial filaments were prepared and ana-
lyzed in the same way. The size separated protein pattern of the
extract (Ext) and the axial filament (AF) is shown for the analysis of
silicatein (SILIC). Western blots ⁄ SDS ⁄ PAGE studies from left to
right show the results for selenoprotein M (SelM), spicule-associ-
ated protein (SAP) and silicatein (SILIC). Signals for selenopro-

tein M (14-kDa band on the blot) and spicule-associated protein
(26 kDa) are observed in the axial filament and to a large extent in
the extracts, whereas silicatein (24 kDa) is predominantly found in
the axial filament. Markers were separated in parallel (lane M).
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3844 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
Effect of selenium on silica formation in vitro
From earlier studies [3,4] it is known that silicatein,
also in its recombinant form, catalyzes silica formation
in vitro; in these studies the colorimetric molybdate
assay was applied.
In the present study, the effect of selenium on silica
formation was elucidated using the fluorescence dye
Rhodamine 123 as an indicator. As outlined in Experi-
mental procedures, the reaction was performed under
controlled conditions, using recombinant silicatein. In
the absence of the tetraethoxysilane substrate, only
small aggregates, < 3 lm, were observed under fluor-
escence light (Fig. 11A). If tetraethoxysilane was added
to the assays with recombinant silicatein, the size of
the aggregates increased to 10 lm after an incubation
period of 15 min (Fig. 11B). Longer incubation for 3 h
resulted in further growth of the aggregates to 30–
50 lm (Fig. 11C). If during this 3 h incubation period
1 lm sodium selenite was present in the reaction
mixture, the sizes of the aggregates reached values of
50–100 lm (Fig. 11D).
Discussion

Selenium is a trace element which is essential for meta-
zoans from humans [11] to sponges [9]. In the marine
environment, the selenium concentration varies
between 10 and 100 nm [17]. It is well established that
selenium is found in naturally occurring proteins in
two forms, either – rarely – it is inserted post-trans-
lationally as a dissociable cofactor into molybdenum-
containing enzymes [18], or cotranslationally into
proteins as the amino acid selenocysteine [14].
The experiments described here show that exposure
of primmorphs to selenium results in a significant
increase in spicule formation. After a 7-day exposure
to selenium, a substantial increase in biosilica content
of the primmorphs was measured. The formation
of new spicules in primmorphs can be monitored in
AB
DC
EF
Fig. 10. Immunofluorescence analysis of sili-
catein, selenoprotein M and spicule-associ-
ated protein in tissue from S. domuncula.
The slices were stained with PoAb-SILIC
(A), PoAb-SelM (C) and PoAb-SPaP (E). In
parallel, the corresponding Nomarsky
interference images (B, D and F) are given;
some spicules are marked in the paired
images (>). One axial filament (af) in the
center of a spicule is marked in (F).
W. E. G. Mu
¨

ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3845
parallel. It is shown that the axial filament, which con-
sists of silicatein [2–4], is formed intracellularly in spe-
cial cells termed sclerocytes [19]. After the initial
intracellular synthesis of small spicules  8 lmin
length and 1.5 lm in diameter, the spicules are prob-
ably extruded into the bulky extracellular space. Then
synthesis is completed extracellularly, again through
the enzymic activity of silicatein, by appositional
growth, as demonstrated here by immunohistochemical
analysis. The antibodies against silicatein not only
reacted with the silicatein of the axial filaments but also
with proteins present on the surface of the spicules.
Exposure of primmorphs to selenium results in the
expression of two proteins which were shown to be
associated with spicule formation; selenoprotein M
and the spicule-associated protein. Selenoprotein M
has, until now, only been described in metazoa and the
sponge selenoprotein M is the phylogenetically oldest
member. The size of selenoprotein M, deduced from
the cDNA, is 13 918 Da. Western blotting studies per-
formed here show that the native protein has a size of
14 kDa, suggesting that only a small signal peptide
exists at the sponge protein, if at all. A phylogenetic
analysis revealed the closest similarity of the sponge
molecule to the human selenoprotein M, whereas the
proteins of invertebrates (D. melanogaster and C. ele-
gans) are only distantly related.
The finding that the steady-state concentration of

selenoprotein M is regulated at the level of transcrip-
tion was surprising. In vertebrates the expression of all
selenoproteins is regulated at the level of translation
[20]. In S. domuncula, the increase in selenoprotein M
expression after exposure to 10 lm selenium is large.
The immunochemical analysis shows that a seleno-
cysteine tRNA population exists in the sponge,
because translation to the full-size protein occurs.
The biological role of selenoprotein M in higher
metazoan phyla is not known in detail. In the zebrafish,
selenoprotein M is expressed in the notochord, the
somites, the spinal cord and the axial fin fold [21]. These
data interestingly show that the expression pattern of
the different selenoproteins in the fish is region-specific.
To obtain an insight into the potential function of
selenoprotein M in S. domuncula, antibodies were
raised. Surprisingly the subsequent immunohistochemi-
cal analysis revealed that selenoprotein M is localized in
the axial filament and on the surface of the spicules.
Metabolic labeling experiments with
75
Se revealed that
a 14-kDa protein also exists in the spicules.
It had previously been shown that selenocysteine in
selenoproteins participates in redox reactions, especi-
ally if selenocysteine has a close cysteine partner [11].
Exactly this constellation exists in selenoprotein M;
one cysteine residue is present three amino acids along
from selenocysteine towards the N-terminus. This
intriguing finding may suggest that selenoprotein M

functions as an enzyme.
Electron microscopic images document that forma-
tion of siliceous spicules starts with the synthesis of
90–260-nm silica granules [22]. Granular silica of
AB
DC
Fig. 11. Influence of selenium on the size of
silica aggregates, formed in vitro. Recombin-
ant silicatein was incubated in the standard
reaction assay, in the absence of the tetra-
ethoxysilane substrate for 3 h (A). If tetra-
ethoxysilane was added to the reaction for
15 min (B) or 3 h (C), the size of the aggre-
gates increased. (D) Larger aggregates of
silica were formed if, during the 3 h
incubation period, 1 l
M sodium selenite was
present in the reaction mixture. The silica
formed was stained with Rhodamine 123 as
described in Experimental procedures and
inspected by fluorescence microscopy.
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3846 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
exactly this size, 70–300 nm, has recently been synthes-
ized in vitro, using recombinant silicatein from S. do-
muncula [23]. It is still unclear by which process(es)
these silica granules are assembled into silica layers.
Based on experimental data, it has been proposed that

several units of silicatein a form a regular repeating
structure of 5–8 nm [24] or 17 nm [2] periodicity. This
useful information explains the 2D orientation of the
growth of the silica granules ⁄ sheet. For the next phase
in spicule production, forming and shaping has to
occur. Additional factors, e.g. low molecular mass
organic molecules or larger size polypeptides, must be
postulated that guide the packaging of the silica gran-
ules into concentric layers. In elegant studies in diatoms
it was shown that the sophisticated construction of the
diatom shell is the result of high molecular mass pro-
teins, frustulins [25], forming a coat around the dia-
toms, and lower molecular mass proteins present in the
silica shell, the silaffins [26]. The latter 4 to 17-kDa pro-
teins are post-translationally modified at their lysines
and serines. The serine units carry the phosphate
groups, and the lysines are modified at their a-amino
groups by methylpropylamine units [27].
It must be stressed that the silica formation in diatoms
proceeds nonenzymatically, in contrast with sponges
which form the spicules enzymatically using silicatein.
However, up until now, nobody has been able to iden-
tify the proteins within the silica sheets in the spicules.
In fact, published data strongly suggest that the thicken-
ing of the spicules is performed enzymatically by apposi-
tion [16]. The immunohistological data presented here
show that silicatein is also present at the surface of the
spicules, supporting this assumption. As the axial fila-
ments of the spicules are not homogeneous and, especi-
ally intracellularly, are associated with membranes and

fibrils, a high-resolution protein analysis of the axial fila-
ments is the only way to further identify (non)enzymatic
proteins involved in spicule synthesis. Here we used
differential display of transcripts to identify the axial
filament-associated protein. The inorganic element
selenium was chosen as inducer, because of its essential
role in the growth of animals ⁄ sponges, its quasi-enzy-
matic function, and its chemical property of existing in
different valences (II, IV and VI). The transition of the
different co-ordination states allows incorporation into
organic molecules [28].
Using differential display of transcripts, we identified
a second protein that is up-regulated in primmorphs
after incubation with selenium. The deduced protein
shares no significant sequence similarity to any protein
in the database. The characteristic feature of this poly-
peptide is the presence of 10 highly similar repeats,
composed of nine amino acids. In the center are
hydrophobic amino acids surrounded by the polar
amino acids glutamine, lysine, serine and threonine.
Taking into account this polar ⁄ hydrophobic ⁄ polar
composition, this protein can be expected to form a
tight association with membranes. This sponge-specific
protein was termed spicule-associated protein because
it exists in the axial filament and also on the surface of
the spicules. The expression of this protein is not affec-
ted by a change in the concentration of silicic acid in
the surrounding milieu (not shown). In contrast, as
demonstrated here, the expression of the silica-poly-
merizing enzyme silicatein is not regulated by selenium

but by silicic acid [4]. Hence, it becomes evident that
morphogenesis of sponges is to a considerable extent
dependent on outside inorganic factors. Both elements,
silicon and selenium, can be considered morphogenetic
factors which control spicule formation in sponges and
in turn skeleton formation in these animals.
The final interesting finding for the biotechnological
application of silicatein, especially with respect to the
understanding of biosilica formation in sponges, is
that, in the presence of selenium, the size of the poly-
merized silica particles formed from recombinant sili-
catein is substantially larger.
Taken together, the data reported show that selen-
ium has a stimulatory effect on formation of siliceous
spicules in sponges. They may also shed new light on
the factors involved in biosilica formation in metazoa.
Experiments to elucidate potential catalytic effects of
both free and protein-bound selenium in the polymer-
ization process of silica are in progress. Our hypothesis
is that selenium is not only involved in protein ⁄ silica-
tein-controlled silica formation in sponges but func-
tions also as a novel morphogenetic factor during
body plan formation in this oldest metazoan phylum.
Experimental procedures
Chemicals and enzymes
The sources of chemicals and enzymes used were given pre-
viously [4,29]. Natural sterile filtered seawater and sodium
metasilicate were obtained from Sigma-Aldrich (Taufkir-
chen, Germany). Sodium selenite (Na
2

SeO
3
) came from
Sigma (Taufkirchen, Germany). Na
2
[
75
Se]O
3
was from
Amersham Corp. [Little Chalford, Buckinghamshire, UK;
370 MBqÆ(mg Se)
)1
] or Polatom (Otwock Swierk, Poland;
1500 MBqÆmg
)1
).
Sponges
Live specimens of S. domuncula (Porifera, Demospongiae,
Hadromerida) were collected near Rovinj (Croatia) and
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3847
then kept in aquaria in Mainz (Germany) for more than
2 years before their use.
Preparation of spicules and axial filaments
Spicules and their axial filaments were prepared as des-
cribed [30]. Briefly, tissue samples were treated first with
H

2
SO
4
⁄ HNO
3
and then with butan-1-ol ⁄ water ⁄ SDS. The
spicules were collected by sedimentation. The axial fila-
ments were isolated from the spicules by dissolving the sil-
ica with hydrofluoric acid (2 m HF ⁄ 8 m NH
4
F; pH 5) at
room temperature overnight. The axial filaments were
collected, dialyzed against distilled water, and collected by
centrifugation [2].
Formation of primmorphs, incubation conditions
and extracts
The procedure for the formation of primmorphs (3D cell
cultures) from single cells was as described previously
[12,13]. Starting from single cells, primmorphs of 3–7 mm
are formed after 5 days; they were cultivated in natural sea-
water, supplemented with 1% RPMI 1640 medium and in
30 lm silicate (as sodium metasilicate) and 10 lm ferric cit-
rate [31]. Seven-day-old, nonattached, 3D cell cultures were
transferred into natural seawater, supplemented with 0.25%
RPMI 1640 ⁄ 0.75% DMEM and 30 lm silicate and 10 lm
Fe(III); incubation proceeded for the indicated periods of
time. In this medium the primmorphs attached to the glass
slide after 1 day. Sodium selenite was added as indicated;
in one series of experiments, silicate and iron were omitted
from the medium. An inverted-stage Olympus IX70 micro-

scope was used for images.
Extracts from primmorphs were prepared by homogeni-
zation in lysis buffer [1· Tris-buffered saline, pH 7.5, 1 mm
EDTA, 1% Nonidet P40, protease inhibitor cocktail
(Roche, Nutley, NJ, USA; one tablet per 10 mL)], centri-
fuged, and the supernatants subjected to western blot ana-
lysis.
Transmission electron microscopy
For transmission electron microscopic analyses, the sponge
samples were cut into pieces (2 mm
3
), incubated in 0.1 m
phosphate buffer (supplemented with 2.5% glutaraldehyde,
0.82% NaCl, pH 7.4) and washed in 0.1 m phosphate buf-
fer (1.75% NaCl) at room temperature. After the samples
had been treated with 1.25% NaHCO
3
, 2% OsO
4
and 1%
NaCl, they were dehydrated with ethanol. The dried sam-
ples were incubated with propylene oxide, fixed in propyl-
ene oxide ⁄ araldite (2 : 1, w ⁄ w), covered with pure araldite,
and hardened at 60 °C for 2 days before being cut
into 60-nm ultrathin slices (Ultracut S; Leica, Wetzlar,
Germany). The samples were transferred to coated copper
grids and analyzed with a Tecnai 12 microscope (FEI Elec-
tron Optics, Eindhoven, the Netherlands).
Silica content of primmorphs
To determine semiquantitatively the silica content in prim-

morphs, these 3D cell aggregates were disintegrated with
sodium hypochlorite (10% for 24 h at room temperature).
The particles, mainly spicules and their precursors, were
obtained by centrifugation, treated with NaOH, and the
amount of released silicic acid determined colorimetrically
by the molybdate assay [3], using the ‘Silicon Test’ as des-
cribed by the manufacturer (Merck, Darmstadt, Germany).
Differential display technique
These studies were performed with either primmorphs kept
for 7 days in the absence of additional selenium in the
RPMI 1640 ⁄ seawater medium or primmorphs that had
been cultivated for 72 h in RPMI 1640 ⁄ seawater, supple-
mented with 10–1000 lm sodium selenite.
The technique of differential display was applied to iden-
tify the changes in the expression levels of transcripts as
described [32,33]. Total RNA (1 lg) from controls and
selenium-treated samples was reverse-transcribed using an
equimolar oligonucleotide mixture of dT
23
N (Sigma; N rep-
resents G, C or A) as 3¢ primer in a reaction with the Mol-
oney murine leukemia virus reverse transcriptase (Promega,
Madison, WI, USA) at a reaction temperature of 37 °C.
The resulting cDNA was used as template for the subse-
quent PCR with the 5¢-IRD 800 infrared labeled arbitrary
primer 5¢-AGTGAATGCG-3¢ and one of the dT
23
N prim-
ers in the assay. To avoid DNA-based contamination from
the reagents, the assay was incubated (before the addition

of the template) with the restriction enzyme Sau3AI for
20 min at 37 °C followed by a final inactivation step at
72 °C for 10 min. The PCR parameters used were: 94 °C
for 3 min, 45 cycles of 94 °C for 30 s, 40 °C for 2 min, and
72 °C for 30 s, with an additional extension step at 72 °C
for 10 min. After amplification, the labeled fragments were
separated on a 6% polyacrylamide sequencing gel using a
DNA sequenator (Li-Cor 4000S). The sequencing run was
stopped after 3 h, and the gel was transferred to millimeter
paper and vacuum dried. The differences in the banding
pattern on the gel were detected by an infrared scanning
device (Odyssey; LiCor, Lincoln, NE, USA). The major
DNA bands differentially seen in the assays with selenite
were excised and diluted in distilled water. The resulting
fragments were subsequently re-amplified, subcloned in
pCRII-Topo vector (Invitrogen, Carsbad, CA, USA), and
sequenced using the ‘Thermo-Sequenase Fluorescent Labe-
led Primer Cycle Sequencing Kit’ (Amersham ⁄ Pharmacia,
Little Chalfont, Buckinghamshire, UK) and the sequencing
device (LongReadIR 4200; Li-Cor).
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3848 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
More than 30 sequences were obtained; among them
three partial cDNAs were isolated, the deduced polypep-
tides of which showed sequence similarity to selenopro-
tein M and two partial cDNAs that showed very low
homology to human mucin. The complete cDNA was
obtained by primer walking [34,35]. The selenoprotein M

cDNA was 652 nucleotides long and was termed SDSelM.
The second cDNA, the deduced polypeptide of which dis-
played no striking homology to proteins reported in the
databases, was termed spicule-associated protein, SDSPIaP,
and had a size of 860 bp.
Sequence analysis
The sequences were analyzed using computer programs
blast and fasta. Multiple alignments were performed with
clustal w version 1.6 [36]. Phylogenetic trees were construc-
ted on the basis of amino-acid sequence alignments by neigh-
bor-joining, as implemented in the ‘Neighbor’ program from
the phylip package [37]. The distance matrices were calcula-
ted using the Dayhoff PAM matrix model as described [38].
The degree of support for internal branches was further
assessed by bootstrapping [37]. The graphic presentations of
the alignments were prepared with genedoc [39].
Preparation of recombinant proteins
selenoprotein M and spicule-associated protein
Expression of selenoprotein M
A fragment of the sponge SDSelM sequence was expressed in
Escherichia coli (TOP10). The ORF (nucleotides 210–462;
corresponding to amino acids 40–123 of the mature protein)
was isolated by PCR using one forward primer
(5¢-
CCATGGCAACCAAACCCTCTGGTCCTA-3¢ (the
NcoI restriction site is underlined) and one reverse primer
(5¢-
AAGCTTAGATACTTCTTGACTTTCACC-3¢)(Hin-
dIII). Hence, the TGA stop codon which is located at nucleo-
tides 187–189 (amino acid 32 in the deduced protein) was not

included. The 252-bp part was cloned into the expression vec-
tor pBAD ⁄ gIIIA (Invitrogen), which contained at the 3¢-ter-
minus the myc epitope and the polyhistidine region. The
insert was expressed overnight at 30 °C in the presence of
0.0002% l-arabinose. The fusion protein was extracted and
purified with the Histag purification kit (Novagen, Madison,
WI, USA). The purity of the material was checked by electro-
phoresis on 10% polyacrylamide gels containing 0.1% SDS
as described by Laemmli [40]. The recombinant selenopro-
tein M, r-SelM, was dialyzed against 25 mm Tris ⁄ HCl buffer
(pH 7.2), supplemented with 10 mmdl-dithiothreitol.
Spicule-associated protein
For the preparation of the recombinant protein of spicule-
associated protein, r-SPaP, the same vector with the
complete ORF of the SDSPIaP cDNA was used; again the
cDNA was inserted into the vector at the same restriction
sites, NcoI and HindIII. Expression and purification proce-
dures were the same as for selenoprotein M.
Raising of antibodies
Polyclonal antibodies (PoAbs) were raised against recom-
binant r-SelM and r-SPaP in female rabbits (New Zealand
White) as described [41]; the PoAbs were termed PoAb-
SelM and PoAb-SPaP, respectively. In control experiments,
100 lL PoAb-SelM or PoAb-SPaP was adsorbed to 20 lg
r-SelM or r-SPaP (30 min; 4 °C) before use. The prepar-
ation of antibodies against recombinant silicatein was as
described previously [16]; they were termed PoAb-SILIC
and raised against the long form of the r-silicatein
(35 kDa).
Western blotting

Protein samples, extracts from primmorphs, and axial fila-
ments were analyzed. Samples of 5 lg protein (axial fil-
ament) or 30 lg extract were dissolved in loading buffer
(Roti-Load; Roth, Karlsruhe, Germany), boiled for 5 min
and subjected to SDS ⁄ PAGE (10% polyacrylamide and
0.1% SDS) [40]. The gels were stained with Coomassie Bril-
liant Blue. In parallel, the proteins were transferred and the
poly(vinylidene difluoride) membranes (Millipore-Roth)
were incubated with PoAb-SILIC, PoAb-SelM or PoAb-
SPaP (1 : 500 dilution each); the immune complexes were
visualized by incubation with anti-rabbit IgG (alkaline
phosphatase conjugated), followed by staining with
4-chloro-1-naphthol. Multi-Tag-Markers (Roche, Mann-
heim, Germany) were used as size markers.
Selenium in spicules
The incorporation of
75
Se, as sodium selenite [370
MBqÆmg
)1
(10 mCiÆmg
)1
)], into axial filaments of spicules
was determined in primmorphs. Samples of 2 g each were
cultured for 0, 1–3 days in seawater [42] supplemented with
200 nCiÆmL
)175
Se. Then spicules were isolated with
H
2

SO
4
⁄ HNO
3
, as described above. They were washed thor-
oughly with distilled water by centrifugation and mechanic-
ally pulverized to separate the silica ‘shell’ and the axial
filaments. The amount of
75
Se-labeled material was deter-
mined, and the c.p.m. obtained was correlated with 5 mg
spicules, before the disintegration. In a second series of
experiments, the primmorphs (2 g) were incubated for
3 days with 200 nCiÆmL
)175
Se. Then spicules were isolated,
pulverized and extracted with lysis buffer. After centrifuga-
tion, the 2000 g supernatant was collected, concentrated
with the Micron-Amicon Centrifugal Filter Device System
(Millipore, Bedford, MA, USA), and subjected to
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3849
SDS ⁄ PAGE (10% gel ⁄ 0.1% SDS). The dried gel was
exposed to Kodak X-Omat XAR X-ray film for 2 days at
)80 °C.
RNA preparation and northern blot analysis
RNA was extracted from liquid-nitrogen pulverized tissue
with Trizol reagent (Gibco-BRL, Grand Island, NY, USA)

as described [43]. Then 5 lg total RNA was electrophoresed
and blotted on to Hybond-N
+
nylon membrane (Amer-
sham). Hybridization was performed with a 250-nucleotide
segment of the SDSelM cDNA, a 300-nucleotide segment
of SDSPIaP and a 400-nucleotide fragment from the silica-
tein-A gene SDSILIC (accession number AJ272013 [4]).
The probes were labeled with the PCR-DIG probe synthesis
kit (Roche). After being washed DIG-labeled nucleic acid
was detected with anti-DIG FAB fragments and visualized
by chemiluminescence technique using CDP (Roche).
Immunohistochemistry
Fresh sponge tissue from the Adriatic Sea was fixed in 2%
paraformaldehyde [44]; after dehydration, the samples were
embedded in Technovit 8100 [45], according to the instruc-
tions of the manufacturer as outlined in Pancer et al. [46].
Sections of 8 lm thickness were prepared. The sponge tis-
sue was deliberately not treated with HF ⁄ NaHF solution
so as not to remove the siliceous spicules. The sections were
blocked with 1% BSA in NaCl ⁄ P
i
and incubated with
PoAb-SelM, PoAb-SPaP or PoAb-SILIC overnight at 4 °C.
The antibodies were used at a 1 : 200 dilution in 0.5% BSA
in NaCl ⁄ P
i
buffer. Rhodamine-conjugated goat anti-rabbit
immunoglobulin (Dako, Carpinteria, CA, USA) was used
as a secondary antibody. The preimmune rabbit serum was

used as a control; it did not react with any structures in the
sections. For light microscopic studies (immunofluorescence
analysis), an Olympus AHBT3 microscope was used. The
fluorescence studies were performed with an AH3-RFC
reflected light fluorescence attachment using the excitation
light wave-length 334 ⁄ 365 nm. In parallel, the nonstained
sections were inspected directly using Nomarsky interfer-
ence contrast optics.
Staining of polymerized silica with Rhodamine
123
Enzymatic silica formation was studied using r-silicatein
[31]. The reactions were performed on a glass slide in a vol-
ume of 100 lL. The reaction mixture contained 10 lg r-sili-
catein in NaCl ⁄ P
i
(pH 7.2) containing 5 mm Fe(III) and
1 lm ZnCl
2
. As substrate, 4.5 mm tetraethoxysilane (Sigma)
was used. The reaction was performed at 22 °C for up to
3 h; the mixture was then washed three times with NaCl ⁄ P
i
.
Where mentioned, 1 lm sodium selenite was added to the
reaction. Rhodamine 123 (Sigma) was used to stain silica
aggregates formed, as described [47]. Finally, the specimens
were analyzed with an Olympus AHBT3 microscope ⁄ AH3-
RFC light fluorescence using the excitation light wavelength
490 nm.
Analytical ⁄ biochemical techniques

Protein concentrations were determined as described by
Lowry et al. [48] using BSA as standard. Glutathione per-
oxidase activity was determined as described [49]. In brief,
primmorphs were extracted with 10 mm Tris ⁄ HCl buffer
(containing 2 mm EDTA, 100 mm NaCl, pH 7.2). After
centrifugation (10 000 g; 10 min; 4 °C), the supernatant
was collected and subjected to the coupled NADPH oxida-
tion enzyme reaction. Hydrogen peroxide was used as a
substrate. The enzyme unit was defined as 1 nmol GSH
oxidized per min. Five parallel experiments were performed,
and mean ± SD values are given.
Acknowledgements
We thank Ms. E. Sehn (Zoological Institute, Univer-
sity of Mainz) for valuable technical assistance. This
work was supported by grants from the Deutsche Fors-
chungsgemeinschaft, the Bundesministerium fu
¨
r Bildung
und Forschung Germany (project: Center of Excellence
BIOTECmarin) and the International Human Frontier
Science Program.
References
1 Lowenstam HA (1981) Minerals formed by organisms.
Science 211, 1126–1131.
2 Shimizu K, Cha J, Stucky GD & Morse DE (1998) Sili-
catein alpha: cathepsin 1-like protein in sponge biosilica.
Proc Natl Acad Sci USA 95, 6234–6238.
3 Cha JN, Shimizu K, Zhou Y, Christianssen SC, Chm-
elka BF, Stucky GD & Morse DE (1999) Silicatein fila-
ments and subunits from a marine sponge direct the

polymerization of silica and silicones in vitro. Proc Natl
Acad Sci USA 96, 361–365.
4 Krasko A, Batel R, Schro
¨
der HC, Mu
¨
ller IM & Mu
¨
ller
WEG (2000) Expression of silicatein and collagen genes
in the marine sponge Suberites domuncula is controlled by
silicate and myotrophin. Eur J Biochem 267, 4878–4887.
5Mu
¨
ller WEG (2005) Spatial and temporal expression
patterns in animals. In Encyclopedia of Molecular Cell
Biology and Molecular Medicine (Meyers, RA, ed.).
Wiley-VCH GmbH, Weinheim. (in press)
6Mu
¨
ller WEG, Wiens M, Adell T, Gamulin V, Schro
¨
der
HC & Mu
¨
ller IM (2004) The bauplan of the Urmeta-
zoa: the basis of the genetic complexity of Metazoa
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.

3850 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS
using the siliceous sponges [Porifera] as living fossils. Int
Rev Cytol 235, 53–92.
7 Weissenfels N (1989) Biologie und Mikroskopische Anat-
omie der Su
¨
ßwasserschwa
¨
mme (Spongillidae). Gustav
Fischer Verlag, Stuttgart.
8 Simpson TL, Gil M, Connes R, Diaz JP & Paris J
(1985) Effects of germanium (Ge) on the silica spicules
of the marine sponge Suberites domuncula: transforma-
tion of the spicule type. J Morphol 183, 117–128.
9 Willoughby R & Pomponi SA (2000) Quantitative
assessment of marine sponge cells in vitro: development
of improved growth medium. In Vitro Cell Dev Biol 36,
194–200.
10 Wu ZP & Hilvert D (1989) Selenosubtilisin as a glu-
tathione peroxidase mimic. J Am Chem Soc 112, 5647–
5648.
11 Hatfield DL & Gladyshev VN (2002) How selenium has
altered our understanding of the genetic code. Mol Cell
Biol 22, 3565–3576.
12 Custodio MR, Prokic I, Steffen R, Koziol C, Borojevic
R, Bru
¨
mmer F, Nickel M & Mu
¨
ller WEG (1998) Prim-

morphs generated from dissociated cells of the sponge
Suberites domuncula: a model system for studies of cell
proliferation and cell death. Mech Ageing Dev 105,
45–59.
13 Mu
¨
ller WEG, Wiens M, Batel R, Steffen R, Borojevic
R & Custodio MR (1999) Establishment of a primary
cell culture from a sponge: primmorphs from Suberites
domuncula. Mar Ecol Progr Series 178, 205–219.
14 Low SC & Berry MJ (1996) Knowing when not to stop:
selenocysteine incorporation in eukaryotes. Trends Bio-
chem Sci 21, 203–208.
15 Korotkov KV, Novoselov SV, Hatfield DL & Glady-
shev VN (2002) Mammalian selenoprotein in which sele-
nocysteine (Sec) incorporation is supported by a new
form of Sec insertion sequence element. Mol Cell Biol
22, 1402–1411.
16 Mu
¨
ller WEG, Krasko A, Le Pennec G, Steffen R,
Ammar MSA, Wiens M, Mu
¨
ller IM & Schro
¨
der HC
(2003) Molecular mechanism of spicule formation in the
demosponge Suberites domuncula: silicatein – collagen –
myotrophin. Prog Mol Subcell Biol 33, 195–122.
17 Schroeder RA, Orem WH & Kharaka YK (2002)

Chemical evolution of the Salton Sea, California:
nutrient and selenium dynamics. Hydrobiologia 473,
23–45.
18 Gladyshev VN, Khangulov SV & Stadtman TC (1994)
Nicotinic acid hydroxylase from Clostridium barkeri:
electron paramagnetic resonance studies show that sele-
nium is coordinated with molybdenum in the catalyti-
cally active selenium-dependent enzyme. Proc Natl Acad
Sci USA 91, 232–236.
19 Uriz MJ, Turon X & Becerro MA (2000) Silica deposi-
tion in Demospongiae: spiculogenesis in Crambe crambe.
Cell Tissue Res 301, 299–309.
20 Sunde RA (2001) Regulation of selenoprotein expres-
sion. In Selenium: its Molecular Biology and Role in
Human Health (Hatfield, DL, ed.), pp. 81–96. Kluwer
Academic Publishers, Boston.
21 Thisse C, Degrave A, Kryukov GV, Gladyshev VN,
Obrecht-Pfumio S, Krol A, Thisse B & Lescure A
(2003) Spatial and temporal expression patterns of sele-
noprotein genes during embryogenesis in zebrafish. Gene
Expr Patterns 3, 525–532.
22 Pisera A (2003) Some aspects of silica deposition in
Lithistid demosponge desmas. Microsc Res Tech 62,
312–326.
23 Tahir MN, The
´
ato P, Mu
¨
ller WEG, Schro
¨

der HC,
Janshoff A, Zhang J, Huth J & Tremel W (2004) For-
mation of biosilica by immobilization of silicatein on
novel functionalized surfaces. Chem Commun (Camb)
24, 2848–2849.
24 Croce G, Frache A, Milanesio M, Marchese L, Causa
`
M, Viterbo D, Barbaglia A, Bolis V, Bavestrello G,
Cerrano C, et al. (2004) Structural characterization of
siliceous spicules from marine sponges. Biophys J 86,
526–534.
25 Kro
¨
ger N & Sumper M (2000) The biochemistry of
silica formation. In Biomineralization from Biology to
Biotechnology and Medical Application (Ba
¨
uerlein E,
ed.), pp. 151–170. Wiley-VCH, Weinheim.
26 Kro
¨
ger N, Lorenz S, Brunner E & Sumper M (2002)
Self-assembly of highly phosphorylated silaffins and
their function in biosilica morphogenesis. Science 298,
584–586.
27 Kro
¨
ger N, Deutzmann R & Sumper M (2001) Silica-
precipitating peptides from diatoms: the chemical struc-
ture of silaffin-1a from Cylindotheca fusiformis. J Biol

Chem 276, 26066–26070.
28 Bo
¨
ck A (2001) Selenium metabolism in bacteria. In
Selenium: its Molecular Biology and Role in Human
Health (Hatfield, DL, ed.), pp. 8–32 Kluwer Academic
Publishers, Boston.
29 Kruse M, Mu
¨
ller IM & Mu
¨
ller WEG (1997) Early evo-
lution of metazoan serine ⁄ threonine- and tyrosine
kinases: identification of selected kinases in marine
sponges. Mol Biol Evol 14, 1326–1334.
30 Kaluzhnaya OV, Belikov SI, Schro
¨
der HC, Rothenber-
ger M, Zapf S, Kaandorp JA, Borejko A, Mu
¨
ller IM &
Mu
¨
ller WEG (2005) Dynamics of skeletal formation in
the Lake Baikal sponge Lubomirskia baicalensis. Part I.
Biological and biochemical studies. Naturwissenschaften
92, 128–133.
31 Krasko A, Schro
¨
der HC, Batel R, Grebenjuk VA, Stef-

fen R, Mu
¨
ller IM & Mu
¨
ller WEG (2002) Iron induces
proliferation and morphogenesis in primmorphs from
the marine sponge Suberites domuncula. DNA Cell Biol
21, 67–80.
32 Mu
¨
ller WEG, Krasko A, Skorokhod A, Bu
¨
nz C, Grebe-
njuk VA, Steffen R, Batel R, Mu
¨
ller IM & Schro
¨
der
W. E. G. Mu
¨
ller et al. Effect of selenium on spicule formation
FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS 3851
HC (2002) Histocompatibility reaction in the sponge
Suberites domuncula on tissue and cellular level: central
role of the allograft inflammatory factor 1. Immuno-
genetics 54, 48–58.
33 Mu
¨
ller WEG, Klemt M, Thakur NL, Schro
¨

der HC,
Aiello A, D’Esposito M, Menna M & Fattorusso E
(2003) Molecular ⁄ chemical ecology in sponges: evidence
for an adaptive antibacterial response in Suberites
domuncula. Mar Biol 144, 19–29.
34 Ausubel FM, Brent R, Kingston RE, Moore DD, Smith
JA, Seidmann JG & Struhl K (1995) Current Protocols
in Molecular Biology. John Wiley and Sons, New York.
35 Wiens M, Koziol C, Hassanein HMA, Batel R, Schro
¨
-
der HC & Mu
¨
ller WEG (1998) Induction of gene
expression of the chaperones 14–3)3 and HSP70
induced by PCB 118 (2,3¢,4,4¢,5-pentachlorobiphenyl) in
the marine sponge Geodia cydonium: novel biomarkers
for polychlorinated biphenyls. Mar Ecol Prog Series
165, 247–257.
36 Thompson JD, Higgins DG & Gibson TJ (1994) CLUS-
TAL W: improving the sensitivity of progressive multi-
ple sequence alignment through sequence weighting,
positions-specific gap penalties and weight matrix
choice. Nucleic Acids Res 22, 4673–4680.
37 Felsenstein J (1993) Phylip, version 3.5. University of
Washington, Seattle.
38 Dayhoff MO, Schwartz RM & Orcutt BC (1978) A
model of evolutionary change in protein. In Atlas of
Protein Sequence and Structure (Dayhoff, MO, ed.), pp.
345–352. National Biomedical Research Foundation,

Washington, DC.
39 Nicholas KB & Nicholas HB Jr (1997) Genedoc: a Tool
for Editing and Annotating Multiple Sequence Align-
ments, Version 1.1.004. Distributed by the author;
/>40 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
41 Schu
¨
tze J, Krasko A, Diehl-Seifert B & Mu
¨
ller WEG
(2001) Cloning and expression of the putative aggrega-
tion factor from the marine sponge Geodia cydonium.
J Cell Sci 114, 3189–3198.
42 Bo
¨
hm M, Schro
¨
der HC, Mu
¨
ller IM, Mu
¨
ller WEG &
Gamulin V (2000) The mitogen-activated protein
kinase p38 pathway is conserved in metazoans: cloning
and activation of p38 of the SAPK2 subfamily from
the sponge Suberites domuncula. Biol Cell 92,
95–104.
43 Grebenjuk VA, Kuusksalu A, Kelve M, Schu

¨
tze J, Schro
¨
-
der HC & Mu
¨
ller WEG (2002) Induction of (2¢-5¢) oligo-
adenylate synthetase in the marine sponges Suberites
domuncula and Geodia cydonium by the bacterial endo-
toxin lipopolysaccharide. Eur J Biochem 269, 1382–1392.
44 Romeis B (1989) Mikroskopische Technik. Urban ⁄
Schwarzenberg, Mu
¨
nchen.
45 Beckstead JH (1985) Optimal antigen localization in
human tissues using aldehydefixed plastic-embedded sec-
tions. J Histochem Cytochem 9, 954–958.
46 Pancer Z, Kruse M, Scha
¨
cke H, Scheffer U, Steffen R,
Kova
´
cs P & Mu
¨
ller WEG (1996) Polymorphism in the
immunoglobulin-like domains of the receptor tyrosine
kinase from the sponge Geodia cydonium. Cell Adhes
Commun 4, 327–339.
47 Li CW, Chu S & Lee M (1989) Characterizing the sili-
con deposition vesicle of diatoms. Protoplasma 151,

158–163.
48 Lowry OH, Rosebrough NJ, Farr AL & Randall RJ
(1951) Protein measurement with the folin phenol
reagent. J Biol Chem 193, 265–275.
49 Cheng WH, Ho YS, Ross DA, Han Y, Combs GF &
Lei XG (1997) Overexpression of cellular glutathione
peroxide does not affect expression of plasma glu-
tathione peroxide or phospholipids peroxide glutathione
peroxide in mice offered diets adequate or deficient in
selenium. J Nutr 127, 405–417.
50 Garnier J, Osguthorpe DJ & Robson B (1978) Predict-
ing the secondary structure of globular proteins. J Mol
Biol 120, 97–120.
51 Kyte J & Doolittle RF (1982) A simple method for dis-
playing the hydrophobic character of a protein. J Mol
Biol 157, 105–132.
52 Rao J & Argos P (1986) A conformational preference
parameter to predict helices in integral membrane pro-
teins. Biochim Biophys Acta 869, 197–214.
Effect of selenium on spicule formation W. E. G. Mu
¨
ller et al.
3852 FEBS Journal 272 (2005) 3838–3852 ª 2005 FEBS