Identification of a novel matrix protein contained in a
protein aggregate associated with collagen in fish otoliths
Hidekazu Tohse
1,2
, Yasuaki Takagi
2
and Hiromichi Nagasawa
1
1 Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences, University of Tokyo, Japan
2 Division of Marine Biosciences, Graduate School of Fisheries Science, Hokkaido University, Japan
Organisms can design and shape minerals to the desired
conformation and orientation. Such mineral structures
are called biominerals and cannot be formed by any
non-biological environments. Calcium carbonate is one
of the most common biominerals, formed mainly by
invertebrates, and has three crystal phases: calcite, ara-
gonite and vaterite. Although calcite is the most stable
crystal thermodynamically, many organisms can form
metastable aragonite crystals with desired morphologies
under normal environments of pressure and tempera-
ture. It is thought that the morphology and polymor-
phism of biominerals can be controlled by the proteins,
polysaccharides and complexes (organic matrices)
within the biominerals themselves [1,2].
In the past decade, many proteins have been isolated
from various calcium carbonate biominerals, and their
roles in the formation of crystal morphologies have
been discussed. These isolated single proteins have
some activity in changing crystal morphologies; how-
ever, analyses of the single proteins has not led
to insights into how these morphologies and

¨
llner C, Burgham-
mer M, Busch-Nentwich E, Berger J, Schwarz H, Riekel C & Nicolson T
(2003) Science 302, 282–286].
45
Ca overlay analysis revealed that the
Glu-rich region has calcium-binding activity. Combined analysis by western
blotting and deglycosylation suggested that OMM-64 is present in an
HMW aggregate with heparan sulfate chains. Histological observations
revealed that OMM-64 is expressed specifically in otolith matrix-producing
cells and deposited onto the otolith. Moreover, the HMW aggregate binds
to the inner ear-specific short-chain collagen otolin-1, and the resulting
complex forms ring-like structures in the otolith matrix. Overall, OMM-64,
by forming a calcium-binding aggregate that binds to otolin-1 and forming
matrix protein architectures, may be involved in the control of crystal
morphology during otolith biomineralization.
Abbreviations
GST, glutathione S-transferase; HMW, high molecular weight; IPTG, isopropyl-b-
D-thiogalactopyranoside; OMM-64, otolith matrix
macromolecule-64; PVDF, polyvinylidene difluoride; TFMS, trifluoromethanesulfonic acid.
2512 FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS
organic matrices are thought to be formed from a
complex of individual matrix proteins. For example, in
biomineralization of mollusk shells, which have chitin
as the structural molecule in their EDTA-insoluble
fraction [3], isolated proteins from the EDTA-insoluble
fraction exhibit different actions on the crystal forma-
tion when they are applied to crystal induction systems
with framework organic substrates [4–7], indicating
that such proteins may interact with the framework

and may construct water-insoluble, gel-like structures in
the biomineral matrices. Identifying the proteins that
construct the aggregates is extremely difficult, however,
because these proteins are not separable by gel electro-
phoresis or liquid chromatography. In the present study,
we have examined and characterized the proteins that
form these aggregates in fish otoliths. We had previously
raised an antiserum against whole otolith matrix con-
taining mainly HMW (> 100 kDa) proteins [13], and
here we used this antiserum to screen an inner ear
cDNA library and thereby clone a cDNA encoding a
protein, named otolith matrix macromolecule-64
(OMM-64), that is contained in a HMW aggregate
in the otolith matrix. During characterization of this
protein, we revealed that the aggregate also contains the
inner ear-specific collagen otolin-1 [9].
Results
Cloning of cDNA and DNA encoding OMM-64
To obtain cDNA clones encoding proteins contained in
the HMW aggregate, immunoscreening was performed
using an antiserum that reacts mainly with the aggregate
in the otolith matrix [13]. After screening, clones conta-
ining omm-64 cDNA were obtained, but the sequence of
the 5¢ end could not be determined. Therefore, 5¢ RACE
was performed. In addition, genomic DNA encoding
OMM-64 was also obtained by genome walking.
Structures of OMM-64 protein and DNA
The cDNA cloned had a length of 2776 bp and
encodes a protein of 628 amino acids (Fig. 1 and sup-
plementary Fig. S1). The open reading frame is

nines and one tyrosine).
A blastp search using the amino acid sequence of
OMM-64 identified starmaker, a zebrafish otolith
matrix protein that contributes to the regulation of oto-
lith crystal polymorphism [14]. Although the identity
between these proteins was only 25%, some distinctive
domains of starmaker are conserved in OMM-64 (sup-
plementary Fig. S2): an N-terminal sequence containing
signal peptides (Met1–Ala36) is highly conserved, and
two (V ⁄ G)TTD sequences found in the tandem repeats
of starmaker are also found in OMM-64. By contrast, a
distinctive sequence that is rich in serine and aspartic
acid in starmaker is not conserved in OMM-64, which
has a glutamic acid-rich sequence instead.
A partial sequence of omm-64 mRNA was found in
the GenBank EST database (accession number
CX067293). This rainbow trout mRNA had been iden-
tified by random sequencing analysis of a cDNA
library constructed by suppressive subtraction of
whole-embryo mRNA at late neurogenesis stages
(hindbrain swelling + heart tube with peristalsis) from
that at early neurogenesis stages (neural groove +
50% epiboly), suggesting that omm-64 is expressed in
the early neurogenesis stage of the embryo and is
involved in inner ear development.
In omm-64 gene, the sequence encoding OMM-64 is
divided into 23 exons, including two large exons in the
middle region of the ORF and the 3¢ UTR (Fig. 1).
This exon ⁄ intron structure is highly similar to that of
the starmaker gene (supplementary Fig. S2): many

stain positive with eosin. Sense-strand probes did not hybridize
to any regions of the sacculus (data not shown). CT, connective
tissue; EL, endolymph region; SqE, squamous epithelial cells.
A novel protein in the otolith matrix framework H. Tohse et al.
2514 FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS
other types of cells owing to their large size and shape
and positive eosin staining (Fig. 2E), and which, like
chloride cells, have Na
+
⁄ K
+
-ATPase activity [15].
Weak expression of omm-64 mRNA was detected in the
sensory epithelium (macula). In the ventral, dorsal and
distal areas of the sacculus, by contrast, mRNA hybrid-
ization signals were barely detectable (Fig. 2C,D).
Identification of the calcium-binding domain in
OMM-64
To determine the regions that have calcium-binding
activity, six fusions of GST with recombinant proteins
of OMM-64 (rOMM-64) were produced and applied
to a
45
Ca overlay assay (Fig. 3). Of these recombinant
proteins, rOMM-64-I, III, IV and V, which include the
Glu-rich domain, were found to have calcium-binding
activity. rOMM-64-II and -C and GST were stained
red using ‘Stains-all’ and were not detected by
45
Ca.

the protein. (A) Schematic drawing of the recombinant proteins. Six
GST-fused recombinant proteins containing the three distinctive
domains of tandem repeat 1 (R1), the Glu-rich domain (E-rich)
and ⁄ or tandem repeat 2 (R2) were synthesized. SP, signal peptide
of the OMM-64 precursor. (B) ‘Stains-all’ staining of the recombi-
nant OMM-64 variants separated by SDS–PAGE to detect nega-
tively charged proteins as blue bands (left) and
45
Ca overlay
analysis of the proteins (right). I–V, C and G indicate the respective
recombinant proteins. G, GST. Calmodulin (C), used as a positive
control, was detected at approximately 17 kDa.
Fig. 4. Detection of OMM-64 in the inner ear tissues by western
blotting using anti-rOMM-64-C serum. In the saccular extract (S)
and endolymph (E), OMM-64 bands were observed by both ‘Stains-
all staining’ and western blotting (arrowheads). All proteins in the
EDTA-soluble (O
S
) and -insoluble (O
I
) otolith matrix were stained
blue using ‘Stains-all’. In these matrices, strong immunoreactions
were detected in the high-molecular-weight region (arrows).
H. Tohse et al. A novel protein in the otolith matrix framework
FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS 2515
of the sugar chains using trifluoromethanesulfonic acid
(TMSF), the HMW aggregate was not completely
digested even after 30 min of treatment (Fig. 5B).
After 15 min, many protein bands were detected by
silver staining, suggesting that several proteins in the

glycine, the beads were found to bind a HMW protein
and two proteins of approximately 95 and 140 kDa
(Fig. 7). The HMW protein band reacted strongly with
anti-rOMM-64 serum, whereas the other two bands of
95 and 140 kDa immunoreacted with anti-recombinant
otolin-1 (rOtolin-1) serum, as previously reported
[9,11]. These results suggest that the HMW aggregate
contains OMM-64 and otolin-1 within the otolith
Fig. 5. OMM-64 is contained in the HMW aggregate in the otolith
matrix and is excised from the aggregate by deglycosylation using
TFMS or heparitinase II. (A) Western blotting of EDTA-soluble
otolith matrix proteins (OSM) after digestion of polysaccharides by
glycopeptidase A (0.5 munits, G), chondroitinase ABC (0.5 units, C),
heparitinase II (10 munits, H), hyaluronidase SD (25 munits, Y) and
endo-a-N-acethylagalactosaminidase (70 munits, E). The HMW
aggregate was digested only by heparitinase II (arrowhead), and a
64 kDa protein band appeared instead (arrow). Some non-specific
binding was observed when these enzymes alone were subjected
to SDS–PAGE (Enzyme). (B) Time course of the effect of TFMS
treatment on the aggregate and free OMM-64. Although the
64 kDa band was observed by western blotting after treatment
with TFMS for at least 5 min (arrow) (aOMM-64), ‘Stains-all’ stain-
ing showed that the HMW aggregate was not digested completely
even after 30 min of treatment (arrowhead). Silver staining indi-
cated that the other proteins may be damaged by the 30 min TFMS
treatment. (C) Heparitinase II digests the HMW aggregate (arrow-
head) and separates free OMM-64 (arrow) in a concentration-
dependent manner. Bovine serum albumin, which was contained in
the enzyme solution, was observed at 66 kDa by both silver and
‘Stains-all’ staining.

(arrowheads) [15]. (C) Localization of OMM-64 in the otolith region
observed by differential interference contrast microscopy. OMM-
64 was localized in the ring-like structures in the otolith. (D) No
immunoreaction was observed in the negative control section of
the otolith region incubated with preimmune serum altered to pri-
mary antibody. CT, connective tissue; EL, endolymph region; M,
macula; O, otolith; TE, transitional epithelium. (E) Schematic of the
inner ear sacculus containing the otolith, indicating the sections in
(A)–(D).
Fig. 7. Separation of native OMM-64, otolin-1 and their complex by
co-immunoprecipitation. Anti-rOMM-64 or anti-rOtolin-1 affinity
beads were incubated with NaCl ⁄ P
i
(N), saccular extract (S) or
EDTA-soluble otolith matrix (O), and specifically bound proteins
were subjected to electrophoresis and staining using ‘Stains-all’.
Western blotting using anti-rOMM-64 and anti-rOtolin-1 antisera
was also performed. When the affinity beads were incubated with
saccular extract, OMM-64 (arrows) and otolin-1 (arrowheads) bound
separately to the beads. By contrast, incubation with otolith extract
resulted in binding of a complex of the HMW aggregate containing
OMM-64 and otolin-1 to the beads.
H. Tohse et al. A novel protein in the otolith matrix framework
FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS 2517
associated directly. In addition, using anti-rOMM-64
beads, a 64 kDa protein, which was not detected by
gel staining, was detected by anti-rOMM-64 serum.
Overall, these data suggest that both OMM-64 and
otolin-1 are contained in the HMW protein aggregate
in the otolith matrix.

lated, and the Glu-rich region and repeat 2 contain
many acidic residues, OMM-64 may be very acidic
overall and may function in interactions with calcium
and subsequent mineral crystallization. Although the
putative isoelectric point of the OMM-64 was calcu-
lated to be 3.5, the mature form of OMM-64 may be
more acidic because it may be highly phosphorylated.
Although we determined that the Glu-rich region of
the protein has calcium-binding activity, we could not
confirm whether repeat 1 also has activity because we
used non-phosphorylated recombinant proteins for
the calcium-binding assay. Therefore, the functions of
the two tandem repeat domains remain unknown
at present. We found starmaker and human dentin
sialophosphoprotein to be homologous proteins to
OMM-64 by blast search (blastp and tblastn). The
relationship between starmaker and dentin sialophos-
phoprotein has been discussed in detail in a previous
report [14]. Although some structural similarities in
the protein and gene were found between OMM-64
and starmaker (see Results), they may not be ortho-
logs because of their relatively low identity (25%).
However, their structural similarities may lead to
similar functions. In fact, knocking down starmaker
expression induces a variation in the polymorphism
of otolith crystals, from aragonite to calcite [14].
Therefore, OMM-64 and starmaker are thought to be
related proteins in terms of both structure and func-
tion. Although we carried out various blast searches
using amino acid, mRNA and genomic DNA

construct the extracellular matrices in the inner ear by
binding to glycosaminoglycans, which can hold water
and form gels.
The tissue-specific and proximal side-specific distri-
bution of mRNA expression and immunolocalization
suggests the potential function of OMM-64. In the
inner ear sacculus, the otolith is close to the proximal
side of the sacculus, and calcification of the otolith
occurs mainly at the proximal surface [19]. In addition,
proteins that may be involved in otolith calcification
A novel protein in the otolith matrix framework H. Tohse et al.
2518 FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS
are concentrated in the proximal endolymph [20].
Therefore, the proximal region of the sacculus pro-
duces the otolith matrix proteins [13] and forms the
environment for otolith mineralization. In particular,
the cells located at the periphery of the macula may be
specialized for production of the otolith matrix pro-
teins, because these cells are rich in rough endoplasmic
reticulum [13], and two otolith matrix proteins, otolith
matrix protein-1 and otolin-1, are also localized in
these cells [11]. Similar to the other otolith proteins,
OMM-64 may contribute to the heterogeneity of the
endolymph chemistry and otolith biomineralization. In
the otolith matrix, OMM-64 was localized in ring-like
structures, indicating that OMM-64 is periodically
incorporated into the otoliths. The manner of incorpo-
ration may be regulated by the binding activity of
OMM-64 to otolin-1, because periodic expression of
omm-64 mRNA was not observed (data not shown),

In summary, we have identified a novel protein,
OMM-64, contained in the HMW aggregate in the
otolith matrix, and shown that the aggregate also con-
tains ear-specific collagen, otolin-1, and forms frame-
work mineral constructs. The two proteins, OMM-64
and otolin-1, are expressed in the same cells in the
inner ear sacculus and are secreted into the extracellu-
lar matrices of the inner ear. In the otoliths, they are
both localized in the ring-like structures. These find-
ings identify for the first time proteins with these func-
tions that construct matrix aggregates in calcium
carbonate biominerals.
Experimental procedures
Animals
Rainbow trout (Oncorhynchus mykiss) weighing approxi-
mately 1000 g were used. They were reared in outdoor
ponds at 10–15 °C under natural light for at least 10 days
before collection of the samples.
Cloning of cDNA and DNA encoding OMM-64
As described previously [13], we detected HMW proteins
that may be aggregated in the otolith matrix by western
blotting using an antiserum raised against whole water-sol-
uble otolith matrix. To identify proteins contained in these
aggregates, immunoscreening of a cDNA library was per-
formed using this antiserum. Approximately 200 000 clones
contained in a kZAP II (Stratagene, La Jolla, CA, USA)
inner ear cDNA library, constructed according to the
method described by Murayama et al. [26], were grown on
LB agar ⁄ LB top agarose plates. Recombinant proteins in
each clone were induced and transferred to poly(vinylidene

gene). After growth and harvest of the Escherichia coli cells,
the amplified DNAs were recovered using a QIAprep mini-
prep kit (Qiagen, Hilden, Germany) and sequenced using a
DNA sequencer (3130xl Genetic Analyzer; Applied Biosys-
tems, Foster City, CA, USA).
Expression analyses of omm-64 mRNA
Total RNA was isolated from various organs (see Fig. 2)
using ISOGEN (Nippon Gene, Tokyo, Japan), and treated
with 4 unitsÆlL
)1
of DNase I (Takara, Kyoto, Japan) at
37 °C overnight. Complete digestion of genomic DNA con-
tamination in the total RNA was confirmed by lack of
amplification of a b-actin mRNA fragment by PCR using a
pair of primers (5¢ -ATCACCATCGGCAACGAGAG-3¢
and 5¢-TGGAGTTGTAGGTGGTCTCGTG-3¢) without
reverse transcription. After purification using phenol ⁄ chlo-
roform, 1 lg of the total RNA was reverse-transcribed
using a first-strand cDNA synthesis kit (Amersham Bio-
sciences, Little Chalfont, UK). Using 1 ⁄ 100 aliquots of the
first-strand cDNAs as templates, PCR was performed using
primers 5¢-GCTATGTTTCTGCAGGGTTCCTA-3¢ (nt
2385–2407) and 5¢-GCGTCATTAAACGTATGTACACT-3¢
(nt 2600–2578). Expression of b-actin mRNA was verified
using the primers described above.
For in situ hybridization, a 216 bp fragment (nt 2385–
2600) of omm-64 cDNA was amplified by PCR as described
above and ligated into pGEM-T vector (Promega). The
plasmid DNA was digested with NotIorNcoI, and anti-
sense and sense probes labeled with digoxigenin were pro-

Asp544–Ser628) were synthesized. The corresponding
regions of the omm-64 cDNA were amplified by RT-PCR
using six pairs of primers (rOMM-64-I, 5¢-CGCGGATCC
ACCGTAGACACTTATGATATA-3¢ and 5¢-CGCCTCCA
CCTAAGAGGCATCCTTGTCCAC-3¢; rOMM-64-II, 5¢-
CGCGGATCCACCGTAGACACTTATGATATA-3¢ and
5¢-CGCCTCGAG CTAAGAGTCAG CTTGCACGTC-3 ¢;
rOMM-64-III, 5¢-CGCGGATCCGCTGATGTGACCAGT
GATGAC-3¢ and 5¢-CGCCTCGAGCTATTTGGGCTCTT
TCATCAT-3¢; rOMM-64-IV, 5¢-CGCGGATCCGCCCCT
GTTAATGATGGAACC-3¢ and 5¢-CGCCTCGAGCTAA
GAAGACTGGGCTGCCAG-3¢; rOMM-64-V, 5¢-CGCGG
ATCCAGGCAAGATTTTAAGCATCCA-3¢ and 5 ¢-CGCC
TCCACCTAAGAGGCATCCTTGTCCAC-3¢; rOMM-64-
C, 5¢-CGCGGATCCGACTCAGTGGATGACCAATCC-3¢
and 5¢-CGCCTCGAGCTAAGAAGACTGGGCTGCC
AG-3¢) that had 5¢ adapters corresponding to BamHI
(GGATCC) and XhoI (CTCGAG) restriction sites, respec-
tively. In the reverse primers, stop codons (TAG) before the
XhoI sites were also added. PCR products were doubly
digested by the restriction enzymes, purified using a
QIAquick PCR purification kit (Qiagen), and ligated
into pGEX 6p-1 vector (Amersham Biosciences), which
had been digested and purified in the same way as the
PCR products. After transformation into XL1-blue cells and
confirmation of the sequences, the plasmid DNA was trans-
formed again into BL21 E. coli cells (Amersham Biosciences).
The cells were grown in LB medium containing
50 lgÆmL
)1

directly applied to SDS–PAGE under reducing conditions.
A novel protein in the otolith matrix framework H. Tohse et al.
2520 FEBS Journal 275 (2008) 2512–2523 ª 2008 The Authors Journal compilation ª 2008 FEBS
Separated proteins were stained with ‘Stains-all’ (Sigma,
St Louis, MO, USA) [28] or were blotted onto a PVDF
membrane to detect
45
Ca
2+
-binding activity [29].
Production of antibody against recombinant
OMM-64
The recombinant rOMM-64-C protein that bound to
glutathione–Sepharose beads was digested from GST by Pre-
scission protease (80 unitsÆmL
)1
gel bed; Amersham Bio-
sciences) at 4 °C for 2 days, eluted with 1 mL of NaCl ⁄ P
i
,
and concentrated and desalted using Ultrafree cartridges
(Millipore, 5000 Da cut-off). The digested rOMM-64-C was
completely separated from GST using a Sep-Pak Cartridge
C18 column (Waters, Millford, MA, USA) by stepwise
elution with acetonitrile.
After the purity and molecular mass (m ⁄ z 9227) of
rOMM-64-C had been confirmed by MALDI-TOF mass
spectrometry (4700 Proteomics Analyzer, Applied Biosys-
tems), the buffer of the protein was changed to NaCl ⁄ P
i

‘Stains-all’ [28] or silver to detect negatively charged pro-
teins and all proteins, respectively.
To detect OMM-64 and otolin-1 by western blotting, anti-
rOMM-64-C and anti-recombinant otolin-1-C [9] sera were
used. Ten micrograms of protein extracted from inner ear
was separated by SDS–PAGE and blotted onto a PVDF
membrane. The membrane was incubated first in 5% fat-free
dried milk in NaCl ⁄ Tris for 1 h, and then in the same solu-
tion containing the antibodies (1 : 1000 dilution) overnight.
After washing the membrane twice (10 mins each) with
NaCl ⁄ Tris containing 0.1% Tween-20 and once with NaCl ⁄
Tris, specific binding of the antibodies was detected by using
Supersignal West Femto Maximum Sensitivity Substrate
(Pierce, Rockford, IL, USA), and the corresponding second-
ary antibody (horseradish peroxidase-conjugated anti-rabbit
IgG, 1 : 5000), according to the manufacturer’s protocol.
Deglycosylation of proteins
Ten micrograms of otolith matrix protein were desalted in
an Ultrafree cartridge (5000 Da cut-off) and completely
dried in a centrifugal concentrator (VC-96W, Taitec,
Saitama, Japan). Chemical deglycosylation of the proteins
was performed by incubation with 50 lL of trifluorome-
thanesulfonic acid (TFMS) at 0 °C for 0, 5, 15 and 30 min.
The solutions were neutralized by adding 500 lL of ice-cold
buffer (1 m Tris). The sample solvent was changed to
20 mm Tris ⁄ HCl pH 8.0 using an Ultrafree cartridge
(5000 Da cut-off). For enzymatic digestion, 10 lg of water-
soluble otolith matrix protein, completely dried in a centri-
fugal concentrator, was incubated at 37 °C overnight with
10 lL of the following deglycosylation enzymes dissolved in

)1
)or
anti-rOtolin-1 serum was allowed to bind to a 0.5 mL bed
of Protein A–Sepharose beads (Amersham Biosciences)
according to the manufacturer’s protocol, and then the
antibodies and beads were covalently bound by incubation
with 20 mm dimethylpimelimidate (MP Biochemicals,
Solon, OH, USA) for 1 h at room temperature. After wash-
ing the beads with 10 mm NaCl ⁄ P
i
(pH 7.4), complete
binding of the antibody was confirmed by the detection of
no bands in the supernatant SDS–PAGE. Subsequently,
100 lL of saccular extract or water-soluble otolith matrix
proteins were bound to 10 l L of the affinity beads at 4 ° C
for 2 h, the beads were then washed twice each by voltexing
10 sec and centrifuged at 5000 g for 30 sec with 1 mL of
0.5 mm NaCl in 10 mm NaCl ⁄ P
i
(pH 7.4), 0.1 m glycine
(pH 2.5) and 20 mm Tris ⁄ HCl (pH 8.0). Proteins that
bound to the affinity beads were analyzed by applying a
5 lL bed of the beads directly to SDS–PAGE.
Tandem mass spectrometry
To identify the two proteins (95 and 130 kDa) bound to
recombinant otolin-1-C beads, tryptic peptides of these pro-
teins were applied to MS ⁄ MS analysis as follows. After gel
electrophoresis, the two protein bands were excised and
destained using a solution of 50% acetonitrile, 25 mm
NH

versity. This study was financially supported in part by
the Ministry of Education, Science, Sports and Culture
(Grants-in-Aid for Creative Basic Research numbers
12NP0201 and 17GS0311 and a Scientific Research for
Young Scientists Start-Up Grant number 18880001).
H. T. was supported by research fellowships from the
Japan Society for Promotion of Science for Young Sci-
entists (number 15-10657) and Akiyama Memorial Life
Science Foundation (number 18-6).
References
1 Belcher AM, Wu XH, Christense RJ, Hansma PK,
Stucky GD & Morse DE (1996) Control of crystal
phase switching and orientation by soluble mollusc-shell
proteins. Nature 381, 56–58.
2 Galini G, Albeck S, Weiner S & Addadi L (1996) Con-
trol of aragonite or calcite polymorphism by mollusk
shell macromolecules. Science 271, 67–69.
3 Addadi L, Joester D, Nudelman F & Weiner S (2006)
Mollusk shell formation: a source of new concepts for
understanding biomineralization processes. Chemistry
12, 980–987.
4 Samata T, Hayashi N, Kono M, Hasegawa K, Horita
C & Akera S (1999) A new matrix protein family
related to the nacreous layer formation of Pinctada
fucata. FEBS Lett 462, 225–229.
5 Kono M, Hayashi N & Samata T (2000) Molecular
mechanism of the nacreous layer formation in Pinctada
maxima. Biochem Biophys Res Commun 269, 213–218.
6 Matsushiro A, Miyashita T, Miyamoto H, Morimoto
K, Tonomura B, Tanaka A & Sato K (2003) Presence

otolith soluble-matrix producing cells in the saccular
epithelium of rainbow trout (Oncorhynchus mykiss)
inner ear. Anat Rec 254, 322–329.
14 So
¨
llner C, Burghammer M, Busch-Nentwich E, Berger
J, Schwarz H, Riekel C & Nicolson T (2003) Control of
crystal size and lattice formation by starmaker in otolith
biomineralization. Science 302, 282–286.
15 Takagi Y (1997) Meshwork arrangement of mitochon-
dria-rich, Na
+
,K
+
-ATPase-rich cells in the saccular
epithelium of rainbow trout (Oncorhynchus mykiss)
inner ear. Anat Rec 248, 483–489.
16 Perkins DN, Pappin DJC, Creasy DM & d Cottrell JS
(1999) Probability-based protein identification by
searching sequence databases using mass spectrometry
data. Electrophoresis 20, 3551–3567.
17 Arias JL, Neira-Carrillo A, Arias JI, Escobar C, Bodero
M, David M & Ferna
´
ndez MS (2004) Sulfated polymers
in biological mineralization: a plausible source for bio-
inspired engineering. J Mater Chem 14, 2154–2160.
18 Jiao X, Billings PC, O’Connell MP, Kaplan FS, Shore
EM & Glaser DL (2007) Heparan sulfate proteoglycans
(HSPGs) modulate BMP2 osteogenic bioactivity in

otolith matrix protein cDNA from the rainbow trout,
Oncorhynchus mykiss. Comp Biochem Physiol B 126,
511–520.
27 Tohse H, Murayama E, Ohira T, Takagi Y & Nagasa-
wa H (2006) Localization and diurnal variations of car-
bonic anhydrase mRNA expression in the inner ear of
the rainbow trout Oncorhynchus mykiss. Comp Biochem
Physiol B 145, 257–264.
28 Campbell KP, MacLennan DH & Jorgensen AO (1983)
Staining of the Ca
2+
-binding proteins, calsequestrin,
calmodulin, troponin C, and S–100, with the cationic
carbocyanine dye ‘Stains-all’. J Biol Chem 258, 11267–
11273.
29 Maruyama K, Mikawa T & Ebashi S (1984) Detection
of calcium binding proteins by
45
Ca autoradiography
on nitrocellulose membrane after sodium dodecyl
sulfate gel electrophoresis. J Biochem 95, 511–519.
30 Tohse H & Mugiya Y (2001) Effects of enzyme and
anion transport inhibitors on in vitro incorporation of
inorganic carbon and calcium into endolymph and oto-
liths in salmon Oncorhynchus masou. Comp Biochem
Physiol A 128, 177–184.
Supplementary material
The following supplementary material is available
online:
Fig. S1. Sequences of the omm-64 cDNA and its