By contrast to the widespread occurrence of aragonite, calcite is limited to several taxa with species-dependent microstructures composed of Keywords domain structure; foliated layer; oy
Trang 1A novel phosphorylated glycoprotein in the shell matrix of the oyster Crassostrea nippona
Tetsuro Samata, Daisuke Ikeda, Aya Kajikawa, Hideyoshi Sato, Chihiro Nogawa, Daishi Yamada, Ryo Yamazaki and Takahiro Akiyama
Laboratory of Cell Biology, Faculty of Environmental Health, Azabu University, Sagamihara, Japan
Subsequent to the pioneering work of Miyamoto et al
[1], Sudo et al [2] and Shen et al [3], more than 20
genes encoding the organic matrix (OM) components
of molluscan shells have been determined and their
deduced amino acid sequences clarified [4–12]
How-ever, the information available to date has been
restricted to the nacreous and prismatic layers of pearl
oysters, leaving the other shell layers poorly
investi-gated at the molecular level One exception is the
find-ing of acidic glycoprotein MSP-1 in the foliated layer
of Patinopecten yessoensis [4]
Through their control of nucleation, growth,
mor-phology and polymorphism of CaCO3 crystals, these
OMs are commonly assumed to be intimately
associ-ated with every phase of molluscan biomineralization, and thus with the overall regulation of the shell micro-structure More recent investigations have primarily involved in vitro measurement of OM activities related
to crystal formation [13–18] Although these studies have clearly shown that OM modulates molluscan bio-mineralization, the results nevertheless demonstrate marked methodology-dependent variation The func-tion of OM thus remains unclear, even in vitro, and is
a topic of future research
Molluscan shells are composed of either aragonite
or calcite By contrast to the widespread occurrence
of aragonite, calcite is limited to several taxa with species-dependent microstructures composed of
Keywords
domain structure; foliated layer; oyster shell;
phosphorylated matrix protein; poly-Asp
sequences
Correspondence
T Samata, Laboratory of Cell Biology,
Faculty of Environmental Health, Azabu
University, 1-17-71 Fuchinobe, Sagamihara,
Kanagawa 229-0006, Japan
Fax: +81 42 769 2560
Tel: +81 42 769 2560
E-mail: samata@azabu-u.ac.jp
Database
The nucleotide sequences have been
sub-mitted to DDBJ with the accession number
AB207821–AB207826
(Received 11 January 2008, revised 31
March 2008, accepted 7 April 2008)
doi:10.1111/j.1742-4658.2008.06453.x
We found a novel 52 kDa matrix glycoprotein MPP1 in the shell of Cras-sostrea nippona that was unusually acidic and heavily phosphorylated Deduced from the nucleotide sequence of 1.9 kb cDNA, which is likely to encode MPP1 with high probability, the primary structure of this protein shows a modular structure characterized by repeat sequences rich in Asp, Ser and Gly The most remarkable of these is the DE-rich sequence, in which continuous repeats of Asp are interrupted by a single Cys residue Disulfide-dependent MPP1 polymers occurring in the form of multimeric insoluble gels are estimated to contain repetitive locations of the anionic molecules of phosphates and acidic amino acids, particularly Asp Thus, MPP1 and its polymers possess characteristic features of a charged mole-cule for oyster biomineralization, namely accumulation and trapping of
Ca2+ In addition, MPP1 is the first organic matrix component considered
to be expressed in both the foliated and prismatic layers of the molluscan shell microstructure In vitro crystallization assays demonstrate the induc-tion of tabular crystals with a completely different morphology from those formed spontaneously, indicating that MPP1 and its polymers are poten-tially the agent that controls crystal growth and shell microstructure
Abbreviations
CBB, Coomassie brilliant blue; GISM, translucent gelatinous insoluble organic matrix; ISM, insoluble organic matrix; MPP1, molluscan phosphorylated protein 1; ntp, nucleotide position; OM, organic matrix; SM, soluble organic matrix.
Trang 2prismatic, foliated, chalky and granular structures In
particular, the shells of oyster species are composed of
a highly complex microstructure consisting of the
chalky layer in addition to the foliated and prismatic
layers The foliated layer is formed by the aggregation
of units termed lath, each with a width of 2 lm and
length of 10 lm [19], whereas the chalky layer has a
homogeneous morphology composed of tiny calcite
granules [19] A variety of studies, mostly based on
amino acid analysis of bulk soluble matrix (SM) and
insoluble matrix (ISM) [20–24], have shown the
pres-ence of OM in oyster species with particularly highly
acidic properties This high acidity is due to Asp and
phosphoserin (p-Ser) [25] Much of the accumulating
data on oyster shell biomineralization were obtained by
Wheeler et al., who have provided summaries of their
extensive studies [26,27] Their investigation of the
frac-tionation and functional analysis of the OM
compo-nents highlighted the inhibitory activity of the OM
against crystal formation in vitro Immunocytochemical
studies of the OM in the prismatic layer of C virginica,
as reported by Kawaguchi and Watabe [28], revealed
that the ISM constituted the framework of the OM
and SM, which comprised several phosphorylated
pro-teins and might be distributed on the surface of the
ISM and surrounded calcite crystals Atomic force
microscope and scanning electron microscope
observa-tions of foliar chips after pyrolysis and their
subse-quent crystallization revealed that crystal formation
occurred on the surface of the laths under the
regula-tion of the OM, which showed pulsed secreregula-tion [19]
As noted above, the primary structure of oyster OM
has yet to be precisely determined In the present
study, we aimed to elucidate the overall picture of the
OM components of Crassostrea nippona by a
combina-tion of biochemical and genetical analyses For gene
analysis, given the close similarity of the shell structure
and amino acid composition of the OM of the oyster
and scallop, we started with the isolation of cDNA
clones homologous with MSP-1 gene Additional
in vitro crystallization assays were then performed to
investigate the function of the OM components
Results
Biochemical characterization of the OM
components extracted from oyster shell
Fractionation of the bulk OM separated two fractions,
namely the SM at approximately 20 mg per 50 g of
shell and the ISM, which was further sub-divided into
two components: a predominant translucent gelatinous
insoluble organic matrix (GISM) pellet at
approxi-mately 120 mg per 50 g of shell and a small quantity
of fibrous precipitate at approximately 5 mg per 50 g
of shell
After SDS⁄ PAGE of GISM, which was largely-solubilized in a sample buffer containing 2-mercaptoeth-anol after boiling, and subsequent staining procedures with negative staining, Stains-all and Methyl green visu-alized an exclusive band of approximately 52 kDa, which showed a negative reaction with Coomassie brilliant blue (CBB) (Fig 1)
SDS⁄ PAGE of the 52 kDa component after enzy-matic deglycosylation and dephosphorylation showed apparent downward shifts in molecular masses of 2.5 and 3.5 kDa, respectively (Fig 2)
Table 1 shows that the 52 kDa component in GISM exhibits an amino acid composition, strikingly domi-nated by Asx (aspartic acid plus asparagine), which, together with Ser and Gly, accounted for more than 80% of the total residue By contrast, the bulk SM showed a different amino acid composition, which comprised large amounts of Asx, Glx (glutamic acid plus glutamine) and Gly, and a much smaller amount
of Ser than that of GISM
Amino acid sequence analysis using a peptide sequencer failed to determine the N-terminal sequence
of the 52 kDa component Likewise, LC⁄ MS ⁄ MS anal-ysis of the V8 protease digests of the 52 kDa compo-nent did not reveal any peptide with sequences corresponding to those of the 44 kDa deduced protein
66.2 45 (kDa)
Fig 1 SDS ⁄ PAGE electrophoretogram of GISM in the OM of
C nippona The same amount of sample was applied to each lane Lane M, molecular mass standards; lane A, CBB staining; lane B, Stains-all staining; lane C, negative staining; lane D, Methyl green staining Arrows on the right side of the lanes indicate the position
of the 52 kDa component A weakly stained band in lane A does not correspond to the 52 kDa component but a minor component with a molecular mass of 45 kDa.
Trang 3encoded by the 1.9 kb cDNA and those reported so
far On the other hand, the deduced 44 kDa protein
was identified as a top protein score of 55 (probability
based mowse score) using the Mascot search engine
for the fragments digested with endoproteinase Asp-N
of the 52 kDa component, whereas scores of the other
proteins in the database were lower than 20 Among each peptide sequence with high peptide scores, a short but specific sequence of DCGVDCGYYEPV (score of 19) at the N-terminal region of the deduced 44 kDa protein and an additional sequence of DNNGDGNG (score of 16) in the NG repeat sequence at the C-ter-minal region were characteristic The same result was obtained using the Sequest search engine The most appropriate condition for Asp-N digestion was the addition of 75 ng of enzyme to 15 lg of protein FTIR analysis of GISM showed the most intensive absorption peaks at 1654 cm)1 and 1561 cm)1, corre-sponding to amides I and II, respectively, characteristic
in protein moiety (Fig 3) [29] The small peak at
1243 cm)1 may represent amide III, sulfates or phos-phates and that at 1408 cm)1 may be associated with carboxylate [29,30] An additional large absorption peak occurred at around 1097 cm)1, which was consid-ered to be associated with carbohydrates [29,30]
Cloning and sequencing of cDNA encoding the
OM component in the foliated layer
A nucleotide fragment of approximately 320 bp was amplified using the primer pairs of F1 and R1 Nucle-otide sequences of the primer positions in this frag-ment (fragfrag-ment A) completely matched that of MSP-1 gene corresponding to F1 and mismatched at five nucleotide positions corresponding to R1, and the deduced amino acid sequences at the N-terminal and C-terminal regions were SGSSSSS and GGDGGDG 3¢-Rapid amplification of cDNA ends (3¢-RACE) using the set of primers of the adaptor primer and the
66.2
45
(kDa)
66.2
45 (kDa)
Fig 2 SDS ⁄ PAGE electrophoretogram of (A) dephosphorylated
GISM and (B) deglycosylated GISM Lane M, molecular mass
stan-dards; lane 1, dephosphorylated GISM (A) and deglycosylated
GISM (B); lane 2, native GISM The same amount of the sample
was applied to each lane Arrows on the right side of the lanes
indi-cate the position of the native and the (A) dephosphorylated and
(B) deglycosylated 52 kDa components.
Fig 3 FTIR spectrum of the GISM fraction The vertical scale shows the intensity exhibited by %T.
Table 1 Amino acid compositions of the 52 kDa component and
the SM of C nippona together with that of the deduced 44 kDa
protein Values were calculated by mole percentage Asx,
Asp + Asn; Glx, Glu + Gln; Ser, Ser + p-Ser Any amount of amino
acids lower than approximately 1% in the 52 kDa component and
the SM may not be accurate because of contamination from the
poly(vinylidene difluoride) membrane.
44 kDa
deduced
protein
52 kDa
Trang 4gene-specific primer F2 based on the nucleotide
sequence of fragment A amplified a fragment of
approximately 1040 bp (fragment B), in which the F1
primer annealed with the same sequence, located
345 bp upstream of the primer position The 5¢-region
of the nucleotide sequence completely matched that of
fragment A
Next, 5¢-rapid amplification of cDNA ends
(5¢-RACE) revealed the presence of one positive clone
(fragment C) with a length of approximately 1050 bp
The 3¢-region of the nucleotide sequence completely
matched that of fragment B
Using PCR employing two gene-specific primers of
F3 and R3 to obtain the full-length cDNA, a fragment
of approximately 1.7 kb was amplified, which included
sequences consistent with those of the
above-men-tioned fragments, A, B and C After addition of the
remaining sequences, namely approximately 100 bp of the 5¢-region and 110 bp of the 3¢-region, the full length of the obtained clone was determined to be approximately 1.9 kb An additional two clones that lacked nucleotide sequences between nucleotide posi-tion (ntp) 913–1011 and ntp 754–1083 of the 1.9 kb clone were amplified The full lengths of these two clones were approximately 1.8 and 1.56 kb, respec-tively The nucleotide sequences reported here have been submitted to the GenBank TM⁄ EBI Data Bank with accession numbers AB207821–AB207826
The cDNA preserved the fundamental structure necessary for an ORF such as the start and stop codons, poly A signal and polyA tail An in-frame stop codon TAG was located at ntp 1696–1698 with
a putative polyadenylation signal (AATAAA) located
at ntp 1866–1871 of the 1.9 kb cDNA The relevant
Fig 4 Nucleotide sequence of the 1.9 kb cDNA and deduced amino acid sequence Numbers on the left indicate the nucleotide positions in the 1.9 kb cDNA sequence (upper) and positions of the amino acid resi-dues in the deduced protein (lower) The putative signal peptide is underlined The start codon (ATG), stop codon (TAG) and putative polyadenylation signal (AATAAA) are boxed.
Trang 5nucleotide and deduced amino acid sequences are
shown in Fig 4
Deduced protein structure encoded by the 1.9 kb
cDNA
The deduced protein encoded by the 1.9 kb cDNA
fragment encompassed 516 amino acid residues and
had a calculated molecular mass before
post-transla-tional modification of 46561.41 Da Following the
typ-ical sequence for signal peptide, comprising 19 amino
acids, the N-terminal amino acid of the mature protein
was expected to be Ala based on the prediction using
neutral networks and hidden Markov models
Eventu-ally, the molecular mass of the mature protein was
estimated to be 44490.85 Da, containing 497 amino
acid residues
The amino acid composition of the deduced protein
was characterized by a high proportions of Ser
(33.80%), Gly (28.97%) and Asp (26.77%), which
together accounted for more than 80% of the total
amino acid residues (Table 1) By contrast, the
occur-rence of basic amino acids was markedly low, with
only two Lys residues, resulting in a much higher
pro-portion of acidic to basic amino acids than in MSP-1
The deduced 44 kDa protein revealed a modular
structure with a domain characterized by repeat
sequences rich in Ser and Gly, named the SG domain
This was segmented eight times by comparatively short
repeats of a DE-rich sequence (Fig 5) The sequence
of N-terminal region was followed by an NGD
domain rich in Asn, Gly and Asp, which formed nine segments of NGD Another NGD domain, containing seven segments of NGD, was characterized by five sets
of GDYNGN⁄ A occurring at the C-terminal region Similar short sequences of GGDGGDGDN occurred twice at the C-terminal side The NGD domain at the N-terminal region was connected by an SDG-rich sequence comprised mainly of an irregular arrange-ment of Ser, Gly and Asp A similar sequence repeated twice at the C-terminal region with nine repeats of SD The subsequent SG domain was dominated by sequences of (Ser)n–(Gly), where n = 1–4 The DE-rich sequence predominantly contained the acidic amino acids, which appeared in a characteristic manner as (DEDCED), (DDGDEDCEDE),
(DDDDDCD-DDD) In the sequence, Asp was contained preferably over Glu, and a single Cys residue was located at its center
A search of the nonredundant GenBank CDS data-base using blast (protein–protein blast and Search for short, nearly exact matches) showed a similarity of 34.4% between the sequence throughout the molecules
of the deduced 44 kDa protein and MSP-1, with only exceptional high similarity between the SG domain of them (Fig 6) Partially high correspondence with phos-phophorin, a dentin Ca-binding phosphoprotein [31], and Lustrin A [3], a molluscan OM protein from a gastropod Haliotis rufescens, was observed over the 50 amino acids comprising the SG domain of this protein
No clear homology with any other protein occurring
in the database
Motif analyses by scanprosite (provided by Swiss Institute of Bioinformatics, SIB, Geneva, Switzerland) and netphosk (provided by Center for Biological Sequence Analysis BioCentrum-DTU Technical Uni-versity of Denmark, Lyngby, Denmark) suggested that
35 and 45 casein kinase II phosphorylation sites were present, respectively A motif of an N-glycosylation site was detected at two positions of the molecule An additional motif of GAGs (glucose aminoglycans)-binding indicated as DGSD was confirmed at two positions of the C-terminal region
With consideration of the phosphorylation sites and excluding the putative signal peptide, use of the scansite tools of the ExPASy server showed that the deduced 44 kDa protein had a very low theoretical pI
of 1.21 considering the 35 casein kinase II phosphory-lation sites
The additional two proteins encoded by the 1.8 and 1.56 kb cDNAs lacked amino acid residues between
256 and 288 corresponding to the SG domain in the second unit of the deduced 44 kDa protein and
Fig 5 Schematic representation of the domain structures of
MPP1 and MSP-1 The SG domain, DE-rich sequence and NGD
domain of MPP1 are arranged to constitute the unit structure
twice, named unit-1 and unit-2 The sequence between these two
units was completely conserved.
Trang 6residues 203 to 316 corresponding to the whole second
unit of the protein, respectively
Tissue specific expression of a transcript of the
1.9 kb cDNA
As shown in Fig 7, northern blot hybridization
showed that a transcript of approximately 1.5–2.0 kb
was detected solely in the RNA from mantle pallial,
where it contributed to the formation of the foliated
layer A band slightly smaller in size, and which had
a much weaker intensity of the chemiluminescence
reaction than the former band, was detected in the
mRNA from the mantle edge, where it contributed to
the formation of the prismatic layer By contrast,
they were not expressed in gill or adductor muscle
In vitro assay of OM activity
In the systems used for the ‘CaCO3 crystal growth assay’, characteristic inhibitory efficiency against crystal formation was recognized after addition of the SM, GISM and the 52 kDa component to the crystallizing solution Inhibition was observed as a change in crystal morphology, from a characteristic rhombohedral shape
to a poor crystalline habit with rounded edges for calcite crystals, and from a spherical shape with needle-like structure to a spherulite shape with smooth surfaces for aragonite crystals, and the complete loss of crystal shape for both in an additive volume-dependent manner One interesting result obtained by contrast interference microscopy was the induction of tabular crystals of oval
to quadrangular shape with rough edges and very fine parallel stria along the bottom face when the three above mentioned components were added to the arago-nitic crystallizing solution with the underlying GISM-derived membrane These crystals were observed to be tightly adhered to the membrane in a manner com-pletely different from those inorganically formed or those formed without fixative (Fig 8A-1, 2) Scanning electron microscopy of the edge of the crystals revealed the presence of rod-like rectangular structures with
a striking morphological appearance and dimensions closely comparable to those of the folia (Fig 8B) Consistent with the findings of Wheeler et al [15], an instantaneous decrease in pH was seen in the ‘CaCO3 precipitation assay’ when CaCl2was added to the bicar-bonate solution, followed by an additional downward trend intercalated by relatively stable periods The dura-tion of the stable periods was increased and the rate of
pH decrease was attenuated in a volume-dependent
Fig 6 Alignment of the amino acid sequences of MPP1 and MSP-1 Asterisks show identical amino acids, and dashes correspond to deletions Numbers on the right and left indicate the number of the amino acid residues in the MPP1 (upper) and MSP-1 (lower) sequences.
Fig 7 Electrophoretogram of a transcript of the 1.9 kb cDNA by
northern hybridization Samples of total RNA were isolated from
different oyster tissues: lane A, mantle edge, responsible for
pris-matic layer formation; lane B, mantle pallial, responsible for foliated
layer formation; lane C, adductor muscle; lane D, gill; lane M,
molecular weight standard of RNA The arrow indicates the 2.0 kb
RNA marker.
Trang 7manner with respect to the additive (Fig 9) Notably,
this tendency toward the inhibition of crystal nucleation
was more intensive with the 52 kDa component than
with the same amount of phosphovitin used for
refer-ence (Fig 9)
Discussion
A common feature of the oyster OM as reported in a
number of studies is the overall similarity of amino
acid composition among the bulk SM, ISM and even
several purified components that comprise the SM, as
described above [20–24] Because no other component
exhibits the same composition or is stained with both
negative staining and Methyl green, we assume that the 52 kDa component, which accounts for a consider-able part of the gelatinous material in the foliated layer, is the main phosphorylated glycoprotein A sec-ond key component in oyster biomineralization might
be the polyanionic components contained in the SM, although their primary structures are still unclear The predicted amino acid composition of the deduced 44 kDa protein agrees well with that of the
52 kDa component in the foliated layer of C nippona (Table 1) and the 54 kDa phospholylated component (RP-1) in the same layer of C virginica [26], as well as those of the bulk OMs reported from several oyster species described to date [20–24] In addition,
LC⁄ MS ⁄ MS analysis of the endoproteinase Asp-N digest of the 52 kDa component revealed the presence
of several peptides with amino acid sequences corre-sponding to those in the sequence of the genetically determined 44 kDa protein, although amino acid sequence analyses using the peptide sequencer failed to determine the N-terminal sequence of the 52 kDa component, strongly suggestive of the presence of N-terminal block As noted in the present study, FTIR, amino acid composition and motif analyses all suggest that the size discrepancy between the deduced
A1
B
A2
Fig 8 Surface views of crystals induced by ‘CaCO 3 crystal
growth assay’ (A1) Spontaneously formed aragonite crystals,
Scale bar = 100 lm (A2) Tabular crystals of oval to quadrangular
shape with rough edges induced on the GISM-derived membrane
after addition of the 52 kDa component at 5 lg Scale
bar = 100 lm (B) Scanning electron microscopy of the edge of
the tabular crystals Scale bar = 100 lm The presence of residual
undissolved CaCO3 crystals was carefully checked by scanning
electron microscopy, an energy dispersive X-ray spectrometer,
FTIR and an X-ray diffractometer Experiments were repeated at
least 10 times for each batch.
A B
C D E
0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00
(min)
(pH) 8.8
8.6
8.4
8.2
8.0
7.8
7.6
7.4
7.2
7.0
Fig 9 Recordings of CaCO 3 precipitation by ‘CaCO 3 precipitation assay’ (A) Reference experiment performed by addition of distilled water (DW) to the crystallizing solution (B, D, E) Addition of the
52 kDa component to the crystallizing solution at 2.5 lg (B), 10 lg (D) and 50 lg (E) (C) Addition of phosphovitin to the crystallizing solution at 50 lg.
Trang 844 kDa protein and the 52 kDa component may be
attributed to post-translational phosphorylation and
glycosylation This assumption was supported by the
results obtained for the enzymatic dephosphorylation
and deglycosylation experiments of the 52 kDa
com-ponent These data indicate with high probability that
the 1.9 kb cDNA is the gene encoding the 52 kDa
protein Finally, we conclude that the 52 kDa
compo-nent is a main novel phosphorylated glycoprotein that
is intimately involved in shell formation of C nippona
and thus can be designated: MPP1 (molluscan
phosph-orylated protein 1) Although MPP1 shares high
homology with MSP-1 as a whole, the differences
between them are obvious with respect to the presence
of a DE-rich sequence and the lack of a K domain,
together with the relatively high amount of Cys in
MPP1 (Fig.5), their respective molecular masses
(52 kDa for MPP1 versus 74.5 kDa for MSP-1), the
number of potential phosphorylation sites (35–45 sites
in MPP1 versus 9–10 sites in MSP-1) and their
respec-tive pI (1.21 for MPP1 versus 3.15 for MSP-1,
consid-ering ten casein kinase II phosphorylation sites)
The complete primary structures of two highly acidic
OM proteins from the prismatic layer and one from
the foliated layer have been reported, namely Aspein,
with a GS(D)5 repeat [9]; Asprich, whose D block has
a maximum 10 Asp repeat [11]; and MSP-1 in the
foli-ated layer, which lacks the poly-D sequences [4] In
addition, a 17 kDa protein, caspartin, isolated from
the prismatic layer of Pinna nobilis [32], had Asp as
the first of 75 N-terminal amino acid residues;
how-ever, its complete primary structure has not been
revealed Among these three genetically determined
proteins, only MSP-1 has been confirmed as being
dis-tributed in the shell, as demonstrated by the
N-termi-nal amino acid sequence of the OM component
matching that deduced from the nucleotide sequence
of the MSP-1 gene, although a band with a
compara-ble molecular size as that of MSP-1 could not be
vali-dated by SDS⁄ PAGE
Regarding the modular structure of MPP1, the
remarkable DE-rich sequence appears to be
anoma-lous, in that the continuous repeats of Asp are
inter-rupted by a single Cys residue, which is conserved in
all DE-rich sequences except one This sequence
con-servation of Cys hints at its functional significance,
namely that it is incorporated in the formation of
intra- or inter-molecular disulfide bonds In the latter
case, MPP1 monomer may be self-assembled to a
poly-mer, converting them to an insoluble form, although
the mechanism of this insolubility is unknown
The secondary structure of MPP1 estimated by the
method of Chau and Fasman [33] consists
predomi-nantly of a loop structure, which mainly corresponds
to the repeated arrangement of the SG domain with densely distributed phosphorylation sites inserted by the DE-rich sequence In turn, this gives rise to the regular arrangement of the anionic molecules of phosphates and acidic amino acids Given this assumption, disulfide-dependent MPP1 polymers occurring in the form of multimeric insoluble gels can be estimated to contain a massively repeating acidic region MPP1 polymers may thus participate in oyster shell formation by accumulating Ca2+ through
an ionotropic effect of phosphates, analogous to that with sulfates [34], which extend from the peptide chain Further binding of Ca2+ to carboxyl groups
of Asp or Glu arranged in the DE-rich sequence occurs, followed by the subsequent reaction of the
Ca2+ with CO3 ), which may be concentrated by the specific function of nacrein whose presence in oyster shells has been genetically determined [35] In this way, subsequent sequential reaction of the anionic and cationic ions may result in the nucleation of CaCO3 crystals With regard to the biochemistry of the reactions between the OM and Ca2+, Weiner and Hood [22] and Weiner and Traub [36] proposed that the regular spacing of the carboxyl side chains of Asp is a close reflection of that of Ca2+ in CaCO3 crystal lattices, and thus controls crystal polymor-phism However, it should be noted that highly acidic proteins have been associated with calcitic shell lay-ers, indicating the potential involvement of the Asp-and⁄ or p-Ser rich components in calcite formation not only in the prismatic layers, but also in the foli-ated layers This notion is supported by the results of the present study
By contrast to this notion, however, our in vitro crystallization assay showed that the OM compo-nents had an inhibitory effect against CaCO3 crystal formation This does not necessarily imply a negative role for the OM components in oyster shell biomin-eralization because, although the soluble and the additive components inhibited crystal formation when present in the isolated state, the same molecules induced tabular crystals with a completely different morphology from spontaneously formed crystals when pre-mixed with underlying GISM-derived mem-brane Unfortunately, X-ray diffractional analysis of the tabular crystals failed to determine their mineral-ogy due to their small quantities, which were far less than the minimum detectable quantity The basement membrane is an artificial material, which is prepared from the gelatinous pellet by clumping together on drying The surface area of the membrane may
be hydrated again and returned to the form of a
Trang 9concentrated gel in the crystallizing solution,
imply-ing that oyster shell formation may occur in a
gelati-nous environment containing a multimeric complex
of the MPP1 molecule A similar environment was
envisaged in the case of the formation of the
nacre-ous layer, to which jelly component comprising
MSI60 might be related [37,38] For formation of
the multimeric complex, GAGs that were estimated
to be in close contact with GISM [28] might be
responsible because the potential binding sites of
GAG were found in the deduced 44 kDa protein
As an additional but decisive contributor to calcite
induction, we emphasize the role of phosphate, which
has been specifically identified as the accessory
mole-cule of p-Ser in the OM of the foliated layer This
notion is supported by the fact that phosphate content
in the foliated layer far exceeds that in the aragonitic
shell layers [20] One study identified phosphate as
favorably controlling calcite formation when added to
the calcium carbonate solution in trace amounts [39]
The precise effect of phosphate in polymorphism
con-trol awaits future study
Additional identification of the MPP1-related
com-ponent in the prismatic layer of C nippona, as well
as in vitro crystallization assays using recombinant
proteins or synthesized peptides, will initiate a new
phase in the elucidation of oyster shell formation,
and highlight the control of CaCO3 polymorphism
and shell microstructure in molluscs In further trials
to obtain a whole figure of molluscan shell
biominer-alization, several additional factors must be taken
into consideration; namely, the behaviour of cells, the
composition of extrapallial fluids, functions of the
signal molecules regulating expression of the OM
component, as well as environmental factors, as
described by Kuboki et al [40] Genetical research
combined with an analyses of these factors may
com-prise a potential tool for the elucidation of molluscan
biomineralization in the future
Experimental procedures
Molluscan materials
We used live individuals of C nippona cultured at the
hatch-ery of Shimane Technology Center for Fisheries, Japan
Extraction and purification of the organic matrix
proteins
Shell surfaces were cleaned with an electric rotary grinder
(JOY-ROBO, Cannock, UK) to roughly remove
perios-tracum and adherent hard tissues Pieces of folia were
carefully separated from the powder of chalky material and then immersed in 5% NaClO for 30 min to remove organic contaminants After rinsing with distilled water (DW) and air-drying, folia were ground into powder with
a ball mill (ITO Manufacturing, Nagano, Japan) The powdered folia was decalcified with 5% acetic acid for
3 days at 4C under constant stirring and with pH regu-lated at over 4.5, followed by dialysis against DW The dialyzed solution was centrifuged at 15 000 g for 30 min
to obtain separation of the supernatant SM and precipi-tated GISM These two fractions were boiled in a sample buffer containing 5% 2-mercaptoethanol for 1 min and then subjected to SDS⁄ PAGE using Pagel (gradient gel
of 5–20%; ATTO, Tokyo, Japan) under reducing condi-tions in a Dual Mini Slab Chamber (ATTO) After elec-trophoresis, bands were stained with CBB (Sigma-Aldrich Chemie, Steinheim, Germany), Stains-all (BDH, Dorset, UK) [41], Methyl green (CHROMA, Rockingham, VT, USA) [42] and negative staining [43], all as previously described
Amino acid composition and N-terminal sequence analysis
Following separation with SDS⁄ PAGE, OM components were electro-blotted onto a poly(vinylidene difluoride) membrane (Immobilon Transfer Membranes; Millipore, Bedford, MA, USA) by a semi-dry blotting system (Nihon Eidou, Tokyo, Japan) and then stained with CBB To determine the amino terminal sequence, the target protein bands were cut from the membrane and subjected directly
to an automated amino acid sequence analyzer LF3000 (Beckman Coulter, Fullerton, CA, USA) To determine amino acid composition, membrane pieces corresponding
to the protein bands were hydrolyzed in 5.7 m HCl at
110C for 24 h Hydrolyzed samples were analyzed with
an L-8500 automated amino acid analyzer (Hitachi, Tokyo, Japan) using ion-exchange, post-column Ninhydrin detection
Enzymatic digestion and LC⁄ MS ⁄ MS analysis V8 protease (Pierce, Rockford, IL, USA) and endoprotein-ase Asp-N proteendoprotein-ase (Roche, Bendoprotein-asel, Switzerland) were added
to the gel pieces, which contained the 52 kDa component dissolved in 50 mm sodium phosphate buffer (pH 7.8) The amounts of the enzymes and proteins were changed at a ratio between 1 : 50 and 1 : 200 After incubation at 37C for
18 h, the protease digests were dried and dissolved in 10 lL
of trifluoroacetic acid, and then cleaned up by Zip-tip (Milli-pore) Purified digests were subjected to LC⁄ MS ⁄ MS anal-ysis on a Paradigm MS4 LC System coupled to a model LCQ ion trap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an electrospray
Trang 10inter-face utilizing a C18 column (Michrom Bioresources, Auburn,
CA, USA)
Deglycosylation and dephosphorylation
experiments
PNgase F (Roche) digestion of GISM was carried out as
described below After addition of 100 lL of incubation
buffer [50 mm sodium phosphate buffer (pH 7.8), 10 mm
EDTA (pH 8.0), 0.5% (v⁄ v) Nonidet P40, 0.2% (w ⁄ v)
SDS, 1% (v⁄ v) 2-mercaptoethanol] to an equivalent volume
of GISM, the mixture was incubated for 18 h at 37C with
2 units of PNgase F
Alkaline phosphatase (Roche) digestion of GISM was
carried out according to the manufacturer’s instructions
After addition of 5 lL of 10· phosphatase buffer to 45 lL
of GISM, the reaction mixture was incubated for 1.5 h at
37C with 4 units of alkaline phosphatase
FTIR analysis
Samples were mixed with KBR and analyzed by FTIR
(Magna-IR 750, Thermo Fisher Scientific)
cDNA cloning
Tissue collection for RNA extraction
The outer mantle epithelial tissue responsible for secretion
of the foliated layer was carefully separated from that part
of the mantle edge responsible for secretion of the prismatic
layer and immediately frozen in liquid nitrogen
Total RNA extraction
Total RNA was extracted from 300 mg of mantle epithelial
tissue using Isogen (Nippongene, Tokyo, Japan) and purified
with a SV RNA Isolation System (Promega, Madison, WI,
USA) The total amount of RNA was calculated with a
spectrophotometer (GeneQuant; GE Healthcare Bioscience,
Quebec, Canada)
PCR amplification
Single-stranded cDNA was synthesized with SuperScript III
RNase H) Reverse Transcriptase (Invitrogen, Carlsbad,
CA, USA), and purified after transcription using a Wizard
SV Gel and PCR Clean-Up System (Promega) A cDNA
fragment encoding the oyster OM protein was amplified
using a set of gene-specific primers of F1 (forward 953,
3¢-end corresponding to ntp 953 of the MSP-1 gene)
(5¢-TCC GGC TCA AGC TCT AGC TCT-3¢) and R1 (reverse
1369 of the MSP-1 gene) (5¢-TCC ATC ACC TCC ATT
GCC TCC-3¢), corresponding to the amino acid sequences
of the SGSSSSS and GGNGGDG of the MSP-1 gene,
respectively Primers were supplied by Texas Genomics Japan (Tokyo, Japan) PCR amplification was performed using KOD-Plus as an enzyme for extensive reaction with a thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) 3¢-RACE was carried out using a set of primers of an adaptor primer (TCG AAT TCG GAT CCG AGC TCT) and the gene-specific primer of F2 (forward 918) (5¢-TGC GAT GAT GAT GAC AGC GGA-3¢), based on the nucle-otide sequence of the cDNA fragment obtained from the first PCR
5¢-RACE was primed using a Smart Race Kit (Clontech, Mountain View, CA, USA) using a set of an adaptor UPM and the gene-specific primer of R2 (reverse 1056) (5¢-TGC GAG GAT GGT GGT GAT GGA-3¢), designed from the nucleotide sequence of the cDNA fragment amplified by 3¢-RACE
The full length of the cDNA encoding the oyster OM protein was amplified using a set of the gene-specific prim-ers of F3 (forward 136) (5¢-CCT AGA AGA ATA CAT CGG GGT-3¢), and R3 (reverse 1827) (5¢-TCT GGC ATG AAA CAC GAC AAC-3¢), based on the nucleotide sequences of the 5¢ and 3¢ terminal regions, respectively
TA cloning After purification and A-tailing, the PCR products were used for ligation with pGEM-T Easy Vectors (Promega), and catalyzed with T4 DNA ligase at 4C for 16 h The ligation products were supplied for transformation of JM109 high-efficiency competent cells (Promega) Positive clones were selected by blue⁄ white colour screening and standard ampicillin selection, followed by purification using
a Qiaprep Spin Miniprep Kit (Qiagen, Tokyo, Japan)
Sequencing The purified clones were labelled with a Thermo Sequence Primer Cycle Sequencing Kit (GE Healthcare Bioscience) and sequenced with an automated DNA sequence analyzer DSQ-1000L (Shimadzu, Kyoto, Japan)
Northern blot hybridization Total RNA was extracted with Isogen (Nippongene) from each tissue (mantle edge, mantle pallial, gill and adductor muscle) of C nippona and purified using a SV RNA Isola-tion System (Promega) RNA samples were segregated by electrophoresis on a 1% (w⁄ v) formaldehyde agarose gel and transferred to a positively-charged nylon membrane (GE Healthcare Bioscience) Hybridization was performed
at 58C using an Alkali Phos Direct Labelling and Detec-tion Kit (GE Healthcare Bioscience) Probes for analysis were designed in correspondence to ntp 4–124 of the 1.9 kb cDNA CDP-star was used for detection and