shell nacre of Nautilus macromphalusBenjamin Marie1,2, Isabelle Zanella-Cle´on3, Marion Corneillat4, Michel Becchi3, Ge´rard Alcaraz2,4, Laurent Plasseraud2,5, Gilles Luquet1,2and Fre´de
Trang 1shell nacre of Nautilus macromphalus
Benjamin Marie1,2, Isabelle Zanella-Cle´on3, Marion Corneillat4, Michel Becchi3, Ge´rard Alcaraz2,4, Laurent Plasseraud2,5, Gilles Luquet1,2and Fre´de´ric Marin1,2
1 UMR 5561 CNRS, Bioge´osciences, Dijon, France
2 Universite´ de Bourgogne, Dijon, France
3 IFR 128 BioSciences, UMR 5086 CNRS, IBCP, Universite´ de Lyon 1, Lyon, France
4 UPSP PROXISS, De´partement Agronomie Environnement, AgroSup, Dijon, France
5 ICMUB, UMR CNRS 5260, Faculte´ des Sciences Mirande, Dijon, France
Keywords
biomineralization; de novo sequencing;
immunolocalization; mollusc shell nacre;
organic matrix
Correspondence
B Marie or F Marin, UMR 5561 CNRS
Bioge´osciences, Universite´ de Bourgogne,
6 Boulevard Gabriel Dijon 21000, France
Fax: +33 3 80 39 63 87
Tel: +33 3 80 39 63 72
E-mail: benjamin.marie@u-bourgogne.fr;
frederic.marin@u-bourgogne.fr
(Received 29 September 2010, revised 18
March 2011, accepted 11 April 2011)
doi:10.1111/j.1742-4658.2011.08129.x
In molluscs, and more generally in metazoan organisms, the production of
a calcified skeleton is a complex molecular process that is regulated by the secretion of an extracellular organic matrix This matrix constitutes a cohe-sive and functional macromolecular assemblage, containing mainly pro-teins, glycoproteins and polysaccharides that, together, control the biomineral formation These macromolecules interact with the extruded precursor mineral ions, mainly calcium and bicarbonate, to form complex organo-mineral composites of well-defined microstructures For several rea-sons related to its remarkable mechanical properties and to its high value
in jewelry, nacre is by far the most studied molluscan shell microstructure and constitutes a key model in biomineralization research To understand the molecular mechanism that controls the formation of the shell nacreous layer, we have investigated the biochemistry of Nautilin-63, one of the main nacre matrix proteins of the cephalopod Nautilus macromphalus After purification of Nautilin-63 by preparative electrophoresis, we demon-strate that this soluble protein is glycine-aspartate-rich, that it is highly gly-cosylated, that its sugar moieties are acidic, and that it is able to bind chitin in vitro Interestingly, Nautilin-63 strongly interacts with the mor-phology of CaCO3crystals precipitated in vitro but, unexpectedly, it exhib-its an extremely weak ability to inhibit in vitro the precipitation of CaCO3 The partial resolution of its amino acid sequence by de novo sequencing of its tryptic peptides indicates that Nautilin-63 exhibits short collagenous-like domains Owing to specific polyclonal antibodies raised against the purified protein, Nautilin-63 was immunolocalized mainly in the intertabular nacre matrix In conclusion, Nautilin-63 exhibits ‘hybrid’ biochemical properties that are found both in the soluble and insoluble proteins, rendering it diffi-cult to classify according to the standard view on nacre proteins
Database The protein sequences of N63 appear on the UniProt Knowledgebase under accession number
P86702
Abbreviations
AIM, acid-insoluble matrix; ASM, acid-soluble matrix; CBB, Coomassie brillant blue; EST, expressed sequence tag; HPAE-PAD, high performance anion exchange-pulsed amperometric detection; LSB, Laemmli sample buffer; N63, Nautilin-63; SEM, scanning electron microscopy; TFMS, trifluoromethanesulfonic acid.
Trang 2The calcified shells, which protect the mollusc soft
tis-sues, comprise layered structures that are produced
extracellularly by the calcifying epithelium of the
man-tle The shell layers are composites of calcium
carbon-ate crystals, which are densely packed together with an
array of biomacromolecules that form a 3D
frame-work Although the organic shell matrix (comprising
mainly proteins, glycoproteins and polysaccharides)
represents only a very small part of the CaCO3 shell
weight (between 1% and 5% for the nacreous layer), it
is now well known to be essential for the control of
the biomineral formation [1] In particular, it is
assumed to interact in different ways with the mineral
phase at the nano- to microscale Indeed, the organic
shell matrix is considered to create a suitable
environ-ment for mineralization to occur [1–3], to promote or
inhibit crystal nucleation [4], to select calcium
carbon-ate polymorph (aragonite and⁄ or calcite) [5], to allow
crystals to grow in privileged directions [3] and to
con-tribute to the spatial arrangement of crystals to form
well-defined microstructures [2,3] At the atomic scale,
this matrix slightly modifies the crystal lattice
parame-ters, although this effect is poorly understood [6]
Because of its admirable biomechanical properties [7],
its use in pearl industry and, finally, its potential use in
dentistry and bone surgery [8,9], nacre is by far the most
studied nonhuman organo-mineral biocomposite It has
a remarkable regular lamellar structure consisting of
uniformly thick layers (approximately 0.5 lm) of
tablet-like aragonite crystals separated by interlamellar layers
of organic matrix This apparent simple geometry
facili-tates various structural investigations from micro- to
nanoscales [10–12] Nacre, or its precursor, ‘foliated
ara-gonite’, appeared early in mollusc history, somewhere in
the Cambrian [13] It constitutes the inner layer of
sev-eral extant mollusc shells, including that of bivalves,
gastropods, cephalopods and monoplacophorans There
are, however, structural differences between
cephalo-pod, gastropod and bivalve nacres Although the bivalve
exhibits a characteristic ‘brick-wall’ nacre
microstruc-ture, those of cephalopods and gastropods are a
contin-uous superimposition of tablets forming characteristic
columnar microstructures Observations of growing
nacre show that each tablet nucleates at a specific
loca-tion on the matrix surface [14] Today, the general
con-sensus is that nacre tablets grow from their center and
expand laterally until reaching the confluence with
neighboring tablets [10] Histochemical observations of
Nautilusnacre [15,16] indicate a concentric distribution
of reactive groups, similar to carboxylates or sulfates,
from the center to the periphery of each single tablet
A recent ultrastructural study has shown that nacre tab-lets are individually coated by a 5 nm thick layer of amorphous calcium carbonate [17] Atomic force microscopy studies by Rousseau et al [18] have shown that each tablet is constituted of nanograins encapsu-lated in a continuous network of an organic intracrystal-line phase Summarizing the different recent advances
on molluscan nacre, Addadi et al [19] have proposed a coherent and dynamic model for nacre formation, as described below
The organic matrix constitutes the framework in which nacre tablets form The major constituents of the matrix are the polysaccharide b-chitin, together with a relatively complex assemblage of hydrophobic and hydrophilic proteins These macromolecules, which control the crystal deposition and microstructure self-assembly, are finally occluded either between the super-imposed parallel lamellae (‘interlamellar matrix’), at the boundary of adjacent mature nacre tablets (‘intertabu-lar matrix’), or within the crystallites (‘intracrystalline matrix’) First, the interlamellar matrix is assumed to
be predominantly constituted of b-chitin fibrils that are aligned with the a-axis of the growing aragonitic tablets [20], suggesting that they can be, directly or indirectly, implicated in the control of the crystal orientation [21]
On the other hand, the b- and c-axes of the nacre tablets are oriented in parallel to the growing front in bivalves, whereas this is not the case in gastropods [22] These data suggest that chitin is therefore either not fulfilling this role in nacre formation or that not all nacres are constructed in the same way, as recently sug-gested by Jackson et al [23] However, structuring the interface between the formed mineral front and the secreting mantle tissues, these different matrices are considered to precede immediately the new minerali-zation [19] Second, the hydrophobic matrix, which contains silk-like proteins [2], is rich in Gly and Ala, or Gly alone, constituting one of the major protein frac-tions of the matrix, which can be extracted by decalcifi-cation of the nacre These hydrophobic silk-like proteins are considered to form a hydrogel phase, supersaturated in calcium ions, between the chitin sheets In this gel, the nacre tablets nucleate and grow [11] During their growth, they push aside and com-press the silk-like protein gel When adjacent tablets come to confluence, the gel polymerizes and remains
‘sandwiched’, forming the intertabular matrix [24] Third, acidic hydrophilic proteins containing carboxyl-ate or sulfcarboxyl-ate reactive groups [25] are dispersed in the gel; they are considered to act as nucleating centers for each tablet; at the same time, they constitute a tenuous
Trang 3organic lace inside which mineral nanograins (initially
amorphous) self-organize, orient and crystallize in a
coordinated manner Once formed, each tablet contains
this intracrystalline matrix
b-chitin, the silk-like proteins and the acidic proteins
are considered to be the three major components of
the nacre organic matrix However, several other nacre
proteins have been identified and characterized over
the past 16 years; a review on mollusc shell proteins is
provided by Marin et al [26] These proteins, which
exhibit diverse putative functionalities according to
their sequences, are not taken in account in the model
Furthermore, from the dozen primary structures of
nacre proteins that have been described so far, most of
them correspond to proteins of the pearl oyster or the
abalone models None of them were described from
the cephalopod nacre
Although the Nautilus nacre has already been the
focus of several ultrastructural [15,16,27–29] or
bio-chemical investigations of the bulk shell matrix [30–32],
only a few studies have dealt with the detailed
charac-terization of its shell proteins or their amino acid
sequence characterization [33,34] Because one of the
keys to elucidating the molecular mechanisms of
bio-mineralization depends on a detailed characterization
of matrix proteins, as well as on the understanding of
their functions, we chose to focus on Nautilin-63
(N63), a major protein of the nacre of the cephalopod
Nautilus macromphalus N63 comprises an acid-soluble
acidic shell matrix glycoprotein, which is specific to the
nacreous layer
Results
N63 purification by preparative electrophoresis
Because N63 was found to be one of the main proteins
of the nacre acid-soluble matrix (ASM) [34], it was
investigated further In our previous study, using a 2D gel, we determined that N63 corresponded to a single acidic protein, and not a mixture of proteins of the same molecular weight but with different isoelectric points: indeed, this protein migrated as a single acidic spot The fractionation of the nacre ASM preparative electrophoresis resulted in the effective one-step purifi-cation of N63 The purity of the N63 extract was checked by monodimensional gel electrophoresis with silver nitrate staining (Fig 1A)
FTIR Figure 1B shows the FTIR profile of N63 and of the nacre ASM Both samples exhibit characteristic bands
of proteinaceous and⁄ or glycoproteinaceous compo-nents [35]: the thick bands around 3270 cm)1 are attributed to the -OH and the amide A groups (N-H bonds), the two small bands at 2915 and 2850 cm)1 were assigned to the C-H bonds, and the two notewor-thy bands near 1640 and 1530 cm)1 were ascribed to the amide I (C=O bond) and the amide II (C-N bond) groups, respectively, which are commonly associated with proteins Carboxylate (COO)) and sulfate (SO4 )) absorption bands are also present in both samples, around 1420 and 1235 cm)1 We note that N63 exhib-its a remarkably strong carbohydrate absorption band (C-O bond) around 1060 cm)1 These observations suggest that N63 is an acidic glycoprotein
Amino acid composition of N63 The purified N63 was analyzed for its amino acid com-position and was compared with those of the nacre ASM, which was obtained previously for the same spe-cies [34] (Table 1) The six dominant amino acid resi-dues are Asx (18%), Gly (17%), Thr (11%), Ala (9%), Glx (8%) and Pro (8%) By comparison with the
Fig 1 Purification and characterization of
N63 (A) 12% SDS ⁄ PAGE of ASM (and of
N63) after its purification by preparative
electrophoresis The gel was stained with
silver nitrate The apparent molecular
weights of the molecular markers (MM) are
indicated on the left (B) Infrared spectra of
the ASM (gray line) and the purified N63
(black line).
Trang 4amino acid composition of the bulk matrix, N63 is
strongly enriched in Thr and Pro residues but depleted
in Gly and Asx residues A search based on the
simi-larity of amino acid composition (AACompIdent:
http://expasy.org/tools/aacomp/) did not produce any
significant hits
Monosaccharide composition of N63
The purified N63 was analyzed for monosaccharide
composition, which was subsequently compared with
that of the whole ASM [34] (Table 2) We note that,
similar to ASM, N63 contains a high amount of
mono-saccharide fraction (total of 275 ngÆlg)1) The five
dominant monosaccharides are glucose (22%),
galac-tose (17%), glucosamine (17%), glucuronic acid (15%)
and galactosamine (12%) By comparison with the
monosaccharide composition of the bulk matrix, N63
appears strongly enriched in glucosamine and
galactos-amine but also strongly depleted in glucose
Interest-ingly, for both samples, we noted an unknown peak on
high performance anion exchange-pulsed amperometric
detection (HPAE-PAD) chromatograms, which eluted
in the ‘acidic monosaccharides’ area, near the expected
galacturonic acid peak [36,37] The identity of this peak
needs to be investigated further
Chemical deglycosylation of N63
In a previous study [34], the periodic acid–Schiff and Alcian blue staining on SDS⁄ PAGE suggested that N63 is an acidic glycoprotein To confirm this finding, the nacre ASM was chemically deglycosylated with tri-fluoromethanesulfonic acid (TFMS) at 0C The ASM and the deglycosylated-ASM were compared on SDS⁄ PAGE gels with double Coomassie brillant blue (CBB)⁄ silver and Alcian blue staining (Fig 2A) This
Table 1 Composition of the nacre ASM and purified N63: amino
acid composition Data are presented as the molar percentage of
total amino acids for each extract Note that Asx = Asn + Asp and
Glx = Gln + Glu Cysteine residues were quantified after oxidation.
Tryptophan residues were not detected (ND) as a result of the
hydrolysis conditions.
Amino acid
% of total amino acids
Table 2 Composition of the nacre ASM and purified N63: mono-saccharidic composition The composition of neutral sugars is obtained by HPAE-PAD Data are represented as ngÆlg)1 of the total matrix and as a percentage of the total identified carbohydrate compounds ND, not detected.
Monosaccharide
ngÆlg)1of matrix (% of total)
a An unattributed band was observed around the galacturonic acid band.
Fig 2 Glycosylation (A) and chitin-binding (B) characterizations of N63 by SDS ⁄ PAGE (A) 12% SDS ⁄ PAGE of ASM and deglycosylat-ed-ASM (Deg-ASM) stained with silver nitrate + CBB (left) and with Alcian blue (right) (B) Chitin-binding ability of N63 (top) and BSA (down, negative control) on 12% SDS ⁄ PAGE stained with silver nitrate Lane 1, water wash; lane 2, 0.2 M NaCl wash; lane 3, extract with LBS For both proteins, the same volume of solution was loaded on the gel.
Trang 5double staining allowed visualization of most of the
macromolecular compounds of the ASM on the same
gel N63 exhibits an important shift of approximately
10 kDa, which represents a loss of apparent molecular
weight of 14% This shift is primarily the result of the
removal of covalently bound polysaccharides
Interest-ingly, the positive Alcian blue staining, observed for
the N63 glycoprotein, was completely lost after
degly-cosylation Because we used Alcian blue under low pH
conditions, this result confirms that an important part
of the polyanionic properties of N63 is a result of the
acidic glycosyl moieties [38,39]
Chitin-binding capability of N63
Framework proteins of the organic nacre matrix are
hypothesized to interact with chitin [19,20] However,
this property has never been tested previously on this
type of matrix The putative chitin-binding ability of
N63 was examined consequently (Fig 2B) The nacre
ASM was incubated in solution with powdered chitin,
and the insoluble mixture was successively washed with
distilled water, saline and finally with hot denaturing
Laemmli sample buffer (LSB) [40] Each washed
sam-ple was analyzed by SDS⁄ PAGE, stained with silver
nitrate BSA, used as a negative control, was
com-pletely washed out with the successive water and saline
treatments, with no band being detected in the LSB
wash (Fig 2B, bottom, lane 3) By contrast, a minor
part of N63 was desorbed after the water and saline
treatments (Fig 2B, top, lanes 1 and 2) and the drastic
LSB wash was required for complete N63 desorption
from chitin (Fig 2B, top, lane 3) This clearly suggests
that N63 has a strong affinity for this insoluble
poly-saccharide, and thus possesses a true chitin-binding
ability
In vitro inhibition of CaCO3precipitation with
N63
The effect of nacre ASM and purified N63 on the
kinetics of CaCO3 precipitation was determined by
monitoring the pH decrease (Fig 3) In the blank
experiment (without sample), the pH decreased
with-out any time lag (approximately 120 s), corresponding
to the rapid precipitation of calcium carbonate When
samples were present in the solution, we observed a
slight inhibition of CaCO3 precipitation First, the
effect of the nacre matrix started to occur above 1 lg
of the ASM and the delay of the reaction was
dose-dependent At approximately 50 lg of nacre ASM, a
complete inhibition of the precipitation of calcium
car-bonate was recorded The observation of the inhibitory
capacity of this matrix is consistent with previous stud-ies on the organic soluble matrix of nacre from differ-ent molluscs [24,41,42] On the other hand, inhibition experiments performed with N63 demonstrate that it presents a five-fold lesser inhibition capacity than the total ASM At 25 lg, the N63 inhibition curve can be superimposed to the 5 lg curve obtained with the total ASM Taken together, our observations indicate that,
if nacre ASM exhibits a moderate capacity of inhibi-tion of CaCO3 precipitation, this effect is not a result
of N63 because the latter presents only a weak inhibi-tion capacity, despite the fact that it carries sulfated (i.e negatively charged) sugars
Interaction with CaCO3crystals precipitated
in vitro The effect of purified N63 on the precipitation and morphology of calcium carbonate crystals grown
in vitro was investigated by scanning electron micros-copy (SEM) (Fig 4) When no protein is added, crys-tals exhibit the typical rhombohedral habitus of calcite with smooth crystal faces (Fig 4A) In the presence
of an increasing amount of purified N63 (0.1–
50 lgÆmL)1), the crystals produced appear mostly as polycrystalline aggregates with foliation and microsteps
at the corners (Fig 4B–F) At the highest concentra-tion (‡ 10 lgÆmL)1), the precipitated CaCO3 crystals exhibit specific linear grooves at the edges of the poly-crystals (Fig 4E,F) FTIR analysis of the calcium carbonate polymorph confirmed that these crystals were only made of calcite Unexpectedly, at the highest concentrations of N63, no inhibition of crystal forma-tion occurs
Fig 3 In vitro inhibition of CaCO3precipitation by nacre ASM and N63 The effects of different concentrations (1–50 lg) of purified N63 and whole ASM were monitored on the pH decrease induced
by the in vitro precipitation of CaCO3in a CaCl2⁄ NaHCO 3 solution [4] Blank tests were performed in the absence of protein.
Trang 6These results indicate that N63 interacts obviously
with the precipitation of CaCO3 because it induces
drastic changes in the crystals morphology but does
not (or very slightly) inhibit their formation, with the
combination of these two effects in a shell matrix
pro-tein being rather unusual
De novo sequencing of N63
Purified N63 was digested with trypsin before analysis
by MS⁄ MS Peptide digests were loaded on a nanoLC
column and analyzed by nanoESI-qQ-TOF Because
no genomic, nor transcriptomic data are available for
Nautilus, the most intensive MS⁄ MS peaks were
manu-ally interpreted (de novo sequencing) after considering
the complexity of the spectra and the numerous ion
combinations For N63, the sequence of 27 peptides,
with lengths comprising between eight and 20 amino
acids, was determined by de novo interpretation of
their respective MS⁄ MS spectra (Table 3) The partial
protein sequences of N63 appear in the UniProt
Knowledgebase under accession number P86702
Among them, two peptides (GPAAVVGVL⁄ IGK and
SFDSWL⁄ ITK) present a sequence similar to two
oth-ers previously obtained by de novo sequencing of the
whole nacre ASM [34]
The MS⁄ MS deduced sequences were individually
submitted to a blastp search against Swiss-Prot nrdb
using the EXPASY website (http:⁄ ⁄ expasy.org ⁄ ), and
to a tblastn search against GenBank and the data-base for expressed sequence tags (EST) (dbEST) using the NCBI online tool (http://www.ncbi.nlm.nih.gov) (Table 3, central and right columns, respectively) Unexpectedly, we did not find any homology with already known mollusc shell proteins, and most of the observed hits concern only one unique peptide and are related to unknown putative proteins or to proteins that are not expected to be components of the mollusc shell matrices These observations should be confirmed
in future works by the use of complementary tech-niques
On the other hand, when the peptide
GPAAVVGV-L⁄ IGK was previously used for blast against an in-house database of mollusk shell matrix [34], we noted that it presents partial similarities with the sequence GPAAVPVAAG of mucoperlin, a shell matrix protein from the Pinna nobilis nacreous layer [24]
MSBLASTsearch The de novo-generated sequences of N63 were also sub-mitted to a MS blast database search, which enables the identification of the proteins or their assignment to
a family of homologous proteins, considering all their internal peptide sequences together (Table 4) Sequence similarities were observed for several peptides of N63
F E
D
Fig 4 SEM micrographs of synthetic calcium carbonate crystals grown in vitro in the presence of N63 at increasing concentrations (lgÆmL)1) (A) Negative control without N63; (B) 0.1 lgÆmL)1; (C) 1 lgÆmL)1; (D) 5 lgÆmL)1; (E) 10 lgÆmL)1; (F) 50 lgÆmL)1 Scale bars are
60, 20 and 2 lm, on the left, center and right, respectively.
Trang 7with vertebrate collagen XI, cuticle collagen from
nem-atoda, and the spidroin-like protein of an arthropod
This sequence similarity with collagen-like proteins is
partially supported by the fact that some peptides pres-ent Gxy repeats Interestingly, we did not find any homology with already known mollusc shell proteins
Table 3 MS ⁄ MS derived sequences of N63 trypsic peptides BLAST search results against Swiss-Prot and mollusc-restricted EST databases (taxid:6447) are presented in the central and right columns, respectively The partial protein sequences of N63 appear in the UniProt Knowl-edgebase under accession number P86702 Alignment results of the BLAST searches are indicated on the peptide sequences: identical and synonymous amino acid positions are underlined or shown in bold for the BLAST search against Swiss-Prot and mollusc ESTs, respectively Respective scores (similar amino acids ⁄ total amino acids) of both BLAST searches are indicated after the name of the matching proteins The
MS ⁄ MS technique does not allowed distinction between L and I residues, which exhibit identical masses.
Expasy BLAST against Swiss-Prot NCBI TBLASTN against mollusc ESTs [sp.]
a
Methylase of Nitrosococcus oceani [Q3JDX6].
Table 4 Results of the MS BLAST search for the identification of N63 using de novo sequenced peptides.
MS BLAST
Identification
Total score
Best score
Peptide matches
Calculated
Swiss-Prot number
Trang 8Immunolocalization of N63
Polyclonal antibodies raised against purified N63 were
produced in rabbit We checked the specificity of these
antibodies by western blotting against both the whole
acid-insoluble matrix (AIM) and ASM extracts
(Fig 5A,B) For these experiments, no immunological
signal was observed for negative controls performed
with pre-immune sera (data not shown) The antibody
raised against the purified N63 showed a specific
immunoreactivity for the 63 kDa band of the nacre
ASM corresponding to N63 This observation
demon-strates that the antibody recognizes exclusively N63
specific epitopes and that N63 protein is exclusively
present in the ASM
To localize N63 directly in both the prismatic and the nacreous layers of the shell of N macromphalus, the im-munogold technique was applied on shell cross-sections, followed by observation by SEM, as described previ-ously [43], using the antibodies raised against the puri-fied N63 (Fig 5C–F) Although a low background was observed for negative control performed with pre-immune serum (Fig 5C), the N63 antibodies exhibit a clear and specific signal on shell nacre (Fig 5E,F), whereas very little signal is observed with the prismatic layer (Fig 5D), testifying that N63 is specific of the nacreous layer Immunolocalization on nacre cross-sec-tions (Fig 5E,F) revealed that N63 is largely distrib-uted inside nacre tablets, and also in the inter-tablet matrix that separates nacre tablets of the same layer
Discussion
In a previous study, we characterized the whole acid soluble matrix extracted from the nacre of the cephalo-pod N macromphalus [34] In particular, we obtained approximately 40 short sequences of different shell proteins, both extracted from the acid-soluble and from the acid-insoluble matrices In the present study,
we focus on one shell protein, which we named Nauti-lin-63 (N63), according to its apparent molecular weight on a 1D electrophoresis gel
N63 is an acidic shell matrix glycoprotein, which is unambiguously specific to the nacreous layer of N mac-romphalus N63 belongs to the acetic acid-soluble frac-tion, and to this fraction exclusively, because no signal was detected on western blotting (Fig 6) and none of its sequenced peptides were found in the acetic acid-insoluble fraction in our previous study [34] In vitro, N63 binds chitin, interacts with the shape of newly-grown calcite crystals but, apparently, has a very limited effect on the precipitation of calcium carbonate Although its glucose-rich glycosyl moieties exhibits sul-fated groups, N63 does not bind calcium ions [34] From the 27 peptidic sequences (of eight to 20 residues
in length) obtained by de novo sequencing, only seven of them exhibit similarities with other putative molluscan proteins, which are not related to calcification One pep-tide is partly similar to a short domain of mucoperlin, a bivalve shell protein At least five obtained peptides have a collagen signature, characterized by Gxy triplets Such a signature has already been found in a short domain of lustrin-A, a nacre protein of the abalone Haliotis rufescens [44] By the immunogold technique, N63 appears to be particularly concentrated inside nacre tablets, as well as between them
The overall composition of all the obtained peptides
of N63, taken together, is enriched in Gly and Pro
F E
Fig 5 Immunodetection of N63 by western blotting (A–B) and the
immunogold technique (C–F) (A) 12% SDS ⁄ PAGE of nacre AIM
and ASM, stained with silver nitrate (B) The nacre AIM and ASM
were tested by western blotting and incubated with the polyclonal
antibodies raised against purified N63 (C) SEM micrographs for the
immunogold negative control performed on nacre without anti-N63
specific sera (D–F) SEM micrographs of the immunogold technique
with anti-N63 specific sera on prismatic (D) and nacreous (E,F) shell
layers Scale bars = 2 lm.
Trang 9residues, whereas the overall amino acid composition
of the isolated protein, after its purification, is enriched
in Asx, Gly and Thr residues, which constitutes a
typi-cal shell protein signature This apparent discrepancy
between de novo sequencing data and the amino acid
composition may be explained in different ways: first,
the de novo sequences represent only approximately
one half of the protein, thus giving a partial picture of
its primary structure Second, the de novo sequencing
by MS presents technical limitations: on the one hand,
it is ineffective on protein domains, which lack
appro-priate cleavage sites (Arg and Lys in the case of trypsic
digestion); on the other hand, the peptides resulting
from the digestion are not detectable, if not ionized
Thus, the technique might introduce a bias in the
rep-resentation of the analyzed peptides for the complete
N63 sequence In the present case, at the very least, it
is likely that some of the nonsequenced peptides
con-stituting the unknown part of N63 are enriched in Asx
and Thr residues
What might be the role of N63 in the formation of
nacre? Our biochemical data, when combined, give an
unusual mosaic picture, which does not simply fit into
the general structural framework given few years ago
by Nudelman et al [16] and Addadi et al [19], and
also recalled in the Introduction of the present study
First, several of the peptides determined by de novo
sequencing are hydrophobic, and might suggest that
N63 is part of the hydrogel where nascent nacre tablets
grow By contrast, N63 is a soluble and acidic protein,
and is present not only around nacre tablets, but also
inside the tablets Second, because N63 contains
sul-fated polysaccharides, it is tempting to assume that
this protein is part of the nacre-nucleating complex
(i.e the central domain observed for each nacre
tab-lets), whatever it is, either a central spot, as suggested
by Crenshaw and Ristedt [15] or a central ring as
reported by Nudelman et al [16] Once again, our
immunogold labeling data do not support this
hypoth-esis because we did not observe spots in the centre of
nacre tablets, but did observe the peripheral
distribu-tion of N63 around the tablets Finally, we clearly
demonstrate that N63 binds chitin in vitro This may
suggest that this protein is able to form
macromolecu-lar complexes with chitin, which, in other words,
means that it should be co-localized with chitin at the
interlamellar interface Obviously, our immunogold
labeling data do not reveal such a location
The fact that N63 can, at the same time, strongly
interact with the shape of CaCO3 crystals precipitated
in vitro without (or slightly) inhibiting their formation
is puzzling Indeed, we observed that, for many
mol-lusc shell proteins, the ability to interact with CaCO3
crystals and the capacity to inhibit the precipitation of CaCO3 were often associated, as observed for P95 or for caspartin [24,26] The present case constitutes the first report indicating that these two properties can be disconnected in a calcifying matrix protein This unu-sual property could be somehow related to the fact that N63 does not bind calcium ions in solution as previously noted [34]
Taken together, these findings suggest that N63 exhibits ‘hybrid’ biochemical properties, some of which are usually found in framework matrix proteins (i.e hydrophobicity, intertabular localization, ‘collagen sig-nature’, absence of calcium-binding, weak ability to inhibit CaCO3 precipitation), whereas others are com-monly met in the soluble matrix components (i.e solu-bility, acidity on 2D gels, enrichment in Asx residues, capacity to interact with CaCO3 crystals) This clearly suggests that the models established in recent years for Nautilus nacre [16,19] must, in some ways, be refined,
by taking in account proteins of ‘intermediate’ bio-chemical properties, such as N63 In the absence of a clear correlation between the structure of N63, its bio-chemical properties and a defined function, we suggest that N63 is a multifunctional protein that plays a key role in binding chitin, and thus in participating in the structuring of the organic framework, at the same time
as finely interacting with the mineral phase It is possi-ble that these two functions are displayed sequentially (chitin-binding, followed by mineral interaction)
We are fully aware that trying to decipher the func-tion of one single component of the nacre matrix will not provide an explanation of the whole process of nacre fabrication As highlighted in a previous study [16],
‘none of the components of the organic matrix functions
in isolation The organic matrix is a structural entity in which the assembly of all components is essential for the correct regulation of crystal nucleation, growth, mor-phology, and polymorph type’ However, we consider that the precise characterization of separate nacre mac-romolecular constituents will provide the complete bio-chemical framework required to precisely analyze the growth of nacre tablets This framework constitutes the prerequisite for studying protein–protein and protein– polysaccharide interactions, and for any attempts aiming to understand the supramolecular chemistry that contributes to the emergence of nacre microstructure
Experimental procedures
Shell preparation and nacre matrix extraction Fresh shells of the cephalopod N macromphalus,
150⁄ 200 mm in length, were collected on the coast of New
Trang 10Caledonia (Pacific) The external prismatic layer was
removed by abrasion under cold water The shells were
mechanically crushed and fragments of the siphon were
removed The nacre fragments were immersed in 1% (v⁄ v)
NaOCl for 24 h to remove superficial contaminants, and
then thoroughly rinsed with water All of the subsequent
extraction procedure was performed at 4C The nacre
powder (< 200 lm) was decalcified overnight in cold dilute
acetic acid (5%, v⁄ v) added by an automatic titrimeter
(Titronic Universal; Schott Instruments GmbH, Mainz,
Germany), until pH 4 was obtained The solution was
cen-trifuged at 3900 g for 30 min The pellet, corresponding to
the AIM, was rinsed six times with MilliQ water (Millipore
Corp., Billerica, MA, USA) and finally freeze-dried The
supernatant comprising the ASM was filtered (5 lm) before
being concentrated with an Amicon ultrafiltration system
on a Millipore membrane (YM10; 10 kDa cut-off) The
concentrated solution (approximately 5–10 mL) was
exten-sively dialyzed against MilliQ water (3 days, several water
changes) before being freeze-dried and weighed
SDS⁄ PAGE and gel staining procedures
The separation of matrix components was performed under
denaturing conditions by monodimensional sodium dodecyl
sulfate polyacrylamide gel electrophoresis (SDS⁄ PAGE)
containing 12% polyacrylamide (Mini-protean 3; Bio-Rad,
Hercules, CA, USA) Protein concentration of the ASM
was estimated by using the BCA-200 Protein Assay kit
(Pierce, Rockford, IL, USA) The nacre matrices were
directly suspended in LSB containing b-mercaptoethanol
and heat-denatured [40] One milligram of AIM was
sus-pended in LBS, heat-denatured (10 min at 100C) and then
centrifuged at 3900 g for 30 s Twenty microlitres of the
supernatant containing the Laemmli-soluble AIM were
loaded onto gel Fifty micrograms of ASM were loaded in
each well Because the classical CBB staining is often
ineffi-cient at revealing all the proteins associated with calcified
tissues, we chose to visualize proteins on the gel with both
silver nitrate [45] and CBB R-250 Glycosylation of ASM
macromolecules was studied qualitatively on denaturing
mini-gels by Alcian Blue 8GX staining [39] at pH 1 for the
detection of sulfated sugars [38]
Deglycosylation with TFMS
Chemical deglycosylation of 5 mg of ASM was performed
with 1.5 mL of TFMS⁄ anisole (2 : 1, v ⁄ v) for 3 h, under a
nitrogen atmosphere, with constant stirring [46] The
tem-perature was maintained at 0C, to preclude peptidic bond
hydrolysis After neutralization with 2 mL of 50% cold
pyridine, the aqueous phase was extracted twice with
diethyl ether and then extensively dialyzed against water
(5 days) before being lyophilized Fetuin was treated
simi-larly and used as a positive control All the deglycosylated
extracts were analyzed on monodimensional SDS⁄ PAGE followed by silver nitrate and Alcian blue staining
Chitin-binding ability
A chitin-binding assay was performed in solution as described previously [47], with some modifications One milligram of nacre ASM and 500 lg of BSA (used as negative control) were dissolved in 200 lL of water and incubated with 10 mg of chitin (C9752; Sigma-Aldrich,
St Louis, MO, USA) for 2 h at 25C under constant stir-ring Samples were centrifuged (13 000 g for 5 min) and the supernatants were taken away and preserved The residues were then rinsed three times with 500 lL of distilled water, before washing with 300 lL of 0.2 m NaCl and centrifuga-tion The precipitates were boiled with LSB for 10 min at
99C Each supernatant and washing solution was ana-lyzed on SDS⁄ PAGE under denaturing conditions After electrophoresis, the gels were stained with silver nitrate [45]
Purification of N63 by preparative SDS⁄ PAGE The nacre ASM was fractionated on a preparative 12% polyacrylamide gel under denaturing conditions as described previously [48] Eighty fractions (5 mL each) were eluted from the preparative gel Aliquots of the fractions were tested by SDS⁄ PAGE with silver nitrate staining Fractions containing the N63 protein were pooled, concentrated, thor-oughly dialyzed against MilliQ water and freeze-dried
Infrared analysis of N63 Infrared spectra were directly recorded on lyophilized sam-ples of nacre ASM and of purified N63 at a 2 cm)1resolution
on a FTIR spectrometer (Vector 22; Bruker, Ettlingen, Ger-many) equipped with a Specac Golden Gate ATR device in the wave number range 4000–500 cm)1 For each extract, we obtained several spectra with a high reproducibility
Amino acid composition of N63 The amino acid composition of the purified N63 was determined by Eurosequence (Groningen, The Nether-lands) Freeze-dried samples were hydrolyzed with 5.7 m HCl in the gas phase for 1.5 h at 150C The resulting hydrolysate was analyzed on an HP 1090 Aminoquant (Hewlett-Packard, Palo Alto, CA, USA) [49] by an auto-mated two-step precolumn derivatization with O-phthalal-dehyde for primary and N-(9-fluorenyl)methoxycarbonyl for secondary amino acids Cysteine residues were quanti-fied after oxidation The hydrolysis procedure does not allow the quantification of tryptophan residues Experi-mentally determined amino acid values may deviate up to approximately 10% For comparison, the amino acid