Protein Expression and Purification 23, 175–179 (2001)doi:10.1006/prep.2001.1487, available online at http://www.idealibrary.com on

Large-Scale Fractionation of Molluscan Shell Matrix

Frederic Marin,*,1 Lucilia Pereira,† and Peter Westbroek**Gaia Science Center, LIC, Gorlaeus Laboratoria, Leiden University Einsteinweg 55, P.O. Box 9502, 2300 RA Leiden,The Netherlands; and †Laboratoire de Physiologie Generale et Comparee, UMR CNRS 8572,

Museum National d’Histoire Naturelle, 7 Rue Cuvier, 75231 P

Received April 9, 2001, and in revised form May 30, 2001; published

The numerous proteins occluded within the mol-luscan shell play a key role in the control of the miner-alization process. Although extensively studied, theseproteins are still poorly known, mainly because theyare difficult to fractionate. In the present paper, wepresent, for the first time, a simple combined strategyfor separating successfully large amounts of molluscanshell proteins. Since shell proteins do not absorb at 280nm, our approach is based on the “blind” separationof these proteins by a preparative denaturing electro-phoresis. They are subsequently detected on dot-blotwith polyclonal antibodies raised against the unfrac-tionated soluble matrix. In the present case, thisapproach allows one to collect enough purified pro-teins to obtain amino-acid composition as well as N-

terminal sequences, and to perform in vitro tests and

soluble matrix gave excellent staining. We decided to

glycosylation studies. Furthermore, this method per-mits one to raise polyclonal antibodies against the iso-lated proteins. q 2001 Academic Press

When a mollusk builds its shell, the calcifying epithe-lium of its mantle secretes a complex amalgamate ofproteins, glycoproteins, and polysaccharides, collec-tively defined as the skeletal matrix. The matrix regu-lates the nucleation, orientation and growth of calcium

carbonate crystals (1, 2). By terminating crystalliza-tion, it also determines the definitive shape of the crys-tal elements of which the shell is composed (3). Becauseof these functions, shell matrices have potential appli-cations in bio and nanotechnology and bone medicine,

1 To whom correspondence should be addressed at present address:IsoTis, bv., Prof. Bronkhorstlaan, 10-D, 3723 MB Bilthoven, The Neth-erlands. Fax: 00 31 30 228 02 55. E-mail: frederic.marin@isotis.com.

1046-5928/01 $35.00Copyright q 2001 by Academic PressAll rights of reproduction in any form reserved.

aris cedex 05, France

online August 20, 2001

and it is for that reason that these materials presentlyreceive a great deal of attention (4, 5).

Progress in this field depends on isolation of the con-stituent macromolecules, followed by structural andfunctional analysis. Part of the interactive material canbe released from the shells by EDTA or acetic acid ex-traction. However, because they are complex mixturesof polydisperse, cross-linked macromolecules con-taining posttranslationally modified proteins, these ex-tracts are extremely difficult to fractionate. In particu-lar, they usually resist chromatographical separation(6).

In numerous cases, polyacrylamide gel electrophore-sis proved to be more successful in resolving the solubi-lized shell matrices in discrete fractions (7–12). In par-ticular, we observed that the use of fresh juvenile shells,acetic acid extraction, and electrophoresis on mini-gelsunder denaturing conditions often gave fast and satis-factory results (13). However, the method has two draw-backs: shell proteins do not stain well with silver, andwhen they do, the staining varies from gel to gel. Fur-thermore, the quantities of matrix used in mini-gelsare low. Thus, the method does not permit subsequentin vitro tests with the fractionated proteins.

Recently, we have observed that staining of blottedgels with antisera raised against the unfractionated

scale up the electrophoresis in combination with theimmunological detection of the eluted proteins. Thepresent paper shows the results obtained with shellproteins of the bivalve Pinna nobilis.

MATERIALS AND METHODS

Shell cleansing. Shells of the Mediterranean bi-valve P. nobilis were kindly provided by the CERAM(Centre d’Etudes des Ressources Animales Marines,

175

developed perpendicularly to the surface of the shell. The internal

teins associated with biominerals (18–20). The negativestaining was particularly pronounced in the second

nacreous layer, which looks homogeneous on the picture, is actuallymade of tiny crystallites of aragonite. A layer of insoluble organicmatrix (thickness around 50 mm) separates the two layers. In thepresent study, the proteinaceous matrix was extracted from the pris-matic layer only.

Marseille; Prof. N. Vicente). They were scrupulouslybrushed and soaked in dilute sodium hypochlorite solu-tion (0.2 g active chlorine/100 ml) for several hours.In our experiments, three different batches of juvenileshells were used. The shells exhibited the classical bi-layered texture, with a calcitic outer prismatic layerand an aragonitic nacreous one (Fig. 1). These layerswere separated mechanically, and the prismatic layerwas used for subsequent tests.

Fragments of cleaned calcitic prisms were crushedunder liquid nitrogen. The powder obtained was treatedwith dilute sodium hypochlorite prior to extensive rins-ing with milli-Q water. This operation led to the com-plete destruction of the organic insoluble sheaths thatsurround the calcitic prisms, without degrading the in-tracrystalline soluble matrix. The material was thendried at 378C.

Matrix extraction. For each batch, between 3 and25 g of the prisms powder was suspended in milli-Qwater in which 20% acetic acid was progressively addeduntil pH 4. Decalcification was performed overnight at48C. The solution was centrifuged 15 min at 5500g. Theinsoluble organic matrix was discarded, whereas thesolubilized fraction was extensively filtered at 48C ona YM10 Amicon membrane (cutoff 10 kDa) and concen-trated to a volume of less than 4 ml.

Polyclonal antibodies and purification of shell pro-teins. The concentrated prisms matrix solution wasdenatured in Laemmli sample buffer (14) at 1008C. Asmall aliquot was tested on 10–12% polyacrylamidemini-SDS–PAGE discontinuous gels (Bio-Rad Protean

AND WESTBROEK

II). One lane of the gel was stained with silver (15). Thesecond lane was blotted (16) onto PVDF2 Immobilon-Pmembrane (Millipore) with the Mini Trans-Blot module(Bio-Rad). The membrane was subsequently incubatedwith polyclonal antibodies used in a previous study (9).These antibodies were raised in two white rabbitsagainst the unfractionated matrix of prismatic and na-creous shell layers of P. nobilis, respectively. After incu-bation with the secondary antibody (GAR/PeroxidaseConjugate, Sigma A6154) and subsequent extensiverinsing, the blot was revealed with the chemolumines-cent substrate Luminol (Sigma A 4685) (17).

The main part of the matrix was fractionated on adiscontinuous preparative gel electrophoresis (Bio-Rad,model 491 Prep Cell) on a 12% polyacrylamide gel. Thegel was run at 250-V. The proteins were eluted at aflow rate of 0.5 ml/min, in 80 fractions of 5 ml each.Because none of the 80 fractions absorbed UV light at280 nm, aliquots of each tube (100 ml) were tested asdot-blots (Bio-Dot apparatus, Bio-Rad) on PVDF mem-branes, which were subsequently incubated with theantibodies raised against the unfractionated matrixand treated as described above. Following the resultof the dot-blot, the tubes containing identified proteinfractions were pooled, and their contents were reducedin volume by ultrafiltration on a 10-ml Amicon cell

176 MARIN, PEREIRA,

FIG. 1. (A) Shell of Pinna nobilis (photo A. Denis, UPS, Orsay). (B)A cross section, when observed with an optical microscope underincident light, shows the bilayered texture of the shell. The outerlayer is made of long calcitic prisms (diameter from 50 to 100 mm)

(cutoff, 10 kDa). Then, these solutions were extensivelydialyzed against Milli-Q water, at 48C, before being ly-ophilized. To verify their purity, the lyophilisates wereanalyzed further on mini-SDS–PAGE and on Westernblots under the same conditions as above.

RESULTS AND DISCUSSION

Protein fractionation on SDS–PAGE. Figure 2shows the results of silver staining of the prisms matrix,after its extraction with acetic acid. On a classical dena-turing mini-gel, intracrystalline molluscan shell pro-teins usually stain poorly with silver even when thelanes are loaded with 20 micrograms of shell matrix(Fig. 2, lanes 2, 4). In the present case, most of theacetic acid-soluble matrix associated with the prismsmigrated as a smear. However, the main discrete con-stituents were represented by two large bands, about17 and 35 kDa, respectively. Other minor blurred bandswere detected as well. The staining varied considerablybetween the different preparations. In the first twobatches, we observed a negative staining of the prismsmatrix, a fact commonly encountered with acidic pro-

batch preparation, between 17 and 35 kDa. Even thedominant 17-kDa protein could be hardly distinguishedin that batch (Fig. 2, lane 4). In contrast, the same band

2 Abbreviations used: PVDF; polyvinylidene fluoride.

nobilis calcitic prisms. The gels were stained with silver. Three

batches were used. Batches 1 and 2, 10% polyacrylamide gel; batch3, 12% polyacrylamide gels. The positions of the two major proteinsfrom prisms matrix (17 and 35 kDa) are indicated by arrows. Lanes1, 3, and 5, molecular mass standards, from top to bottom: 94, 67,43, 30, 20.1, and 14.4 kDa.

was well stained in the third batch preparation (Fig.2, lane 6). The reason of this variation in staining is

FIG. 3. (A) Western blot of CH3CO2H-soluble shell matrix of Pinna no(lane 2) and anti-nacre matrix antibodies (lane 3). Lane 1, molecular maeluted from preparative SDS–PAGE. The dots were stained with the awith the anti-nacre matrix antibodies. Tubes 6 to 9 and tubes 25 to 28

F SHELL MATRIX 177

the resulting image was more contrasted than the oneobtained after silver staining. The two main proteinbands as well as minor components were clearly visible.In the present case, at least 9 bands were distinguished(Fig. 3A, lane 2). In particular, the 17-kDa band was theshell protein with the lowest molecular weight whichexhibited a strong immunogenicity. The Western blotpattern obtained with the anti-nacre matrix antibodies(Fig. 3A, lane 3) was also significant since the two mainproteins of the prisms extract were the single bands tobe stained.

After fractionation of the whole matrix on the prepar-ative gel, the pattern obtained with the dot blotting ofthe 80 fractions was in good agreement with the patternfound on the Western blot. In particular, the 17-kDaproteins eluted in tubes 6 to 9 and the 35-kDa protein,in tubes 25 to 28 (Fig. 3B and 3C). Note that with thedot-blot stained with anti-prism antibodies, an interme-diary fraction was eluted in tubes 16 to 19. This fraction,seen in lane 2 with an apparent molecular weight of26 kDa, was not collected. The tubes corresponding tothe two main proteins were pooled, ultrafiltered andconcentrated. The extracts, when tested on a mini-gel,gave in each case, a single band (Fig. 4A, lanes 2 and 3).A similar result was obtained with the more sensitiveWestern blot (Fig. 4 B), after incubation with the anti-bodies elicited against the prisms matrix.

FRACTIONATION O

FIG. 2. Mini-SDS–PAGE of CH3CO2H-soluble shell matrix of Pinna

Interest of the method. In the present paper, we iso-not known.lated pure intracrystalline macromolecules from theThe blot treated with polyclonal antibodies raisedshell of the bivalve P. nobilis. Obtaining polyclonal anti-against the unfractionated prisms matrix gave a muchbodies raised against the unfractionated matrix is acleaner pattern (Fig. 3A). Because some macromole-

cules contain more immunogenic epitopes than others, prerequisite for the detection of immunogenic discrete

bilis calcitic prisms, and staining with anti-prisms matrix antibodiesss standards, same as for Fig. 2. (B) Dot-blot of the first 48 fractions

nti-prisms matrix antibodies. (C) Same as B. The dots were stainedwere pooled.

spective function in the calcification process. Finally,

FIG. 4. Mini-SDS–PAGE (A) and Western blot (B) of the two lyo-

philisates from pooled tubes 6 to 9 and 25 to 28. The gel was stainedwith silver, and the Western blot with the anti-prisms matrix antibod-ies. Lane 1, unfractionated CH3CO2H-soluble matrix; lanes 2 and 6,17-kDa protein; lanes 3 and 5, 35-kDa protein; lane 4, molecularmass standards.

fractions on blots. When these antibodies are available,the purification procedure is not time-consuming: thepreparative fractionation can be performed overnight,the dot-blot, tested the following day, and the pooledfractions, ultrafiltered and extensively dialyzed duringthe next days. As described, our method does not permitisolation of all the proteins of molluscan shell matrices,particularly minor or poorly immunogenic ones. How-ever, we consider that the technique used constitutes asignificant improvement in comparison to all publishedworks. A large-scale extraction with 3 times 25 g ofcleaned prisms powder yielded around 180 mg of acetic-acid-soluble intracrystalline matrix. After fractionationof this matrix by preparative gel electrophoresis, wewere able to obtain 8.5 mg of pure 17-kDa protein. Notonly are these quantities more than sufficient to obtainN-terminal sequences and amino acid composition andto study the glycosylation of the protein, but the tech-nique also allows in vitro inhibition tests and in vitrocontrolled crystallization. This is an important aspectif we consider that the identification of the few genesencoding molluscan shell proteins (reviewed in Ref. 13)could not be followed until now by a detailed in vitrostudy of the function of the corresponding protein.

By applying this technique, the extraction of largeamount of pure molluscan shell proteins may not be anobstacle anymore. In addition, the amounts of fraction-ated proteins are sufficient to develop polyclonal anti-bodies in a mouse or in a white rabbit (100 and 400

AND WESTBROEK

mg antigens required, respectively) against some of thepurified protein. Thus, a direct immunolocalization ofeach of these proteins in the shell calcium carbonatecrystallites may also contribute to clarifying their re-

178 MARIN, PEREIRA,

this technique for purifying molluscan shell proteinscan also be applied to other mineralizing systems suchas echinoderm spicules, brachiopod shells, or coral skel-etons.

ACKNOWLEDGMENTS

F.M. thanks Elizabeth de Vrind-de Jong (Gorlaeus Lab., Leiden)for critically reading the manuscript and Alain Denis (UPS, Orsay)for providing a photo of Pinna nobilis. This work was entirely sup-ported by a grant from the Fondation Simone et Cino Del Duca(Paris, France).

REFERENCES

1. Weiner, S., and Traub, W. (1984) Macromolecules in mollusc shellsand their functions in biomineralization. Phil. Trans. R. Soc.London B304, 425–434.

2. Weiner, S., and Addadi, L. (1997) Design strategies in mineralizedbiological materials. J. Mater. Chem. 7 (5), 689–702.

3. Wheeler, H. P., Rusenko, K. W., and Sikes, C. S. (1988) in “Chemi-cal Aspects of Regulation of Mineralization” (Sikes, C. S., andWheeler, A. P., Eds.), pp. 9–13, University of South AlabamaPublication Services, Mobile, Alabama.

4. Mann, S. (2000) The chemistry of form. Angew. Chem. Int. Ed.39, 3392–3406.

5. Westbroek, P., and Marin, F. (1998) A marriage of bone and nacre.Nature 392, 861–862.

6. Cuif, J. P., Gautret, P., and Marin, F. (1991) in “Mechanisms andPhylogeny of Mineralization in Biological Systems” (Suga, S.,and Nakahara, H., Eds.), pp. 391–395, Springer-Verlag, Tokyo.

7. Weiner, S. and Lowenstam, H. A. (1977) Discrete molecularweight components of the organic matrices of mollusc shells. J.Exp. Mar. Biol. Ecol. 30, 45–51.

8. Keith, J., Stockwell, S., Ball, D., Remillard, K., Kaplan, D., Than-nhauser, T., and Sherwood, R. (1992) Comparative analysis ofmacromolecules in mollusc shells. Comp. Biochem. Physiol. 105B (3/4), 487–496.

9. Marin, F., Muyzer, G., and Dauphin, Y. (1994) Electrophoreticand immunological characterization of shell soluble matricesfrom two living Pteriomorphid Bivalves, Pinna nobilis L. andPinctada margaritifera (L.). C.R. Acad. Sci. Paris 318, (ser.II), 1653–1659.

10. Marxen, J. C., and Becker, W. (1997) The organic shell matrixof the freshwater snail Biomphalaria glabrata. Comp. Biochem.Physiol. 118 B (1), 23–33.

11. Weiss, I. M., Kaufmann, S., Mann, K., and Fritz, M. (2000) Purifi-cation and characterization of perlucin and perlustrin, two newproteins from the shell of the mollusc Haliotis laevigata. Biochem.Biophys. Res. Commun. 267, 17–21, doi:10.1006/bbrc.1999.1907.

12. Miyashita, T., Takagi, R., Okushima, M., Nakano, S., Miyamoto,H., Nishikawa, E., and Matsushiro, A. (2000) ComplementaryDNA cloning and characterization of pearlin, a new class of ma-trix protein in the nacreous layer of oyster pearls. Mar. Biotech-nol. 2, 409–418, doi:10.1007/s101260000013.

13. Marin, F., Corstjens, P., De Gaulejac, B., De Vrind De Jong, E.,

FRACTIONATION OF SHELL MATRIX 179

and Westbroek, P. (2000) Mucins and molluscan calcification: 17. Leong, M. M. L., Milstein, C., and Pannell, R. (1986) Luminescent

molecular characterization of mucoperlin, a novel mucin-like pro-tein from the nacreous shell-layer of the fan mussel Pinna nobilis(Bivalvia, Pteriomorphia). J. Biol. Chem. 275, 20,667–20,675,doi:10.1074/jbc.M003006200.

14. Laemmli, U. K. (1970) Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 227, 680–685.

15. Morrissey, J. H. (1981) Silver stain for proteins in polyacrylamidegels: A modified procedure with enhanced uniform sensitivity.Anal. Biochem. 117, 307–310.

16. Burnette, W. N. (1981) ‘Western blotting’: Electrophoretictransfer of proteins from sodium dodecyl sulfate–polyacrylamidegels to unmodified nitrocellulose and radiographic detection withantibody and radioiodinated protein A. Anal. Biochem. 112,195–203.

detection method for immunodot, Western, and Southern blots.J. Histochem. Cytochem. 34, 1645–1650.

18. Tsay, T. G., and Veis, A. (1985) Preparation, detection, and charac-terization of an antibody to rat alpha-phosphophoryn. Biochemis-try 24, 6363–6369.

19. Myers, J. M., Veis, A., Sabsay, B., and Wheeler, A. P. (1996) Amethod for enhancing the sensitivity and stability of stains-allfor phosphoproteins separated in sodium dodecyl sulfate-poly-acrylamide gels. Anal. Biochem. 240, 300–302, doi:10.1006/abio.1996.0360.

20. Goldberg, H. A., and Warner, K. J. (1997) The staining of acidicproteins on polyacrylamide gels: Enhanced sensitivity and stabil-ity of “Stains-All” staining in combination with silver nitrate.Anal. Biochem. 251, 227–233, doi:10.1006/abio.1997.2252.