Structure and composition of the boundary zone between aragonitic crossed lamellar and calcitic prism layers in the shell of Concholepas concholepas (Mollusca, Gastropoda)

Structure and composition of the boundary zone between aragonitic crossed lamellar andcalcitic prism layers in the shell of Concholepas concholepas (Mollusca, Gastropoda)

Yannicke Dauphin,1,a Jean-Pierre Cuif,2 Marine Cotte,3 and Murielle Salome3

1 Micropaleontology, Paris VI University, 75252 Paris, France2 Geosciences, Paris Sud University, 91405 Orsay, France

3 European Synchrotron Radiation Facility, 38043 Grenoble, France

Abstract. Mollusc shells are composed of two or three layers. The main layers are well-stud-ied, but the structural and chemical changes at their boundaries are usually neglected. Amicrostructural, mineralogical, and biochemical study of the boundary between the innercrossed lamellar and outer prismatic layers of the shell of Concholepas concholepas showedthat this boundary is not an abrupt transition. Localized structural and chemical analysesshowed that patches of the inner aragonitic crossed lamellar layer persist within the outercalcitic prismatic layer. Moreover, a thin aragonitic layer with a fibrous structure is visiblebetween the two main layers. A three-step biomineralization process is proposed thatinvolves changes in the chemical and biochemical composition of the last growth incrementsof the calcite prisms. The changes in the secretory process in the mantle cells responsible forthe shell layer succession are irregular and discontinuous.

Additional key words: biomineral, organic matrix, FTIR, XANES

The shells of molluscs are typically composed ofseveral structural layers in the form of calcite oraragonite, the two mineral polymorphs of calciumcarbonate. These layers are biocomposite materials,having organic phases in addition to mineral. Mostspecies of molluscs build shells entirely of layers ofaragonitic crossed lamellar material, with each layerhaving different crystal orientation. As Kobayashi(1969, 1983, 2004) has shown for many species ofbivalves, shells of this type may show complexgeometry and large diversity in the spatial arrange-ment of biomineral crystals. For this reason, moststudies on molluscan shell mineralogy and micro-structure have focused on the simpler, “nacro-pris-matic” type of mollusc shell usually found inPteriomorpha such as “oysters.” Shells of this typehave a calcitic, prismatic outer layer, and an arago-nitic nacreous inner layer. Previous studies haveusually concentrated on only one of the two shelllayers, so that the boundary between calcite andaragonite, or prisms and nacre, has been neglected.

Although the shells of gastropods have receivedless attention than those of bivalves, existing datashow that most species have shells consisting of

multiple layers of crossed lamellar aragonite. Somegastropod taxa, however, show a diversity of shellstructures. Most haliotids, for example, have shellswith two basic structural layers: an inner layer ofaragonite (nacre) and an outer layer of calciticprisms. In some species of haliotids, however, theshell comprises two layers of aragonite. The shell ofone species, Haliotis tuberculata (LINNAEUS 1758), iseven more complex, in that aragonitic and calciticprisms are imbricated (Dauphin et al. 1989).

Another example of a gastropod having an arago-nitic-prismatic shell structure is Concholepas con-cholepas (BRUGUIERE 1789). This species is ofcommercial, pharmacological, and (paleo)environ-mental interest. Concholepas concholepas (alsoknown as “loco”) is a popular seafood in SouthAmerica, and overfishing is a threat for wild stocks.The haemocyanin of C. concholepas has an unusualstructure (De Ioannes et al. 2004) and may havetherapeutic value against cancerous tumours (Molte-do et al. 2006). In addition, the shell composition ofC. concholepas is used as a paleoindicator of El Ninoevents and related palaeoclimatic changes (Guzmanet al. 2009). Because of interest in C. concholepas bymany different stakeholders, commercial aquacultureof this species is now under development in SouthAmerica.

aAuthor for correspondence.E-mail: [email protected]

Invertebrate Biology x(x): 1–12.© 2012, The American Microscopical Society, Inc.DOI: 10.1111/j.1744-7410.2012.00265.x

The shell of C. concholepas consists of two mainlayers: an inner aragonitic crossed lamellar layer andan outer calcitic prismatic layer (Dauphin et al.2003). The microstructure, nanostructure, and chemi-cal contents of both layers have been previouslydescribed, but not the boundary between the twolayers. Among the numerous papers dealing withmicrostructure and composition of mollusc shells,few are dedicated to the structural and compositionalshifts between layers; most often, only individuallayers are described. However, microstructural andchemical data on the border zone between aragoniticnacre and calcitic prisms in the bivalve Pinctadamargaritifera (LINNAEUS 1758), the pearl oyster, haveshown that the changes in mineralogy, structure, andcomposition are neither abrupt nor synchronous(Dauphin et al. 2008).

The shell of C. concholepas offers the opportunityto study another example of a “boundary” betweenthe two shell layers having different mineral compo-sition and microstructure. From a biological pointof view, study of the intermediate zone betweenthese two layers may help identify critical processesassociated with shell secretion at the interfacebetween the shell layers. If an intermediate layerdoes exist between the inner and outer shell layers,it likely will be very thin, making in situ analysesnecessary. Here, we use a combination of micro-structural and microanalytical data to correlate thestructural and compositional changes between theinner and outer shell layers.

Methods

Concholepas concholepas (Caenogastropoda, Mu-ricidae) (Fig. 1A), also called the southern hemi-sphere abalone, is a carnivorous gastropodabundant along the coasts of Chile and southernPeru. Its habitat ranges from the intertidal zone to amaximum depth of 15 m (Gallardo 1979; Vermeij &Carlson 2000). Living adults of C. concholepas werecollected in the Antofagasta area (Santa MariaIsland, northern Chile) at depths of ~5 m. The softparts of the animal were removed and shells wererinsed with tap water.

Optical microscopy

Sections (5–8 lm thick) were observed with polar-ized light microscopy and fluorescence microscopy.Polished sections were stained with acridine orange,a fluorescent cationic dye indicative of acidic sulph-ated sugars (Dauphin 2003). A working solution ofacridine orange was made by dissolving 20 mg of

acridine orange in 95 mL of Milli-Q H2O and 5 mLof methanol. Sections were placed in the solutionfor 30 min, and then rinsed with Milli-Q water. Epi-fluorescence of polished stained surfaces wasobserved under ultraviolet light using a Zeiss Stan-dard microscope (Carl Zeiss International, Oberko-chen, Germany) equipped with Neofluor objectives,a Zeiss mercury lamp, a 365 nm excitation filter,and a 400 nm long pass transmission filter.

Scanning electron microscopy

Scanning electron microscopy (SEM) was usedto examine both fractured surfaces and polishedsections of shells. Polished sections were producedby sectioning shells embedded in resin and polish-ing the sections using various grades of diamondpaste down to a final 0.25 lm grade. Polishedsections were rinsed for 1 min in a detergentdissolved in hot water to remove any oil residue,then rinsed in tap water. Various acid or enzy-matic solutions were then used to reveal micro-structural features of the layers, details of whichare given below. Observations were conductedwith a Philips SEM-505, a Philips XL30-SEMmicroscope (Philips, Amsterdam, the Netherlands;both located at Geosciences, Universite Paris Sud),or a Hitachi low vacuum TM-1000 tabletopmicroscope with back scattered electron (BSE)detector (Hitachi, Tokyo, Japan; European Syn-chrotron Radiation Facility, ESRF).

Feigl’s solution was used to identify aragonite.The solution was prepared as follows. A solution of11.8 g MnSO4·7H2O in 100 mL of deionized waterwas heated to boiling, and 1 g of Ag2SO4 was added.After cooling, 1–2 drops of dilute NaOH wereadded. After 1–2 h, the solution was mixed, filtered,and stored in a dark bottle. Polished sections ofshells were immersed in Feigl’s solution for 1–2 min atroom temperature. Aragonite was stained black, butcalcite remained colorless (Dauphin 1992). Stainedsamples were examined using the BSE detector inthe Hitachi TM-1000 microscope. Image contrastwas correlated with different average atomic numbercomposition within regions of the sample. The silverin the Feigl’s solution, with its high atomic number,led to high brightness of the aragonite layers.

Staining of polished sections with lead citrate anduranyl acetate allowed visualization of organic com-ponents when the sections were imaged with a BSEdetector in a low vacuum SEM. For double stainingof polished sections, sections were first fixed andetched with a mixture of formic acid (0.1%) andglutaraldehyde (3%) for 3 min. Sections were then

Invertebrate Biologyvol. x, no. x, xxx 2012

2 Dauphin, Cuif, Cotte, & Salome

immersed in a solution of lead citrate (0.4 g leadcitrate, 100 mL Milli-Q water, 10 mL 1M NaOH)for 30 min at room temperature, then rinsed withMilli-Q water. They were then immersed in a solu-tion of uranyl acetate (6 g uranyl acetate in 90 mL

Milli-Q water-ethanol [50:50] for 20 min at roomtemperature), followed by rinsing in Milli-Q water.The stained samples were coated with carbon andobserved using the BSE detector in the HitachiTM-1000 microscope.

Fig. 1. Structure of the shell of Concholepas concholepas. A. Outer surface of the shell bearing epibionts. B. Polishedand etched surface of a sectioned shell showing the thick outer prismatic calcitic layer (PR) and the inner aragoniticcrossed lamellar layer (CL). Growth lines (dotted arrows) are visible only in the outer layer. C. Polished and etched(using 5% formic acid, 10 s) section showing the irregular thickness and shape of the layers. Dotted arrows indicategrowth lines in the prismatic layer. D. Thin section showing growth lines (dotted arrows) in the outer layer, the pris-matic units, and the sublayers of the crossed lamellar layer. E. Polished and etched (using 2% glutaraldehyde and0.1% formic acid, 2 min at 20°C) section showing the irregular boundary between inner and outer shell layers. F. Pol-ished and etched (using 10% formic acid, 10 s) section showing the first- and second-order lamellae of the crossedlamellar layer, in which growth lines are not visible. A thin layer between the outer and inner layer is present (doublearrow). G. Polished and etched (using 5% formic acid, 10 s) section showing detail of the crossed lamellar layer. First-order lamellae are not in direct contact (white lines). H. Polished, fixed, and decalcified prismatic layer showing theabundance of organic matrix. Mutvei solution (1% acetic acid +25% glutaraldehyde, 14 min at 20°C). I. Polished andetched (using 1% formic acid, 5 s) section showing detail of the boundary between crossed lamellar and prismatic lay-ers, showing the progressive modification of the shell layers. CL, crossed lamellar layer; OM, organic matrix; PR, pris-matic layer.


Aragonite-calcite boundary in a gastropod shell 3

Fourier transform infrared spectroscopy

Fourier transform infrared (FTIR) spectra wereobtained from powdered isolated layers. Smallpieces of crossed lamellar and prismatic layers wereimmersed in 3% NaClO to remove organic contami-nants, rinsed with Milli-Q water, dried, and groundinto powder. The powdered samples were dried at38°C for one night. All spectra were recorded at4 cm�1 resolution with 64 scans using a strong Nor-ton-Beer apodization on a Perkin Elmer model 1600(Perkin Elmer, Waltham, MA, USA), from 4000 to450 cm�1. The spectrometer was equipped with adiffuse reflectance accessory, which permits diffusereflectance infrared Fourier transform (DRIFT)measurements with high sensitivity on powders. Allspectra were corrected using the Kubelka–Munkfunction.

Synchrotron infrared microscopy was performedon the FTIR end-station of the ID21 beam line atthe ESRF. This beam line is equipped with a Con-tinuum IR microscope (Thermo Nicolet, ThermoInstrument Systems Inc., Santa Fe, NM, USA) cou-pled to a Nexus FTIR bench (Thermo Nicolet). Themicroscope operates in confocal mode, where thefocusing Schwarzschild objective and the collectionSchwarzschild objective have a magnification of 329(NA = 0.65). Due to its small source size and highcollimation properties, a synchrotron IR source isefficiently coupled to the low acceptance of the con-focal microscope, while its high spectral lumines-cence (brightness) allows signal-to-noise ratios to bekept at diffraction-limited resolutions. The micro-scope is equipped with a computer-controlled xystage allowing acquisition of profiles or maps of thesample. Maps were scanned with steps of 4 lm inboth directions, with an aperture size of 595 lm2.Spectra were collected in reflection mode, with a res-olution of 8 cm�1. For each spectrum, 256 scanswere accumulated in the wave-number range 4000–700 cm�1. Polished and cleaned surfaces were pre-pared as for the first step of SEM observations,without etching. The final format of the data wasrecorded as absorbance values.

X-ray absorption near edge structure spectroscopy

The ID21 scanning X-ray microscope (ESRF) usesFresnel zone plates as focussing optics to generate asubmicron X-ray probe, which was used to investi-gate the sample with various contrast mechanisms(fluorescence, transmission, phase contrast). Anenergy-dispersive high-purity germanium detector(Princeton Gamma-Tech, Princeton, NJ, USA)

mounted in the horizontal plane perpendicular tothe beam collected the fluorescence emission pho-tons. This geometry minimized the contribution ofelastic scattering. Provided the sample is thinenough, a silicon photodiode can also be mounteddownstream from the sample to exploit the transmis-sion signal. An energy range between 2 and 9 keVwas available, which gave access to the K-edge ofsulphur at 2472 eV. The X-ray absorption near edgestructure (XANES) energy scan around the sulphurK-edge was achieved using a fixed-exit double-crys-tal Si(111) monochromator located upstream fromthe microscope, which offered the necessary energyresolution. This experiment required the X-raymicroscope to be operated under vacuum to avoidthe strong absorption of the sulphur emission linesby air. Chemical maps of sulphur as sulphates andas amino acids were obtained by tuning the incom-ing energy beam at 2.4827 and 2.47373 keV, respec-tively.

Although the primary beam energy was set aroundthat of the sulphur K-edge energy region, elementswith absorption edges at lower energies were alsosubject to excitation and emission of fluorescencephotons, and could therefore be determined. Thus,micro-fluorescence element maps of magnesium(Mg), strontium (Sr), and phosphorus (P) wereobtained simultaneously with the sulphur (S) maps.

To prepare shell samples for XANES, they wereembedded in resin, cut perpendicular to the thick-ness of the shell, and polished using various gradesof diamond paste. The oil residue from the diamondpaste was removed with a detergent diluted in hotwater for 1 min, and the sample was then rinsedwith tap water. Because fluorescence emission wasused as a contrast mechanism, it was essential thatthe sample offered a flat surface to the X-ray beam,in order to avoid signal fluctuations that could havebeen produced by sample topography. Samples werethus etched with 1% (v/v) acetic acid for 5 s. Etch-ing of the sample after polishing also helped to elim-inate potential surface contamination and remainsof the diamond pastes and oil.

Results

Microstructures

Polished and etched sections through an adultshell showed minor and major growth lines; theseare particularly visible in the outer prismatic layer(Fig. 1B,C). Such layers were also visible in thin sec-tions (Fig. 1D). The most visible feature of theouter calcitic layer in polished and etched sections



observed with SEM was the abundance of growthlines (Fig. 1C,D: dotted arrows). Nonetheless, theprismatic arrangement of this layer was not visibleby SEM (Fig. 1B,C), whereas the elongated pris-matic units were visible in thin sections (Fig. 1D).Growth lines were also more or less visible in thecrossed lamellar inner layer (Fig. 1C–G). The com-plex arrangement of the lamellae within the crossedlamellar layer probably obscured these growth lines(Fig. 1G).

Observation of the boundary between the twomain layers showed that the thickness of the outerlayer did not increase regularly (Fig. 1C–F). Oscilla-tions in the thickness of the outer and inner layerswere common but not regular. The boundarybetween the aragonitic crossed lamellar layer andthe calcitic prismatic layer was not abrupt (Fig. 1F).Low magnifications showed a thin layer with afibrous structure (Fig. 1F). The crossed lamellarlayer was similar in structure to those of other mol-lusc shells (Fig. 1G). The calcitic prismatic layerincluded a large quantity of organic components(Fig. 1H). Higher magnifications showed a complex

boundary between the two layers. Major growthbands became visible, and an imperfect crossedlamellar layer appeared within these bands. Irregu-larities in the polished surfaces showed the uneventhickness and structure of this boundary layer(Fig. 1I).

A detailed sketch of the 3D hierarchical arrange-ment of the lamellae of the crossed lamellar layer isshown in Fig. 2. Lamellae from first to third ordersare complex structures. Fourth-order lamellae werecomposed of rounded granules similar to those ofother aragonitic layers of mollusc shells (Dauphinet al. 2003; Dauphin 2008).

Surfaces stained with acridine orange showed thatthe boundary layer between the aragonite and cal-cite was thin, with an irregular shape (Fig. 3A–C).From its intense yellow-orange color, it can be saidthat this layer had a high content of acidic sulphat-ed sugar. The structure of the first-order lamellawas visible in the crossed lamellar layer, whereas theprismatic units were not displayed in the calcitelayer. However, some irregular patches were visiblein the calcite, similar to those observed by SEM

Fig. 2. Three-dimensional reconstruction of the arrangement of the lamellae in the crossed lamellar layer. First-orderlamellae are composed of second-order lamellae, which are composed of third-order lamellae, which in turn are com-posed of fourth-order lamellae. Up to fourth-order lamellae are visible using scanning electron microscopy. Theappearance of the crossed lamellar layer varies according to the orientation of the section.



(Fig. 1C,E). It can be hypothesized that the organiccomponents of these irregular zones were similar tothose of the aragonite crossed lamellar layer.

Feigl-stained surfaces also showed a boundaryzone between the prismatic and crossed lamellar lay-ers, as well as irregular patches (Fig. 3D–F: IL).This zone deflected electrons strongly after Feiglstaining, indicating its aragonitic composition.Again, no structure was visible in the calcite layer,whereas first-order lamellae were visible in the ara-gonite layer.

Thin sections observed in polarized light showedthe calcitic units and growth lines (Fig. 3G). Sectionsstained with lead citrate-uranyl acetate were imagedwith a BSE detector to detect evidence of organic

components within shell layers (Fig. 3H–I). Noorganic envelope surrounding the prismatic unitswas visible in SEM images collected with a second-ary electron detector, despite the fixing and etchingprocedures (Figs. 1B–F, 3C). Nonetheless, stainedsections observed with a BSE detector clearlyshowed growth lines (Fig. 3H); they also revealedelongated structures that corresponded with the ori-entation and size of prismatic units seen in thin sec-tions. However, no organic layer was visible betweenthe prisms. Growth lines were distinct within the ara-gonitic crossed lamellar layer when stained sectionswere imaged with a BSE detector. In most sections,the first-order lamellae were not contiguous; a smallvoid was visible in some zones (Fig. 1F,G: white

Fig. 3. Structure of the shell of Concholepas concholepas. A–C. Acridine orange-stained polished sections, showing thestructure of the crossed lamellar layer (CL), the intermediate layer (IL), and the outer prismatic layer (PR). D–F. Feigl-stained polished sections, showing the irregular boundary (IL) between prismatic and crossed lamellar layers. G. Obliquethin section of the calcitic layer showing the prismatic units and growth lines (double arrow). H. Double-stained (leadcitrate and uranyl acetate) section through the prismatic layer showing growth lines (double arrows) and some prismaticunits not seen in unstained sections. I. Double-stained section through the crossed lamellar layer showing first- andsecond-order lamellae and growth lines. CL, crossed lamellar layer; IL, intermediate layer; PR, prismatic layer.



lines). However, other zones did not show this smallvoid: first-order lamellae showed “white” bound-aries, suggesting the existence of a thin organic mem-brane (Fig. 3I). Thus, this small space was probablya result of etching.

Fourier transform infrared spectroscopy

Both calcitic and aragonitic layers includedorganic matrix (Fig. 4). A mosaic image shows thelocation of infrared maps (Fig. 5A). The mineralogi-cal change was visible at 875 cm�1 (Fig. 5B). Thiswavelength is common for aragonite and calcite as adoublet 877–845 cm�1 (m2 band). However, a partof this doublet was sensitive to Mg and Sr contents:the 877 cm�1 part varied between 863.7 and858.8 cm�1 in aragonite, while the other part of thedoublet (844 cm�1) was not altered. There was nocorrelation between the m2 doublet and the chemicalcontent in calcite. Growth lines and boundaries ofthe prisms were not displayed at this wavelength(Fig. 5B), whereas the first-order lamellae of thearagonitic crossed lamellar layers were displayed.Changes in color within the first-order lamellaesuggested the presence of the second-order lamellae.The various orientations of the sublayers of thecrossed lamellar layers were also visible (Fig. 5A,B).

Again, the boundary between the two layers wasneither abrupt nor linear. A thin zone (Fig. 5B: IL)was present between the crossed lamellar layer andthe calcite. Moreover, patches of this material enterinto the calcite layer (Fig. 5B–E). From the m2band, it can be hypothesized that these zones were

built of aragonite. Results from other wavelengthssuggest that mineralogy was not the only changewithin the intermediate zone. The band at1560 cm�1 is related to the amide II band in pro-teins. The map showed a strong difference betweencalcite and aragonite, the light blue color of calcitebeing indicative of a higher content of these organiccomponents (Fig. 5C). As for the m2 band, somepatches penetrated into the calcite. No structure wasvisible in the crossed lamellar layer. The amide Iband at 1663 cm�1 not only showed the differencebetween calcite and aragonite, but the inner struc-ture of the layers was also displayed (Fig. 5D).There was no strong difference in color between thetwo layers, except for the thin boundary layer andthe patches in the calcitic layer. The band at1734 cm�1 showed a similar pattern (Fig. 5E); thisband is related to the C=O double bond in COOH,as present in aspartic and glutamic acids (1710–1760 cm�1) and in lipids (Dreissig et al. 2009). Theinner structure of the aragonitic layer was also dis-played by the 2900 cm�1 band; this was probablyrelated to lipids. However, the patches were absentor very weak. Thus, the calcite-aragonite boundaryseems more abrupt (Fig. 5F).

XANES maps

The mapped zone of the shell is shown inFig. 6A. Distribution maps of sulphur showed theunequal contents of the crossed lamellar and pris-matic layers (Fig. 6B,C). Thin growth lines were dis-played in the organic sulphate-rich calcitic layer, butthe prismatic structure was not visible (Fig. 6B). Incontrast, the different sublayers and first-orderlamellae of the crossed lamellar layers were visiblein SEM images (Fig. 6A: CL), whereas growth lineswere not. Some patches with organic sulphate con-tent similar to that of the crossed lamellar layerspenetrated into the calcitic layers (Fig. 6A,B,D: dot-ted squares). The thin layer between calcite and ara-gonite, with a distinct structural pattern, had theintensity of the aragonitic layer. The distributionmap of sulphur associated with sulphated sugarsshowed that the calcitic layer had a higher contentthan that of the aragonitic layer (Fig. 6B). Again,growth lines were visible in the prismatic layer, butthe outlines of the prismatic units were not visible.Maps of amino acid sulphur showed growth lines inboth layers. Amino acid sulphur contents were low,as shown by the poor resolution of the inner struc-ture of the layers (Fig. 6C). Maps for Sr, Mg, and Pshowed that the contents of the crossed lamellarlayer were lower than those of the prismatic layer

Fig. 4. Diffuse reflectance infrared Fourier transform(DRIFT) spectra of the two shell layers of Concholepasconcholepas, showing their mineral-organic compositions.



(Fig. 6D–E). Growth lines of the calcitic layer werevisible (Fig. 6A), but the limits of the prismaticunits were not.

Detailed maps of the boundary for S, Sr, and Pshowed growth lines in both crossed lamellar andprismatic layers (Fig. 7). First-order lamellae werevisible in S, but were less distinct in Sr and P. Theprismatic units were not apparent in the calcitelayer.

Discussion

Electron microprobe chemical analyses of the twoshell layers of Concholepas concholepas showed that

calcite and aragonite did not differ in their calciumcontent. The presence of minor elements mightpotentially allow greater discrimination of these twopolymorphs, but the content of Mg and Sr withincalcite and aragonite of molluscan shell differs fromthat of non-biogenic calcite and aragonite. Non-bio-genic calcite is rich in Mg and poor in Sr, whereasnon-biogenic aragonite is rich in Sr and poor in Mg.In mollusc shells, the Mg content of calcite is low,while the Sr content of aragonite is high. Previousenergy-dispersive spectrometry analyses showed thatMg and Sr content in calcite and aragonite layers ofC. concholepas were in the usual range for molluscshells (Dauphin et al. 2003). Strontium and Mg

Fig. 5. Structure of the shell of Concholepas concholepas as indicated by Fourier transform infrared spectrometry. A.Mosaic optical view of the polished surface of a shell section observed with FTIR spectrometry, showing sublayers inthe crossed lamellar layer (CL), and patches of intermediate layer (IL) penetrating into the outer prismatic layer (PR).The box indicates the region illustrated in B–F. B–F. Infrared maps. B. v2 band map at 875 cm�1: the lamellae of thearagonitic layer are visible (white lines), whereas the prismatic units observed in thin sections and SEM images are notseen. C. Amide II map at 1560 cm�1 showing that the calcitic prismatic layer has a higher amide content than the ara-gonitic crossed lamellar layer. The limit is not abrupt, and patches with low amide II content penetrate into the pris-matic outer layer. D. Amide I map at 1663 cm�1. The inner structure of the aragonitic crossed lamellar layer is visible.E. 1734 cm�1 map showing similar patches (IL). F. 2900 cm�1 map, probably indicative of lipids. While the innerstructure of the aragonitic crossed lamellar layer is visible, the patches in the prismatic layer observed in other mapsare not revealed at this wavelength. CL, crossed lamellar layer; IL, intermediate layer; PR, prismatic layer.



contents were low in both prismatic and crossedlamellar layers. Sulphur content was higher in thecalcitic prisms (Fig. 8).

The two main shell layers, crossed lamellar andprismatic, included both structural units and growthlines. The visibility of these features was a function

Fig. 6. Structure of the shell of Concholepas concholepas as indicated by X-ray absorption near edge structure energyscans. A. Scanning electron micrograph of the vertical section used for XANES maps. The dotted square indicatespatches of aragonite in the calcitic prismatic layer. B. Distribution map of sulphate showing the growth lines (dottedline) and high sulphate content within the prismatic layer. The dotted square indicates patches of aragonite in the cal-citic prismatic layer. C. Distribution map of sulphur-containing amino acids (Saa). D. Distribution map of strontium(Sr), showing growth lines in the prismatic layer. The dotted square indicates patches of aragonite in the calcitic pris-matic layer. E. Distribution map of phosphorus (P). The prismatic layer has a higher content than the crossed lamellarlayer, and growth lines are more visible in the prismatic layer. F. Distribution map of magnesium (Mg) showing thehigh content within calcitic prisms. CL, crossed lamellar layer; PR, prismatic layer.



of the method applied. Growth lines in prismaticlayers were visible in SEM images and XANESmaps, but not in FTIR maps. Prismatic units wereclearly apparent in polished sections stained withlead citrate and uranyl acetate, but not in XANESmaps. Observations by SEM also showed that theboundary between the two main layers was irregu-lar, and not sharp. Because of the irregular thick-ness and the general thinness of this boundary layer,only in situ observations and analyses can providedata about its structure and composition. Irregulari-ties in the intermediate zone are not only morpho-logical; staining and analytical maps showed thatmineralogy and organic components were also dis-tinct in this zone. Results from UV fluorescence,FTIR, and XANES maps showed that the composi-tion of the organic matrix within the intermediatezone and the irregular patches was similar to that ofthe aragonite layer. They also showed that the finalgrowth stage of the prismatic layer was different

from that of the main part of this layer. Thus, notonly is there a thin intermediate layer between thetwo main layers, but the structural and chemicalchanges from one layer to another are not regular,as shown by aragonitic patches.

Comparisons with other mollusc shells

Few data on the structure and composition ofgastropod shells obtained using in situ methods areavailable. The shells of Haliotis tuberculata (LINNA-

EUS 1758) and Haliotis rufescens SWAINSON 1822(abalone) were studied using XANES maps (Dau-phin et al. 2005). The structure of the shells of Hali-otis spp. and C. concholepas differ, but in bothgenera, one of the three layers is composed of largecalcitic prismatic units. Despite this apparent simi-larity, electron backscatter diffraction analyses haveshown that in C. concholepas, calcitic prisms are lessordered than those found in Haliotis spp. Moreover,in C. concholepas, the crystals of some shell areasare better ordered than in others (Perez-Huertaet al. 2011). Scanning electron microscopic analysesdid not reveal the presence of organic envelopesaround the prisms of C. concholepas shells (Dauphinet al. 2007). However, XANES maps of organic sul-phates indicate the presence of such envelopes inHaliotis spp. Growth lines are also displayed inthese images. As for C. concholepas, S content washigher in the calcitic layer than in the aragoniticnacreous and prismatic layers of a given shell. Usingimmunolabelling of MSI 31-like and MSI 60-likeproteins in Haliotis spp., Jolly et al. (2004) haveshown that the labelled proteins were distributed inthree specific zones of the external epithelium. Itmust be noted that the tubular zone and the foldedouter epithelium were never observed as positive to

Fig. 7. Detailed X-ray absorption near edge structure maps of the boundary layer. A. Sulphate map shows the growthlines (dotted line) in both layers, and the first-order lamellae of the crossed lamellar layer. B. Strontium map. C. Phos-phate map. Both the strontium and phosphate maps show growth lines, but outlines of the first-order lamellae arefaint. CL, crossed lamellar layer; PR, prismatic layer.

Fig. 8. Minor chemical composition of the calcitic pris-matic and aragonitic crossed lamellar layers of the shellof Concholepas concholepas. Values are given in parts permillion.



immunohistochemical labelling at the same time. Inshells of the gastropod Nerita undata LINNAEUS

1758, the structural and mineralogical shifts fromthe outer calcitic layer towards the inner aragoniticcrossed lamellar layer are also complex (Nouet et al.2012).

Most gastropod shells have no pallial line; themantle is not attached to the shell by small musclesarranged in a narrow line visible on the inner sur-face of the shell. In molluscs, shell areas where mus-cles are attached consist of aragonitic small prismsor fibers. The shell of Pinctada margaritifera, abivalve, also has a boundary layer between the inneraragonitic nacre and the outer calcitic prisms. Inmembers of this genus, like in C. concholepas, themantle is not attached to the shell; there is no pallialline. The boundary layer of P. margaritifera is builtof fibrous aragonite. In situ analyses have shownthat the organic component of this layer is similarto that of the nacreous layer (Dauphin et al. 2008;Farre et al. 2011). The biochemical change is notabrupt; the organic content of the end of the pris-matic layer differs in both sugar and amino acidcontent, so that mineralogical-structural changesand chemical changes are not synchronous. Twomatrix proteins, MSI 31 and MSI 60, are specific tothe organic matrix of the calcitic and the aragoniticlayer, respectively (Sudo et al. 1997). From in situhybridization, these authors reported that the twoepithelial regions corresponding to the calcite andaragonite layers are contiguous, with no commonreaction. Transcript localization was done for twoproteins involved in biomineralization: aspein andN14-N16 pearlin. In situ analysis showed that thetranscripts (pmarg-aspein and pmarg-pearlin) arelocalized in two distinct areas of the outer epithe-lium; no overlap was detected, but transcripts ofonly two genes were characterized in this study(Joubert et al. 2010). In this species, SEM images,XANES, and time-of-flight secondary ion massspectrometry (TOF-SIMS) maps showed the pres-ence of a fibrous aragonitic layer between the twomain layers (Dauphin et al. 2008; Farre et al. 2011).Comparison of these results has shown that thebiomineralization process is not dependent only onproteins. Kitano & Hood (1965) have shown thattrace elements, amino acids, and sulphated sugarsplay a role in the mineralogy of calcium carbonateprecipitation.

Acknowledgments. This work was supported by anESRF grant (EC 545). Samples were collected by theInstitut de Recherche pour le Developpement (PALEO-

TROPIQUE team) during the collaborative programmeCONCHAS (Programme National d’Etude de la Dyna-mique du Climat). The authors are grateful for theinsightful comments of L. Page, B. Pernet, and severalanonymous reviewers, which helped to improve the man-uscript.

References

Dauphin Y 1992. Methode d’identification localisee de ladiagenese des aragonites d’origine biologique a l’echelledu microscope electronique a balayage. Palaont. Z. 66:13–22.

———— 2003. Soluble organic matrices of the calcitic pris-matic shell layers of two pteriomorphid bivalves: Pinnanobilis and Pinctada margaritifera. J. Biol. Chem. 278:15168–15177.

———— 2008. The nanostructural unity of mollusc shells.Mineral. Mag. 72: 243–246.

Dauphin Y, Cuif JP, Mutvei H, & Denis A 1989. Miner-alogy, chemistry and ultrastructure of the external shelllayer in ten species of Haliotis with reference to Halio-tis tuberculata (Mollusca: Archaeogastropoda). Bull.Geol. Inst. Univ. Uppsala, N.S. 15: 7–38.

Dauphin Y, Guzman N, Denis A, Cuif JP, & Ortlieb L2003. Microstructure, nanostructure and compositionof the shell of Concholepas concholepas (Gastropoda,Muricidae). Aquat. Living Resour. 16: 95–103.

Dauphin Y, Cuif JP, Salome M, & Susini J 2005. Specia-tion and distribution of sulfur in a mollusk shell asrevealed by in situ maps using X-ray absorption near-edge structure (XANES) spectroscopy at the S K-edge.Am. Mineral. 90: 1748–1758.

Dauphin Y, Williams CT, Salome M, Susini J, & Cuif JP2007. Microstructures and compositions of multilayeredshells of Haliotis (Mollusca, Gastropoda). In: Biomin-eralization: From Paleontology to Materials Science.Arias JL & Fernandez MS, eds., pp. 265–272. Univer-sity of Santiago, Chile.

Dauphin Y, Ball AD, Cotte M, Cuif JP, Meibom A, Sa-lome M, Susini J, & Williams CT 2008. Structure andcomposition of the nacre – prism transition in the shellof Pinctada margaritifera (Mollusca, Bivalvia). Anal.Bioanal. Chem. 390: 1659–1169.

De Ioannes P, Moltedo B, Oliva H, Pacheco R, FaunesF, de Ioannes AE, & Becker MI 2004. Hemocyanin ofthe molluscan Concholepas concholepas exhibits an unu-sual heterodecameric array of subunits. J. Biol. Chem.279: 26134–26142.

Dreissig I, Machill S, Salzer R, & Krafft C 2009. Quanti-fication of brain lipids by FTIR spectroscopy and par-tial least squares regression. Spectrochim. Acta 71:2069–2075.

Farre B, Brunelle A, Laprevote O, Cuif JP, Williams CT,& Dauphin Y 2011. Shell layers of the black-lip pearloyster Pinctada margaritifera: matching microstructureand composition. Comp. Biochem. Physiol. B 159: 131–139. doi: 10.1016/j.cbpb.2011.03.001.



Gallardo C 1979. El ciclo vital del Muricidae Concholepasconcholepas y consideraciones sobre sus primeras fasesde vida en el bentos. Biol. Pesq. 12: 79–89.

Guzman N, Dauphin Y, Cuif JP, Denis A, & Ortlieb L2009. Diagenetic changes in Concholepas concholepasshells (Gastropoda, Muricidae) in the hyper-aridconditions of Northern Chile – implications for palae-oenvironmental reconstructions. Biogeosciences 6: 197–207.

Jolly C, Berland S, Milet C, Borzeix S, Lopez L, & Dou-menc D 2004. Zonal localization of shell matrix pro-teins in mantle of Haliotis tuberculata (Mollusca,Gastropoda). Mar. Biotechnol. 6: 541–551.

Joubert C, Piquemal D, Marie B, Manchon B, Pierrat F,Zalella-Cleon I, Cochennec-Laureau N, Gueguen Y, &Montagnani C 2010. Transcriptome and proteomeanalysis of Pinctada margaritifera calcifying mantle andshell: focus on biomineralization. BMC Genomics 11:613.

Kitano Y & Hood DW 1965. The influence of organicmaterial on the polymorphic crystallization of calciumcarbonate. Geochim. Cosmochim. Acta 29: 29–42.

Kobayashi I 1969. Internal microstructure of the shell ofbivalve molluscs. Am. Zool. 9: 663–672.

———— 1983. Biomineral formation of gastropods, incomparison with that of pelecypods. In: Biomineraliza-tion and Biological Metal Accumulation. Westbroek P& De Jong EW, eds., pp. 261–266. Reidel PublishingCompany, Dordrecht, the Netherlands.

———— 2004. Nacreous and crossed lamellar layers ofBivalvia: diversity and evolution. In: Biomineralization(BIOM2001): Formation, Diversity, Evolution andApplication. Kobayashi I & Ozawa H, eds., pp. 88–91.Tokai University Press, Kanagawa, Japan.

Moltedo B, Faunes F, Haussmann D, De Ioannes P, deIonnaes AE, Puente J, & Becker MI 2006. Immunothera-pic effect of Concholepas hemocyanin in the murine blad-der cancer model: evidence for conserved antitumorproperties among hemocyanins. J. Urol. 176: 2690–2695.

Nouet J, Cotte M, Cuif JP, Dauphin Y, & Salome M2012. Biochemical change at the setting-up of thecrossed-lamellar layer in Nerita undata (Mollusca, Gas-tropoda). Minerals 2: 85–99.

Perez-Huerta A, Dauphin Y, Cuif JP, & Cusack M 2011.High resolution electron backscatter diffraction (EBSD)data from calcite biominerals in recent gastropod shells.Micron 42: 246–251.

Sudo S, Fujikawa T, Nagakura T, Ohkubo T, SakaguchiK, Tanaka M, Nakhashima K, & Takahashi T 1997.Structures of mollusc shell framework proteins. Nature387: 563–564.

Vermeij GJ & Carlson SJ 2000. The muricid gastropodsubfamily Rapaninae: phylogeny and ecological history.Paleobiology 26: 19–46.

Wang D, Wallace AF, DeYoreo JJ, & Dove PM 2009.Carboxylated biomolecules control magnesium contentof amorphous calcium carbonate: insights for calcifica-tion. Proc. Natl. Acad. Sci. 106: 21511–21516.



Structure and composition of the boundary zone between aragonitic crossed lamellar and calcitic prism layers in the shell of Concholepas concholepas (Mollusca, Gastropoda)

Documents