Supporting online materials:

Materials and Methods:

Samples
Animals were obtained from the sub-population of oysters that occur in the
Northern Gulf of Mexico (Texas to North Florida coasts). These oysters are often
solitary individuals and have the greatest recorded growth rate for the species.
Adult oysters in over-wintering condition (15-18°C) display the greatest degree of
shell growth over the course of a year (S1). Animals were purchased from Bon
Secour Fisheries of Bon Secour, Alabama. Newly harvested oysters were
selected on the basis of size (no more then 10 cm in length) and shipped
overnight on ice. The oysters were scrubbed to remove mud and debris, dipped
in a saturated salt solution to remove worms and allowed to air dry (S2). The
oysters were placed in a depuration tank that contained 60 gallons (~ 227 liters)
of circulating aerated artificial seawater (Instant Ocean, 18 ‰ at 18°C). The
water was filtered for particulates (4 µm and UV sterilized). Oysters were held in
the tank for 10 days prior to experimentation.
Induction of shell regeneration
A v-shaped notch was cut on the right side of the bivalve, in close
proximity to the animal’s adductor muscle (Fig 1A of text). The notch was cut so
that the body of the animal was not harmed. The animal was returned to aquaria
and within 24-48 hours after notching, regenerated shell would appear in the
notched region.
Hemolymph collection
Hemolymph was also collected from notched oysters by use of a 1.5” long
21 gauge needle affixed to a 3 ml sterile plastic syringe. Most collections yielded
from 0.5 – 1 ml of hemolymph and were immediately transferred to 1.5 ml plastic
microfuge tubes and held on ice.
Differential interference contrast (DIC) microscopy of living hemocytes
Volumes from 20 to 40 µl of freshly drawn hemolymph were spotted on 8
well slides (Erie Scientific, Catalog number ER-201) and placed in a humidified
Petri dish for 20 min. A glass cover slip was affixed to the slide and it was
examined with a Zeiss Axiovert 135 fluorescent microscope set up for DIC
through a Zeiss Plan Neofluar 40 x oil 1.3 numerical aperture (N.A.) objective.
Images were collected by a Diagnostic Instruments SPOT RT cooled CCD
camera. A total of five fields for each sample of hemolymph was collected and
enumerated. Hemolymph from a total of 3 oysters was analyzed.
DIC microscopy of living mantle tissue sections
Oysters were relaxed by injection of 1% cocaine solution (dissolved in
osmotically adjusted phosphate buffered saline (PBS, 20mM sodium phosphate,
150mM or higher NaCl, the exact concentration of sodium chloride is dependent
upon the osmolality of the oyster’s holding tank, pH 7.4), into the adductor
muscle. Within 5 minutes the shells would gape (S3). In order to gain access to
the body of the animal, the hinge ligament was popped open and the flat valve
was excised from the adductor muscle. Mantle sections were dissected near the
growing margin of shell. The shell facing side of the mantle was affixed to a glass
cover slip and mounted to a glass slide. The sections were examined with a
Zeiss Axiovert 135 fluorescent microscope set up for DIC through a Zeiss Plan
NeoFluar 40x 0.75 N.A. objective. Images were collected by a Diagnostic
Instruments SPOT RT cooled CCD camera.
Visualization of living hemocytes at the mineralization front
Regenerated shell was collected from induced oysters, approximately 48 h
after notching. The shell pieces were placed into 1 ml sterile osmotically
adjusted PBS to which 10 µl of calcein AM (Molecular Probes, Catalog #C3099,
www.probes.com) was added. The pieces were incubated for 1-2 hours at room
temperature. Shell pieces were washed twice with osmotically adjusted PBS.
Each piece was transferred to a depression well slide and fresh osmotically
adjusted PBS was added to the well. Care was taken to ensure that the mantle
facing side of the shell piece was mounted onto the imaging surface of the cover
slip. The preparation was examined with a Zeiss Axiovert 135 fluorescent
microscope (FITC channel) through a Zeiss Plan NeoFluar 40x 0.75 N.A.
objective. Images were collected by a Diagnostic Instruments SPOT RT cooled
CCD camera and post-processed by Adobe Photoshop 6.0.
Scanning electron microscopy (SEM) of fixed hemocytes
Hemolymph from oysters undergoing shell regeneration was obtained by
the method outlined above. Approximately 20 µl of hemolymph was spotted on a
glass cover slip and placed in a humidified Petri dish for 20 min. The cells were
fixed with 95% ethanol for 10 minutes, and then air dried. Specimens were
sputter coated with gold (Denton Desk II, 45 ma for 90 sec at less then 25 mtorr)
and examined with a Hitachi Model 3500 SEM equipped with an Oxford energy
dispersive X-ray microanalysis system (S4). The EDS was standardized to
copper prior to specimen analysis. Images and spectra data were digital
processed and recorded.
Alternatively, oyster hemolymph was placed on a glass slide and
incubated for 20 min. Adherent cells were fixed in 4% paraformaldehyde for 4 h.
Slides were washed 3X with de-ionized sterile filtered water and dried overnight.
The fixed specimens were carbon coated and imaged by SEM.
SEM observations of hemocytes at the mineralization front
Newly formed shell was collected from the notch of induced oysters 48 h
after induction. The pieces were fixed with 95% ethanol for 10 min. and set out
to air dry. Dried pieces were mounted to aluminum stubs and sputtered coated
with gold (Denton Desk II, 45 ma for 90 sec. at less then 25 mtorr). Specimens
were examined with a Hitachi Model 3500 SEM. Images were digitally processed
and recorded.
For EDS analysis, regenerated shell was fixed in 4% paraformaldehyde
immediately after collection. Shell pieces were washed 3X with de-ionized sterile
filtered water and dried overnight. The specimen was affixed to an aluminum
stub, carbon coated and imaged by SEM. The spectrometer was calibrated to a
cobalt standard prior to sample analysis
Raman micro spectroscopic analysis
Approximately 500 mg of living mantle tissues (see above) were collected
and placed into 1.5 ml microfuge tube. The tissue was digested by the addition of
1 ml of 10 N NaOH, centrifuged at 16K x g (RCF at tip) for 10 min. The
supernatant was discarded and 1 ml of a 50% sucrose solution was added. The
pellet was re-suspended, mixed and centrifuged at 20K x g (RCF at tip) for 20
min. This step separated cellular debris from crystals thus yielding a pellet that
contained mostly crystals. The pellet was washed three times in a 10 mM NaOH
solution and transferred to aluminum foil covered slide and allowed to air dry.
Raman spectra of the extracted crystals were obtained using a Renishaw
System 1000 Raman spectrometer, coupled to a Leica DM LM microscope. The
other components of the system included a 785 nm (± 2 nm) low power diode
laser (Renishaw, Inc., Hoffman Estates IL), a stigmatic single spectrograph and a
two-dimensional thermoelectrically cooled CCD camera. All the spectra were
acquired using a Leica LM-Plan 50x 0.75 N.A. objective and compared with
Raman spectra of aragonite, calcite and vaterite standards (S5). All data
acquisition was performed using the Renishaw WiRE v1.3 software. Data were
subsequently transferred into Grams/32 (Galactic Inc, Salem, NH) for processing
and analysis.

Fig. S1. Differential Interference Contrast Microscopy (DIC) of the 3 cell types in
oyster hemolymph. A. Agranulocytes compared to REF Granulocytes.
Agranulocytes have a flattened appearance, possess little or no granules and
move very slowly. In contrast, highly motile REF granulocytes appear as
brilliantly lit cells that contain cubodial birefringent granules. B. Granulocytes
compared to REF Granulocytes. In contrast to REF granulocytes, the highly
motile granulocytes have a fan shaped appearance and contain numerous
relatively non-refractile spherical granules.

Fig S2. SEM of an intracellular crystal in an induced hemocyte from oyster

hemolymph. A. Backscattered electron image (BSE) in which a crystal is visible
within a single cell. In BSE mode, elements with higher atomic number scatter
more electrons (such as calcium and silicon) thus appearing lighter while carbon
appears black. The light areas around the cell are due to backscattered electrons
from the silica composition of the slide. B. Secondary electron (SE) image shows
the entire cell. C. Pseudocolor projection shows a crystal that is clearly visible
within the cell. Digitally processed pseudo-color is created by assigning a unique
color to each of the BSE and SE images then merging the two images together.
Fig S3. EDS spot analysis of a calcium containing crystal in an induced
hemocyte from oyster hemolymph. The specimen was prepared in the same
manner as that in figure S2. The spectrometer was calibrated to a cobalt
standard prior to sample analysis. S-1, a spot on the crystal contains 2.6 times
more calcium then spot S-2, a spot on the glass slide.

Fig S4. EDS spot analysis of crystal bearing hemocyte on the surface of
prismatic shell. A. The individual prisms of the prismatic layer beneath the cell
are visible. B. S-1, a spot on the crystal contains 1.34 times more calcium then
spot S-2, a spot on the glass slide. This is a lower ratio then figure S3 due to the
fact that the underlying surface is calcified shell.
Fig S5. Pseudocolor SEM of crystal-bearing hemocytes on the surface of newly
formed prismatic shell. Shell pieces were fixed in 4% paraformaldehyde
immediately after collection. A. Individual prisms (pink) appear to be enclosed in
organic matrix (green). B. SEM of box in A. A hemocyte is visible on surface of
an individual prism. Intracellular crystals are yellow. In contrast to the cells, the
organic matrix surrounding the prisms appears to be devioid of crystals. 5 µm
scale bar.

References and Notes:

S1. P. S. Galtsoff. Fish Bull 64:1(1964)
S2. G. A. Debrosse, S. K. Allen, J. Shellfish Res. 12: 29 (1993).
S3. C. C. Coney. Veliger 36:413 (1993).
S4. J. J. Bozzola, L.D. Russell, Electron Microscopy: Principles and Techniques
for Biologists (Jones & Bartlett Publishers, Inc, Sudbury, MA, 1999).
S5. I. R. Lewis, H.G.M. Edwards Eds., Handbook of Raman Spectroscopy: from
the research laboratory to the process line (Marcel Dekker, New York, 2001).
 REPORTS
refractive granulocytes (REF granulocytes or
Hemocyte-Mediated Shell REF cells), which contains the birefringent
granules (fig. S1). REF granulocytes have

Mineralization in the been reported to be spent granulocytes (21),

whereas others consider them to be the result
of cellular differentiation, thus opening the
Eastern Oyster possibility that they may have some role in
shell formation (19, 22).
Andrew S. Mount,1* A. P. Wheeler,1 In order to observe cellular involvement
Rajesh P. Paradkar,2† D. Snider3 in shell formation, we induced rapid shell
growth by notching the oyster in the shell
The growth of molluscan shell crystals is usually thought to be initiated from margin where the highest rates of biominer-
solution by extracellular organic matrix. We report a class of granulocytic alization are observed (23). Newly formed
hemocytes that may be directly involved in shell crystal production for oysters. shell is observed as a light brown to dark
On the basis of scanning electron microscopy (SEM) and x-ray microanalysis, purple leather-like material that is lightly
these granulocytes contain calcium carbonate crystals, and they increase in mineralized (24). Figure 1A is an image of an
abundance relative to other hemocytes following experimentally induced shell intact oyster, 48 hours after induction. Regen-
regeneration. Hemocytes are observed at the mineralization front using vital erated shell, indicated by the arrow, is visible
fluorescent staining and SEM. Some cells are observed releasing crystals that within the notched region of the shell.
are subsequently remodeled, thereby at least augmenting matrix-mediated There are three distinct shell layers asso-
crystal-forming processes in this system. ciated with an adult oyster shell. The outer-

Downloaded from http://science.sciencemag.org/ on February 4, 2019

Molluscan shell formation is often cited as
resulting in large measure from extracellular
events mediated by organic matrix secreted
from the mantle epithelium (1–3). The ma-
trix-mediated hypothesis states that the com-
plex organic matrix induces heterogeneous
nucleation of calcium carbonate crystals on
its surface and regulates crystal growth,
thereby forming the crystal morphologies that
are unique to the various layers of molluscan
shell. In vitro studies using natural and syn-
thetic molecular systems have shown that
matrix control of crystal morphology is pos-
sible (3–10). However, these experiments are
often conducted at levels of calcium carbon-
ate supersaturation that are far higher than
values reported for the extracellular environ-
ment in which shell is formed, leaving the
extent of matrix involvement in crystal induc-
tion an open question. When in vitro experi-
ments were conducted using insoluble matrix
extracts (IM) from the shell of the oyster
Crassostrea virginica, the level of supersatu-
ration required to induce mineralization was
no lower than that of controls (11). Further-
more, analysis using flow-cell atomic-force
microscopy of oyster shell foliated pieces
revealed that in situ matrix-coated crystals do
not induce secondary nucleation events on
their surfaces (12).
One alternative to the matrix-mediated
hypothesis is that crystal nucleation is intra-
cellular and that crystallogenic cells supply
1
Department of Biological Sciences, Clemson Univer-
sity, Clemson, SC 29634, USA. 2Center for Advanced
Engineering Fibers and Films, Clemson University,
Clemson, SC 29634, USA. 31581 Regimental Lane,
Johns Island, SC 29455, USA.
*To whom correspondence should be addressed. E-
mail: mount@clemson.edu
†Present address: Dow Chemical Company, Analytical
Sciences, 2301 North Brazosport Boulevard, Building
B 1470, Freeport, TX 77541, USA.
most layer, or periostracum, is composed
nascent crystals to the mineralization front. chiefly of organic matrix, which has a smooth
Earlier radioisotope analyses of calcium up- appearance when examined by SEM (23–25).
take in the oyster support this hypothesis. Immediately below the periostracum is the
These studies have shown that only 2.4% of prismatic layer of the shell. It is a thin sheet
the calcium of the shell-forming mantle tissue of calcite composed of crystals in the form of
turns over rapidly (13–15), revealing a large, prisms surrounded by organic matrix (Fig. 1,
nonexchangeable pool of calcium seques- B and D) (24). The innermost foliated layer
tered in these tissues. Although the source of covers the underlying prismatic layer of shell
the nonexchangeable pool has not been iden- and constitutes most of the mineralized shell
tified, one candidate is the circulating amoe- volume (Fig. 1C). It is composed of layers of
boid hemocytes. These cells can migrate to thin laths of calcite.
the surface of the shell-facing outer mantle A common criticism of shell induction
epithelium (OME) (16, 17), and one class of experiments is that the shell layers of re-
the hemocytes contains granules that are bi- generated shell do not resemble those asso-
refringent when examined by polarized light ciated with normal shell growth (15, 26).
microscopy. We reasoned that some of these Our SEM observations of regenerated outer
granules might be calcium carbonate crystals. prismatic and inner foliated layers from
There are two major forms of amoeboid notched oysters, however, do resemble
cells in oyster hemolymph: agranulocytes and their normal counterparts with some minor
granulocytes (17–19). Agranulocytes tend to exceptions. Comparison of regenerated
spread out thinly across a glass cover slip and prismatic shell (Fig. 1D) to adult prismatic
secrete collagen fibers (20). Granulocytes are shell (Fig. 1B) shows newly formed prisms
highly motile cells that appear to have a growing across the surface of the periostra-
macrophage-like function (19). Some of cum. These prisms have not yet completely
these latter cells form a subclass known as covered the outer membrane but have the
Fig. 1. (A) Eastern
Oyster C. virginica at
48 hours after induc-
tion. The arrow points
to new shell that has
regenerated within
the cut or notched re-
gion of the mollusk.
Bar, 1 cm. (B) SEM of
mature prismatic shell.
The prisms range from
5 to 25 !m in edge
length. Bar, 20 !m. (C)
Mature foliated shell showing individual laths that appear to be coated with organic material. Bar,
2.5 !m. (D) SEM of regenerated prismatic shell 48 hours after induction. The prisms are nearly the
same size as in (B) yet have not completely grown together. Periostracum is visible as a smooth
unmineralized sheet between the crystals. Bar, 50 !m. (E) Regenerated foliated shell 48 hours after
induction. The individual laths appear free of debris and are arranged as in (C). Laths range from
0.25 to 5.3 !m in width (23). Bar, 5 !m.

www.sciencemag.org SCIENCE VOL 304 9 APRIL 2004 297

 REPORTS
same irregular shapes and sizes as in adult
shell (Fig. 1B). A comparison of regener-
ated folia (Fig. 1E) to adult folia (Fig. 1C)
reveals an identical microstructure of the
crystals with the exception that newly
formed folia have smoother surfaces and
sharper edges than those of mature shell.
Several observations suggest that REF
granulocytes play a role in the induced
biomineralization response of oysters.
First, comparisons of hemocytes collected
at time of induction to those collected at 48
hours after induction indicate that there is
an increase of REF granulocytes from 5 to
15% of the total hemocyte population (Ta-
ble 1). Second, SEM of REF granulocytes
reveal crystal-shaped inclusions in these
cells (Fig. 2A; fig. S2). The edge length of
the crystals ranged from 0.5 to 1 !m (Fig.
2B). It is likely that the inclusions are
calcium carbonate crystals, given their
rhombohedral appearance and the fact that
line scan and spot analyses by x-ray micro-
analysis (SEM-EDS) confirmed that they
contain calcium at higher levels than sur-
rounding regions of the cell (Fig. 2B; fig
S3). Third, crystal-bearing REF granulo-
cytes appear to release their crystals at the
mineralization front. SEM observations of
regenerated (newly formed) shell pieces re-
vealed REF cells on the prismatic shell
surface in association with lines of fibrous
materials and crystals (Fig. 3A; fig. S4).
Several of these cells are clustered in close
association with secreted crystals (Fig. 3, A
and B). Some cells contain just one or two
crystals (Fig. 3C), whereas others contain
several (Fig. 3D; fig. S5). In Fig. 3D, crys-
tals in varying stages of cellular release are
evident. These observations of directed ac-
tivity at the mineralization front suggest
that the REF granulocytes represent a de-
velopmental stage that delivers crystals to
the site of shell formation.
Further evidence that living hemocytes
are present at the mineralization front was
obtained using the vital fluorescent stain, cal-
cein AM. Hemocytes can be observed (Fig.
4) on newly regenerated prismatic shell. The
labeled cells ranged in size from 14 to 20 !m.
Table 1. Hemocyte morphology at the time of
induction (0 hours) and 48 hours after induction of
shell in C. virginica. Data were normalized to per-
cent from a mean of three different oysters. A
minimum of five fields from each oyster’s hemo-
lymph sample was enumerated.
Percent
Hemocyte morphology
0 hours 48 hours
Agranulocytes 55 54
Non-REF granulocytes 40 31
REF granulocytes 5 15
Total 100 100
298 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org
One cell had filopodia-like extensions around centers and terminates when only the outer
its periphery (Fig. 4A), whereas the other walls remain (Fig. 3F). A single remodeled
appeared more elongated with pseudopod- crystal may yield several plates, all of uni-
like projections (Fig. 4B), resembling the cell form thickness and with dimensions similar
in Fig. 3D. Both cells appeared motile and to those of the foliated laths of regenerated
shared the same morphological features as shell (Fig. 1E). Preliminary indications are
circulating oyster hemocytes. that hemocytes actively participate in these
Crystals released at the regeneration front remodeling and assembly processes.
have a striking appearance in that their mor- The supply of crystalline calcium carbon-
phology resembles inorganic calcite and their ate by hemocytes can occur at a rate suffi-
surfaces appear clean with sharply defined cient to support normal shell formation in
edges. However, they are quickly remodeled. oysters. Radioisotope incorporation studies
Crystals presumably secreted by hemocytes estimate an average of 6 !g per cm2 of
at the margin of the newly formed shell are mantle per hour calcium deposited in oysters
depicted in Fig. 3E. The crystals have grown as shell (2, 14, 27). Based on the average
in size, and holes can be seen near the centers crystal size of secreted crystals detected by
of several individual crystals. They appear to SEM (Fig. 3) and assuming four crystals are
be dissolved by a process that starts near their secreted per cell, it is estimated that 200,000
Fig. 2. SEM of crystal-
bearing REF granulo-
Downloaded from http://science.sciencemag.org/ on February 4, 2019
cytes isolated from
hemolymph, 48 hours
after induction. (A)
The peripheries of four
cells are outlined in
white. The membrane
has been removed
from the largest cell,
revealing many cyto-
plasmic granules and
the presence of a sin-
gle crystal in the cell.
The box highlights the crystal within the cell. Bar, 10 !m. (B) Enlargement of the box in (A). The
polyhederal shape resembles a calcite crystal. The crystal was subjected to SEM-EDS by line scan.
The scanned region is indicated by the straight line. The jagged line shows count intensity obtained
from the calcium main alpha K line. Bar, 1 !m. The crystal contains a higher relative abundance of
calcium in comparison to the cellular background, including the spherical cytoplasmic granule that
was scanned to its left (23).
Fig. 3. (A) SEM of crystal-bearing hemocytes (REF granulocytes) of newly formed prismatic shell on
the mineralization front, 48 hours after induction. The cells are oriented along a line of filamentous
secretions and at least three crystals are visible. Bar, 10 !m. (B) Enlargement of the box in (A). Two
crystals associated with two different hemocytes are visible. Bar, 5 !m. (C) Enlargement of the box
in (B). A single crystal-bearing hemocyte is visible and the crystal size is 2 !m by 2 !m. Bar, 5 !m.
(D) Crystal-bearing hemocyte disgorging several crystals at once. Crystal sizes range from 1 !m by
2 !m to 2 !m by 4 !m edge length. The crystals have a polyhederal shape that resembles calcite.
Bar, 10 !m. (E) Crystals that appear to be undergoing remodeling. These crystals occur in the
transition between prismatic and foliated shell layers. Bar, 50 !m. (F) Enlargement of the box in (E).
The walls of the dissolved crystals have formed plate-like structures with dimensions more similar
to foliated laths (arrow) (23). Bar, 10 !m.

REF cells would be required to sustain this
level of mineralization. Typical C. virginica
hemocyte counts averaged approximately 10
million cells per ml, which falls within the
range of published values (18, 28). Assuming
conservatively that an oyster has about 3 ml
of hemolymph and that during periods of
active shell growth 10% of the hemal cell
population is crystal-bearing REF granulo-
cytes (Table 1), then only approximately 7%
of the REF cell population would need to be
Fig. 4. Fluorescent microsco-
py of living cells on a regen-
erated prismatic shell piece,
48 hours after induction.
Shell pieces are incubated in
calcein AM, a fluorescent re-
agent that determines cell vi-
ability. The esterase activity
of living cells cleaves the dye,
producing a charged green
regenerated per hour to sustain the reported
average rate of calcium incorporation into
shell. In addition, intracellular deposits of
calcite in REF granulocytes could account for
the large (approximately 97%) nonexchange-
able pool of bivalve mantle tissue calcium
that has been reported in several radioisotope
studies (13, 14, 29).
A separate class of calcium carbonate
crystals has been observed on the shell-
forming surface of the OME (Fig. 5A).
REPORTS
Although the general OME is not as active
as the gland located in the periostracal
groove (25, 30), it is involved in shell
thickening and its surface is accessible to
hemocytes (16, 17). Raman microspectros-
copy of OME-extracted crystals revealed
their mineralogy to be calcite, consistent
with the known mineralogy of prismatic
and foliated shell layers (23) (Fig. 5C).
Although OME crystals have similar ap-
pearances and shapes as crystals released
by hemocytes, they are somewhat larger in
size (5 to 7 !m edge length) than the newly
secreted crystals (1 to 4 !m edge length)
(Fig. 5, A and B). We speculate that the OME-
deposited crystals arise from nascent crystal
secretions of REF cells and grow after depo-
sition onto the surface of the OME.
A cellular basis for oyster shell forma-
tion is consistent with many other biomin-
eralization processes, such as spicule for-

Downloaded from http://science.sciencemag.org/ on February 4, 2019

fluorescent product (23). The mation. Spicules occur in most phyla, in-
organic matrix that forms the cluding notably the Mollusca (3, 31). In
walls of the prisms is auto- most of these taxa, spicule mineralization is
fluorescent in the FITC (fluo- usually mediated by a single cell type
rescein isothiocyanate) channel and is visible in the background. (A) Granulocyte exhibiting
filopodia-like extensions around its periphery. Esterase-positive granules can be seen. Bar, 10 !m. (scleroblast) and involves filamentous col-
(B) Calcein-positive hemocyte that resembles the cell in Fig. 3D. Several pseudopodia are visible. lagen-like proteins, and mature spicules
Bar, 10 !m. form extracellular structures that may un-
dergo further growth and remodeling.
The involvement of cells in shell forma-
tion may have evolutionary significance for
the Mollusca. Scheltema’s phylogeny of
extant Mollusca proposes two separate evo-
lutionary lineages: Aculifera (spicule form-
ers) and Conchifera (shell formers) (32,
33). Although any cellular processes of
mineralization in the two groups may be
strictly homologous, it is possible that they
have a common origin.
The discovery that hemocytes can initiate
mineral growth on secreted organic sheets
reduces the difficulties associated with nucle-
ation of crystals from oyster extrapallial flu-
id, which at best is marginally supersaturated
(10, 11). Apparently, crystal formation in-
volves complex interactions between organic
phases and cells.

References and Notes

1. K. Simkiss, K. M. Wilbur, Biomineralization, Cell Biol-
ogy and Mineral Deposition (Academic Press, New
York, 1989).
2. A. P. Wheeler, in Calcification in Biological Systems, E.
Bonucci, Ed. (CRC Press, Boca Raton, FL, 1992), pp.
179 –216.
3. H. A. Lowenstam, S. Weiner, On Biomineralization
(Oxford Univ. Press, Oxford, 1989), pp. 74 –175.
4. J. Aizenberg, D. A. Muller, J. L. Grazul, D. R. Hamann,
Science 299, 1205 (2003).
5. A. M. Belcher et al., Nature 381, 56 (1996).
Fig. 5. (A) DIC (differential interference contrast) light micrograph of shell facing (OME) with cell 6. S. Mann, B. R. Heywood, S. Rajam, J. D. Birchall,
junctions clearly visible, as indicated by the arrow. Note the individual crystals (in the upper right Nature 334, 692 (1988).
corner of the panel), which have edge lengths of approximately 5 to 7 !m. Bar, 10 !m. (B) 7. G. S. Falini, S. Albeck, S. Wiener, L. Addadi, Science
Extracted crystal from the surface of the OME, shown in the crosshairs of the Raman confocal 271, 67 (1996).
8. L. A. Addadi, J. Moradian, E. Shay, N. G. Maroudes, S.
microscope, before excitation with the laser. (C) Raman spectra of the crystal showing peak Weiner, Proc. Natl. Acad. Sci. U.S.A. 84, 2732 (1987).
intensities at 1087, 712, 280, and 154 cm–1 Raman shift. The inset figure is a calcite standard that 9. E. M. Greenfield, D. C. Wilson, M. A. Crenshaw, Am.
has the same peak intensities. From these data it was concluded that the crystals on the Zool. 24, 925 (1984).
shell-facing surface of the OME are calcite crystals, consistent with the mineralogy of both 10. M. A. Crenshaw, Biol. Bull. 143, 506 (1972).
prismatic and foliated shell layers (23). 11. A. S. Mount, thesis, Clemson University (1999).

www.sciencemag.org SCIENCE VOL 304 9 APRIL 2004 299

 REPORTS
12. C. S. Sikes, A. P. Wheeler, A. Wierzbicki, A. S. Mount,
R. A. Dillaman, Biol. Bull. 198, 50 (2000).
13. L. H. Jodrey, Biol. Bull. 104, 398 (1953).
14. K. M. Wilbur, L. H. Jodrey, Biol. Bull. 103, 269 (1952).
15. K. M. Wilbur, in Physiology of Mollusca, K. M. Wilbur,
C. M. Yonge, Eds. (Academic Press, New York, 1964),
vol. 1, pp. 242–282.
16. A. F. Eble, in The Eastern Oyster, Crassostrea virginica,
V. S. Kennedy, R. I. E. Newell, A. F. Eble, Eds. (Mary-
land Sea Grant College, College Park, MD, 1996), pp.
271–298.
17. T. C. Cheng, in Invertebrate Blood Cells, N. A. Rat-
cliffe, A. F. Rowley, Eds. (Academic Press, London,
1981), pp. 233–300.
18. E. Gosling, Bivalve Molluscs, Biology, Ecology and
Culture (Blackwell Publishing, Oxford, 2003).
19. T. C. Cheng, in The Eastern Oyster, Crassostrea vir-
ginica, V. S. Kennedy, R. I. E. Newell, A. F. Eble, Eds.
(Maryland Sea Grant College, College Park, MD,
1996), pp. 299 –333.
20. M. Justus, S. Patel, A. S. Mount, poster presented at
the 6th International Conference on Shellfish Resto- 30. N. Watabe, in The Mollusca, K. M. Wilbur, Ed. (Academic
ration, Charleston, SC, 20 to 24 November 2002. Press, New York, 1983), vol. 4, part 1, pp. 289–317.
21. M. G. McCormick-Ray, T. Howard, J. Invertebr. Pathol. 31. R. J. Kingsley, Am. Zool. 24, 833 (1984).
58, 219 (1991). 32. A. H. Scheltema, Biol. Bull. 184, 57 (1993).
22. P. M. Hine, Fish Shellfish Immunol. 9, 367 (1999). 33. A. H. Scheltema, C. Schander, Biol. Bull. 198, 121 (2000).
23. Materials and methods are available as supporting 34. We thank E. Ruppert, R. Montanucci, and D. Bushek
material on Science Online. for careful reading and suggestions that have im-
proved the quality of this manuscript. We also thank
24. M. R. Carriker, R. E. Palmer, R. S. Prezant, Proc. Natl.
J. Hatchell for his efforts in cell enumeration and
Shellfish Assoc. 70, 139 (1980).
construction of the oyster holding facility; B. Kay for
25. M. R. Carriker, in The Eastern Oyster, Crassostrea technical help with EM; J. Mount for assistance in
virginica, V. S. Kennedy, R. I. E. Newell, A. F. Eble, Eds. revised figure preparation; and T. McNutt Scott, M.
(Maryland Sea Grant College, College Park, MD, Johnstone, S. Patel, and M. Justus for invaluable dis-
1996), pp. 75–167. cussions. We gratefully acknowledge support from
26. P. S. Galtsoff, Fish Bull. 64, 1 (1964). NSF under award no. 20020445.
27. K. M. Wilbur, A. S. M. Saleuddin, in The Mollusca, K. M.
Supporting Online Material
Wilbur, Ed. (Academic Press, New York, 1983), vol. 4,
www.sciencemag.org/cgi/content/full/304/5668/297/DC1
part 1, pp. 236 –288.
Materials and Methods
28. S. E. Ford, S. A. Kanaley, D. T. Littlewood, J. Invertebr. Figs. S1 to S5
Pathol. 61, 49 (1993). References
29. A. P. Wheeler, P. L. Blackwelder, K. M. Wilbur, Biol.
Bull. 148, 472 (1975). 15 August 2003; accepted 2 March 2004

Characterization of a Common in psoriasis (8, 9). In this case, the carrier

frequency should be higher than a few per-

Downloaded from http://science.sciencemag.org/ on February 4, 2019

cent, with only a short conserved haplotype
Susceptibility Locus for ("200 kb) detected (8). Third, numerous mu-
tations might exist in the putative gene, in
Asthma-Related Traits which case only weak or absent haplotype
associations might be detectable.
Tarja Laitinen,1,2 Anne Polvi,2 Pia Rydman,1 Johanna Vendelin,1,2 To distinguish between these hypotheses, we
Ville Pulkkinen,1 Paula Salmikangas,1 Siru Mäkelä,2 Marko Rehn,1 adopted a genotyping scheme whereby we in-
Asta Pirskanen,1 Anna Rautanen,4 Marco Zucchelli,5 creased the density of markers used, with inter-
mediate analyses to guide further genotyping
Harriet Gullstén,2,5 Marina Leino,6 Harri Alenius,6 Tuula Petäys,7 (Fig. 1A). More specifically, if genetic associa-
Tari Haahtela,7 Annika Laitinen,3 Catherine Laprise,9 tion analysis suggested that a haplotype occurred
Thomas J. Hudson,10 Lauri A. Laitinen,8 Juha Kere*2,5 in patients more often than in controls, additional
markers were genotyped to either exclude or
Susceptibility to asthma depends on variation at an unknown number of genetic support the identity-by-descent of the haplotypes
loci. To identify susceptibility genes on chromosome 7p, we adopted a hierarchical observed in unrelated patients. For the haplo-
genotyping design, leading to the identification of a 133-kilobase risk-conferring types to be identical by descent, all newly typed
segment containing two genes. One of these coded for an orphan G protein– markers would have to be identically shared
coupled receptor named GPRA (G protein–coupled receptor for asthma suscepti- between them. The genotyping was done on 86
bility), which showed distinct distribution of protein isoforms between bronchial original genome scan families and an additional
biopsies from healthy and asthmatic individuals. In three cohorts from Finland and 103 trios (all together, 874 subjects) (10). Suc-
Canada, single nucleotide polymorphism–tagged haplotypes associated with high cessive rounds of genotyping and analysis by the
serum immunoglobulin E or asthma. The murine ortholog of GPRA was up-regu- haplotype pattern mining (HPM) algorithm (11)
lated in a mouse model of ovalbumin-induced inflammation. Together, these data suggested the strong association of a conserved
implicate GPRA in the pathogenesis of atopy and asthma. haplotype pattern spanning between NM51 and
SNP563704, separated by 46 kb (Fig. 1A). The
Asthma is a complex phenotype with a prov- lated traits have been undertaken (1). The HPM algorithm searches for allele patterns
en genetic component, and several projects to first published genome-wide scan in asthma shared between several haplotypes among large
map susceptibility genes for asthma and re- suggested six tentative genetic loci, among sets of unrelated haplotypes.
them chromosome 7p, which was then To fully explore the genetic variation in
1
GeneOS Limited, 00251 Helsinki, Finland. 2Depart- strongly implicated in a study of Finnish and associated haplotypes, we sequenced nonre-
ment of Medical Genetics, 3Department of Anatomy, Canadian families and confirmed in West petitive DNA segments in this interval (from
Biomedicum Helsinki, University of Helsinki, 00014
Helsinki, Finland. 4Finnish Genome Center, University Australian families (2– 4). To positionally position 506,401 to 638,799 in the public
of Helsinki, 00014 Helsinki, Finland. 5Department of clone asthma candidate genes on chromo- sequence NT_000380; all positions are given
Biosciences at Novum and Clinical Research Centre, some 7p, we studied the Finnish Kainuu sub- with reference to this sequence) in one patient
Karolinska Institutet, 14157 Huddinge, Sweden. 6De- population and considered three alternative homozygous for the susceptibility haplotype
partment of Industrial Hygiene and Toxicology, Finn-
ish Institute of Occupational Health, 00251 Helsinki, hypotheses. First, that only one copy of the and one control subject homozygous for
Finland. 7Department of Allergological Diseases, 8De- susceptibility allele may have survived in this the most common (nonrisk) haplotype. These
partment of Medicine, Helsinki University Central population, with long-conserved haplotypes sequences were then compared to the public
Hospital, 00029 Helsinki, Finland. 9Community being observed, as would be consistent with sequence (NT_000380). Two observations
Genomic Medicine Centre, Université du Québec à
Chicoutimi, Saguenay, Quebec G7H 2B1 Canada.
some previous findings (5–7). Second, there emerged from these analyses: first, the
10
McGill University and Genome Quebec Innovation might exist a founder effect, common to patient did not reveal a single instance of
Centre, Montreal, Quebec H3A 1A4 Canada. many European populations. This would be heterozygosity, confirming the identity-
*To whom correspondence should be addressed. E- consistent with the common disease/common by-descent of this chromosome segment.
mail: juha.kere@biosci.ki.se variant hypothesis, as observed for example Second, comparison of the susceptibility

300 9 APRIL 2004 VOL 304 SCIENCE www.sciencemag.org

Hemocyte-Mediated Shell Mineralization in the Eastern Oyster
Andrew S. Mount, A. P. Wheeler, Rajesh P. Paradkar and D. Snider

Science 304 (5668), 297-300.

DOI: 10.1126/science.1090506

Downloaded from http://science.sciencemag.org/ on February 4, 2019

ARTICLE TOOLS http://science.sciencemag.org/content/304/5668/297

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2004/04/07/304.5668.297.DC1
MATERIALS

RELATED file:/contentpending:yes
CONTENT

REFERENCES This article cites 22 articles, 3 of which you can access for free
http://science.sciencemag.org/content/304/5668/297#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.