From the Departments of Chemistry and
§ Biochemistry & Molecular Biology, University of New
Hampshire, Durham, New Hampshire 03824
Received for publication, July 28, 2000, and in revised form, November 13, 2000
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the bivalve mollusc Mytilus
edulis shell thickening occurs from the extrapallial (EP) fluid
wherein secreted shell matrix macromolecules are thought to
self-assemble into a framework that regulates the growth of
CaCO3 crystals, which eventually constitute ~95% of the
mature shell. Herein is the initial report on the purification and
characterization of a novel EP fluid glycoprotein, which is likely a
building block of the shell-soluble organic matrix. This primary EP
fluid protein comprises 56% of the total protein in the fluid and is
shown to be a dimer of 28,340 Da monomers estimated to be 14.3% by
weight carbohydrate. The protein is acidic (pI = 4.43) and rich in
histidine content (11.14%) as well as in Asx and Glx residues (25.15%
total). The N terminus exhibits an unusual repeat sequence of histidine
and aspartate residues that occur in pairs:
NPVDDHHDDHHDAPIVEHHD~. Ultracentrifugation and polyacrylamide
gel electrophoresis demonstrate that the protein binds calcium and in
so doing assembles into a series of higher order protomers, which
appear to have extended structures. Circular dichroism shows that the
protein-calcium binding/protomer formation is coupled to a significant
rearrangement in the protein's secondary structure in which there is a
major reduction in Biomineralization refers to the biological regulation of inorganic
mineral deposition (1-3). Generally, organisms use proteins and
polysaccharides to regulate mineralization by affecting the nucleation,
growth regulation, and growth cessation of the attendant mineral
crystals (4-9). Typically, mineralization takes place from a fluid
medium, which is biologically regulated in its content, supersaturated
with the ions being deposited, and spatially separated from its
surroundings (10, 11).
In the mollusc Mytilus edulis, biomineralization manifests
itself in exoskeletal shell formation. As a whole, mollusc shells are
95-99.9% by weight CaCO3 with the residual mass being
composed of biological macromolecules (11, 12). The shell of M. edulis is not homogeneous in its spatial distribution of
CaCO3. Instead, the shell has an outer prismatic and an
inner pearl-like nacreous layer, which contain CaCO3
deposited as crystals of calcite and aragonite, respectively. These two
mineralized layers are continuous, ultrastructurally unique, and reside
one on top of the other along the long axis of the shell (1).
Anatomically, the extrapallial
(EP)1 fluid fills the cavity
between the most outer visceral organ (the mantle) and the external shell. The EP fluid resides inside of the pallial line (site of mantle
attachment near the shell perimeter) and is the medium from which
prismatic layer thickening and nacre layer nucleation and growth occur
(13-15). The location and contents of the EP fluid implies that it
plays an essential role in mineralization/demineralization processes
in vivo. Despite this belief, the EP fluid has received little study, because most investigations of shell formation have focused on the shell itself (1).
Molluscan shell elongation and thickening occur throughout the life of
the animal and is thought to be an organic matrix-mediated process
controlled by a network of macromolecules composed of protein,
carbohydrate, and glycoprotein (1, 16-19). In M. edulis these macromolecules are thought to be synthesized in the mantle. Exterior to the pallial line, the mantle is in intimate contact with
the shell margin and as such is thought to secrete matrix material
directly onto the growing shell edge. Interior to the pallial line, the
mantle and shell are spatially separated by the EP cavity and
constituent EP fluid. At this location, organic material is secreted
from the mantle into the EP fluid where it is thought to self-assemble
into a matrix prior to mineral deposition (15, 17, 20).
Qualitative analysis of EP fluid has shown it to contain the
biomacromolecular materials found in the mature shell (protein, glycoprotein, carbohydrate, amino acids) (21, 22). The protein component of the EP fluid is heterogeneous, and certain fluid extracts
have demonstrated the ability to bind calcium (21, 22) and to inhibit
in vitro CaCO3 crystallization in the same manner as fractions of shell matrix protein (13). Studies of the ion
content of the EP fluid show that its composition differs from that of
the animal blood and the surrounding sea water (23-25). Examination of
the pH, [CO2], and [Ca2+] as a function of
shell opening and closure reveal the EP fluid to be a dynamic
physiological medium, which may play a role in pH regulation within the
animal (26).
Given the implied significance of the EP fluid in shell mineralization,
the present study was undertaken to purify and characterize its major
protein component. We show here for the first time that the major
protein component of the EP fluid is a 56,000 molecular weight
glycoprotein that is a homodimer composed of 14.3% carbohydrate. This
primary EP fluid protein reversibly binds calcium in a manner that
leads to significant changes in its secondary structure and to the
formation of distinct multimeric species composed of the constituent
monomers. The major EP fluid protein appears to be a building block of
the soluble organic matrix of the shell. A hypothesis for the mode of
matrix self-association is presented.
EP Fluid Extraction--
Mussels were obtained from coastal New
Hampshire waters or from Great Eastern Mussel aquaculture (Tenants
Harbor, ME). Live animals were kept on ice for <1 h during
extrapallial fluid extraction. The shell cavity was accessed by slicing
the adductor muscle with a scalpel. With the shell cavity open, excess
water was removed by tilting the animal upright and letting it drain
onto a paper towel. The needle of a 500-µl gas-tight syringe was
placed bevel down on the shell margin and slid beneath the mantle at
the pallial line into the extrapallial space for EP fluid extraction.
Care was taken not to disturb the mantle/shell attachment. Once
extracted, the EP fluid was placed on ice in a polypropylene vial. To
ensure that the EP fluid was not contaminated by blood from the animal, fluid was also extracted directly from the extrapallial space by the
careful drilling of a hole through the shell. Identical protein banding
patterns on PAGE were observed for fluid obtained by either extraction
method. Individual animals yielded an average of 300 µl of fluid;
however, variation from practically nothing to in excess of 500 µl
was observed.
Unless otherwise mentioned, the protein buffer solution used throughout
was 20 mM (3-(N-morpholine)propanesulfonic acid
(MOPS), pH 7.5, in 0.1 M KCl, and all steps were performed
at 4 °C. The buffer was passed through a 1.5- × 25-cm column of
Chelex 100 metal ion chelation resin (Sigma) and then filtered/degassed
by suction filtration through a 0.45-µm pore Whatman filter.
Immediately after extraction, the EP fluid was centrifuged (3500 × g) and the supernatant was retained and dialyzed against 1 liter of buffer 85% saturated with ammonium sulfate (Sigma) for at
least 4 h using Spectra/Por 6000-8000 molecular weight cut off
cellulose dialysis tubing. Precipitated protein was collected by
centrifugation at 12,500 × g for 10 min and dissolved
in a minimum volume of buffer followed by dialysis against buffer to remove residual (NH4)2SO4.
Chromatography--
Cation exchange chromatography was performed
on a 2- × 50-cm column of Sephadex CM-50 resin (Sigma) using a protein
buffer mobile phase and a gravity flow rate of 0.5 ml/min without a
salt gradient. Anion exchange chromatography was then performed on three 0.7- × 2.5-cm Amersham Pharmacia Biotech HiTrap Q columns connected in series using a buffer mobile phase with a flow rate of 1.5 ml/min and a 200-ml, 0.1-0.5 M KCl linear salt gradient for protein elution. Finally, size exclusion chromatography was performed on 1.5- × 57-cm Sepharose 6B-CL resin (Sigma) with a buffer
mobile phase and gravity flow rate of 15 ml/h. All chromatography was
performed at room temperature with column effluent monitored at 280 nm.
Fractions of 5 ml were collected.
Protein Analysis--
Electrophoresis was carried out on a
Hoefer Mighty Small II electrophoresis unit using electrophoresis grade
reagents from Sigma, reagent grade ammonium persulfate from J. T. Baker Chemical, low molecular weight SDS-PAGE markers from Bio-Rad, and
ampholines from Amersham Pharmacia Biotech. Electrophoretic protocols
(native PAGE, SDS-PAGE, isoelectric focusing) using Coomassie Brilliant Blue R-250 protein staining were taken from the Hoefer Scientific Instruments Manual 1992-1993. Glycoprotein-specific gel staining was
done by the method of Rauchsen (27). Gels were 1.5 mm thick, and
experiments were conducted using a limiting current of 20 mA/gel. Gel
scanning was performed on a Hoefer Scientific GS 300 transmittance/reflectance scanning densitometer interfaced to a MINC-23
computer (Digital Equipment Corp.). Photographs of gels were scanned
using a UMAX UC300 color scanner interfaced to a MacIntosh Quadra 700 computer. Adobe Photoshop 3.0 was used for image processing.
Protein concentration was estimated by a Bio-Rad protein assay using
bovine serum albumin as a standard following the method of Bradford
(28). Determination/quantitation of carbohydrate moieties was done by
the method of Dubois et al. (29) using D-galactose in buffer to construct 5- to 30-µg standard
calibration curves and SDS-PAGE analysis of EP protein before and after
digestion with
peptide-N4-(N-acetyl-
The protein molecular weight was determined by matrix-assisted, laser
desorption ionization, time-of-flight mass spectrometry on a Perceptive
Biosystems, Voyager-Elite biospectrometry workstation. Instrument
calibration was achieved using an insulin mass standard. Analysis of
the NH4HCO3 lyophilized EP fluid protein sample
was performed using an accelerating voltage of 30,000 V, a pressure of
9.50 × 10
Amino acid analysis samples were dialyzed against 5.0 mM
MOPS, pH 7.5, 0.1 M KCl and then hydrolyzed for 20 h
in 6 N HCl, 0.05% mercaptoethanol, 0.02% phenol, at
115 °C. At AAA Laboratory (Mercer Island, WA), amino acid
separation and quantitation were performed on a Beckman 7300 analyzer
using System Gold software and Beckman buffers following the
ion-exchange method of Moore and Stein (32). At the University of
Michigan laboratory, the analysis was performed on an Applied
Biosystems 420H amino acid analyzer. The sample for N-terminal amino
acid sequence analysis was prepared by dialysis of purified protein
against a 10% (v/v) solution of acetic acid. The analysis was done on
an Applied Biosystems Inc., 475A protein sequencer, by the method of
automated Edman degradation (33).
Ultracentrifugation Analysis--
Sedimentation equilibrium and
velocity experiments were performed on a Beckman XLA analytical
centrifuge equipped with Rayleigh interference optics (34). Experiments
were conducted at 20 °C using a four-hole titanium rotor spinning at
40,000 rpm. The sample cells were two-channel, charcoal-filled, and
equipped with Epon centerpieces and sapphire windows. The buffer used
in all sedimentation velocity analyses was 20 mM Tris, pH
7.5, 0.1 M KCl; where mentioned, samples were analyzed with
the addition of CaCl2 to the buffer solution. Protein
concentrations ranging from 0.25 to 2.0 mg/ml were employed to
investigate possible mass action effects in the centrifugation data.
The method of Stafford (35) was used to obtain sedimentation
coefficient distributions (g(s*)) from the time derivative
of the concentration distributions (dc/dt), and sedimentation and diffusion coefficients were calculated from the line
widths and positions of Gaussian fits to the g(s*)
versus s data (35). The partial specific volume
was calculated from the amino acid (vide infra) composition
of the protein using the software program Sednterp (36) and
assuming a glycan composed of hexoses (30, 31).
Circular Dichroism--
Circular dichroism (CD) experiments were
done on a JASCO J700 circular dichroism spectropolarimeter calibrated
using a 0.06% (+)-10-camphorsulfonic acid solution (Aldrich). CD
spectra of EP fluid protein (0.65 mg/ml) in buffer with and without 10 mM Ca2+, along with matching buffers, were
obtained using a 0.05-mm sample cell. Instrument optics and sample
chamber were continually flushed with 30 liters/min of dry
N2 gas. The instrument settings were; scan range, 180-260
nm; scan rate, 1 nm/min; wavelength step, 0.5 nm; sensitivity, 5 millidegrees; response time, 16 s. Protein secondary
structural analysis of the CD spectra (182-260 nm) was done using the
Self-Consistent Method (SELCON) program of Sreerama and Woody (37)
using the fitting constraints recommended by the authors.
Raw EP Fluid Analysis and Purification of Primary Protein
Component--
Protein and carbohydrate concentrations in untreated EP
fluid samples were determined for nine batches of fluid yielding an average of 4.3 ± 1.8 mg/ml for protein and 3.8 ± 1.2 mg/ml
for carbohydrate. The distribution of fluid proteins in samples taken from both batch extracts and individual animals was studied by native
and SDS-PAGE. Fig. 1A shows
the consistent distribution of at least six protein bands with the
persistent appearance of a major protein band (marked by an
arrow in lanes 1-6, Fig. 1A), corresponding to 56 ± 15% (n = 3) of the total
fluid protein based on the integrated band intensity. This major
protein fraction, hereafter designated the EP protein, was the
component targeted for purification.
Chromatographically separated protein fractions (see "Experimental
Procedures") were analyzed by electrophoresis as a means of tracking
the protein of interest during purification. Fig. 1B shows a
15% native PAGE gel loaded with duplicate samples of protein, at the
various stages of purification, and a duplicate loading of protein
standards. (A 15% acrylamide gel is shown in Fig. 1B, where
diffusion of the EP glycoprotein is minimized; however, similar band
patterns are seen in 7 and 10% gels.) Upon completion of
electrophoresis, the gel was sliced in half and one half (lanes
6-10) was stained for protein, while the other half (lanes
1-5) was stained specifically for glycoprotein. The right
half of the gel in Fig. 1B demonstrates the
effective purification of the major EP protein to a single band
(lane 7), and the left half (lanes
1-4) shows that the major EP protein stains positive for
glycoprotein. Of the standard proteins used (lanes 5 and
6), only transferrin, which is ~5% by weight
carbohydrate, stained in both gel halves. The purified EP protein was
determined to be 14.3% carbohydrate (see "Experimental
Procedures").
N-terminal Amino Acid Sequence and Trypsin Digest--
The
20-amino acid sequence of the N terminus of the EP protein reveals an
interesting repeat pattern of histidine and aspartate residues,
viz.: NPVDDHHDDHHDAPIVEHHD~.
The sequence was entered into several protein data
bases,2 none of which
produced a protein match containing a compatible sequence. Mass
fragments of 943.8, 1056.9, 1093.8, 1112.2, 1150.1, 1241.3, 1299.6, 1337.7, 1429.8, 1584.0, 1700.0, 1731.4, 1805.8, 1984.6, 2131.2, 2164.7, 2201.9, 2260.2, 2685.2, and 3047.3 were produced by trypsin digestion.
Amino Acid Analysis--
The averaged results from the amino acid
analyses of three separate EP protein purifications are shown in Table
I. The data are presented as percent
molar composition for each residue, as well as in residues/protein
based on a protein molecular weight of 28,340 (vide infra)
containing 14.3% carbohydrate. Table I shows that EP protein contains
a significant amount of acidic (Asx, Glx) residues. This finding is
consistent with of those of shell-soluble organic matrix protein
fractions (18, 38-40). The protein is also rich in histidine content
(11.14%), a property also reflected in the N-terminal sequence where 6 of the 20 amino acids are histidine residues that occur in pairs
(vide supra).
Isoelectric Focusing--
Results from isoelectric focusing showed
the EP protein to be acidic with at least six distinguishable isoforms,
which are focused as a tight stack of bands with isoelectric point (pI) values ranging from 4.08 to 4.67 with a median pI value of 4.43. Although the source of this charge heterogeneity is not known, such
occurrences are common for glycoproteins (41).
Molecular Weight Determination--
Fig.
2 shows the time-of-flight mass spectrum
of an EP protein sample. The peak centered at 28,340 mass units is the
ionized EP protein (M/1+). The peak centered at 14,200 mass
units is the doubly ionized protein species (M/2+). The
peak breadth indicates microheterogeneity with minor peaks at 27,000, 27,840, and 28,960 Da, in addition to the major 28,340-Da peak. The
observation of heterogeneity in the extent of glycosylation is common
for glycoproteins (41), and the addition or deletion of monosaccharide
moieties may account for the observed mass heterogeneity of the EP
protein. The same protein electrophoretic migration was observed on
15% SDS-PAGE under either reducing or nonreducing conditions (data not
shown), indicating a single type of subunit and lack of intersubunit
disulfide bonds.3
Fig. 3 shows an overlay of a size
exclusion chromatogram of purified EP protein with that of a set of
protein standards. A plot of the log molecular weight versus
the size exclusion elution coefficient Kav is
shown in the Fig. 3 (inset) from which a molecular weight of
52,600 ± 1100 (n = 8) is estimated for the native
protein. The resultant molecular weight is approximately double the
value from mass spectrometry, a result suggesting that the EP protein in its native state is a homodimer of the 28,350-Da subunits.
To confirm this finding, sedimentation velocity and equilibrium
measurements were also carried out on the native protein. Fig.
4 shows the sedimentation distribution
coefficient, g(s), versus the apparent
sedimentation coefficient, s, for the protein in Tris
buffer. Two components are present, a major species at s
A sedimentation coefficient of 4.7 is incompatible with a protein of
Mr
Sedimentation equilibrium experiments produced Z average
apparent molecular weights that were observed to decrease with
increasing rotor speed due to the presence of the two nonequilibrating
species. Nevertheless, at higher rotor speeds, where the heavier
s = 10.5 component sediments out, the apparent
molecular weight reached a value of Mr ~ 52,000 (Fig. 5). This value, although
clearly less accurate, corresponds well with the molecular weights
derived from sedimentation velocity and size exclusion chromatography. Therefore, we conclude that the s Protein Interaction with Calcium--
Protein behavior in the
presence of calcium was examined by electrophoresis experiments wherein
all solutions used in the production and running of gels were brought
to 10 mM Ca2+, the physiological calcium
concentration in the EP fluid (22). Fig.
6 shows a 10% native gel (gel
A) with the EP protein migrating two-thirds of the way through the
gel as a single band; gel B corresponds to an identical set
of samples run on a Ca2+-doped gel. Gel B shows
that in the presence of the Ca2+ the single major EP fluid
protein band in gel A is converted into several more slowly
migrating protein bands, which remain in the 3% stacking gel of
gel B. The gels demonstrate that, in the presence of
calcium, the single EP protein band is converted into several larger,
more slowly migrating species. Other divalent metal ions, including
Mg2+, Mn2+, and Cd2+, have similar
effects to those of Ca2+.
Sedimentation Velocity in the Presence of Calcium--
Fig.
7A shows traces of EP protein
sample (0.9 mg/ml) analyzed before and immediately after the addition
of 10 mM calcium. The formation of multiple protomeric
species is evident. It is curious that peak A from the dimer shifts
from 4.7 to ~4.0 s in the presence of calcium, an
observation suggesting that Ca2+ binding produces a less
compact structure in the intermediate ~4.0-s species prior
to its assembly into protomeric species as shown (Fig. 7B).
Fig. 7B also shows the result of the multiple Gaussian curve
fit to the species distribution pattern of a sedimentation velocity
cell containing 0.6 mg/ml protein in 20 mM Tris, pH 7.5, brought to 10 mM Ca2+ and allowed to stand for
24 h. Both Figs. 7A and 7B clearly
demonstrate the presence of at least four, distinguishable, protomeric
species having s values of ~10, 15, 19, and 22 (labeled B-E) as well as the intermediate
~4.0-s species. Also, these figures show that the
stoichiometry of protomer formation is finite, leading to species that
are distinguishable and reasonably uniform in molecular size and that,
given sufficient time and a physiological concentration of
Ca2+, the parent peak (peak A) fully converts
protomers. It is also apparent from the data in these figures that the
breadth of the peaks from B through E is
increasing, which is opposite of what would be expected if these peaks
represented pure components. Because there was no definitive evidence
for mass action equilibrium between species (vide infra),
the excess peak spreading probably reflects increased
microheterogeneity in the structures of the more rapidly sedimenting
components. Such heterogeneity with respect to self-association may
reflect the trapping of stable or mesostable protomer conformation
states upon calcification of the protein. That some of the protomeric
species represent different structural assemblies involving the same
number of subunits is also a possibility. The anomalous peak widths of
the protomers preclude the determination of molecular weights of these
species by sedimentation velocity measurements.
Although protomer assembly is calcium-dependent, analysis
at higher calcium concentration (100 or 500 mM) did not
change the species distribution profile from that in Fig.
7B. Additionally, Fig. 8 shows
that the relative abundances of protomers, analyzed from samples taken
from a single preparation and run at different concentrations, remains
essentially constant regardless concentration (see Fig. 8 legend).
Repeated attempts to demonstrate simple mass action behavior through
diluting and reconcentrating samples failed to clearly exhibit the
reversible redistribution of species expected. Therefore, although
there may be some small portion of the protein capable of participating
in reversible association, our evidence indicates that the bulk of the
material forms stable complexes on the time scale of our experiments.
However, some variation in protomer species distribution was seen
between different protein preparations for reasons that are not
understood at this time.
Fig. 8 also shows a pronounced dependence of the sedimentation
coefficient on the concentration of the protomeric species. The
apparent values of the sedimentation coefficient s as a
function of protein concentration for species A through D is
illustrated in Fig. 9 from which
s20,w0 values of 4.71 ± 0.06, 11.8 ± 0.3, 17.4 ± 0.3, and 25.2 ± 0.3 for the
four species were obtained after correcting for the density and
viscosity of the buffer. The data for the A species were obtained in
the absence of calcium and for the B through D species in the presence
of calcium. (The drop in sedimentation coefficients at lowest
concentration points in Fig. 9 reflect the difficulty in accurately
measuring s values of the most dilute solutions; in other
experiments a small increase in s is seen at the lowest concentrations.)
The rather large variation in s with protein concentration
(Fig. 9) implies that the protomeric species have elongated structures, more so than the parent A species. For spherical structures, the dependence of s on concentration is predicted to be small,
<1%/mg/ml (42), contrary to what is observed. The dependence of
s on concentration for the C and D species is rather
pronounced, namely
Attempts to separate the multimeric species, B through E, were made by
incubating the purified EP protein in 10 mM
Ca2+ buffer and then running the sample over a Sepharose
6B-CL size exclusion chromatography column using a 10 mM
Ca2+ buffered mobile phase. Analysis on 5% native PAGE
gels of the peaks produced by this separation showed that, although
enrichment of a given multimer band could be observed, isolation of
distinct multimers was never realized. This result may be due to the
inability to fully separate the protomers on size exclusion, because of their elongated structures or because of heterogeneity in the associated species (vide supra).
Circular Dichroism--
CD spectra of EP fluid protein in the
presence and absence of calcium demonstrate a significant change in
secondary structure accompanying Ca2+ binding (Fig.
10). Analysis of secondary structure
rearrangement is summarized in Table II
and shows that Ca2+ binding causes an 18% increase in the
estimated amount of This study reinforces and extends previous work (22, 24, 25)
in finding the EP fluid rich in macromolecular content (~4
mg/ml in both protein and carbohydrate). Inventory of the fluid
proteins (Fig. 1) was consistent, showing a single, dominant glycoprotein (14.3% carbohydrate) in all preparations. This
glycoprotein was purified to homogeneity on the basis of PAGE and a
single N-terminal amino acid sequence. The native state of the EP
glycoprotein in the absence of calcium is a dimer (A species) of the
28,340-Da monomers (Figs. 2-5) and has a pI of 4.43.
Sedimentation velocity analysis in the absence of added calcium shows
that the EP protein species A associates to form a limited amount of B
(Fig. 4). However, upon addition of calcium, the EP protein forms a
finite number of larger protomers, B through E, which appear to have a
discrete stoichiometry in their assembly (Figs. 6 and 7). Circular
dichroic spectra (Fig. 10) of the EP protein taken in calcium-free and
calcium-containing (10 mM, the physiological concentration)
buffer solutions show that the protein undergoes a significant
structural change upon calcium binding and that the protein has a much
greater Interestingly, although calcium binding is critical to significant
protomer assembly, a fixed concentration of protein (0.6 mg/ml) in a
10, 100, or a 500 mM Ca2+ solution,
respectively, produced the same species distribution profile
as that seen in Fig. 7B, indicating that a 50-fold increase in calcium concentration does not affect the multimeric distribution pattern. This phenomenon implies that the formation of higher order
multimers (peaks C, D, and E)
is not the result of additional calcium-protein interaction but rather
the result of protein-protein interactions (Fig. 7B).
Similarly, the distribution of protomer species at a fixed
Ca2+ concentration of 10 mM is relatively
unaffected by the protein concentration (Fig. 8).
The ability to self-assemble is one of the most intriguing properties
ascribed to the macromolecules of the organic matrix. Little is known
about how newly secreted matrix materials assemble into a functioning,
precisely arranged network, capable of directing the construction of
mineral formations with exacting detail. The novel EP fluid
glycoprotein described here may provide important clues regarding how
the organic matrix is developed. Soluble EP protein secreted into the
10 mM calcium environment of the EP fluid take on a
definitive secondary structure as a result of the calcium binding as
demonstrated by the CD measurements (Fig. 10). This characteristic
structural arrangement enables the Ca-protein complexes to form
protein-protein associations allowing the protein to organize into
large protomeric assembles (Fig. 7). Furthermore, the centrifugation
data (Fig. 7) implies that Ca2+ binding is largely involved
in the conversion of the A species to the B species. Taken together,
these results suggest that the alteration in secondary structure that
results in, or accompanies, the conversion from A to B also allows for
B-B or A-B (protein-protein) interactions, resulting in the formation
of the C, D, and E protomers. This hypothesis is supported by the fact
that removal of Ca2+ by dialysis results in protomer
conversion back to the A species. Much further work is needed to
elucidate the structural properties of these very interesting and
complex protein assemblies and their interaction with calcium. Such
studies are currently underway.
Similarities between the properties of the EP protein described here
and general characteristics of soluble matrix proteins from other
systems are also noted. It is largely accepted that the soluble organic
matrix (SOM) is comprised of proteins rich in aspartic acid and
glutamic acid (18, 38-40) with acidic glycoproteins considered to be a
prominent matrix component (17, 20, 43). Asx and Glx are the most
abundant residues in the EP protein (Table I) and based on the
N-terminal amino acid sequence data and the acidic isoelectric point of
4.43, both residues seem to be present in predominantly acidic form.
Analyses for phosphorylation and sulfation of the EP protein were both
negative, indicating that the EP protein is not likely to be a
homologue of the calcium-sequestering phosphoprotein particles found at
the inner shell lamella of some bivalve species. This finding is in
agreement with prior studies on this subject (44, 45).
A SOM primary protein sequence based on amino acid analysis which has
received much attention is the repeat:
DXDXDX (X = glycine or
serine). This sequence is proposed to act as a template for CaCO3 crystal growth by epitaxy (46, 47). Although not an exact match, the N-terminal sequence of the EP fluid protein (see "Results") also shows a pattern of consistently spaced acidic residues. The fact that the two sequences are not present in sequence data banks but seem to occur in SOM proteins and now in the EP fluid
protein strengthens the notion that the EP fluid protein ultimately
functions as a matrix protein. This idea is consistent with the
anatomical location of the protein, its relatively large abundance, and
by its behavior in the presence of calcium. Work is currently underway
in an attempt to identify the EP protein within the SOM; however, the
cross-linked multicomponent nature the SOM makes it a difficult
material to study.
The EP protein resembles, in some aspects, a recently reported blood
glycoprotein from M. edulis (48) in that both are rich in
histidine, have similar dimer molecular weights (~56
versus ~61 kDa), bind Ca2+, and are acidic
proteins. Their amino acid compositions are also similar. However, the
blood protein is composed of two types of subunits of molecular
masses, ~29 and ~35-39 kDa, whereas the EP protein consists
of only one type of 26-kDa subunit. The blood protein has been
postulated to be a metal ion transporter involved in the accumulation
of metals by the animal; it binds Cd2+ in addition to
Ca2+ (48), as does the EP protein (vide
supra).
In summary, the extrapallial fluid of M. edulis contains a
structurally unique dimeric glycoprotein (A), which has been purified and partially characterized. This glycoprotein binds calcium to form a
protomer (B), a process that is accompanied by a significant secondary
structural rearrangement in the protein. This structural rearrangement
appears to trigger a self association into higher order protomers (C,
D, and E) that follows a stoichiometric pattern yet to be determined
and is fully reversible upon the removal of calcium. Based on its
characteristic features, inherent similarities with soluble organic
matrix proteins and its anatomical location, this study suggests that
the EP protein is a precursor or building block to the soluble organic
matrix of the shell.
-sheet with an associated increase in
-helical
content of the protein. A model for shell organic matrix self-assembly
is proposed.
INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
-glucosaminyl)asparagine
amidase F. A value of 14.3% carbohydrate was determined (30) and
confirmed by mass spectrometry (31). The N-linked glycan has
a mass of ~4000 Da and is composed of hexoses (30, 31).
8 torr, and a low mass cutoff of
500.0.
RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
View larger version (107K):
[in a new window]
Fig. 1.
A, 15% native PAGE gel of fluid samples
extracted in October, August, July 1993. Lanes in A:
1, 20 µl of Raw Fluid (RF) Oct.; 2, 10 µl of
RF Oct.; 3, 20 µl of RF Aug.; 4, 10 µl of RF
Aug.; 5, 20 µl of RF Jul.; 6, 10 µl of RF
Jul.; 7, protein standards (from top to
bottom), transferrin (80,000); albumin (67,000); carbonic
anhydrase (34,000); myoglobin (17,500). The arrow denotes
the major protein band. B, 15% native PAGE gel stained for
glycoprotein (left) and indiscriminately for protein
(right). Lanes in B: 1 and
10, after cation exchange; 2 and 9,
Raw fluid; 3 and 8, after anion exchange;
4 and 7, after size exclusion (most purified
protein); 5 and 6, protein standards (same as in
A).
Amino acid composition
View larger version (15K):
[in a new window]
Fig. 2.
Time-of-flight mass spectrum of purified EP
fluid protein.
View larger version (21K):
[in a new window]
Fig. 3.
Overlay of a size exclusion chromatogram of
purified EP protein and size exclusion standards. Peak
1, thyroglobulin (670,000); peak 2, gamma globulin
(158,000); peak 3, ovalbumin (44,000); peak 4,
myoglobin (17,000); peak 5, vitamin B-12 (1,350).
Inset, plot of log molecular weight versus
elution coefficient Kav for protein standards.
EP protein (open circle) estimated to have molecular weight
of 52,600.
4.7 and a minor one near s
10.5 that we ascribe to an assembly of the s
4.7 species (vide infra). The minor species typically represented only 7-20% of the total, depending on the protein preparation. Lack of mass action effects over a 16-fold concentration range of protein indicated that equilibration between the major and
minor species is slow during the 24-h time period between dilution of
the sample and acquisition of the centrifugation data. That both
species are derived from a single type of subunit is evidenced by the
presence of a single band in SDS-PAGE over a wide range of sample
loadings on the gel.
View larger version (24K):
[in a new window]
Fig. 4.
Distribution g(s*)
versus s of the EP protein in the absence
of added calcium. Experimental data (solid line);
Gaussian fit (dotted line). Inset, apparent
molecular weight versus protein concentration. Conditions:
0.45 mg of protein/ml, 20 mM Tris, 0.1 M KCl,
pH 7.5.
28,000. Accordingly, the apparent
molecular weight of the 4.7-s component was determined.
Gaussian curve fitting of the data (e.g. Fig. 4) for protein
concentrations ranging from 0.055 to 0.9 mg/ml produced a diffusion
coefficients for the s = 4.7 species of D = 7.3 × 10
11 to 8.1 × 10
11
m2/s. Apparent molecular weights were calculated from the
Svedberg equation using the apparent diffusion coefficients,
D, at each protein concentration and the partial specific
volume v = 0.7089 cm3/g for the
protein (see "Experimental Procedures"). Over a wide range of
concentrations (Fig. 4, inset), the molecular weights determined from the 4.7-s peak were ~52,000. Although
there appears to be a very slight decrease in the apparent molecular
weight with decreasing concentration, the uncertainties in the fitting of the lower concentration data preclude interpretation of this observation. Thus, the molecular weights determined in this manner agree with the value from size exclusion chromatography and are in
accord with the protein being a dimer in its native state. (A similar
determination of the molecular weight of the s
10.5 component by sedimentation velocity was not possible due to variability in the experimental value of D obtained from curve fitting
of the small peak of this minor species.)
4.7 species is a
dimer.
View larger version (14K):
[in a new window]
Fig. 5.
Apparent molecular weight from sedimentation
equilibrium as a function of rotor speed. Conditions are the same
as in Fig. 4.
View larger version (45K):
[in a new window]
Fig. 6.
Native PAGE gels in the absence
(A) and presence (B) of 10 mM Ca2+. All aspects of the gels are
identical. Lanes (for both gels): 1, protein
standards (from top to bottom, transferrin
(80,000); albumin (67,000); carbonic anhydrase (34,000); myoglobin (17,
500)); 2-9, EP fluid protein. SG and
RG denote the 3% stacking and 10% running gels,
respectively.
View larger version (15K):
[in a new window]
Fig. 7.
Distribution g(s*)
versus s, the apparent sedimentation
coefficient. A, sedimentation velocity analysis
of samples containing 0.9 mg/ml without calcium (dotted
line) and in the presence of 10 mM Ca2+
(solid line) added just prior to analysis. B,
0.65 mg/ml protein sample in 10 mM Ca2+
equilibrated for 24 h prior to analysis. Solvent for all cells was
20 mM Tris, pH 7.5, 0.1 M KCL. Gaussian
components are shown.
View larger version (18K):
[in a new window]
Fig. 8.
Distribution g(s*)
versus s, the apparent sedimentation
coefficient as a function of protein concentration. The protein
concentrations were 0.25, 0.5, 1.0, and 2.0 mg/ml in 20 mM
Tris, 0.1 M KCl, pH 7.5, with 10 mM
CaCl2. From Gaussian curve fitting, the percentages of the
A, B, C, and D components, respectively, are: 2.6 ± 0.3, 34.7 ± 0.2, 30.4 ± 0.3, and 35.0 ± 0.5% for the 2.0 mg/ml sample; 2.1 ± 0.1, 32.4 ± 0.1, 35.4 ± 0.2, and
33.2 ± 0.2% for the 1.0 mg/ml sample; and 2.5 ± 0.4, 29.0 ± 0.3, 28.8 ± 0.6, and 44.0 ± 0.7% for the 0.5 mg/ml sample.
View larger version (16K):
[in a new window]
Fig. 9.
Apparent sedimentation coefficient
s as a function of protein concentration for the A, B,
C, and D species. Conditions are given in Fig. 8. The data for
species A is in the absence of 10 mM CaCl2; the
others contain 10 mM calcium.
7.0 ± 1.2 and
6.4 ± 0.8%/mg/ml
for the two components, respectively (Fig. 9). The slopes of the
lines for the A and B species are zero within experimental error. If we
assume that the A species dimer has a molecular weight of 56,000 (twice
the value from mass spectrometry) and a partial specific volume of
0.7089 cm3/g, we calculate a Stoke's radius of 3.05 ± 0.10 nm. For hydrations of 0.3 to 0.4 H2O/g protein, an
axial ratio r
3 is predicted for either a prolate or
oblate ellipsoid. We conclude that the A species is probably somewhat
asymmetric but much less than the C and D protomers.
-helix with concurrent reduction in the
-sheet. The increase in a negative absorption at 222 nm and positive
adsorption at 190 nm are hallmark features of
-helix formation. Thus
the protomeric forms (B through E) contain more
-helical structure
than the parent ~4.7-s species (A species).
View larger version (18K):
[in a new window]
Fig. 10.
CD spectra of a 0.65 mg/ml protein solution
of 20 mm MOPS, pH 7.5, 0.1 M KCl in the absence
(dotted line), and in the presence (solid
line) of 10 mM Ca2+ after incubating
the sample for more than 24 h. The 10 mM
Ca2+ sample represents complete conversion to protomeric
species B through E.
Secondary structure composition from CD spectral analysis
DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES
-helical content with concurrent reduction in the amount of
beta sheet (Table II).
![]() |
ACKNOWLEDGEMENTS |
---|
We thank Dr. Paul Matsudaira for his assistance with the protein sequence determination, Drs. Robert MacColl and John Osterhout for their help with the circular dichroism measurements, John K. Grady and Kari L. Hartman for preparing some of the figures and helping with the centrifuge data analysis, and Dr. Robert Trimble for comments on the manuscript.
![]() |
FOOTNOTES |
---|
* This work was supported by Grant R37-GM20194 from the National Institute of General Medical Sciences (to N. D. C) and by Grants BIR-9314040 and DBI-9876582 from the National Science Foundation (to T. M. L.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence should be addressed: Dept. of Chemistry, Parsons Hall, University of New Hampshire, Durham, NH 03824. Tel./Fax: 603-862-2520; E-mail: ndc@cisunix.unh.edu.
Published, JBC Papers in Press, November 17, 2000, DOI 10.1074/jbc.M006803200
2 Data bases of NCBI, Rutgers Protein Data Bank, Swiss-Prot, TrEMBL, Genpept, GenBankTM, and Kabat Sequences.
3 An apparent molecular weight of 37,300 ± 1,800 was obtained from SDS-PAGE using the protein standards: lysozyme (14,400), trypsin inhibitor (21,500), carbonic anhydrase (31,000), ovalbumin (45,000), and serum albumin (66,200). Because glycans do not bind SDS, erroneously high glycoprotein molecular weights are commonly observed by SDS-PAGE (41).
![]() |
ABBREVIATIONS |
---|
The abbreviations used are: EP, extrapallial; PAGE, polyacrylamide gel electrophoresis; pI, isoelectric point; g(s*), sedimentation coefficient distribution; s, apparent sedimentation coefficient; s20, w0 sedimentation coefficient corrected for density, viscosity, and solute concentration; CD, circular dichroism; SOM, soluble organic matrix; MOPS, (3-(N-morpholine)propanesulfonic acid.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
1. | Lowenstam, H. A., and Weiner, S. (1989) On Biomineralization , pp. 7-24, Oxford University Press, New York |
2. | Weiner, S. (1986) CRC Crit. Rev. Biochem. 20, 365-408[Medline] [Order article via Infotrieve] |
3. | Wheeler, A. P., and Sikes, C. S. (1989) in Biomineralization (Mann, S. , Webb, J. , and Williams, R. J. P., eds) , pp. 95-131, VCH Publishers, New York |
4. | Belcher, A. M., Wu, X. H., Christensen, R. J., Hansma, P. K., Stucky, G. D., and Morse, D. E. (1996) Nature 381, 56-57[CrossRef] |
5. | Wierzbicki, A., Sikes, C. S., Madura, J. D., and Drake, B. (1994) Calcif. Tissue Int. 54, 133-141[Medline] [Order article via Infotrieve] |
6. | Addadi, L., and Weiner, S. (1992) Angew. Chem. Int. Ed. Engl. 31, 153-169 |
7. | Mann, S., Heywood, B. R., Rajam, S., and Birchall, J. D. (1988) Nature 334, 692-695[CrossRef] |
8. | Addadi, L., and Weiner, S. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 4110-4114[Abstract] |
9. | Watabe, N., and Wilbur, K. M. (1960) Nature 188, 334 |
10. | Mann, S. (1988) Nature 332, 119-124 |
11. | Wilbur, K. M. (1984) Am. Zool. 24, 839-845 |
12. | Hare, P. E. (1963) Science 139, 216-217 |
13. | Wilbur, K. M., and Bernhardt, A. M. (1984) Biol. Bull. 166, 251-259 |
14. | Eyster, L. S., and Morse, M. P. (1984) Am. Zool. 24, 871-882 |
15. | Young, S. D., Crenshaw, M. A., and King, D. B. (1977) Marine Biol. 41, 253-257 |
16. |
Gunthrope, M. E.,
Sikes, C. S.,
and Wheeler, A. P.
(1990)
Biol. Bull.
179,
191-200 |
17. | Krampitz, G. (1982) in Biological Mineralization and Demineralization (Nancollas, G. H., ed) , pp. 219-232, Springer-Verlag, New York |
18. | Weiner, S. (1983) Biochemistry 22, 4139-4145 |
19. | Wheeler, A. P., and Sikes, C. S. (1984) Am. Zool. 24, 933-944 |
20. | Weiner, S. (1984) Am. Zool. 24, 945-951 |
21. | Weiner, S., Lowenstam, H. A., and Hood, L. (1977) J. Exp. Mar. Biol. Ecol. 30, 45-51 |
22. | Misogaines, M., and Chasteen, N. D. (1979) Anal. Biochem. 100, 324-334[Medline] [Order article via Infotrieve] |
23. | Wada, K., and Fujijuki, T. (1980) in Mechanisms of Mineralization in the Invertebrates and Plants (Wataba, N. , and Wlibur, K. M., eds) , pp. 175-190, University of South Carolina Press, Columbia, SC |
24. | Crenshaw, M. A. (1972) Biol. Bull. 143, 506-512 |
25. | Peitrzak, J. E., Bates, J. M., and Scott, R. M. (1976) Hydrobiologia 50, 89-93 |
26. | Crenshaw, M. A., and Neff, J. M. (1969) Am. Zool. 9, 881-885 |
27. | Gerard, C. (1990) Methods Enzymol. 182, 536-537 |
28. | Bradford, M. (1976) Anal. Biochem. 72, 248-254[CrossRef][Medline] [Order article via Infotrieve] |
29. | Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F. (1956) Anal. Chem. 28, 350-356 |
30. | Hattan, S. J., and Trimble, R. B. (1996) Glycobiol. 6, 755 (abstr.) |
31. | Naggar, E., Ye, S., Chasteen, N. D., Reinhold, V., and Reinhold, B. (2000) Glycobiology 10, (Abstr. 157), 1119 |
32. | Ozols, J. (1990) Methods Enzymol. 182, 587-601[Medline] [Order article via Infotrieve] |
33. | Matsudaira, P. (1990) Methods Enzymol. 182, 602-613[Medline] [Order article via Infotrieve] |
34. | Laue, T. M. (1992) Analytical Ultracentrifugation in Biochemistry and Polymer Science , pp. 63-89, The Royal Society of Chemistry, Cambridge, UK |
35. | Stafford, W. F., III (1992) Analytical Ultracentrifugation in Biochemistry and Polymer Science , pp. 359-393, The Royal Society of Chemistry, Cambridge, UK |
36. | Laue, T. M., Shah, B. D., Ridgeway, T. M., and Pelletier, S. L. (1992) Analytical Ultracentrifugation in Biochemistry and Polymer Science , pp. 90-125, The Royal Society of Chemistry, Cambridge, UK |
37. | Sreerama, N., and Woody, R. W. (1993) Anal. Biochem. 209, 32-44[CrossRef][Medline] [Order article via Infotrieve] |
38. | Crenshaw, M. A., and Ristedt, H. (1980) in Mechanisms of Mineralization in the Invertebrates and Plants (Wataba, N. , and Wlibur, K. M., eds) , pp. 355-367, University of South Carolina Press, Columbia, SC |
39. | Greenfield, E. M., Wilson, D. C., and Crenshaw, M. A. (1984) Am. Zool. 24, 925-932 |
40. | Halloran, B. A., and Donachy, J. E. (1995) Comp. Biochem. Physiol. 111B, 221-231 |
41. | Beeley, J. G. (1985) Glycoprotein and Proteoglycan Techniques (Laboratory Techniques in Biochemistry and Molecular Biology, Vol. 16) , pp. 24-25, Elsevier Science Publishers B. V., Amsterdam |
42. | Rowe, A. J. (1977) Biopolymers 16, 2595-2611 |
43. | Krampitz, G., Drolshagen, H., Hausle, J., and Hof-Irmscher, K. (1983) in Biomineralization and Biological Metal Accumulation (Westbroek, P. , and de Jong, E. W., eds) , pp. 231-247, D. Reidel, Publishing Co., Dordrecht, Holland |
44. | Marsh, M. E., and Sass, R. L. (1983) J. Exp. Biol. 226, 193-203 |
45. | Marsh, M. E., and Sass, R. L. (1985) J. Exp. Biol. 234, 237-242 |
46. | Weiner, S., and Traub, W. (1984) Phil. Trans. R. Soc. Lond. B. Biol. Sci. 304, 425-434 |
47. | Weiner, S., and Hood, L. (1975) Science 190, 987-989[Medline] [Order article via Infotrieve] |
48. | Nair, P. S., and Robinson, W. E. (1999) Arch. Biochem. Biophys. 366, 8-14[CrossRef][Medline] [Order article via Infotrieve] |