Expression and characterization of recombinant vitamin K-dependent
c-glutamyl carboxylase from an invertebrate,
Conus textile
Eva Czerwiec
1
, Gail S. Begley
1
, Mila Bronstein
2
, Johan Stenflo
1,3
, Kevin Taylor
1
, Barbara C. Furie
1,2
and Bruce Furie
1,2
1
Marine Biological Laboratory, Woods Hole, MA, USA;
2
Center for Hemostasis and Thrombosis Research, Beth Israel Deaconess
Medical Center and Harvard Medical School, Boston, MA, USA;
3
Department of Clinical Chemistry, Lund University,
University Hospital, Malmo, Sweden
The marine snail Conus is the sole invertebrate wherein both
the vitamin K-dependent carboxylase and its product,
c-carboxyglutamic acid, have been identified. To examine its
biosynthesis of c-carboxyglutamic acid, we studied the
carboxylase from Conus venom ducts. The carboxylase
cDNA from Conus textile has an ORF that encodes a 811-

,75l
M
and 74 l
M
, respect-
ively. The recombinant Conus carboxylase, in the absence
of endogenous substrates, is stimulated up to fivefold by
vertebrate propeptides but not by Conus propeptides.
These results suggest two propeptide-binding sites in the
carboxylase, one that binds the Conus and vertebrate
propeptides and is required for substrate binding, and the
other that binds only the vertebrate propeptide and is
required for enzyme stimulation. The marked functional
and structural similarities between the Conus carboxylase
and vertebrate vitamin K-dependent c-carboxylases
argue for conservation of a vitamin K-dependent carb-
oxylase across animal species and the importance of
c-carboxyglutamic acid synthesis in diverse biological
systems.
Keywords: blood coagulation; conotoxins; hemophilia; post-
translational processing; vitamin K.
The vitamin K-dependent carboxylase catalyzes the post-
translational conversion of glutamic acid into c-carboxy-
glutamic acid in prothrombin, other blood coagulation
proteins, and various vitamin K-dependent proteins [1,2]. In
this reaction, CO
2
replaces the c-proton on specific glutamic
acid residues of the peptide substrate to yield c-carboxy-
glutamic acid. This enzymatic reaction is unique in that it

product, c-carboxyglutamic acid, have been identified
Correspondence to B. Furie, Center for Hemostasis and Thrombosis
Research, Beth Israel Deaconess Medical Center and Harvard
Medical School, Boston, MA 02215, USA.
E-mail:
Abbreviations: proPT18, residues )18 to )1 of proprothrombin;
proPT28, residues )18 to +10 of proprothrombin; proFIX18, resi-
dues )18 to )1 of proFactor IX; proFIX28, residues )18 to +10 of
proFactor IX; pro-e-TxIX/12, residues )12 to )1ofe-TxIX precursor;
e-TxIX12, residues 1–12 of e-TxIX; pro-e-TxIX/24, residues )12 to
+12ofe-TxIX precursor; pro-e-TxIX/41, residues )29 to +12 of
e-TxIX precursor.
(Received 12 September 2002, accepted 24 October 2002)
Eur. J. Biochem. 269, 6162–6172 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03335.x
[23,24]. These marine gastropods use small biologically
active peptides (conotoxins) to paralyze fish, marine worms
and molluscs [25,26]. Many c-carboxyglutamic acid-con-
taining conotoxins have been identified [27–31]. The metal-
binding properties [32–35] and the 3D structures of some of
these conopeptides suggest a specific structural role for
c-carboxyglutamic acid [36–40]. Experiments with crude
preparations of Conus carboxylase have shown that this
enzymatic reaction requires vitamin K [24,41]. Efficient
carboxylation requires a carboxylation-recognition site
located on a precursor form of the conotoxin [42,43].
However, the Conus carboxylation-recognition site is dif-
ferent from the carboxylation-recognition site in mamma-
lian carboxylase substrates.
We have previously isolated a highly conserved region
from the Conus carboxylase gene that exhibits marked

properties of this unique enzyme.
EXPERIMENTAL METHODS
Materials
Live cone snails were obtained from Fiji, and frozen
specimens of C. textile were obtained from Vietnam.
FastTrack kits, TA cloning kits (with pCR2.1-TOPO and
pCR4-TOPO vectors), the pcDNA 3.1/V5-His cloning kit,
the pIB/V5-His TOPO TA cloning kit and anti-V5 horse-
radish peroxidase-conjugated antibody were from Invitro-
gen (Carlsbad, CA, USA). A kZAPII Custom cDNA
library from C. textile venom duct was prepared by
Stratagene (La Jolla, CA, USA). TRIzol reagent, Thermo-
Script RT-PCR System, Platinum PCR Supermix, Platinum
Pfx Polymerase, restriction enzymes, synthetic oligonucleo-
tide primers, serum-free adapted Sf21 cells and Sf 900-II
SFM medium were obtained from Gibco–BRL Life Tech-
nologies (Grand Island, NY, USA). RACE kits and
Advantage cDNA polymerase mix were from Clontech
(Palo Alto, CA, USA). AmpliTaq Gold polymerase and
buffer were from Perkin–Elmer (Branchburg, NJ, USA).
Reagents for DNA purification were from Qiagen (Santa
Clarita, CA, USA). Reagents for digoxygenin labeling of
DNA and detection were obtained from Roche Biochem-
icals (Indianapolis, IN, USA). Superose 12 was from
Pharmacia (Piscataway, NJ, USA). NaH
14
CO
3
(55 mCiÆ
mmol

microsomal preparations and cell homogenates
Snails were extricated from their shell and laid flat on a
cooled glass plate. Venom ducts were removed and homo-
genized using a Tissue Tearor mixer for 10 s in 5 : 1 to
10 : 1 (w/v) buffer A (250 m
M
sucrose, 500 m
M
KCl,
25 m
M
imidazole/HCl, pH 7.2) containing 0.1% (w/v)
Chaps and 1 · PIC (2 m
M
dithiothreitol, 2 m
M
EDTA,
0.5 lgÆmL
)1
leupeptin, 1 lgÆmL
)1
pepstatin A, 2 lgÆmL
)1
aprotinin). Homogenates were centrifuged at 12 000 g for
5 min, and supernatants were subsequently centrifuged at
100 000 g for 3 h at 4 °C to separate the microsomal
fraction. The supernatant was discarded, and the pellet was
resuspended in buffer B [25 m
M
Mops (pH 7.0), 500 m

(pH 7.2)
Ó FEBS 2002 Conus vitamin K-dependent c-carboxylase (Eur. J. Biochem. 269) 6163
containing 0.1% (w/v) Chaps, 0.1% (w/v) phosphatidyl-
choline, 0.1 m
M
phenylmethanesulfonyl fluoride and 20%
(v/v) glycerol and sonicated. COS7 cells (5 · 10
6
cells) were
washed with NaCl/P
i
, trypsinized and collected in NaCl/P
i
(pH 7.2) containing 20% glycerol and 1 · PIC. Cells were
homogenized in a glass homogenizer (3 · 10 strokes) and
centrifuged at 500 g. The pellet was rehomogenized and
washed 3 times with the same buffer. Pooled supernatants
were centrifuged at 100 000 g for 3 h at 4 °Ctoseparatethe
microsomal fraction. The pellet was resuspended in NaCl/P
i
(pH 7.2) containing 0.5% (w/v) Chaps, 0.2% (w/v) phos-
phatidylcholine, 1 · PIC and 20% (v/v) glycerol by soni-
cation.
Enzyme assays
The amount of
14
CO
2
incorporated into exogenous sub-
strates was measured in reaction mixtures of 125 lL

propeptide were performed at a constant concentration of
enzyme and substrate (3.6 m
M
FLEEL or 1.6 m
M
e-TxIX12) and increasing concentrations of the propeptide
proPT18, proFIX18 or pro-e-TxIX/12, as indicated. Vita-
min K epoxidase activity was assayed by HPLC as previ-
ously described [21].
Molecular cloning of
C. textile
vitamin K-dependent
carboxylase
All PCRs were performed in a PE Applied Biosystems 9700
thermocycler. Degenerate primers were used at a final
concentration of 1 l
M
, and gene-specific primers at a final
concentration of 0.2 l
M
. Sequences of PCR products were
obtained after cloning into the pCR2.1-TOPO or pCR4.0-
TOPO vector. Ligation reactions were subsequently used to
transform chemically competent Escherichia coli TOP10
cells. Transformants were selected on Luria–Bertani agar
plates containing 50 lgÆmL
)1
kanamycin and 5-bromo-4-
chloro-3-indolyl-
D

The degenerate primer [F(L/I)(L/I/S)(P/S)YWY(V/I)F
(L/F)LDK(T/P)(S/T/A)WNNHSYL] was designed based
on the sequence of the region that was identified in
vertebrate and invertebrate carboxylases (residues 142–
163). The degenerate primer in combination with a gene-
specific primer yielded a specific product that encodes 260
residues of a carboxylase homolog (homologous to region
164–401 of bovine carboxylase). This sequence informa-
tion was used to design a new probe (Probe 2, 537 bp),
complementary to the region ending 186 bp 5¢ (C. textile
sequence) of the codon for Gly386. This probe identified a
clone with an insert of 1740 bp that contained the start
codon of the Conus carboxylase at position 67. The insert
encodes a protein of 557 amino acids that is homologous
to the region 1–518 of bovine carboxylase. The full length
of the Conus carboxylase was obtained by assembling the
sequences from the phage library clones and RACE-PCR
reaction products.
Expression of vitamin K-dependent carboxylase cDNA
Gene fusion constructs encoding a C-terminal V5-tagged
and His-tagged enzyme were made in the pIB-V5/His
TOPO vector and in the pcDNA3.1-V5/His vector. The
Table 1. Amino-acid sequences of synthetic substrates and propeptides. Bold type ¼ mature sequence.
Name Sequence
proPT18 HVFLAPQQARSLLQRVRR
proPT28 HVFLAPQQARSLLQRVRRANTFLEEVRK
proFIX18 TVFLDHENANKILNRPKR
proFIX28 TVFLDHENANKILNRPKRYNSGKLEEFV
pro-e-TxIX/12 LKRTIRTRLNIR
e-TxIX12 ECCEDGWCCTAA

vitamin K-dependent carboxylase
The cell homogenate preparations were evaluated for
recombinant carboxylase by Western-blot analysis after
transfer to a poly(vinylidene difluoride) membrane after
electrophoresis on a 10% SDS/polyacrylamide gel. The
expressed protein was detected using the horseradish
peroxidase-conjugated anti-V5 Ig (1 lgÆmL
)1
). The CAT-
V5/His protein was used as a positive control. Positive
bands were detected by chemiluminescence. Quantitative
Western-blot analysis was performed using Positope as a
protein standard. Carboxylase was quantitated in micro-
somal preparations from transfected Sf21 cells and COS7
cells. Densitometric analysis was performed using the
GELPRO ANALYZER
program (Media Cybernetics, North
Reading, MA, USA).
RESULTS
Cloning of the vitamin K-dependent carboxylase cDNA
Vitamin K-dependent carboxylase activity was measured in
venom duct homogenates from Conus bandanus, Conus
geographus, Conus leopardus, Conus marmoreus, Conus
striatus, C. textile and Conus virgo.Incontrastwith
mammalian tissue preparations, crude cone snail venom
duct homogenates contain large amounts of endogenous
substrates which become labeled with
14
CO
2

Æ(30 min)
)1
], and the
lowest in venom duct homogenates from C. striatus
[9 · 10
3
c.p.m.Æ(mg protein)
)1
Æ(30 min)
)1
]. Relative amo-
unts of carboxylation occurring on endogenous substrates
varied from as high as 32% of total activity in C. textile
venom duct homogenate to as low as 0.5% in C. geographus
homogenate. Because of the high specific activity and the
availability of the species, we chose to clone and express the
Conus carboxylase from C. textile.
The full-length cDNA encoding the vitamin K-depend-
ent carboxylase from C. textile was assembled, as described
in Experimental methods. The entire cDNA sequence
includes 3795 bp, with a 5¢ UTR of 66 nucleotides and a
3¢ UTR of 1296 nucleotides. The translational start site
begins at nucleotide 67 and the stop site (TAA) is at
nucleotide 2499. The 5¢ untranslated region, which was not
mapped, is presumably incomplete. The 3¢ untranslated
sequence includes a polyadenylation consensus sequence
(AATAAA) located 17 nucleotides upstream of the polyA
tail. An ORF of 2433 bp encoding an 811-amino-acid
protein is predicted (Supplementary material; GenBank
accession number AF382823).

due insertions, one three-residue insertion, one six-residue
insertion, and one 19-residue insertion; there are two one-
residue deletions. From this alignment, seven conserved
regions (CR) were identified. These include CR1
(33–317), CR2 (356–415), CR3 (420–451), CR4 (465–
519), CR5 (528–544), CR6 (555–567), and CR7 (581–609).
Residues that are identical among the Conus carboxylase
and the vertebrate carboxylases are highlighted in deep
Ó FEBS 2002 Conus vitamin K-dependent c-carboxylase (Eur. J. Biochem. 269) 6165
yellow in Fig. 1. These identical residues are widely
distributed within the conserved sequences of the Conus
carboxylase. The most extensive regions of high sequence
identity among all of the carboxylases include residues
118–126 and 157–167 in CR1, residues 195–241 in CR1,
residues 390–407 in CR2, and residues 528–544 in CR5.
The bovine carboxylase and Conus carboxylase sequence
share 42% identity (
CLUSTAL
method with the
MEGALIGN
program). Of the amino acids between residues 33 and
610, 52% are identical comparing the bovine and Conus
carboxylase sequences (
CLUSTAL
method with
MEGALIGN
program), and 65% are conserved using
BLAST
analysis
comparing all of the vitamin K-dependent carboxylases.

incorporation into FLEEL in the absence of
vitamin K was 483 c.p.m.Æ(30 min)
)1
with the same micro-
somal fraction. As COS7 cells have endogenous carboxylase
activity, COS7 cells transfected with a plasmid vector
lacking the Conus carboxylase cDNA were tested for
carboxylase activity for comparison. In these experiments,
14
CO
2
incorporation into FLEEL was 6274 c.p.m.Æ(30
min)
)1
. These experiments indicate increased expression of
carboxylase in COS cells of about sevenfold over endo-
genous carboxylase levels.
To ensure that the increased caboxylase activity observed
in COS cells transfected with a plasmid vector containing
carboxylase cDNA arose from expression of carboxylase
from this cDNA, we expressed recombinant Conus car-
boxylase in Sf21 insect cells. These cells do not express
endogenous carboxylase activity. Cells transfected with the
construct containing the Conus carboxylase cDNA showed
significant carboxylase activity. This activity had an abso-
lute requirement for vitamin K in that the carboxylase
activity was 1237 c.p.m.Æ(30 min)
)1
in the presence of
vitamin K and 181 c.p.m.Æ(30 min)

K-dependent carboxylase containing a C-terminal V5 and
His tag was determined using antibodies to the V5 epitope.
Cell homogenates from Conus carboxylase cDNA-trans-
fected cells, nontransfected cells and cells transfected with
chloramphenicol acetyltransferase (CAT) cDNA were an-
alyzed by Western blot (Fig. 2). Sf21 cells transfected with
the carboxylase cDNA-containing plasmid show a major
band at  130 kDa (Fig. 2, lane B). Cells transfected with
the CAT-V5/His plasmid show the expected band of
 33 kDa (Fig. 2, lane C). In contrast, no bands from
homogenates from nontransfected cells and a preparation
of purified flag-tagged bovine carboxylase are detected using
the anti-V5 antibody (Fig. 2, lanes A and D). When
expressed in COS7 cells, the Conus carboxylase migrates
with an apparent molecular mass of 130 kDa (data not
shown).
Specific carboxylase activity of the recombinant
Conus
carboxylase
The concentration of expressed Conus carboxylase in
microsomes from transfected Sf21 cells and COS7 cells
was determined by quantitative Western-blot analysis using
the 53-kDa Positope protein as a standard (Fig. 3A,B).
Microsomal preparations from transfected Sf21 or COS7
cells show the carboxylase at 130 kDa (Fig. 3A,B). Micro-
somal preparations from nontransfected Sf21 cells or mock-
transfected COS7 cells do not contain protein that can be
detected by antibody to V5 (data not shown). Using
densitometric analysis, the concentration of recombinant
Conus carboxylase was determined to be 2 ± 0.3 lgÆmL

with the recombinant bovine carboxylase, the recombinant
Conus carboxylase activity has a 10-fold lower specific
activity at maximum stimulation of both carboxylases by
proPT18.
Enzymatic properties of the recombinant
Conus
carboxylase
The recombinant carboxylase has functional properties that
are similar to those of bovine carboxylase, with the
exception of propeptide binding and stimulation. In
contrast with previous reports of Conus carboxylase enzy-
matic properties [41,43], our assays were performed in the
absence of endogenous carboxylase substrates, thus elimin-
ating interference of carboxylation of exogenous substrates
by endogenous substrates. The apparent K
m
of the recom-
binant Conus carboxylase for reduced vitamin K was
52 ± 10 l
M
; this can be compared with a value of 23 l
M
for the bovine carboxylase [19]. The K
m
values for peptides
based on mammalian vitamin K-dependent precursors,
including FLEEL, proPT28 and proFIX28, were
430 ± 100 l
M
,1.7±0.02l

Our work [43] and that of Bandyopadhyay et al.[42]
demonstrated the importance of the propeptide in directing
c-carboxylation of Conus precursor substrates by the Conus
carboxylase. The propeptide of these substrates binds tightly
to the carboxylase and, as with the bovine carboxylase,
represents all or almost all of the binding energy for the
enzyme–substrate interaction. This is also the case for the
recombinant Conus carboxylase. The K
m
values for pro-
peptide-containing substrates is decreased in parallel with
both recombinant Conus carboxylase and recombinant
bovine carboxylase (Table 3). In addition, the activity of
mammalian carboxylases operating on nonpropeptide-con-
taining substrates such as FLEEL is stimulated by the
addition of synthetic peptides based on the sequences of
residues )18 to )1 of the propeptides from blood coagu-
lation precursors, including bovine Factor X, human
proFactor IX and human proprothrombin [49]. Whether
the propeptide-binding site that directs carboxylation and
the site that stimulates carboxylase activity are identical or
separate remains unresolved. To study propeptide stimula-
tion of the Conus carboxylase activity on FLEEL, we used
recombinant Conus carboxylase expressed in Sf21 cells as
these cells do not contain endogenous carboxylase activity
or carboxylase substrates. Carboxylation of FLEEL by
recombinant Conus carboxylase is increased about fivefold
Fig. 3. Quantitative Western-blot analysis. (A) Microsomal prepar-
ation from Sf21 cells expressing recombinant Conus carboxylase.
Known amounts of Positope protein (lane 1, 2.5 ng; lane 2, 5 ng; lane

+ proPT18 ND 56 ± 7
Sf21 cells
–proPT18 48 ± 7 2.1 ± 0.6
+ proPT18 120 ± 20 29 ± 6
Table 3. Comparison of kinetic properties of recombinant Conus and
recombinant bovine c-carboxylases. Recombinant Conus carboxylase
was expressed in Sf21 cells.
K
m
(l
M
)
Conus Bovine
Vitamin KH
2
52 ± 10 23
a
FLEEL 430 ± 100 2200
b
proFIX28 6 ± 2 3.1
a
proPT28 1.7 ± 0.02 3.6
c
Pro-e-TxIX/12 565 ± 60 1500
d
Pro-e-TxIX/24 75 ± 20 69
d
Pro-e-TxIX/41 74 ± 18 117
d
a

and 5.5 l
M
, respectively. These results
show that propeptides based on human proprothrombin
and human proFactor IX bind the Conus carboxylase.
The propeptide that directs c-carboxylation of the
conotoxin precursor and lowers the apparent K
m
[43] ,
pro-e-TxIX/12, does not stimulate the carboxylation of
FLEEL by either the bovine carboxylase or the Conus
carboxylase (Fig. 4). Identical results were obtained when
e-TxIX12 was used as the substrate instead of FLEEL
(Fig. 4 inset). Masking of stimulation by the Conus
propeptides resulting from the presence of high concentra-
tions of the stimulator ammonium sulfate [50] was ruled out
by performing experiments in the absence of ammonium
sulfate (data not shown).
The bovine vitamin K-dependent carboxylase expresses
vitamin K epoxidase activity. The recombinant Conus
carboxylase also functions as an epoxidase. Formation of
vitamin K epoxide associated with the formation of
c-carboxyglutamic acid was measured by detection of the
epoxide by HPLC. In the absence of carboxylase, substrate
or reduced vitamin K, no vitamin K epoxide was measured
(Table 4). The addition of proPT18, the propeptide from
human proprothrombin, stimulated epoxidation to about
the same extent as it stimulated carboxylation.
DISCUSSION
The sole known function of vitamin K, an essential

vitamin K-dependent carboxylase that is broadly represen-
ted in animal phyla [10]. This 38-residue motif was identified
in the human, bovine, rat, mouse, whale, toadfish, hagfish,
horseshoe crab and cone snail carboxylase gene. The cone
snail motif was obtained by RT-PCR using primers based
on conserved vertebrate sequences. The C-terminal region
of this motif differs from the Conus carboxylase cDNA
obtained by library screening.
Fig. 4. Effect of propeptides on carboxylase activity. The effect of
proFIX18, proPT18 and pro-e-TxIX/12 on FLEEL carboxylation by
the recombinant bovine carboxylase (closed symbols) and the recom-
binant Conus carboxylase (open symbols) was determined with
increasing concentrations of propeptide. The results were analyzed
with the
GRAPHPAD PRISM
3 program using nonlinear curve fitting. The
data are the mean of three experiments and the error bars represent
standard deviation. ProFIX18 (h, j), proPT18 (n, m) and pro-e-
TxIX/12 (s, d). Inset: Effect of pro-e-TxIX/12 on carboxylation of
e-TxIX12 (1.6 m
M
) by the recombinant bovine carboxylase (closed
symbols) and the recombinant Conus carboxylase (open symbols).
Incorporation of
14
CO
2
into e-TxIX12 was measured in the presence of
increasing concentrations of pro-e-TxIX/12.
Table 4. Epoxidase activity from recombinant Conus carboxylase.

C-terminal region is not involved in propeptide binding [19].
Comparison of our C. textile vitamin K-dependent car-
boxylase cDNA and that of Bandyopadhyay et al.[51]
reveals near sequence identity, but with several differences.
We report the entire 3¢ untranslated region, including
 1.2 kb, and partial 5¢ untranslated sequence. In the ORF,
there are six single-base nucleotide differences in the two
cDNA clones, five of which encode a different amino acid
and one that is a silent substitution. Using the bovine
numbering system (Fig. 1), we observe Arg179 rather than
histidine, Thr430 rather than alanine, Pro654 rather than
serine, Met726 rather than threonine, and Gly743 rather
than valine. At the present time, we do not know whether
these are polymorphisms or sequencing artefacts in either of
the clones.
To prove that the cloned cDNA encoded a vitamin K-
dependent carboxylase, we expressed this coding sequence
in vertebrate and invertebrate cells. In the absence of a
molluscan heterologous expression system, we transfected
Sf21 insect cells with an expression plasmid containing the
Conus carboxylase coding sequence. Conus carboxylase
expressed by the transfected cells has a molecular mass of
 130 kDa. The Conus carboxylase contains 53 more amino
acids than bovine carboxylase; the glycosylation state of this
enzyme is not known nor can we comment on whether the
recombinant carboxylase expressed in insects reflects the
glycosylation state of the native protein. As the Sf21 cells
have no endogenous carboxylase activity, epoxidase activity
or endogenous carboxylase substrates, recombinant Conus
carboxylase can be analyzed without ambiguity or interfer-

propeptides.
From the current data as well as from previous studies
[24,41–43], the Conus and mammalian enzymes (a) all
require vitamin K for c-carboxylation and (b) recognize
their substrate via the carboxylation-recognition site enco-
ded in the amino-acid sequence of the substrate. In addition,
we demonstrate in this study that the recombinant Conus
carboxylase has epoxidase activity. The carboxylation-
recognition site is most often found in a precursor form of
the c-carboxyglutamic acid-containing substrates in both
the Conus and mammalian systems, although uncarboxyl-
ated osteocalcin is a low-K
m
substrate that lacks such an
external site but instead contains a unique internal site [52].
Despite these major similarities, several differences between
the Conus and mammalian vitamin K-dependent carboxy-
lases are noteworthy. First, the propeptide sequence that
directs c-carboxylation in conotoxins is different from the
propeptide sequences of blood coagulation proteins. Sec-
ond, the stimulatory effect of the mammalian propeptides is
less potent on the recombinant Conus carboxylase than for
the mammalian carboxylases. Most importantly, the Conus
propeptide does not stimulate the carboxylation of FLEEL
by the Conus or bovine carboxylase.
Since the discovery that the propeptide of the precursor
forms of vitamin K-dependent proteins contain a
c-carboxylation-recognition site that directs carboxylation
[5] and that the free propeptide greatly stimulates the
carboxylation of FLEEL by the carboxylase [49], it has

reduced form of vitamin K, molecular oxygen, carbon
dioxide, a vitamin K-dependent carboxylase that also
co-ordinately oxidizes vitamin K to the vitamin K epox-
ide, and a salvage enzyme, the vitamin K epoxide
reductase, to cycle vitamin K epoxide to vitamin K. The
presence of a Conus vitamin K-dependent carboxylase
with functional and structural similarity to the mammalian
carboxylase has broad and important biological implica-
tions. However, it would seem that this system was not
conserved in invertebrates and vertebrates to post-trans-
lationally modify glutamic acids on blood coagulation
proteins and conotoxins during c-carboxyglutamic acid
synthesis. Rather, it seems that this system, which
developed early in evolution, has a more fundamental
purpose that may or may not involve c-carboxyglutamic
6170 E. Czerwiec et al.(Eur. J. Biochem. 269) Ó FEBS 2002
acid synthesis. Indeed, synthesis of blood coagulation
proteins in vertebrates and toxins in the cone snail may be
secondary functions of this enzyme.
ACKNOWLEDGEMENTS
We especially appreciate the efforts of Tony Nahacky in providing us
with cone snails, and Drs David Roth and Takako Hirata for helpful
discussions. This work was supported by grants (HL38216 and
HL42443) from the National Institutes of Health.
REFERENCES
1. Furie, B., Bouchard, B.A. & Furie, B.C. (1999) Vitamin K-
dependent biosynthesis of gamma-carboxyglutamic acid. Blood
93, 1798–1808.
2. Stenflo, J. (1999) Contributions of Gla and EFG-like domains to
the function of vitamin K-dependent coagulation factors. Crit.

10. Begley, G.S., Furie, B.C., Czerwiec, E., Taylor, K.L., Furie, G.L.,
Bronstein, L., Stenflo, J. & Furie, B. (2000) A conserved motif
within the vitamin K-dependent carboxylase gene is widely dis-
tributed across animal phyla. J. Biol. Chem. 275, 36245–36249.
11. Romero, E.E., Deo, R., Velazquez-Estades, L.J. & Roth, D.A.
(1998) Cloning, structural organization, and transcriptional
activity of the rat vitamin K-dependent gamma-glutamyl car-
boxylase gene. Biochem. Biophys. Res. Commun. 248, 783–788.
12. Wu, S M., Morris, D.P. & Stafford, D.W. (1991) Identification
and purification to near homogeneity of the vitamin K-dependent
carboxylase. Proc. Natl Acad. Sci. USA 88, 2236–2240.
13. Carlisle, T.L. & Suttie, J.W. (1980) Vitamin K dependent car-
boxylase: subcellular localization of the carboxylase and enzymes
involved in vitamin K metabloism in rat liver. Biochemistry 19,
1161–1167.
14. Bristol, J.A., Ratcliffe, J.V., Roth, D.A., Jacobs, M.A., Furie, B.C.
& Furie, B. (1996) Biosynthesis of prothrombin: intracellular
localization of the vitamin K-dependent carboxylase and the sites
of gamma-carboxylation. Blood 88, 2585–2593.
15. Kuliopulos, A., Nelson, N.P., Yamada, M., Walsh, C.T., Furie,
B., Furie, B.C. & Roth, D.A. (1994) Localization of the affinity
peptide-substrate inactivator site on recombinant vitamin K-
dependent carboxylase. J. Biol. Chem. 269, 21364–21370.
16. Sugiura, I., Furie, B., Walsh, C.T. & Furie, B.C. (1996) Profactor
IX propeptide and glutamate substrate binding sites on the vita-
min K-dependent carboxylase identified by site-directed muta-
genesis. J. Biol. Chem. 271, 17837–17844.
17. Yamada, M., Kuliopulos, A., Nelson, N.P., Roth, D.A., Furie, B.,
Furie, B.C. & Walsh, C.T. (1995) Localization of the Factor IX
propeptide binding site on recombinant vitamin K dependent

New York.
25. Olivera, B.M., Gray, W.R., Zeikus, R., McIntosh, J.M., Varga, J.,
Rivier, J., de Santos, V. & Cruz, L.J. (1985) Peptide neurotoxins
from fish-hunting cone snails. Science 230, 1338–1343.
26. Olivera, B.M., Rivier, J., Scott, J.K., Hillyard, D.R. & Cruz, L.J.
(1991) Conotoxins. J. Biol. Chem. 266, 22067–22070.
27. Haack,J.A.,Rivier,J.,Parks,T.N.,Mena,E.E.,Cruz,L.J.&
Olivera, B.M. (1990) Conantokin-T. A gamma-carboxyglutamate
containing peptide with N-methyl-
D
-aspartate antagonist activity.
J. Biol. Chem. 265, 6025–6029.
28. Fainzilber, M., Gordon, D., Hasson, A., Spira, M.E. & Zlotkin, E.
(1991) Mollusc-specific toxins from the venom of Conus textile
neovicarius. Eur. J. Biochem. 202, 589–595.
29. Nakamura, T., Yu. Z., Fainzilber, M. & Burlingame, A.L. (1996)
Mass spectrometric-based revision of the structure of a cysteine-
rich peptide toxin with gamma-carboxyglutamic acid, TxVIIA,
from the sea snail, Conus textile. Protein Sci. 5, 524–530.
30. Kalume, D.E., Stenflo, J., Czerwiec, E., Hambe, B., Furie, B.C.,
Furie, B. & Roepstorff, P. (2000) Structure determination of two
conotoxins from Conus textile by a combination of matrix-assisted
laser desorption/ionization time-of-flight and electrospray ioniza-
tion mass spectrometry and biochemical methods. J. Mass Spec-
trom. 35, 145–156.
31. Fainzilber, M., Nakamura, T., Lodder, J.C., Zlotkin, E., Kits,
K.S. & Burlingame, A.L. (1998) gamma-Conotoxin-PnVIIA,
a gamma-carboxyglutamate-containing peptide agonist of
Ó FEBS 2002 Conus vitamin K-dependent c-carboxylase (Eur. J. Biochem. 269) 6171
neuronal pacemaker cation currents. Biochemistry 37,

273, 16248–16258.
39. Rigby, A.C., Baleja, J.D., Li, L., Pedersen, L.G., Furie, B.C. &
Furie, B. (1997) Role of gamma-carboxyglutamic acid in the cal-
cium-induced structural transition of conantokin G, a conotoxin
from the marine snail Conus geographus. Biochemistry 36, 15677–
15684.
40. Rigby, A.C., Lucas-Meunier, E., Kalume, D.E., Czerwiec, E.,
Hambe,B.,Dahlqvist,I.,Fossier,P.,Baux,G.,Roepstorff,P.,
Baleja,J.D.,Furie,B.C.,Furie,B.&Stenflo,J.(1999)Acon-
otoxin from Conus textile with unusual posttranslational mod-
ifications reduces presynaptic Ca
2+
influx. Proc. Natl Acad. Sci.
USA 96, 5758–5763.
41. Stanley, T.B., Stafford, D.W., Olivera, B.M. & Bandyopadhyay,
P.K. (1997) Identification of a vitamin K-dependent carboxylase
in the venom duct of a Conus snail. FEBS Lett. 407, 85–88.
42. Bandyopadhyay, P.K., Colledge, C.J., Walker, C.S., Zhou, L.M.,
Hillyard, D.R. & Olivera, B.M. (1998) Conantokin-G precursor
and its role in gamma-carboxylation by a vitamin K-dependent
carboxylase from a Conus snail. J. Biol. Chem. 273, 5447–5450.
43. Bush, K.A., Stenflo, J., Roth, D.A., Czerwiec, E., Harrist, A.,
Begley, G.S., Furie, B.C. & Furie, B. (1999) Hydrophobic amino
acids define the carboxylation recognition site in the precursor of
the gamma-carboxyglutamic-acid-containing conotoxin epsilon-
TxIXfromthemarineconesnailConus textile. Biochemistry 38,
14660–14666.
44. Li, T., Yang, C. & Stafford, D. (2000) Identification of a Droso-
phila vitamin K-dependent gamma-glutamyl carboxylase. J. Biol.
Chem. 275, 18291–18296.

carboxylase-binding domain on osteocalcin. Biochem. J. 364,323–
328.
53. Stenflo, J., Ferlund, P., Egan, W. & Roepstorff, P. (1974) Vitamin
K dependent modifications of glutamic acid residues in pro-
thrombin. Proc. Natl Acad. Sci. USA 71, 2730–2733.
54. Nelsestuen, G.L., Zytkovicz, T.H. & Howard, J.B. (1974) The
mode of action of vitamin K. Identification of gamma- carbo-
xyglutamic acid as a component of prothrombin. J. Biol. Chem.
249, 6347–6350.
6172 E. Czerwiec et al.(Eur. J. Biochem. 269) Ó FEBS 2002