Expression experiments of the FXI–Val371Ile recombinant protein, followed by activation assays, showed both a different time course in FXI activation and a slight delay in factor IX acti
Trang 1activation loop is responsible for a case of mild type II
factor XI deficiency
Cristina Bozzao1*, Valeria Rimoldi1*, Rosanna Asselta1, Meytal Landau2, Rossella Ghiotto3,
Maria L Tenchini1, Raimondo De Cristofaro4, Giancarlo Castaman3and Stefano Duga1
1 Department of Biology and Genetics for Medical Sciences, University of Milan, Italy
2 Department of Biochemistry, George S Wise Faculty of Life Science, Tel Aviv University, Israel
3 Department of Hematology and Hemophilia and Thrombosis Center, San Bortolo Hospital, Vicenza, Italy
4 Haemostasis Research Centre, Catholic University School of Medicine, Rome, Italy
Coagulation factor XI (FXI) is the precursor of a
tryp-sin-like serine protease that catalyzes, upon activation,
the conversion of factor IX (FIX) to activated FIX
(FIXa) [1,2] Human FXI, primarily produced by
hepatocytes, is a glycoprotein of 160 kDa circulating
in plasma in a noncovalent complex with high
mole-cular mass kininogen [3] Structurally, FXI zymogen
comprises four N-terminal tandem repeats of about
90 residues, named apple domains (Ap1–4), followed
by a catalytic serine protease domain located at the
C-terminal end Uniquely among coagulation serine
proteases, FXI is secreted as a homodimer composed
of two identical polypeptide chains linked by non-covalent interactions and by a Cys321–Cys321 disulfide bond between the Ap4 domains [4,5]
Among serine proteases that can activate FXI, i.e activated factor XII (FXIIa), FXIa, and thrombin, the main physiologic activator is actually thrombin formed
on the surface of activated platelets [6–8] Cleavage at the Arg369–Ile370 bond in each monomer produces both an N-terminal heavy chain, which binds FIX and high molecular mass kininogen [9], and a C-terminal
Keywords
coagulation factor XI deficiency; functional
characterization; missense mutation;
mutational screening; type II defect
Correspondence
S Duga, Department of Biology and
Genetics for Medical Sciences, University of
Milan, Via Viotti, 3 ⁄ 5-20133 Milan, Italy
Fax: +39 02 5031 5864
Tel: +39 02 5031 5823
E-mail: stefano.duga@unimi.it
*These authors contributed equally to this
study
(Received 21 May 2007, revised 11
Septem-ber 2007, accepted 8 OctoSeptem-ber 2007)
doi:10.1111/j.1742-4658.2007.06134.x
Coagulation factor XI (FXI) is the zymogen of a serine protease that, when converted to its active form, contributes to blood coagulation through proteolytic activation of factor IX FXI deficiency is typically an autosomal recessive disorder, characterized by bleeding symptoms mainly associated with injury or surgery Of the more than 100 FXI gene muta-tions reported in FXI-deficient patients, most are associated with a propor-tional decrease in FXI funcpropor-tional and immunologic levels (type I defects), whereas only a few mutations leading to the presence of dysfunctional molecules in plasma have been molecularly analyzed to date (type II defi-ciencies) We report the functional and molecular characterization of a missense mutation (Val371Ile) identified, in the heterozygous state, in a 25-year-old Italian male with mild FXI deficiency Laboratory analysis revealed reduced functional FXI levels (34%), but normal antigen levels (102%), distinctive of a type II defect Given the proximity of Val371 to the FXI activation site, a possible interference with zymogen activation was postulated Expression experiments of the FXI–Val371Ile recombinant protein, followed by activation assays, showed both a different time course
in FXI activation and a slight delay in factor IX activation by thrombin-activated FXI
Abbreviations
FIX, coagulation factor IX; FIXa, activated factor IX; FXI, coagulation factor XI; FXIa, activated factor XI; FXI:Ag, antigen FXI level;
FXI:C, functional FXI level; FXIIa, activated factor XII.
Trang 2light chain, containing the catalytic domain [10] These
two chains are held together by three disulfide bonds
and both are essential for FIX activation [5] FXI
acti-vation generates a new N-terminus of the catalytic
serine protease domain, which is called the activation
loop (residues 370–376) Following cleavage, the
acti-vation loop undergoes large movement towards the
activation pocket in the protease domain, stabilizing
the FXIa active site [11]
FXI deficiency is an autosomal recessive bleeding
disorder that is rare in most populations (prevalence
1 : 106) but is particularly common among Ashkenazi
Jews, in whom a heterozygote frequency of 8% has
been reported [12] This rare coagulopathy is
charac-terized by decreased FXI functional activity, usually
associated with low levels of FXI antigen (type I
defects) By contrast, type II defects, characterized by
the presence of dysfunctional molecules in plasma, are
rare [13,14] Usually, homozygous and compound
het-erozygous patients have severe FXI deficiency (FXI
activity < 20%), whereas heterozygotes have mild⁄
par-tial FXI deficiency (20–50%) [15] Bleeding tendency
in FXI-deficient patients seems to poorly correlate with
plasma FXI levels and hemorrhagic episodes are
usu-ally associated with injury or surgery, but may be so
severe as to demand replacement therapy [16]
The genetic basis of this rare coagulation disorder is
invariantly represented by mutations within the FXI
gene (F11), which is located on chromosome 4q35.2
and consists of 15 exons spread over a genomic region
of 23 kb To date, > 100 mutations responsible for
FXI deficiency have been described [13,17] Among
Ashkenazi Jewish patients, two prevalent mutations
(Glu117stop and Phe283Leu, also called type II and
type III mutations) account for 95% of cases of FXI
deficiency [12] However, in patients belonging to other
ethnic groups a significantly higher level of allelic
het-erogeneity has been reported Remarkable exceptions
are represented by French Basques, French patients
from Nantes, and English patients, in whom different
prevalent ancestral mutations were found [18–20] An
unusual dominant transmission of FXI deficiency has
been described in some families, in which four different
missense mutations exert a dominant-negative effect on
wild-type FXI secretion through intracellular
hetero-dimer formation [21,22]
The aim of this study was the molecular
character-ization of the F11 germline missense mutation
Val371Ile identified in the heterozygous state in an
Italian patient affected by mild FXI deficiency, who
had normal immunologic FXI levels associated with a
reduced activity of the factor, distinctive of a type II
defect
Results
Patient data The propositus was a male born in 1981, who was referred in 2001 for the evaluation of a prolonged par-tial thromboplastin time discovered prior to a surgical procedure An appendectomy, adenoidectomy, and right-knee arthroscopy carried out previously had been without mishap Laboratory analysis revealed a reduced functional FXI level (FXI:C¼ 34%), although the antigen FXI level was normal (FXI:Ag ¼ 102%), suggestive of a mild type II FXI deficiency His mother, born in 1949, also had reduced FXI:C (43%) associated with normal FXI:Ag (132%), whereas the father, born in 1947, had both functional and antigen FXI levels within the range of normality (96 and 134%, respectively) Both parents were asymptomatic
Mutational screening The entire coding region, including exon–intron boundaries and 300 bp of the promoter region of F11, was sequenced Sequence analysis identified a het-erozygous Gfi A transition in exon 11 correspond-ing to cDNA position 1165 (numbercorrespond-ing accordcorrespond-ing to GenBank accession number NM_000128, starting from the first nucleotide of the ATG start codon), which causes a Val371Ile substitution (numbering omits the signal peptide) The same mutation was found in the heterozygous state in the proband’s mother Amino acid substitution involves the second residue of the FXI light chain after the proteolytic cleavage that leads to FXI activation Residue Val371 is located one residue following the cleavage site of FXI (P2¢ tion, according to the convention of numbering posi-tion around the scissile bond), thus is part of the activation loop
One hundred haploid genomes from unrelated Ital-ian control individuals were also analyzed by direct sequencing and the Val371Ile mutation was absent in all of them (data not shown)
Expression of wild-type and Val371Ile recombinant FXI in COS-1 cells
To evaluate the pathogenic role of the Val371Ile muta-tion, both the wild-type and mutant protein were expressed in COS-1 cells To this end, mutagenesis was performed on the pCDNA3⁄ FXI plasmid to produce the pCDNA3⁄ FXI–Val371Ile vector as des-cribed in Experimental procedures Following tran-sient transfection with either pCDNA3⁄ FXI or
Trang 3pCDNA3⁄ FXI–Val371Ile or with equimolar amounts
of both expression plasmids (to mimic the
heterozy-gous condition), serum-free conditioned media and cell
extracts were analyzed for the presence of FXI antigen
using ELISA FXI antigen levels, measured in both
conditioned media and lysates of cells expressing the
mutant protein (in either the heterozygous or
homozy-gous state), were not significantly different from those
measured in wild-type samples (Fig 1A) In particular,
in media conditioned by cells expressing either
wild-type or mutant FXI, antigen levels ranged from 300 to
500 ngÆmL)1, whereas, levels of immunoreactive FXI
were between 20 and 40 ngÆmL)1in the corresponding
lysates
FXI specific activity was measured in conditioned
media as the ratio between FXI:C and FXI:Ag levels
The specific activity of the FXI–Val371Ile protein was
significantly reduced when compared with the
wild-type one ( 30 and 80% for the heterozygous and the
homozygous condition, respectively; Fig 1B)
Our results are consistent with the FXI:C and
FXI:Ag measured in the patient’s plasma, further
sup-porting the hypothesis that the Val371Ile mutation is a
type II defect leading to the production of a defective FXI molecule
FXI activation Three biologically relevant proteases can activate FXI (FXIIa, FXIa, and thrombin) [8,23], all of which cleave FXI at the Arg369–Ile370 bond and expose the active-site catalytic triad Given the proximity of Val371 to the FXI activation site, a possible interfer-ence on FXI activation was hypothesized To this end, equal amounts of wild-type and mutant recombinant FXI molecules from conditioned media were directly activated by either thrombin or FXIIa
As shown in Fig 2, time-course experiments of FXI activation by either thrombin or FXIIa were per-formed Both proteases were able to cut wild-type and FXI–Val371Ile precursors into two fragments corre-sponding to the heavy (48 kDa) and light (32 kDa) chains The Val371Ile substitution causes a delay in FXI activation time following both types of activation protocols, as is clearly appreciable in Fig 2A,B Indeed, after 16 h digestion with FXIIa at 37 C,
0 0 0 0 0 0 0 0 0 0 0 1
0 1
0 1
0
s n
d d
s
C A
I D E M D E N I T I D
Y T I V I T C A C I F I C E S L
V E N E G I T N A
e l I 1 3 l a V s
u g z o r e t e h e
p t -d l
0 0 0 0 0 0 0 0 0 0
0 1
0 1
*
*
*
*
d
A I D E M D E N I T I D O
Fig 1 Transient expression of wild-type and mutant FXI protein in COS-1 cells pCDNA3 ⁄ FXI, pCDNA3 ⁄ FXI–Val371Ile or equimolar amounts
of both plasmids (heterozygous condition) were transiently transfected in COS-1 cells Equal numbers of cells and equal amounts of plas-mids were used in transfection experiments, as described in Experimental procedures (A) Antigen levels of recombinant FXI were measured
in both conditioned media and the corresponding cell lysates using an ELISA assay Bars represent relative concentrations of protein in media and cell lysates compared with the mean antigen level measured in the wild-type Results are given as mean ± SD (B) The specific activities of recombinant proteins were determined by calculating the ratio between FXI activity (measured using a one-stage method based
on a modified partial thromboplastin time) and FXI antigen levels Bars represent mean ± SD of four independent experiments, each per-formed in duplicate The mean value of wild-type FXI was set as 100% The results were analyzed by unpaired t-test (*P < 0.05;
**P < 0.01; ***P < 0.001), ns, not significant; nd, not determined.
Trang 4about half of the total mutated protein remains uncut,
while the wild-type FXI is almost entirely activated
(Fig 2A)
In order to give a more accurate quantitative
description of the FXI activation process, a
chromo-genic assay was used to compare cleavage of the
sub-strate S-2366 by the wild-type and mutant FXI,
previously activated by thrombin in absence of dextran
sulfate Activation of both proteins by human
a-thrombin followed pseudo-first-order kinetics, as
shown in Fig 3 This was in agreement with a
stochas-tic model of FXI activation by thrombin, whereby the
latter cleaves either of the two chains of zymogen FXI
independently according to simple first-order kinetics
If this were not the case, a double exponential or a
sigmoidal curve would have been observed in the
acti-vation kinetics Under the experimental conditions
used in this study, after 2 h incubation 88% of
wild-type FXI and 60% of mutant FXI was activated by
thrombin The kcat⁄ Km value of FXI activation was
9.8 ± 0.6· 104 and 4.8 ± 0.8· 104m)1Æs)1 for
wild-type and FXI–Val371Ile, respectively These findings
showed that the Val371Ile mutation reduces by
approximately twofold the specificity of thrombin
interaction with the FXI–Val371Ile
Activation of FIX by FXIa The functional properties of activated FXI–Val371Ile were explored both by a proteolytic assay using a com-mercially available FIX and by measuring Michaelis parameters of S-2366 hydrolysis To this purpose, wild-type FXIa and FXIa–Val371Ile, completely acti-vated by thrombin (as described in Experimental pro-cedures) were incubated for different periods with commercial FIX Upon FXI activation, FIX is cleaved
at two sites, releasing an activation peptide, and pro-ducing the protease FIXa [10,24] As shown in Fig 4, incubation of FIX with wild-type FXIa results in almost complete activation after 30 min, whereas FXIa–Val371Ile causes a dramatic reduction in the uncleaved FIX form only after 60 min of incubation
A possible effect of dextran sulfate on FIX activation was ruled out by performing the same experiment in the absence of FXI No activation of FIX was detect-able after 60 min of incubation (data not shown) The observed delay in FIX activation may be due to
a decrease in the catalytic activity of mutant FXIa, possibly caused by a perturbed conformational state of FXIa linked to the Val371Ile mutation A moderate but significant reduction in the catalytic competence of
A
B
Fig 2 Time course of wild-type and Val371Ile FXI activation SDS ⁄ PAGE of wild-type FXI and FXI–Val371Ile (1.5 ng of protein) incubated with FXIIa (1 lg) (A) or thrombin (0.5 U) (B) At various time points, indicated at the top of the panels, samples were stopped in reducing sample buffer and eventually separated onto a 10% polyacrylamide gel FXI activation was evaluated by western blotting using polyclonal goat anti-human IgG recognizing both uncleaved FXI and FXI heavy and light chains The estimated molecular masses of monomeric uncut FXI (80 kDa), FXIa heavy chain (48 kDa), and FXIa light chain (32 kDa) are indicated.
Trang 5Val371Ile FXIa was confirmed by investigating the
catalytic competence of FXIa towards the synthetic
substrate S-2366 (Fig 5) The kcat and Km values of
S-2366 hydrolysis by wild-type FXI were 49.8 ± 3 s)1
and 595 ± 63 lm, respectively, with kcat⁄ Km¼
8.37· 104m)1Æs)1 The same parameter values were
45 ± 4 s)1 and 739 ± 100 lm for FXI–Val371Ile,
with kcat⁄ Km¼ 6.09 · 104m)1Æs)1 The reduction in
the kcat⁄ Kmvalue for S-2366 hydrolysis was significant,
but the effect of the mutation on FIX activation was
even more evident, as shown in Fig 4 This suggests
that the mutation may alter molecular recognition
between FXIa and FIX, which necessarily involves, in
addition to the catalytic residues, a more extended sur-face area of FXIa The observed increase in Km for S-2366 of the mutant FXI may arise from allosteric effects, and thus may be generated from structural per-turbations located far from the catalytic pocket
Discussion
In this study, a novel missense mutation in F11 was identified in a proband with mild type II FXI defi-ciency In vitro expression of the FXI–Val371Ile recombinant protein, followed by activation assays, showed slight differences in both FXI activation and FIX activation by thrombin-activated FXI The func-tional defect evidenced by in vitro assays is compatible with the deficiency observed in the two analyzed Val371Ile carriers, even though the specific activity cal-culated for the recombinant mutant protein is some-what higher than expected We cannot exclude that differences in the dimerization and⁄ or secretion effi-ciency of mutant versus wild-type FXI might explain,
at least in part, this discrepancy
Evolutionary conservation analysis of serine prote-ase sequences shows that the position corresponding to FXI–Val371 is highly conserved For example, among serine protease coagulation factors (i.e FVII, FIX,
FX, FXII, plasminogen, and thrombin, showing an overall amino acid sequence identity of 30–45%), this position is occupied solely by a valine This conserved amino acid is replaced by isoleucine in the Val371Ile FXI mutant Interestingly, an isoleucine residue natu-rally occupies the position corresponding to FXI Val371 in some other serine proteases, such as vita-min-K-dependent protein C, hepatocyte growth factor activator, and neurotrypsin
The Val371Ile mutation in FXI results in a relatively mild physicochemical difference, because valine and isoleucine are both highly hydrophobic, b-branched
X I F
0 0 5 5 0
I X F e l I 1 3 l a V I
X F e p y t d l w
) n i m ( )
n i m
a X I F
X I F
a X I F
Fig 4 Time course of FIX activation Commercially available FIX (12.5 ng) was activated with 1.5 ng of recombinant FXI, either wild-type or FXI–Val371Ile, both in turn activated by thrombin (0.5 U for 135 min; complete activation was assessed by western blot analysis) At differ-ent time points (indicated at the top of each panel) digestions were stopped and proteins were resolved by Laemmli SDS ⁄ PAGE using 12% (w ⁄ v) acrylamide gels.
0 1 0 1 0 0 0 0 0
0
8
6
4
2
0
e m i
T ( m ) i n
I X F -T W
I X F -1 3 V
Fig 3 FXI activation by thrombin Purified wild-type (d) and FXI–
Val371Ile (s) (10 n M , final concentration) were activated by
throm-bin (3 n M , final concentration) At different time points hirudin
(10 n M , final concentration) was added to inhibit thrombin activity,
so that the chromogenic substrate S-2366 (500 l M , final
concentra-tion) was hydrolyzed solely by activated FXI The velocity of S-2236
hydrolysis by FXIa at each time point was converted into FXIa
con-centration by means of Eqn (2) Error bars indicate SEM.
Trang 6amino acids (the b-carbon has two substitutions);
nev-ertheless, the mutation is not isosteric, isoleucine being
larger than valine, and having an additional methyl on
its side chain
In the structure of the FXI zymogen [11], Val371 is
located on the linker region between the Ap4 and
pro-tease domains, and its surface area is 77% exposed to
the solvent After activation of FXI, the activation
loop (residues 370–376), which is located at the new
N-terminus of the protease domain, undergoes a large
movement towards the activation pocket of FXIa As
a result, in the structure of FXIa [25], the surface area
of Val371 is 92% buried within the protein, contacting
residues Arg144, Gly188, Asp189, Cys219, and Ala220
Given that Val371 is buried in the structure of FXIa,
the introduction of a larger residue in this position
most likely causes some degree of structural change;
this is especially true in the case of the introduction of
an isoleucine, a b-branched amino acid that is not
flex-ible In the active conformation, Val371 forms contacts
with neighboring residues that are important for
stabi-lizing the active state (e.g Asp189, which is part of the
S1 pocket responsible for the binding specificity of the
substrate) [26] Consequently, substitution of Val371 to
isoleucine might prevent the full development of the
active conformation This hypothesis is further
con-firmed by the results of FIX proteolytic assays, which
showed a slight delay in FIX activation by FXIa
acti-vated by thrombin (Fig 4); moreover the kcat and Km
values of S-2366 hydrolysis showed that the Val371Ile
mutation has only minor conformational effects on the geometry of the catalytic site of the enzyme (Fig 5)
In contrast to the activated FXI, in the structure of the FXI zymogen, Val371 is located on a loop region, exposed to solvent, and does not form many contacts with other residues (Fig 6) Therefore, the additional methyl in the Val371Ile mutant probably does not disturb the structure and the domain rearrangement
in the zymogen FXI Nevertheless, recombinant FXI– Val371Ile activation was slower than that of the wild-type protein (Fig 2) suggesting a small activation defect This might be explained by the proximity of the mutation to the cleavage site, probably resulting in a small interference with the binding of the activator to the FXI zymogen
There are some examples of inherited coagulation disorders in which one of the peptide linkages required for the proteolytic zymogen activation cannot be cleaved by the physiological activator In most cases, the mutated residue corresponds to the P1 site (i.e the C-terminal residue of the activation peptide) [27–33] However, some mutations involving the P1¢ and P2¢ positions (i.e the two first residues from the N-termi-nal end of the catalytic domain) were previously reported to cause mild to severe FIX deficiency either
6 1 2
1 8
0 4
0 0
0
0
0
0
0
mM) ( 6 3 2 -S
I X F -1 3 V
Fig 5 Determination of Michaelis parameters of S-2366 hydrolysis.
Steady-state kinetics of S-2366 hydrolysis by wild-type (d) or FXI–
Val371Ile (s), under the conditions reported in Experimental
proce-dures The continuous lines were drawn according to Michaelis
equation using the best fit parameters: (d) kcat¼ 49.8 ± 3 s)1,
K m ¼ 595 ± 63 l M ; (s) k cat ¼ 45 ± 4 s)1, K m ¼ 739 ± 100 l M
Error bars indicate SEM.
Fig 6 Structural consequences of the Val371Ile substitution Rib-bon representation of the superimposition between the structures
of the catalytic serine protease domain of the zymogen (red) and activated (green) FXI The Ile371 residue, in both structures, is dis-played by space-filled atoms The catalytic triad (blue space-filled atoms) is also shown The conformational movements of Ile371, located in the activation loop at the N-terminus of the catalytic domain, are notable In the zymogen FXI, Ile371 is exposed to the solvent, while in the activated FXI it is inserted into the protein The picture was drawn with PYMOL (DeLano Scientific, San Carlos, CA; http://www.pymol.org).
Trang 7by altering the functional properties of FIXa or
by delaying its activation by FXIa In particular,
four different amino acid substitutions (Val182Leu,
Val182Phe, Val182Ala, and Val182Gly) corresponding
to the here-reported Val371Ile in FXI, were found in
hemophilia B patients [34–37] The phenotypic
conse-quences of these missense mutations were variable,
ranging from the complete loss of function of FIX
Kashihara (Val182Phe) to a residual 15% of
procoagu-lant activity of FIX Cardiff (Val182Leu) [37]
In conclusion, the Val371Ile mutation, identified and
characterized here, brings the number of naturally
occurring FXI variants responsible for type II
deficien-cies to seven [13] Of these, three have been
character-ized in-depth, showing different mechanisms
underlying the pathologic phenotype, i.e a reduction
in affinity for platelets (Ser248Asn) [38], a modest
reduction of FXI catalytic activity (Pro520Leu) [39],
and a greatly reduced rate of FIX activation associated
with resistance to antithrombin inhibition (Gly555Glu)
[40] Uniquely, our mutation is associated with a defect
both in FXI activation (slower than normal), and in
FIX activation (slightly delayed), thus supporting the
role of residues neighboring the active site in
influenc-ing and stabilizinfluenc-ing the enzyme active state
Experimental procedures
Blood collection and genomic DNA extraction
This study was approved by the Institutional Review Board
of the University of Milan All subjects signed an informed
consent according to the Declaration of Helsinki before
blood withdrawal Peripheral venous blood was collected in
1 : 10 volume of 0.11 m trisodium citrate, pH 7.3 Genomic
DNA was extracted from whole blood using a standard
salting-out procedure
Coagulation studies
Immediately after collection, citrated blood was centrifuged
at 2500 g for 15 min at room temperature FXI activity was
performed by a one-stage method based on a modified
par-tial thromboplastin time, using FXI-deficient plasma as
sub-strate (Hemoliance, Salt Lake City, UT) FXI antigen was
measured by an ELISA based on a goat anti-human FXI
affinity purified IgG as capture antibody and a goat
human FXI peroxidase-conjugated IgG as detecting
anti-body (Affinity Biological Inc., Hamilton, Ontario, Canada)
FXI levels were expressed in both tests as percentages of
pooled normal plasma from 30 normal male and female
individuals The detection limits of the FXI functional and
immunologic assays were 1 and 0.1%, respectively
PCR amplifications and DNA sequencing
PCR were performed on 50–100 ng of genomic DNA in a
25 lL volume, following standard procedures [41] PCR and sequencing primers were designed on the basis of the known genomic sequence of F11 (GenBank accession num-ber NM_000128) The primer couple used to amplify F11 exon 11 and to identify the Val371Ile mutation was ex11-F 5¢-GTCAATTCCATTTTTCATGTGC-3¢ and FXI-ex11-R 5¢-CGTTTTTTACCACTGAAGCAAT-3¢ All other primer sequences, as well as the specific PCR condition for each primer couple, are available on request Sequencing reactions were performed on both strands on PCR products purified by MICROCON 100 columns (Millipore, Bedford, MA) The BigDye Terminator Cycle Sequencing Kit ver-sion 3.1 and an automated ABI-3100 DNA sequencer (Applied Biosystems, Foster City, CA) were used
Site-directed mutagenesis
The pCDNA3⁄ FXI expression plasmid, containing full-length FXI complementary DNA (cDNA), was kindly provided by A Zivelin (Institute of Thrombosis and Hemo-stasis, Chaim Sheba Medical Center, Tel Hashomer, Israel) The identified missense mutation was introduced in pCDNA3⁄ FXI by the QuikChange Site-Directed Muta-genesis Kit (Stratagene, La Jolla, CA), according to the manufacturer’s instructions The mutant plasmid pCDNA3⁄ FXI–Val371Ile was checked by sequencing the whole FXI cDNA insert as well as 200 bp of flanking DNA on both sides of the cloning site Large-scale plasmid preparations were obtained using the EndoFree Plasmid Maxi Kit (Qiagen, Hilden, Germany)
Proteins and antibodies
Thrombin, FIX, FXI, and FXIIa were obtained from Enzyme Research Laboratories (Swansea, UK) The sources
of the antibodies were as follows: rabbit anti-human FIX (catalogue number F 0652) from Sigma (St Louis, MO), goat anti-human FXI (catalogue number GAFXI-IG) from Affin-ity Biologicals Inc., peroxidase-conjugated goat anti-rabbit IgG from Pierce Biotechnology Inc (Rockford, IL), and peroxidase-conjugated donkey anti-goat IgG from Jackson ImmunoResearch Laboratories Inc (West Grove, PA)
Cell culture
African green monkey kidney COS-1 cells were cultured
in DMEM (EuroClone, Wetherby, UK) supplemented with 10% fetal bovine serum (HyClone, South Logan, UT), antibiotics (100 UÆmL)1 penicillin and 100 lgÆmL)1 streptomycin; EuroClone) and glutamine (2 mm; Euro-Clone), and grown at 37C in a humidified atmosphere
Trang 8of 5% CO2 and 95% air, according to standard
proce-dures
Expression of recombinant proteins
In each transfection experiment an equal number of cells
(400 000) were transiently transfected with the
Lipofecta-mine 2000 reagent (Invitrogen, Carlsbad, CA) in six-well
plates with 4 lg of plasmid DNA (pCDNA3⁄ FXI, or
pCDNA3⁄ FXI–Val371Ile, or equimolar amounts of both
plasmids), essentially as described by the manufacturer
Twenty-nine hours after transfection, cells were washed
twice with NaCl⁄ Pi and cultured for additional 48 h in
1 mL of serum-free medium supplemented with glutamine,
antibiotics, and 5 mgÆmL)1BSA For each experiment
(per-formed four times in duplicate) a mock sample, with the
empty pCDNA3 plasmid, was set up
Conditioned media from each well were tested for both
FXI antigen and coagulation activity and used to prepare
FXIa for SDS⁄ PAGE analysis
FXI measurement in conditioned media and
in cell lysates
FXI antigen levels were evaluated by ELISA, as described
above, both in conditioned media and in cell lysates
Stan-dard curves were constructed with reference plasma diluted
1 : 100 to 1 : 6400 in Tris-buffered saline (NaCl⁄ Tris:
50 mm Tris, 150 mm NaCl, pH 7.5) Conditioned media
were collected in prechilled tubes containing a protease
inhibitor mixture (Complete; Roche, Basel, Switzerland),
centrifuged to remove cell debris, and stored at)80 C until
use To obtain cell lysates, cells were washed three times
with prechilled NaCl⁄ Pi and incubated for 1 h on ice with
1· NaCl ⁄ Pi, 1.5% Triton X-100, and 1· Complete Samples
were collected and centrifuged to remove cell debris
FXI coagulant activity was measured in media (collected
without any protease inhibitor) as described above (see
‘Coagulation studies’)
Activation of FXI
FXI was activated either with FXIIa or with thrombin For
each activation experiment, the exact amount of the
recom-binant protein was assessed by an ELISA assay, as
described above; on average, 2.5 lL of conditioned media
corresponded approximately to 1.5 ng of protein
FXIIa (1 lg) and 1.5 ng of recombinant FXI, either
wild-type or mutant, were incubated in NaCl⁄ Tris at 37 C
for different periods Each reaction was carried out in a
final volume of 20 lL Samples were removed into reducing
SDS sample buffer and size-fractionated on 10%
polyacryl-amide SDS gels
Because in vitro activation of FXI by thrombin is highly
enhanced in the presence of polyanions such as dextran
sulfate [42], 1.5 ng of recombinant FXI, either wild-type or mutant, was activated with 0.5 U ( 5 nm) of human thrombin in NaCl⁄ TrisA (NaCl ⁄ Tris supplemented with 0.1 mgÆmL)1 BSA) containing 1 lgÆmL)1 dextran sulfate (500 000 Da) at 37C for different periods The concentra-tion of dextran sulfate (1 lgÆmL)1) used in our experiments was found to be optimal in previous studies [23,42,43] Each reaction was carried out in a final volume of 20 lL Aliquots (each containing 1.5 ng of recombinant FXI) were stopped by adding 10 lL of 3· reducing Laemmli sample buffer, and run on 10% SDS⁄ PAGE
Proteins were then transferred onto 0.45 lm pore-size nitrocellulose membranes (Schleicher & Schuell, Brentford, UK) and analyzed by western blotting, using a polyclonal goat anti-human FXI IgG
Activation of FIX by FXIa
Recombinant FXI, either wild-type or mutant, was acti-vated with 0.5 U thrombin in NaCl⁄ TrisA containing
1 lgÆmL)1 dextran sulfate at 37C for 135 min in a total reaction volume of 20 lL; complete activation was verified
by western blotting (see below) After that, 1.5 ng of FXIa and 12.5 ng of FIX were incubated in NaCl⁄ TrisA with 2.5 mm CaCl2at 37C for different periods Each reaction was carried out in a final volume of 30 lL The ability of residual thrombin and dextran sulfate in the buffer solution
to activate FIX was ruled out in preliminary experiments
At different time points aliquots were removed into reduc-ing SDS sample buffer, and size-fractionated on 12% poly-acrylamide SDS gels
Proteins were transferred onto nitrocellulose membranes and analyzed by western blot using polyclonal rabbit anti-human FIX IgG
Western blotting analysis
Blots were incubated at room temperature for 1 h in NaCl⁄ Tris containing 0.1% Tween 20 and 5% (w ⁄ v) skimmed milk The membranes were then incubated for 2 h with primary antibodies and subsequently for 1 h with donkey anti-goat IgG or goat anti-rabbit IgG horseradish peroxidase-conjugated secondary ones When the anti-human FIX IgG was used, dilutions were performed in NaCl⁄ Tris supplemented with 0.3% BSA at room tempera-ture; all other incubations were done in NaCl⁄ Tris contain-ing 5% milk Proteins were detected uscontain-ing Enhanced Chemioluminescence, SuperSignal West Dura Extended Duration Substrate (Pierce)
Assay of FXI activation
Before activation by thrombin, supernatants from cells expressing recombinant FXI were concentrated
Trang 9approxi-mately four- to fivefold by means of VivaSpin 30
concen-trators (Sartorius Ltd., Epsom, UK) Activation of both
wild-type and FXI–Val371Ile (10 nm) by thrombin (3 nm),
purified as previously detailed [44], was measured by a
chromogenic assay, as follows Incubations were carried
out in 100 lL of 50 mm Tris, 150 mm NaCl, pH 7.5, with
0.1% poly(ethylene glycol) 6000 at 25C In the FXI
acti-vation by thrombin, dextran sulfate was omitted from the
reaction buffer to avoid any spurious effect on FXI
auto-activation At various time intervals, 10 lL of recombinant
hirudin (Sigma) at a final concentration of 10 nm were
added to inhibit thrombin activity Then 50 lL of 500 lm
(final concentration) S-2366 (pyroGlu-Pro-Arg-pNA;
Chro-mogenix, Mo¨lndal, Sweden) were added to the solution,
and the amount of free paranitroaniline released by FXIa
was determined by measuring the change in absorbance at
405 nm in a Benchmark II microplate reader (Bio-Rad
Laboratories, Hercules, CA) To eliminate any scattering
contribution, the absorbance at 620 nm was always
sub-tracted from the reading at 405 nm The initial velocity of
S-2366 hydrolysis obtained at each time point was
consid-ered proportional to FXIa generated by thrombin The
velocity of S-2366 hydrolysis was then analyzed as a
func-tion of time, to calculate the pseudo-first-order rate constant
of both wild-type and mutant FXI cleavage by thrombin
Accordingly:
Vt¼ V1 ð1 expðk tÞÞ
where Vtand V¥are the velocities of S-2366 hydrolysis by
formed FXIa at time t and ¥, respectively, and k is the
pseudo-first order rate constant of FXI activation by
thrombin The best-fit value of k is thus independent from
the intrinsic catalytic activity of both wild-type and mutant
FXIa, but depends only on the specificity of thrombin–FXI
interaction The only assumption made was that the value
of the asymptotic parameter V¥corresponds to the velocity
of the substrate hydrolysis by the FXIa concentration equal
to the nominal concentration of zymogen FXI present in
solution, assuming that the entire amount of zymogen FXI
was converted to FXIa at time¥ The reaction was studied
at a concentration of FXI < Kmof thrombin hydrolysis so
that the rate constant k was proportional to the value of
kcat⁄ Kmof the activation, according to:
where T is the thrombin concentration
Measurement of Michaelis parameters of S-2366
hydrolysis by wild-type and FXI–Val371Ile
After 120 min of FXI activation by thrombin, 88%
(8.8 nm) of wild-type FXI and 63% (6.3 nm) of mutant
FXI were activated, according to [45]:
½FXIa120¼ V120=V1 FXIT ð2Þ where V120is the velocity of S-2366 hydrolysis at 120 min and FXIT is the total concentration of either wild-type or mutant zymogen FXI present in the activation solution The validity of this approach was confirmed in the case of the wild-type form, whose concentration, calculated by Eqn (2) was in agreement within 10% error with that obtained from a reference curve, where the catalytic acti-vity of different concentrations of a purified FXIa prepara-tion (Hematological Technologies Inc., Essex Juncprepara-tion, VT)
in the presence of 500 lm S-2366 were linearly correlated
to the nominal enzyme concentration (supplementary Fig S1)
At time 120 min, chosen to avoid instability or autohy-drolytic damage of thrombin at longer incubation times, an aliquot of the activation solution was taken to measure the Michaelis parameters of S-2366 hydrolysis by FXIa The Michaelis parameters, kcat and Km, were calculated on the basis of known concentration of wild-type and mutant FXIa and using the program grafit (Erithacus Software Ltd., Staines, UK)
Structural analysis
The structural analysis was conducted using the crystal structure of the zymogen FXI (PDB code: 2F83) [11] and FXIa (PDB code: 1XX9) [25] The solvent-accessible area for each residue in both structures was calculated using the surfv program [46] with a probe sphere of radius 1.4 A˚ and default parameters The percentage of the surface-exposure of each residue in the monomer was calculated from the total solvent-accessible area on a Gly-X-Gly tripeptide (where X represents each of the 20 amino acids)
Evolutionary conservation analysis
Evolutionary conservation analysis was carried out using the ConSurf web-server [47] (http://consurf.tau.ac.il/) The calculations were performed using the structure of FXIa (PDB code: 1XX9) [25], based on an alignment of 200 ser-ine protease sequences collected from the SWISSPROT database [48] and default parameters
Acknowledgements
The authors would like to thank Sofia H Giacomelli for excellent technical assistance SD is a recipient of a Bayer Hemophilia Early Career Investigator Award
2006 The financial support of PRIN (Programmi di Ricerca Scientifica di Rilevante Interesse Nazionale, Grant n 2005058307-002) is gratefully acknowledged
Trang 101 Davie EW, Fujikawa K, Kurachi K & Kisiel W (1979)
The role of serine proteases in the blood coagulation
cascade Adv Enzymol Relat Areas Mol Biol 48, 277–318
2 Fujikawa K, Legaz ME, Kato H & Davie EW (1974)
The mechanism of activation of bovine factor IX
(Christmas factor) by bovine factor XIa (activated
plasma thromboplastin antecedent) Biochemistry 13,
4508–4516
3 Saito H (1977) Purification of high molecular weight
kininogen and the role of this agent in blood
coagula-tion J Clin Invest 60, 584–594
4 Dorfman R & Walsh PN (2001) Noncovalent interactions
of the Apple 4 domain that mediate coagulation factor
XI homodimerization J Biol Chem 276, 6429–6438
5 McMullen BA, Fujikawa K & Davie EW (1991)
Loca-tion of the disulfide bonds in human coagulaLoca-tion factor
XI: the presence of tandem apple domains Biochemistry
30, 2056–2060
6 Gailani D, Ho D, Sun MF, Cheng Q & Walsh PN
(2001) Model for a factor IX activation complex on
blood platelets: dimeric conformation of factor XIa is
essential Blood 97, 3117–3122
7 Yun TH, Baglia FA, Myles T, Navaneetham D, Lopez
JA, Walsh PN & Leung LL (2003) Thrombin activation
of factor XI on activated platelets requires the
interac-tion of factor XI and platelet glycoprotein Ib alpha with
thrombin anion-binding exosites I and II, respectively
J Biol Chem 278, 48112–48119
8 Gailani D & Broze GJ Jr (1991) Factor XI activation in
a revised model of blood coagulation Science 253, 909–
912
9 Baglia FA, Sinha D & Walsh PN (1989) Functional
domains in the heavy-chain region of factor XI: a high
molecular weight kininogen-binding site and a
sub-strate-binding site for factor IX Blood 74, 244–251
10 Bouma BN & Griffin JH (1977) Human blood
coagula-tion factor XI Purificacoagula-tion, properties, and mechanism
of activation by activated factor XII J Biol Chem 252,
6432–6437
11 Papagrigoriou E, McEwan PA, Walsh PN & Emsley J
(2006) Crystal structure of the factor XI zymogen
reveals a pathway for transactivation Nat Struct Mol
Biol 13, 557–558
12 Shpilberg O, Peretz H, Zivelin A, Yatuv R, Chetrit A,
Kulka T, Stern C, Weiss E & Seligsohn U (1995) One
of the two common mutations causing factor XI
defi-ciency in Ashkenazi Jews (type II) is also prevalent in
Iraqi Jews, who represent the ancient gene pool of Jews
Blood 8, 429–432
13 Saunders RE, O’Connell NM, Lee CA, Perry DJ &
Perkins SJ (2005) Factor XI deficiency database: an
interactive web database of mutations, phenotypes, and
structural analysis tools Hum Mutat 26, 192–198
14 Salomon O & Seligsohn U (2004) New observations on factor XI deficiency Haemophilia 10 (Suppl 4), 184–187
15 Ragni MV, Sinha D, Seaman F, Lewis JH, Spero JA & Walsh PN (1985) Comparison of bleeding tendency, factor XI coagulant activity, and factor XI antigen in 25 factor XI-deficient kindreds Blood 65, 719–724
16 Peyvandi F, Lak M & Mannucci PM (2002) Factor XI deficiency in Iranians: its clinical manifestations in com-parison with those of classic hemophilia Haematologica
87, 512–514
17 Quelin F, Francois D, d’Oiron R, Guillet B, de Rau-court E & de MazanRau-court P (2005) Factor XI defi-ciency: identification of six novel missense mutations (P23L, P69T, C92G, E243D, W497C and E547K) Haematologica 90, 1149–1150
18 Zivelin A, Bauduer F, Ducout L, Peretz H, Rosenberg
N, Yatuv R & Seligsohn U (2002) Factor XI deficiency
in French Basques is caused predominantly by an ances-tral Cys38Arg mutation in the factor XI gene Blood 99, 2448–2454
19 Quelin F, Trossaert M, Sigaud M, Mazancourt PD & Fressinaud E (2004) Molecular basis of severe factor XI deficiency in seven families from the west of France Seven novel mutations, including an ancient Q88X mutation J Thromb Haemost 2, 71–76
20 Bolton-Maggs PH, Peretz H, Butler R, Mountford R, Keeney S, Zacharski L, Zivelin A & Seligsohn U (2004)
A common ancestral mutation (C128X) occurring in 11 non-Jewish families from the UK with factor XI defi-ciency J Thromb Haemost 2, 918–924
21 Kravtsov DV, Monahan PE & Gailani D (2005) A classification system for cross-reactive material-negative factor XI deficiency Blood 105, 4671–4673
22 Kravtsov DV, Wu W, Meijers JC, Sun MF, Blinder
MA, Dang TP, Wang H & Gailani D (2004) Dominant factor XI deficiency caused by mutations in the factor
XI catalytic domain Blood 104, 128–134
23 Baglia FA & Walsh PN (1998) Prothrombin is a cofac-tor for the binding of faccofac-tor XI to the platelet surface and for platelet-mediated factor XI activation by throm-bin Biochemistry 37, 2271–2281
24 Di Scipio RG, Kurachi K & Davie EW (1978) Activa-tion of human factor IX (Christmas factor) J Clin Invest 61, 1528–1538
25 Jin L, Pandey P, Babine RE, Gorga JC, Seidl KJ, Gelfand E, Weaver DT, Abdel-Meguid SS &
Strickler JE (2005) Crystal structures of the FXIa catalytic domain in complex with ecotin mutants reveal substrate-like interactions J Biol Chem 280, 4704–4712
26 Perona JJ & Craik CS (1995) Structural basis of sub-strate specificity in the serine proteases Protein Sci 4, 337–360
27 Hamaguchi M, Matsushita T, Tanimoto M, Takahashi
I, Yamamoto K, Sugiura I, Takamatsu J, Ogata K,