Characterization of porcine factor VII, X and comparison with human factor VII, X
Younan Chen1, Jianlin Qiao1, Weidong Tan, Yanrong Lu, Shengfang Qin, Jie Zhang,
Shengfu Li, Hong Bu, Jingqiu Cheng⁎
Key Laboratory of Transplant Engineering and Immunology, Ministry of Health, West China Hospital, Sichuan University, 610041, P.R. China
a b s t r a c ta r t i c l ei n f o
Submitted 17 October 2008
Revised 6 February 2009
Available online 16 March 2009
(Communicated by R.I. Handin, M.D.,
6 February 2009)
Objective: Factor VII (FVII) and factor X (FX) are two predominant molecules of coagulation cascade. Whether
porcine FVII and FX could efficiently work in human circulation is important for successful pig to human liver
transplantation. We compared the genetic characterizations and coagulation activities of porcine and human
FVII and FX to shed insight into the further investigation of potential inter-species molecular incompatibility
between porcine FVII, FX and human derived procoagulants and anticoagulants in xenotransplantation.
Methods: Multiple rounds of PCR were used to screen the positive clones from a porcine liver tissue cDNA
library. 5′ RACE and 3′ RACE were conducted to get the full-length cDNA. The three-dimensional structure of
protein was modeled by Swiss-Model program. Prothrombin Time (PT) of porcine and human plasma was
determined by coagulation autoanalyzer. Activities of porcine FVII and FX were detected by adding the
porcine plasma into FVII or FX-deficient human plasma.
Results: We cloned the full-length cDNA of porcine FVII and FX, which contained 1416 bp and 1856 bp, coding
445 and 479 amino acids, respectively. Porcine FVII and FX shared 74.08% and 73.1% amino acid identities
with human FVII and FX. Sequence alignments showed that porcine FVII might have additional γ-
carboxyglutamic acid in Gla domain, and one important variation of Lys62-Glu in light chain. No significant
difference was observed in TF binding region of heavy chain, while 4 variations were identified in the
important functional residues responsible for proteolysis activity, as Gln217-Glu, Thr151-Lys, Glu154-Val and
Gln40-Leu. However, no apparent change was displayed in the 3-D model of the heavy chain of porcine FVII.
When porcine FX was analyzed, great variations have been found at active peptide (Ser143 to Arg194) with
only 11.6% identity. Some important variations at γ-carboxyglutamic acids and Ca2+binding sites were
identified, while high conservations were discovered at other functional sites. Comparisons on 3-D protein
models demonstrated that the protein backbones of porcine and human FX were highly conserved, and little
difference was shown at the molecular surface of anticoagulant binding sites S2 and S3. PT detection of
porcine and human plasma showed similar results, while coagulation activities of porcine FVII and FX were
remarkably higher than that of human.
Conclusion: Porcine FVII and FX showed relatively high homology with human FVII and FX in nucleotide,
amino acid sequences and three-dimensional structure. However, the different affinities to important
macromolecules caused by genetic differences might contribute to the molecular incompatibilities in liver
© 2009 Elsevier Inc. All rights reserved.
Factor VII (FVII) and factor X (FX) are two glycoproteins that are
synthesized in liver and circulate in the blood in the inactive
zymogen conformation . FVII is the initiator of extrinsic coagula-
tion pathway and FX is at the point of convergence of the extrinsic
and intrinsic activation pathways leading to the final stages of
hemostasis . Activation of Factor X by the intrinsic pathway
requires the interaction of factor IXa and factor VIIIa on a
phospholipid surface in the presence of calcium ion. The extrinsic
FX activation complex is composed of Factor VII/VIIa assembled with
tissue factor. Once activated, Factor Xa associates with Factor Va in
another macromolecular membrane complex responsible for the
activation of prothrombin.
Xenotransplantation using pigs as the transplant source has the
potential to resolve the severe shortage of human organ donors .
However, due to the evolutionary distance, a series of immunological
and physiological barriers must be overcome for the successful clinical
application of xenotransplantation . Function compatibility of
porcine coagulation factors synthesized by liver is one of the
determinants for pig to human liver transplantation. Previous studies
demonstrated that porcine coagulation factors in plasma could trigger
human intrinsic, extrinsic and common clotting pathway . Other-
wise, several potential important molecular incompatibilities may
Blood Cells, Molecules, and Diseases 43 (2009) 111–118
⁎ Corresponding author. Fax: +86 28 85164034.
E-mail address: email@example.com (J. Cheng).
1Co-first authors for their equal contribution to this paper.
1079-9796/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
Blood Cells, Molecules, and Diseases
journal homepage: www.elsevier.com/locate/ybcmd
arise when pig coagulation factors interact with human ligands and
contribute to coagulation disorder after transplant , which are far
from clear elucidation.
The relatively ineffective binding of porcine tissue factor pathway
inhibitor (TFPI) expressed on porcine endothelium to human FX has
been clearly proved , while the compatibilities of porcine FX or FVII
with human coagulation related molecules have not been investi-
gated. In this study, wecloned thefull-length cDNAof porcine FVII and
FX, and compared the nucleotide sequences, deduced protein
sequences and three-dimensional structure models with human FVII
and FX, particularly on some important function domains and
interaction sites to give insight into the possibility of molecular
incompatibility caused by porcine FVII and FX when pig to human
Reagents and equipments
Oligotex mRNA Mini Kit, from Qiagen; SMART™RACE cDNA
Amplification Kit, from BD; pGEM®-T Easy Vector Systems, from
Promega; RNA PCR Kit (AWV) Ver.3.0, from TaKaRa; STA-Stargo
autoanalyzer from France; PT assay kit, from Diagnostic Stago,
Asnieres France; and FVII-deficient human plasma and FX-deficient
plasma from Diagnostic Stago, Asnieres France.
Screening a pig liver tissue cDNA library
A pig liver tissue cDNA library was constructed from a healthy
6 years old male pig of Bana Minipig Inbred Line JS 151 substrain (16th
generation) as our previous report . PCR was used to screen the
target gene from the cDNA library, according to the protocol described
previously [9,10]. Before screening the library, a partial fragment of
porcine FVII cDNA was acquired by RT-PCR with the specific primers
designed based on the most conservative sequences from human,
mouse and bovine FVII. The PCR product was sequenced and
confirmed to be the porcine FVII gene by BLAST program of NCBI.
New primers were designed based on the sequence of the above PCR
product. PCR and sequencing were performed again to confirm the
specificity and efficiency of the primers that were used in subsequent
Several rounds of PCR with cDNA library as template were then
conducted to zoom out the screening pool, and pick out the single
clone recombinated with target gene finally. In the first round of PCR,
the templates were aliquots of divided primary cDNA library super-
natant; in the second round, the templates were different dilutions of
the positive aliquot of the first round PCR. Then the sample with the
lowest concentration while positive amplification was dispensed into
96-well plate as 8⁎8 pattern. Aliquots of each row or each column
were collected to serve as PCR templates. The double positive well was
determined to be the subpool for next round of screening. After two
rounds of PCR screening, a subpool with titer less than N⁎103pfu/ml
was obtained, which had small complexity but amplified positive
clones. 100 pfu of the last subpool were cultured on the plate, and
several single clones were randomly selected for PCR to identify the
positive recombinants, then sequencing was performed by ABI 3700
Sequencer. The sequence was named FVII-S.
mRNA isolation from porcine liver tissue
Total RNA from porcine liver tissue was isolated using Trizol
reagents. mRNA was purified from total RNA using Oligotex
polystyrene-latex particles according to the recommendations in
the protocol of Oligotex mRNA Mini Kit. The quantity and integrity
of RNA were determined by optical density measuring and
5′ Rapid amplification of cDNA ends (RACE)
1 μg mRNA of pig liver tissue was used to produce 5′ terminal
enriched cDNA with special primer (BD SMART II™A Oligo). Gene
specific reverse primers were designed according to the 5′ end of the
acquired sequence. The Universal Primer (UPM) is used in the first
round of nested PCR reaction and the Universal Nest Primer (UNP)
was used in the second round of nested PCR reaction. The amplifica-
tion cycle was: 94 °C for 30 s, 72 °C for 3 min, for 5 cycles; 94 °C for
30 s, 70 °C for 30 s, 72 °C for 3 min, for 5 cycles; 94 °C for 30 s, 68 °C for
30 s, 72 °C for 3 min, for 30 cycles. The PCR product was ligated into
the pGEM-T Easy Vector, and was verified by DNA sequencing. The
result was named FVII-R.
3′ Rapid amplification of cDNA ends (RACE)
To get the 3′ end mRNA sequence of FX, 3′ RACE was performed. 3′
RACE-Ready cDNA was produced with 1 μg mRNA and special primer,
3′ CDS primer. Three times of nested PCR reaction with different
combinations of primers were performed to reach the 3′ end. Reverse
primer UPM and UNP were used in the first and second round of
nested PCR reaction respectively as 5′ RACE. Gene specific forward
primers were designed according to the 3′ end of the acquired
sequence. The sequencing results of three round PCR were assembled
to a prolonged 3′ RACE sequence named FX-3R.
Sequence analysis and alignments
Fragment assembling, multiple sequence alignments and further
analysis of the full-length cDNA were performed by DNAMAN
biological software. BLASTN and BLASTX programs of NCBI (http://
www.ncbi.nlm.nih.gov) were used to align the new sequence against
For further analysis of porcine FX protein sequence, several online
ExPASy Proteomics tools (website: http://au.expasy.org/tools/) were
used, such as Simple Modular Architecture Research Tool (SMART),
SignalP, NetOGlyc, and NetNGlyc, to predict function domains, signal
peptide cleavage sites and glycosylation sites, respectively.
Homology modeling of the three-dimensional protein structure of
The 3-D protein model of porcine FX was generated with SWISS-
MODEL program (http://swissmodel.expasy.org/). Figures were
produced using the DeepView program. The 3-D structures using as
the templates to generate the new models and to carry out the
comparative study were obtained from SIB's ExPDB database at the
following website: http://www.rcsb.org/pdb/ .
Activity detection of FVII and FX of healthy pigs
Venous blood samples were collected from 23 individuals of
Chinese Guizhou Minipig bred in Agriculture Institute of Guizhou
University, 6–10 months, 9 males and 14 females. Venous blood was
collected into plastic tubes with 0.11 mol/L trisodium citrate, then the
plasma was separated by centrifugation at 2500 g for 15 min. The
coagulation activity of FVII was detected in a clinical laboratory of
West China Hospital according to the universal protocol used in
clinical samples. Prothrombin Time (PT) of porcine plasma was
determinedusing a STA-Stargo autoanalyzer to analyze the function of
extrinsic coagulation pathway. Porcine plasma was added in FVII-
deficient or FX-deficient human plasma and the PT and activated
partial thromboplastin times (APTT) were tested to determine the
function of porcine factors in human condition. When compared with
human, the PT results of routine pre-operation test from 44 surgery
patients were used, while the reference value of FVII and FX activities
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
of healthy human were provided by Department of Laboratory
Medicine of West China Hospital.
Characterization of porcine FVII
Cloning the full-length cDNA of porcine FVII
Positive clones were enriched and selected after screening the
porcine liver tissue cDNA library. The cDNA insert named FVII-S was
verified to code porcine FVII by multiple alignments with FVII mRNA
or protein sequences from human, bovine, dog and so on. The FVII-S
contained 1343 bp with an incomplete open reading frame (ORF), but
the C-terminal of the predicted amino acid was complete evidenced
by the poly A signal (AATAAA) and comparison with the C-terminal
amino acid sequences of other FVII proteins.
5′ RACE reaction amplified a 490 bp fragment named FVII-R. The
deduced protein sequence was consistent with the N-terminal amino
acids of other FVII proteins. Assembling FVII-S and FVII-R yielded a
1416 bp cDNA coding 445 amino acids with predicted molecular
weight of 49,286.4 Da. Sequence of the full-length porcine FVII cDNA
(Fig. 1) was submitted to GenBank, and the acquired accession
number is DQ50371.
Characterizing the predicted amino acid sequence of porcine FVII
Comparing the amino acid sequences of porcine FVII with that of
other three mammalians demonstrated that the highest homology of
78.3% identity was between porcine FVII and bovine FVII, the next of
74.08% was betweenporcine andhuman,andthelowestof70.94%was
between porcine and mouse.
The functional domains of porcine FVII were deduced by SMART
program as following: signal peptide (1 to 20aa), gamma-carboxyglu-
tamicacid (Gla) domain (39 to 83aa), two epidermal growth factor
(EGF) domains (84 to 120aa, 128–166aa), and catalytic domain
(190 to 425aa, trypsin domain). This pattern is conserved within
vitamin K-dependent coagulation factors of different species. The
individual domain of FVII showed different similarities among the
four species. Table 1 displayed the amino acid identity percents when
porcine individual domain compared with the corresponding region
of bovine, human and mouse, and EGF1 domain showed the highest
identity in the four species. As expected, all the cysteine residues of
each domain were conserved between species, which suggested that
FVII from different species may share similar folding conformation
due to the conserved disulfide bonds.
Sequence analysis of glycosylation sites showed minor variations
in the four species. NetNGlyc 1.0 program had revealed two potential
N-glycosylation sites (N183, N222) in porcine FVII. Similar results
were obtained in bovine and mouse as bovine FVII (N185, N243) and
mouse FVII (N185, N243), while apparent different sites in human FVII
Functional site analysis and comparison with human FVII
The vast majority of FVII in blood is in the form of unactivated
zymogen. The activation of zymogen FVII to enzyme FVIIa is largely
TF-dependent and arises at the site of vascular injury after formation
of TF/FVII complex . Considering the critical roles of TF binding
and proteolysis in the physiological function of FVII, we compared
some important functional amino acids of human FVII with porcine
FVII to predict the potential different processes when porcine FVII
interacted with human molecules.
Fig. 1. Nucleotide and deduced amino acid sequences of porcine FVII (accession number DQ530371) are shown. The initiation codon (atg) and end codon (tag) are underlined,
respectively. The starting and ending points of each domain are indicated by double lines. The signal peptide is 1–20aa, the Gla domain is 39–83aa, the EGF1 domain is 84–120aa, the
EFG2 domain is 128–166aa, and the catalytic domain is 190–425aa, respectively. Two potential N-glycosylation sites (N) are boxed as N183 and N222.
Amino acid identities of each domain in porcine FVII and FX compared with that of
bovine, human and mouse
Gla domainEGF1 domain EGF2 domainTrypsin domain
FVII (%) FX (%) FVII (%)FX (%) FVII (%) FX (%)FVII (%) FX (%)
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
Key residues in light chain. The essential event in activation of FVII is
a cleavage of the peptide bond between Arg15 and Ile16 (using the
chymotrypsinogen numbering convention). The first 152 amino
acids constitute the light chain. Various experimental approaches
indicated that the Gla domain, the EGF1 domain and the protease
domain provided the most significant contributions to the interac-
tion with TF [13,14]. The N-terminal Gla domain is fully loaded with
7 Ca2+ions and has significant contact with the C-terminal domain
of TF. The EGF1 domain contains a single Ca2+ion, and extensive
contacts with TF. Sequence alignments manifested that Gla domain
of porcine FVII had an additional Glu (GLu34) suggesting that there
might be an additional γ-carboxyglutamic acid in porcine FVII after
post-translation modification, which might be ascribed different
Dickinson demonstrated that some surface-exposed residues in
VIIa had played a functionally important role in binding to TF by a
structurally conservative approach of alanine scanning mutagenesis
. As to the light chain, they were Lys62, Gln64, Ile69, Phe71, and
Arg79, and the largest reduction in the calculated free energy binding
was for the mutant at Ile-69. Except for one variation of Lys62-Glu,
these residues are conserved in porcine light chain.
Key residues in heavy chain.
unique protease domain (trypsin domain), is usually considered to
have three allosteric sites — the TF binding region, the active site
binding cleft, and the macromolecular substrate exosite . Among
several residues of protease domain contacting with TF, only Arg134,
Met164, and Asp309 substitutions affected cofactor affinity ,
whereas, other residues (Pro161, Leu163, Arg379) displayed binding
defects but did not contact TF in the crystal structure . These
residues are all conserved in porcine protease domain.
The protease domain has a single high-affinity Ca2+coordinating
site in the loop from Glu70 to Glu80. Binding with Ca2+in this region
is associated with catalytic function of VIIa . Inporcine FVII, Lys75-
Asp76-Glu77 replaced Glu75-His76-Asp77 of human FVII. However,
evidence by mutations only showed the function defects when Glu70
and Glu80 were substituted.
Loops 142–152, 186–194 and 216–223 in combination with the N-
terminus differ structurally between zymogen and enzyme, which are
termed the “activation domain” . The catalytic triad (His57-
The heavy chain of FVII, containing the
Asp102-Ser195) is flanked by two functionally important regions. In
the loop 96–99, residue Thr99 contributes to X activation. Secondly,
Lys192 and Arg148 constitute a prominent basic cluster just below the
catalytic triad. These functional sites are associated with macromo-
lecular substrate exosite, andplaycrucialrolesinproteolysis efficiency
of FVII . Among the residue replacements, the most significant
functional defects were observed for Gln217, Trp215, Leu41, Gln40,
Thr151, Leu73, Glu154, Met156, and Leu144 . Surprisingly, in
contrast to the high conservation observed at a previous region, 4 of
these 9 important functional residues displayed variations, as Gln217-
Glu, Thr151-Lys, Glu154-Val and Gln40-Leu. Additionally, there were
other three variations observed in activation domains, as Ser189-Ala,
Thr221-Ala and Val222-Thr. The distinct difference in the amino acid
sequence of porcine FVIIa for these important sites would rise
potential difference in the interaction with substrates.
Three-dimensional modeling of porcine FVII heavy chain
The heavy chain of FVII contains the most important catalytic
domain responsible for proteolysis function and is subject to
numerous allosteric influences. Here, we preliminarily compared the
conformations of some functional important sites between porcine
and human FVII heavychain, using 3-D protein models constructed by
knowledge-based comparative protein modeling program SWISS-
MODEL. A cluster of crystal structures of homologous proteins were
selected from the Protein Data Bank (PDB) as primary templates to
construct the new model of porcine FVII, such as PDB 2aerH, 1kliH,
2b7dH. The X-ray crystal structure of a complex of human FVIIa and
soluble tissue factor (PDB 1DAN) was chosen from PDB, and the TF
structure was excluded in comparison. As shown in Fig. 2, the
conformations of porcine and human FVII heavy chain were roughly
consensus, characterized by abundant β-strands constructing two
consecutive β-barrels . Structural changes accompanying zymo-
gen conversion to enzyme are restricted to the 2nd-barrel. In spite of
some significant variations in primary sequence, no remarkable
difference was observed in the backbone conformations between
the two structures, especially when focusing on some critical
functional sites such as catalytic cleft, activation domains, Ca2+
binding loop, and TF binding region. A salt bridge had been paid more
attention for its important role in FVII activation. When the cleavage
between Arg15 and Ile16 occurs, the new N-terminal residue Ile16
Fig. 2. Comparison of 3-D structures of porcine FVII heavy chain and that of human. (a) porcine FVII heavy chain; (b) human FVII heavy chain. (i) Ca2+binding site is colored in
orange; (ii) activation domains including three loops (142–152,186–194 and 216–223) are colored in light blue; (iii)catalytic triad (His57-Asp102-Ser195) is colored in rose red; (iv)
the two residues involved in the important salt bridge Ile16-Asp194 are colored in deep blue, and the B2 strand involved in the H-bond registration is colored in red; (v) the four
variations of Gln217-Glu, Thr151-Lys, Glu154-Val and Gln40-Leu are colored in red.
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
should bury its side chain in the hydrophobic environment by
establishing a salt bridge between its amino nitrogen atom and the
side chain of residue Asp194 that is adjacent to the catalytic triad. This
insertion is prerequisite to the complete activation of FVIIa, and it is
accompanying the loss of H-bonds between Glu154 and residues
Val21 and Cys22 via a reregistration of strand B2 . In the model of
porcine FVII heavy chain, similar allosteric features were also
observed. The Ile16 was inserted to the catalytic cleft, and the
identical H-bond registration in strand B2 was shown after H-bonds
were build in this region.
The activity of porcine FVII by plasma detection
The results of PT and FVII activity detection were listed in Table 2.
Briefly, no significant difference was observed when porcine plasma
compared with normal human plasma, nevertheless, the average
porcine FVII activity determined by VII-deficient human plasma was
significantly higher than the average level of human reference value
Characterization of porcine FX
Full-length cDNA cloning of porcine FX
The RT-PCR with homologous primers had amplified a 466 bp
segment, which served as the basic sequence for following RACE
reaction. 839 bp sequence was acquired in 5′ RACE reaction. The
sequencing results of three nested PCR reaction for 3′ RACE were
assembledtoa continuous segmentof600 bp.AssemblingFX-PCR,FX-
5R and FX-3R had generated an integrated sequence of 1856 bp with a
single complete ORF of 1437 bp, a 49 bp 5′ untranslated flanking
region, and a 368 bp 3′ untranslated flanking region. The deduced
protein sequence had 479 amino acids with predicted molecular
weight of 53,076.9 Da. Sequence of the full-length porcine FX cDNA
was submitted to GenBank (accession number DQ530372), and its
basic information was presented in Fig. 3.
Characterizing the predicted amino acid sequence of porcine FX
Multiple alignments of amino acid sequences of FX were
performed between pig and human, bovine and mouse. The results
indicated that porcine FX was highly homologous to bovine FX with
identity of 79.27%. The identity shared by porcine and human FX was
73.10%, and by porcine and mouse FX was 73.51% respectively.
SMART analysis of deduced protein sequence of porcine FX
indicated the conserved functional domains as following: signal
peptide (1 to 23aa), gamma-carboxy glutamic acid (Gla) domain (41
to 85aa), two EGF domains (86 to 122aa, 128–165aa), and catalytic
domain (234 to 462aa, trypsin domain). This pattern is conserved
within FX derived from different species of animals. Comparison of
individual domain of pig FX with corresponding regions in that of
bovine, human and mouse revealed a higher similarity than entire
sequence comparisons (Table 1). As FVII, all cysteine residues of each
domain were conserved between species.
Sequence analysis of glycosylation sites showed variation in the
four species. NetNGlyc 1.0 program had revealed two potential N-
glycosylation sites (N109, N217) in pig FX sequence, while different
sites in human FX (N221, 231), bovine FX (N311) and mouse FX
(187,218). NetOGlyc 3.1 program had identified one potential mucin
type GalNAc O-glycosylation site in porcine FX (T207), while quite
variation in human FX (T199, T211, T483), bovine FX (T16, T55, T301,
T586, T576, T578) and mouse (T208).
Comparison of the coagulation activity of porcine and human FVII, FX
aPT was determined using plasma of human and pig, PN0.01.
bActivities of porcine FVII and FX were determined by adding porcine plasma into
FVII/FX-deficient human plasma, Pb0.01.
cHuman activities of FVII and FX are the reference values provided by the
Department of Laboratory Medicine of West China Hospital.
Fig. 3. Nucleotide and deduced amino acid sequences of porcine FX (accession number DQ530372) are shown. The initiation codon (atg) and end codon (tag) are underlined and
italic, respectively. The signal peptide is underlined; the Gla domain is boxed; the two EGF domains are shaded in grey and the Trypsin domain is both boxed and shaded. Two
potential N-glycosylation sites (N) and one potential mucin type GalNAc O-glycosylation site are in bold and underlined.
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
Functional site analysis and comparison with human FX
FX can interact with several macromolecules to regulate blood
coagulation . Therefore, analyzing the substrate recognition sites
is important for predicting the function of porcine FX. (Following
amino acid numbers was based on mature FX sequence without signal
peptide, which started from the first amino acid of Gla domain.)
Sequence alignments suggested that Gla domain of porcine FX had
ten glutamic acid residues, while additional one (Glu 36) was found in
that of human. It suggested that there might be an additional γ-
carboxyglutamic acid in porcine FX after post-translation
Two patches were identified to be involved in membrane binding
of FX bycrystallographic study . The hydrophilic patch 1 of the Gla
domain, including Glu25, Arg28, Glu29 and Glu32, is the major
contributor toward the ionic interaction with phospholipid and the
Ca2+ion bridging. Hydrophobic residues Phe4, Leu5 and Val8 of
hydrophobic patch 2 of the Gla domain of FX are another contributor
to membrane binding. These two binding sites were found highly
conserved in porcine FX, except for one variation of Leu5 to Trp.
Asp63 in EGF1 domain presents as β-hydroxyaspartic acid of
human FX, FVII and FIX, which is the target for Asp/Asn-hydroxylase
to generate mature protein from prepeptide . The same pattern
was observed in porcine FX.
Sequence alignments between protein sequences of porcine and
human FX had revealed a remarkable variant segment of Ser143 to
Arg194 between EGF2 domain and Trypsin domain. Only 6 of total 51
amino acids were identical, which was much lower than conservation
of other position. However, this segment might be cut off when
activation process with minor influence on protein function.
As other vitamin K-dependent coagulation factors, Gla and EGF
domains of FX are the major functional parts to ligate Ca2+, otherwise,
catalytic domain of heavy chain is also involved in Ca2+ligation to
generate Xa-Va-prothrombin complex . Compared with human
FX, the Ca2+ligation patch of porcine FX heavy chain presented a
relatively low identity of 66%, with four amino acids variation of total
eleven (human: DRNTEQEEGGE; pig: DHNLEKEEGDE).
Crystallographic study suggested that Ala183–Asp194 of human FX
was the primaryactive site, involved in constructing substrate binding
pocket (using the chymotrypsinogen numbering convention). Ser195,
His57, Asp102, Lys96, Tyr99, Trp215, Ser214 and Gln61, located beside
it, have also played important roles in proteolysis function of FX, and
might be the target of anticoagulation molecules such as tissue factor
pathway inhibitor (TFPI) and AT-III . Most of these amino acids
were found conserved in porcine FX, except for some differences in
position Lys96 and Tyr185-Lys186-Gln187.
Three-dimensional modeling of porcine FX heavy chain
A cluster of known crystal structures was selected from the Protein
Data Bank (PDB) as primary templates to construct the 3-D model of
porcine FX, such as 1kigH,1c5mD and 1iqhA. The crystal structure of a
complex of human Xa and inhibitor RPR209685 (PDB NFW ) was
chosen from PDB too. The modeling results (Fig. 4) demonstrated a
remarkable conservation of FX heavy chain in backbone conformation
between the two species. As shown in human model (Fig. 4b), the FX
heavy chain formed four α-helixes and more than ten β-strands, and
the same pattern was observed in porcine model (Fig. 4a). Ca2+
binding site of human and porcine FX had shown some difference in
primary structure, however, no apparent difference was seen in the
loop structure of the two 3-D models. The catalytic triad DHS, His57,
Asp102, Ser195, was shown in the center of the molecule, which was
completely conserved in the two models. The up left side surface of
the model was active site described before. The loop structures of the
two models had presented no significant difference, while when the
molecular surfaces were created, some obvious differences arose.
In the family of vitamin K-dependent coagulation factors, some
resembling macromolecular interaction sites have been disclosed,
which have become the recognition sites of many anticoagulant
chemical medicine and peptides. Crystallographic studies of human
Xa heavy chain has discovered four pocket-like sites named S1, S2, S3
and S4, which are potential targets for anticoagulant reagents . S1
has been located in the active site Gln192; S2 is quite associated with
the conformation of Tyr60, Lys90, Lys96 and Tyr99; S3, the medicine
design target of serine protease inhibitor, is located at Gly216; Glu97,
Lys96, Tyr99, Phe174 and Trp215 are involved in S4 (ary-binding site),
which are mainly combined with aromatic amino acid or lipid
molecules. Though the primary sequences of these sites were highly
conserved in porcine FX heavy chain, some important differences had
been distinguished when the molecular surfaces of these sites were
creased, especially at S2 and S3. Porcine S2 and S3 sites had
constructed a close pocket-like surface, while the pocket structure
between human S2 and S3 was open because the Try210 and Gly211
could not generate hydrophobic surface.
Comparison of FX activity between pig and human
The detection of porcine FX activity by X-deficient human plasma
showed that the average level of FX activity of 23 pigs was 142.00±
Fig. 4. Comparison of 3-D structure of porcine FX heavy chain and that of human. (a) porcine FX heavy chain; (b) human FX heavy chain. Active site is colored in pink; Ca2+binding
site is colored in deep blue; Na+binding site is colored in light blue; activity triad DHS is colored in red; the macromolecular inhibitor binding sites S1, S2, S3 and S4 are colored in
green, yellow, purple and blue, respectively. The backbone conformations of porcine and human FX heavy chains are remarkably conserved at 3-D level. Some differences are
observed at the molecular surface of active site and S2 and S3 macromolecular inhibitor binding sites.
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
26.14, while the reference value of healthy human is 70–120 (Pb0.05).
So porcine FX represented a 1.49 fold activity to human FX.
In this study, full-length cDNA of porcine FVII has been cloned and
sequenced. The comparison showed relatively high homology of
porcine FVII with human FVII in amino acid sequence and three-
dimensional structure. No significant evidence was disclosed in the
difference of TF binding affinity, while the variations at important
functional sites might account for the different coagulation activities
between porcine and human FVII, which may contribute to stronger
clotting after xeno-liver transplantation.
In pig to human liver transplantation, pig synthesized FVII would
be secreted into human peripheral circulation, thus the efficient
function of porcine FVII has to depend on the interaction with TF
expressed on human endothelia cells. Conversely, aberrant interac-
tion between porcine TF and human FVII will result in coagulopathy
after pig to human heart or kidney transplantation. The compatibility
of porcine TF and human FVIIa was demonstrated through an in vitro
experiment by Kopp et al.  Their results showed that when TNFα-
activated porcine aortic endothelia cells (PAEC) were incubated with
human factor X and with the catalytic amount of activated human
coagulation factor VII, increased TF activity was observed when
compared with human aortic endothelia cells (HAEC), indicating
effective cofactor activity of porcine TF for human factor VIIa. In
contrast to the efficient coagulation activity, molecular incompat-
ibility between porcine tissue factor pathway inhibitor (TFPI) and
activated human factor X have been comprehensively documented.
Otherwise, there is poor knowledge of interaction of porcine FVII with
human TF. Our comparison of the amino acids of TF binding region
between porcine and human FVII showed a quite similarity. EGF1
domain, the most important contributor to TF binding, presented the
highest identity among the total domains, and these key amino acids
for binding energy mentioned above were all consensus in porcine
sequence. The three-dimensional structure model of porcine FVII
light chain fitted the human FVII light chain structure quite well in
ribbon (data not shown). These data will provide beneficial
complements to our knowledge about porcine FVII, and shed insight
into the further investigation of the compatibility between porcine
FVII and human TF.
In our study, coagulation activity of FVII was determined by a
universal method for clinic examination. The PTassay is considered as
an important reference for FVII activity. The normal PT of porcine
plasma compared with human suggested porcine FVII had similar
physiological function as human FVII in its auto condition. When
porcine plasma added into FVII-deficient human plasma, porcine FVII
in plasma has interacted with both auto and xeno macromolecular
substrates, thus the stronger clotting capacity reflected the higher
proteolytic activity of porcine FVII to xenosubstrates. This result is
consistent with our previous report in Zhang, L et al. , in which 4.7-
fold FVII activity was observed whenporcine FVII was added into FVII-
deficient human plasma. When comparison of porcine proteolysis
domainwith that of humanwas performed, some genetic evidence for
the different activity were identified. Catalytic cleft and macromole-
cular substrate binding exosite are predominantly responsible for the
proteolytic function of FVII. A potential mechanism by which TF may
activate VIIa is through stabilization of the active conformation of the
protease domain . 97 Ala mutations in the protease domain
discovered 33 substitutions resulting in a reduction of proteolytic
function without affecting TF binding . Among them 9 residue
mutations could cause most significant functional defects. 4 variations
in these 9 residues and another 3 variations in the activation domains
observed in porcine FVII protease domain might contribute to
different allosteric influences on the interaction with substrate.
Although obvious conformation changes were not identified in the
ribbon structure of three-dimensional model of porcine FVII, different
binding affinity, catalytic rate and conformation stability would arise
from minor changes in single amino acid. Considering the potential
strongeractivityof porcine FVIIin humanenvironment,thereis a clear
need for an extensive study of molecular incompatibility between pig
FVII and human molecules to elucidate the possible role of porcine
FVII in xenocoagulation.
Previous studies have disclosed that porcine TF in association with
its cofactor, human factor VIIa, was highly efficient at converting
human factor X to Xa . However, porcine tissue factor pathway
inhibitor (TFPI) was relatively ineffective at binding human factor Xa
. Consequently, it failed to inhibit Xa and efficiently generate the
TFPI/Xa complex which is needed to inhibit the VIIa/TF complex .
When pig to human liver transplantation, FX synthesized by porcine
liver would be secreted into peripheral blood and contact with human
original coagulation factors and anticoagulants. The incompatibilities
between porcine FX and human molecules would contribute to
coagulation disorder after transplant.
In spite of the relatively high conservations in sequences of porcine
FX, the variations at positions of glycosylation sites were considerably
worth to pay attention, for the important role of glycosylation in
physiological function of coagulation factors . These sites were
predicted by bioinformatics and should be tested by experiments
when further researches on their influences on protein function.
Gla domain provides critical regions for Ca2+binding and
membrane binding [21,30]. Porcine FX Gla domain showed significant
higher similarity with that of human than bovine and mouse.
Consistently, the key functional residues involved were highly
conserved, suggesting less trouble when porcine FX attach human
Ser143 to Arg194 is located just between EGF1 domain and Trypsin
domain, which exhibited remarkably lower homology level than any
of the other domains. This is not surprising, since this segment would
be cleaved out upon activation of FX and is therefore probably not a
critical component for the coagulant activity. When the single chain
form of prepeptide is converted to mature FX with double chains,
cleavage occurred at Arg139–Arg140, and a short peptide of Arg140-
Lys141-Arg142 is thrown out. Subsequently, when FX is activated by
TF/VII or IXa/VIIIa complex, cleavage at Arg194-Ile195 creates
disulfide bond at 132aa and 302aa while the segment of Arg139–
Arg140 is leaved out . Though little impact onprotein function, the
great variation reflected the difference of evolution, and has been
considered as a specific feature of different species.
The heavy chain (trypsin domain), containing essential binding
sites for substrate and anticoagulants, is the most important
functional part in all of the coagulation factors. Therefore, the trypsin
domain of FX is notonlycriticalinprocoagulant process, but alsois the
target for many physical or artificial anticoagulant molecules to break
the normal configuration and hinder the binding of procoagulants
. Previous in vivo and in vitro studies have demonstrated that
porcine TFPI has failed to efficiently inhibit human FXa, but the
cloning of porcine TFPI did not found significant difference from
human TFPI . It is reasonable to suppose that there will have some
differences in the TFPI binding sites on porcine and human FX.
Nevertheless, high conservations were found in the active sites of
Trypsin domain with only 4aa variations, which didn't give us helpful
evidence for the possible differences at TFPI binding sites between
Chattopadhyay and Fair reported that three segments were
important in Xa binding with hydrolysis substrates such as IXa/VIIIa
and TF/VIIa, which were identified by inhibition assay with
synthesized peptides . Sequence alignments suggested porcine
and human FX only shared 50% identity in 267–277aa that is included
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118
in the most efficient segment mentioned above. This result suggested Download full-text
that some potential incompatibility may occur when porcine FX
interacts with human substrates.
In our study, the analysis of 3-D protein model had concentrated
on the backbone structures of heavy chains of porcine and human FX.
Only slight difference had been discovered, while great similarity had
presented in most important functional sites. This result may provide
evidence for the possible replaceable function when porcine FX
was released into peripheral blood after pig to human liver
When the porcine FX activities were detected by X-deficient
humanplasma,ourresult suggested thattheactivityof porcine FX was
1.49 fold of average human FX activity. This result agrees with our
previous study on Chinese Bana Minipig Inbred Line, another strain of
Chinese pig, reported by Zhang L. et al. , who found that the porcine
FX activity was 2.9 fold of that of human. However, in that paper, no
apparent difference in PT and APTT of porcine or human plasma was
described. The stronger activity of porcine FX suggested that porcine
FX could trigger the clotting pathway in human plasma through
efficient cross-species interaction with human other coagulation
factors, but the stronger function may impact the balance of
coagulation and anticoagulation resulting in hypercoagulable state
after pig to human liver transplantation. Because the detection was
based on a mixed reaction, which molecule is the decisive factor of
different activity between pig and human is still unclear.
This study was supported by National Basic Research Program of
China No. 2009CB522401; National Natural Science Foundation of
China No. 30772037; National Program for High Technology Research
and Development of China No. 2006AA02A117; Program for Chang-
jiang Scholars and Innovative Research Team in University, Ministry of
Education; China Postdoctoral Science Foundation No. 20080430195.
We are grateful to Professor Liu Peiqiong of the Agriculture Institute of
Guizhou University, who kindly provided us the Guizhou Minipig
 K.G. Mann, Factor VII-activating protease: coagulation, fibrinolysis, and athero-
thrombosis? Circulation 107 (2003) 654–655.
 E.W. Davie, K. Fujikawa, W. Kisiel, The coagulation cascade: initiation, main-
tenance, and regulation, Biochemistry 30 (1991) 10363–10370.
 J.Y. Deschamps, F.A. Roux, P. Sai, E. Gouin, History of xenotransplantation,
Xenotransplantation 12 (2005) 91–109.
 D.K. Cooper, B. Gollackner, C. Knosalla, K. Teranishi, Xenotransplantation—how far
have we come? Transpl. Immunol. 9 (2002) 251–256.
 L. Zhang, Y. Li, H. Jiang, et al., Comparison of hepatic coagulant, fibrinolytic, and
anticoagulant functions between Banna Minipig Inbred line and humans,
Transplantation 79 (2005) 1128–1131.
 F.H. Bach, H. Winkler, C. Ferran, et al., Delayed xenograft rejection, Immunol. Today
17 (1996) 379–384.
 C.W. Kopp, J.B. Siegel, W.W. Hancock, et al., Effect of porcine endothelial tissue
factor pathway inhibitoron human coagulation factors, Transplantation 63 (1997)
 W. Tan, Y. Chen, L. Zhang, et al., Construction and characterization of a cDNA
library from liver tissue of Chinese Banna Minipig inbred line, Transplant. Proc. 38
 Y. Chen, W. Tan, X. Lu, et al., Full-length cDNA cloning and protein three-
dimensional structure modeling of porcine prothrombin, Blood Cells Mol.
Diseases 38 (2007) 93–99.
 J. Wang, W. Tan, Y. Chen, et al., Full-length cDNA cloning and protein three-
dimensional structure modeling of factor VII of rhesus monkey, Macaca mulatta,
Blood Cells Mol. Diseases 40 (2008) 237–243.
 D.R. Grimm, M.B. Colter, M. Braunschweig, et al., Porcine factor V: cDNA cloning,
gene mapping, three-dimensional protein modeling of membrane binding sites
and comparative anatomy of domains, Cell. Mol. Life Sci. 58 (2001) 148–159.
 T.S. Edgington, C.D. Dickinson, W. Ruf, The structural basis of function of the TF.
VIIa complex in the cellular initiation of coagulation, Thromb. Haemost. 78 (1997)
 L.V. Rao, T. Williams, S.I. Rapaport, Studies of the activation of factor VII bound to
tissue factor, Blood 87 (1996) 3738–3748.
 W. Ruf, The interaction of activated factor VII with tissue factor: insight into the
mechanism of cofactor-mediated activation of activated factor VII, Blood Coagul.
Fibrinolysis 9 (Suppl. 1) (1998) S73–S78.
 C.D. Dickinson, C.R. Kelly, W. Ruf, Identification of surface residues mediating
tissue factor binding and catalytic function of the serine protease factor VIIa, Proc
Natl. Acad. Sci. U. S. A. 93 (1996) 14379–14384.
 C. Eigenbrot, Structure, function, and activation of coagulation factor VII, Curr.
Protein Pept. Sci. 3 (2002) 287–299.
 D.W. Banner, A. D'Arcy, C. Chene, et al., The crystal structure of the complex of
blood coagulation factor VIIa with soluble tissue factor, Nature 380 (1996) 41–46.
 V.J. Bom, R.M. Bertina, he contributions of Ca2+, phospholipids and tissue-factor
apoprotein to the activation of human blood-coagulation factor X by activated
factor VII, Biochem. J. 265 (1990) 327–336.
 R. Huber, W. Bode, Structural basis of the activation and action of trypsin, Acc.
Chem. Res. 11 (1978) 114–122.
 J. Jesty, A.K. Spencer, Y. Nakashima, et al., The activation of coagulation factor X.
Identity of cleavage sites in the alternative activation pathways and characteriza-
tion of the COOH-terminal peptide, J. Biol. Chem. 250 (1975) 4497–4504.
 H. Mizuno, Z. Fujimoto, H. Atoda, T. Morita, Crystal structure of an anticoagulant
protein in complex with the Gla domain of factor X, Proc. Natl. Acad. Sci. U. S. A. 98
 F.S. Hagen, C.L. Gray, P. O'Hara, et al., Characterization of a cDNA coding for human
factor VII, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) 2412–2416.
 K. Mertens, R.M. Bertina, The contribution of Ca2+ and phospholipids to the
activation of human blood-coagulation Factor X by activated Factor IX, Biochem. J.
223 (1984) 607–615.
 D. Venkateswarlu, L. Perera, T. Darden, L.G. Pedersen, Structure and dynamics of
zymogen human blood coagulation factor X, Biophys. J. 82 (2002) 1190–1206.
 H. Brandstetter, A. Kuhne, W. Bode, et al., X-ray structure of active site-inhibited
clotting factor Xa. Implications for drug design and substrate recognition, J. Biol.
Chem. 271 (1996) 29988–29992.
 C. van 't Veer, T.M. Hackeng, C. Delahaye, et al., Activated factor X and thrombin
formation triggered by tissue factor on endothelial cell matrix in a flow model:
effect of the tissue factor pathway inhibitor, Blood 84 (1994) 1132–1142.
 S. Crikis, P.J. Cowan, A.J. d'Apice, Intravascular thrombosis in discordant
xenotransplantation, Transplantation 82 (2006) 1119–1123.
 T. Hamamoto, M. Yamamoto, O. Nordfang, et al., Inhibitory properties of full-length
and truncated recombinant tissue factor pathway inhibitor (TFPI). Evidence that the
third Kunitz-type domain of TFPI is not essential for the inhibition of factor VIIa-
tissue factor complexes on cell surfaces, J. Biol. Chem. 268 (1993) 8704–8710.
 J. Schulte Am Esch II, S.C. Robson, et al., O-linked glycosylation and functional
incompatibility of porcine von Willebrand factor for human platelet GPIb
receptors, Xenotransplantation 12 (2005) 30–37.
 F. Thiec, G. Cherel, O.D. Christophe, Role of the Gla and first epidermal growth
factor-like domains of factor X in the prothrombinase and tissue factor-factor VIIa
complexes, J. Biol. Chem. 278 (2003) 10393–10399.
 T. Lindhout, G. Willems, R. Blezer, H.C. Hemker, Kinetics of the inhibition of human
factor Xa by full-length and truncated recombinant tissue factor pathway
inhibitor, Biochem. J. 297 (Pt 1) (1994) 131–136.
 A. Chattopadhyay, D.S. Fair, Molecular recognition in the activation of human
blood coagulation factor X, J. Biol. Chem. 264 (1989) 11035–11043.
Y. Chen et al. / Blood Cells, Molecules, and Diseases 43 (2009) 111–118