Given the nature of COX-1, a housekeeping enzyme that is consistently expressed in cells, we hypothesize that a Trip-cat enzyme, constructed by linking COX-1 to PGIS, is likely to demons
Trang 1by linking cyclo-oxygenase isoform-1 to prostacyclin
synthase that can constantly biosynthesize prostacyclin, the vascular protector
Ke-He Ruan, Shui-Ping So, Vanessa Cervantes, Hanjing Wu*, Cori Wijaya and Rebecca R Jentzen* Department of Pharmacological and Pharmaceutical Sciences, Center for Experimental Therapeutics and PharmacoInformatics,
University of Houston, TX, USA
Prostacyclin (prostaglandin I2, PGI2) [1], which has
strong antiplatelet aggregation and vasodilation
prop-erties [1–4], and is synthesized from endothelial and
vascular smooth muscle cells, has been identified as
one of the most important vascular protectors against thrombosis and heart disease [5] Recently, there have been many new studies that have confirmed the impor-tance of PGI2 in vascular protection For instance, it
Keywords
COX; cyclo-oxygenase; PG12; prostacyclin;
prostaglandin 12
Correspondence
K.-H Ruan, Department of Pharmacological
and Pharmaceutical Sciences, Center for
Experimental Therapeutics and
PharmacoInformatics, University of
Houston, Room 521, Science & Research 2
Building, Houston, TX 77204-5037, USA
Fax: +1 713 743 1884
Tel: +1 713 743 1771
E-mail: khruan@uh.edu
*Present address
The University of Texas Health Science
Center, Houston, TX, USA
(Received 15 July 2008,
revised 23 September 2008,
accepted 25 September 2008)
doi:10.1111/j.1742-4658.2008.06703.x
It remains a challenge to achieve the stable and long-term expression (in human cell lines) of a previously engineered hybrid enzyme [triple-catalytic (Trip-cat) enzyme-2; Ruan KH, Deng H & So SP (2006) Biochemistry 45, 14003–14011], which links cyclo-oxygenase isoform-2 (COX-2) to prostacy-clin (PGI2) synthase (PGIS) for the direct conversion of arachidonic acid into PGI2through the enzyme’s Trip-cat functions The stable upregulation
of the biosynthesis of the vascular protector, PGI2, in cells is an ideal model for the prevention and treatment of thromboxane A2 (TXA2 )-medi-ated thrombosis and vasoconstriction, both of which cause stroke, myo-cardial infarction, and hypertension Here, we report another case of engineering of the Trip-cat enzyme, in which human cyclo-oxygenase iso-form-1, which has a different C-terminal sequence from COX-2, was linked
to PGI2 synthase and called Trip-cat enzyme-1 Transient expression of recombinant Trip-cat enzyme-1 in HEK293 cells led to 3–5-fold higher expression capacity and better PGI2-synthesizing activity as compared to that of the previously engineered Trip-cat enzyme-2 Furthermore, an HEK293 cell line that can stably express the active new Trip-cat enzyme-1 and constantly synthesize the bioactive PGI2was established by a screening approach In addition, the stable HEK293 cell line, with constant produc-tion of PGI2, revealed strong antiplatelet aggregation properties through its unique dual functions (increasing PGI2 production while decreasing TXA2 production) in TXA2 synthase-rich plasma This study has optimized engi-neering of the active Trip-cat enzyme, allowing it to become the first to stably upregulate PGI2 biosynthesis in a human cell line, which provides a basis for developing a PGI2-producing therapeutic cell line for use against vascular diseases
Abbreviations
AA, arachidonic acid; COX, cyclo-oxygenase; COX-1, cyclo-oxygenase isoform-1; COX-2, cyclo-oxygenase isoform-2; ER, endoplasmic reticulum; FITC, fluorescein isothiocyanate; IP,PGI2receptor; PGE2, prostaglandin E2; PGF2, prostaglandin F2; PGG2,prostaglandin G2;PGH2, prostaglandin H 2; PGI 2, prostaglandin I 2 (prostacyclin); PGIS, prostaglandin I 2 (prostacyclin) synthase; SLO, streptolysin-O; TM,
transmembrane domain; TXA2,thromboxane A2;TXAS, thromboxane A2synthase.
Trang 2was discovered that PGI2 receptor (IP) -knockout mice
showed an increase in thrombosis tendency [6] Also,
the suppression of PGI2 biosynthesis by
cyclo-oxygen-ase isoform-2 (COX-2) inhibitors was linked to
increased rates of heart disease in human clinical trials
[7] Thus, increasing the biosynthesis of PGI2would be
very useful for protection of the vascular system It is
known that the biosynthesis of prostanoids through
the arachidonate– cyclo-oxygenase (COX) pathway
occurs when arachidonic acid (AA) is first converted
into prostaglandin G2 (PGG2, catalytic step 1), and
then to prostaglandin endoperoxide [prostaglandin H2
(PGH2)] (catalytic step 2) by COX isoform-1 (COX-1)
or COX-2 in cells [8] The PGH2then serves as a
com-mon substrate for downstream synthases, and is
isom-erized to prostaglandin D2, prostaglandin E2 (PGE2),
prostaglandin F2 (PGF2), and prostaglandin I2 (PGI2)
or thromboxane A2 (TXA2) by individual synthases
(catalytic step 3) The overproduction of TXA2, a
pro-aggregatory and vasoconstricting mediator, has been
identified as one of the key factors causing thrombosis,
stroke, and heart disease [1,2] PGI2is the primary AA
metabolite in vascular walls, and has opposite
biolo-gical properties to that of TXA2; it therefore represents
the most potent endogenous vascular protector, acting
as an inhibitor of platelet aggregation and a strong
vasodilator on vascular beds [9–12] Specifically
increasing PGI2 biosynthesis requires a highly efficient
chain reaction between COX and PGI2 synthase
(PGIS), which consists of triple catalytic (Trip-cat)
functions
Recently, we engineered a hybrid enzymatic protein
with the ability to perform the Trip-cat functions by
linking the inducible COX-2 to PGIS through a
trans-membrane (TM) domain [13,14] Here, we refer to this
previously engineered enzyme as Trip-cat enzyme-2
Transient expression of active Trip-cat enzyme-2 in
HEK293 and COS-7 cells has been demonstrated
However, there are concerns in using Trip-cat
enzyme-2 in vivo, because COX-enzyme-2 has an inducible nature, has
a lower capacity to be stably expressed, and may also
lead to numerous pathological processes, such as
cancers and inflammation Given the nature of COX-1,
a housekeeping enzyme that is consistently expressed
in cells, we hypothesize that a Trip-cat enzyme,
constructed by linking COX-1 to PGIS, is likely to
demonstrate stable expression in cells and therefore
lead to constant production of the vascular protective
prostanoid PGI2 To test this hypothesis, in this article
we report the construction of a new Trip-cat enzyme
linking COX-1 to PGIS, which we call Trip-cat
enzyme-1 Our studies have confirmed that Trip-cat
enzyme-1 can be stably expressed in HEK293 cells and
therefore lead to the generation of a cell line that con-stantly delivers the vascular protector PGI2 This study has provided a fundamental step towards specifically and stably upregulating PGI2 biosynthesis in thera-peutic cells for the prevention and treatment of throm-bosis and heart disease
Results
Design of a new-generation Trip-cat enzyme (COX-1 linked to PGIS) that directly converts
AA to the vascular protector PGI2
As described above, we recently invented an approach for engineering an active hybrid enzyme (Trip-cat enzyme-2), by linking human COX-2 to PGIS (COX-2– linker–PGIS), which demonstrated Trip-cat activities in converting AA to PGG2, PGH2, and finally PGI2[13,14] (Fig 1) This finding provided great potential for specif-ically upregulating PGI2biosynthesis in ischemic tissues through the introduction of the Trip-cat enzyme-1 gene into these target tissues On the other hand, there is the COX-1 enzyme, which is well known to have a similar function (coupling to PGIS to synthesize PGI2 in vitro and in vivo) to that of COX-2 The housekeeping enzyme COX-1, which has less pathological impact, could be safer for gene and cell therapies than COX-2, which is involved in the pathological processes of
PGI 2
PGH 2
PGG 2
3 rd
Catalytic reaction
1 st
Catalytic reaction
2 nd
Catalytic reaction
PGIS
Substrate
AA
TM linker
COX-1
Fig 1 A model of the newly designed Trip-cat enzyme-1 Trip-cat enzyme-1 was created by linking COX-1 to PGIS through an opti-mized TM linker (10 amino acid residues) without alteration of the protein topologies in the ER membrane The three catalytic sites in and reaction products of COX-1 and PGIS are shown.
Trang 3inflammation and cancers, and shows inducible
tran-sient expression This suggested that the Trip-cat
enzyme containing COX-1 (Fig 1) may have better
therapeutic potential than that containing COX-2 in
terms of stable expression in cells and pathogenic
prop-erties Also, the X-ray crystal structure shows that the
membrane orientation and the membrane anchor
domain of COX-1 are similar to those of COX-2 This
led us to design a single molecule containing the cDNA
of human COX-1 and PGIS with a connecting TM
lin-ker derived from human bovine rhodopsin [15] (Fig 1)
Cloning of Trip-cat enzyme-1 by linking COX-1
to PGIS
A PCR approach was used to link the C-terminus of
human COX-1 (NCBI GenBank ID: NM_080591) to
human PGIS (NCBI GenBank ID: D38145) by a
heli-cal linker with 10 residues
(His-Ala-Ile-Met-Gly-Val-Ala-Phe-Thr-Trp) derived from human rhodopsin
The resultant cDNA sequence encoding the novel
Trip-cat enzyme-1 (COX-1–10aa–PGIS) was then
sub-cloned into the pcDNA3.1 vector for mammalian cell
expression [13] Note that the entire cDNA sequence
of Trip-cat enzyme-1 encodes a single human protein
sequence, which could be used for therapeutics
Expression of the engineered Trip-cat enzyme-1
in HEK293 cells
Despite the many similarities between human COX-1
and COX-2, there are several important differences
For example, it has been reported that the C-terminal
Leu and the last six residues of COX-1 are important
for the enzyme’s activity [16] However, they are not
identical to those of COX-2 Therefore, it was
interest-ing to investigate whether the linkage (from the
C-ter-minal Leu of COX-1 to the N-terminus of PGIS) in
Trip-cat enzyme-1 would affect its expression, protein
folding, and enzyme activity Using the constructed
pcDNA3.1 COX-1–10aa–PGIS plasmid, the
recombi-nant COX-1–10aa–PGIS protein was successfully
over-expressed in the HEK293 cell line, showing the correct
molecular mass of approximately 130 kDa in western
blot analysis (Fig 2A, lane 1) This indicated that the
linkage from the C-terminal Leu of COX-1 to the
N-terminus of PGIS had no effect on Trip-cat enzyme
expression In addition, a comparison of the expression
levels between COX-1–10aa–PGIS and COX-2–10aa–
PGIS revealed that the transfected HEK293 cells
expressed approximately three-fold more COX-1–
10aa–PGIS protein than COX-2–10aa–PGIS protein
under identical conditions (Fig 2A, lane 2)
Subcellular localization of COX-1–10aa–PGIS
To determine whether the linkage of the C-terminal Leu of COX-1 to PGIS had any effects on the sub-cellular localization of Trip-cat enzyme-1, HEK293 cells expressing the enzyme COX-1–10aa–PGIS were permeabilized and stained Nonsignificant differences were observed in the endoplasmic reticulum (ER) staining patterns for the cells treated with
streptolysin-O (SLstreptolysin-O), which selectively permeabilized the cell membrane, and with saponin, which generally permea-bilized both the cell and the ER membranes (Fig 2B) The results indicated that the modification of the link-age between the COX-1 Leu residue and the PGIS N-terminus had no significant effect on the subcellular localization of COX-1–10aa–PGIS in the cells The idea that the PGIS domain is located on the cytoplas-mic side of the ER and that the COX-1 domain is located on the ER lumen for the overexpressed COX-1–10aa–PGIS was also supported by immunostaining Antibody against PGIS was used to stain the cells trea-ted with SLO or saponin, but antibody against COX-1 would only stain the cells treated with saponin (Fig 2B) These data further confirmed that the 10 amino acid linkage between COX-1 to PGIS had no significant effects on the subcellular localization of COX-1 and PGIS in the ER membrane
Trip-cat activities of Trip-cat enzyme-1 in directly converting AA to the vascular protector PGI2 The biological activities of HEK293 cells expressing the different eicosanoid-synthesizing enzymes that con-vert AA to PGI2 were assayed by the addition of [14C]AA The resultant [14C]eicosanoids, metabolized
by the enzymes in the cells, were profiled by HPLC analysis (HPLC separation linked to an automatic scintillation analyzer; Fig 3) The Trip-cat activities that occur during the conversion of [14C]AA to [14 C]6-keto-PGF1a (degraded PGI2) require two individual enzymes, COX-1 and PGIS, in HEK293 cells (Fig 3A), because neither COX-1 (Fig 3B) nor PGIS (Fig 3C) alone could produce [14C]6-keto-PGF1afrom [14C]AA in HEK293 cells However, the cells express-ing cat enzyme-1 were able to integrate the Trip-cat activities of COX-1 and PGIS by converting the added [14C]AA to the end-product, [14C]6-keto-PGF1a (Fig 3D) It should be noted that in HEK293 cells expressing Trip-cat enzyme-1, most of the added [14C]AA was converted to [14C]6-keto-PGF1a, with very low amounts of byproducts In contrast, the cells coexpressing COX-1 and PGIS synthesized less PGI2 and produced significant amounts of other unidentified
Trang 4lipid molecules These data clearly indicated that the
enzymatic conversion of AA to PGI2 is more efficient
with Trip-cat enzyme-1 than with coexpressed
individ-ual COX-1 and PGIS
Enzyme kinetics of Trip-cat enzyme-1 compared
to those of its parent enzymes
In cells coexpressing COX-1 and PGIS, the
coordina-tion of COX-1 and PGIS in the ER membrane (for
the biosynthesis of PGI2 from AA) is very fast Only
120 s were required for 50% of the maximum activity
to be reached (Fig 4A, triangles) The reaction was
almost saturated after approximately 5 min The
amount of PGI2 produced when the reaction was
extended from 5 min to 15 min increased by only 5%
On the other hand, cells expressing the engineered
Trip-cat enzyme-1 (Fig 4A, closed circles) showed the
same time-course pattern as that of the coexpressed
wild-type COX-1 and PGIS In addition, Trip-cat enzyme-1 also showed an identical dose-dependent response to that of the parent enzymes in the biosyn-thesis of PGI2 (Fig 4B) The Km and Vmax values for Trip-cat enzyme-1 were approximately 5 and 400 lm, respectively; these are almost identical to those of the coexpressed COX-1 and PGIS This study has indi-cated that the expressed Trip-cat enzyme-1 in the cells has correct protein folding, subcellular localization and native enzymatic functions in a single folded protein, similar to to its parent enzymes
Establishing stable expression of Trip-cat enzyme-1 in cells
Stable expression of the engineered Trip-cat enzyme-1
in cells is the basis for having the cells constantly pro-duce PGI2 In this study, an HEK293 cell line was used as the model for testing After G418 screening for
b
B A
Fig 2 (A) Western blot analysis for overexpressed COX-1–10aa–PGIS and COX-2–10aa–PGIS in HEK293 cells HEK293 cells transiently trans-fected with cDNA of COX-1–10aa–PGIS (lane 1) or COX-2–10aa–PGIS (lane 2), or the pcDNA3.1 vector alone (lane 3), were solubilized and separated by 7% SDS ⁄ PAGE, and then transferred to a nitrocellulose membrane The expressed Trip-cat enzymes were stained with antibody against PGIS The molecular mass (130 kDa) of the engineered enzymes is indicated by an arrow (B) Immunofluorescence micrographs of HEK293 cells In brief, the cells were grown on coverslides and transfected with the cDNA plasmid(s) of COX-1–10aa–PGIS (row 1), cotrans-fected COX-1 and PGIS (row 2), or transcotrans-fected with the pcDNA3.1 vector alone (row 3) The cells were permeabilized by SLO (columns a and b) or saponin (columns c and d), and then incubated with affinity-purified rabbit antibody against PGIS peptide (columns a and c) or mouse antibody against COX-1 (columns b and d) [13] The bound antibodies were stained with FITC-labeled goat anti-(rabbit IgG) (columns a and c)
or rhodamine-labeled goat anti-(mouse IgG) (columns b and d) The stained cells were then examined by fluorescence microscopy [13].
Trang 5the transiently transfected HEK293 cells containing the
cat enzyme-1 cDNA, cells stably expressing
Trip-cat enzyme-1 were successfully created, as indiTrip-cated by
the enzyme activity assays showing continuous
[14C]PGI2 production after the addition of [14C]AA
(Fig 5, black squares) However, the same cells
trans-fected with COX-2–10aa–PGIS cDNA could only
pro-duce PGI2 for a few days (Fig 5, open squares), due
to a failure in the stable expression of Trip-cat
enzyme-2 This study indicated that the engineered
Trip-cat enzyme-1 most likely adopted the
housekeep-ing properties of COX-1, which produced constant
expression in the cells, whereas Trip-cat enzyme-2
mainly adopted the properties of inducible COX-2,
which expressed the protein for only a short period of
time
Antiplatelet aggregation
The effects of HEK293 cells expressing COX-1–10aa–
PGIS on antiplatelet aggregation were explored It is
known that platelets contain large amounts of COX-1
and thromboxane A2synthase (TXAS) When AA was
added to the platelet-rich plasma, the platelets began
to aggregate in minutes (Fig 6A, line a) However, this
aggregation was completely blocked in the presence of
cells expressing COX-1–10aa–PGIS (Fig 6A, line b)
In contrast, the aggregation was only partially blocked
in the presence of cells coexpressing COX-1 and PGIS
(Fig 6A, line c) This indicated that AA was not only
converted into PGI2 (by COX-1 and PGIS), to act against platelet aggregation, but also converted into TXA2, promoting platelet aggregation by the abundant TXAS in the platelets In contrast, no effects were observed with the nontransfected, control HEK293 cells (Fig 6A, line d) From these observations, it is clear that the engineered Trip-cat enzyme-1 has supe-rior antiplatelet aggregation activity to coexpressed COX-1 and PGIS
To test whether Trip-cat enzyme-1 can indirectly inhibit platelet aggregation induced by other factors, such as collagen (through non-COX pathways), it is necessary to compare the effects of HEK293 cells (expressing Trip-cat enzyme-1) on human platelets induced by collagen (Fig 6B, bars 1 and 2) with those
of the AA-induced platelets (Fig 6B, bars 3 and 4) It
is clear that cells expressing Trip-cat enzyme-1 could not only directly inhibit AA-induced platelet aggre-gation (Fig 6B, bar 4), but also significantly inhibit collagen-induced platelet aggregation by up to 50% (Fig 6B, bar 2)
Competitively upregulating PGI2biosynthesis in the presence of platelets
To further demonstrate the competitive upregulation
of PGI2 biosynthesis by COX-1–10aa–PGIS in the presence of TXAS, [14C]AA was added to platelet-rich plasma containing endogenous COX-1 and TXAS, in the presence and absence of cells stably expressing
0
100
200
300
400 A
0 100 200 300
400
D
0
100
200
300
400
C
0
100
200
300
400
B
Non specific peak
Fig 3 Determination of the Trip-cat activi-ties of the fusion enzymes for directly con-verting AA to PGI2, using an isotope-HPLC method for HEK293 cells Briefly, the cells ( 0.1 · 10 6 ) transfected with the cDNA(s)
of both COX-1 and PGIS (A), COX-1 (B), PGIS (C) and COX-1–10aa–PGIS (D) were washed and then incubated with [14C]AA (10 lM) for 5 min The metabolized [ 14 C] eicosanoids produced from the [ 14 C]AA
in the supernatant were analyzed by HPLC
on a C18 column (4.5 · 250 mm) connected
to a liquid scintillation analyzer The total counts for the specific peaks in each assay are approximately: 400 counts in (A); 550 counts in (B); 600 counts in (C); and 750 counts in (D).
Trang 6COX-1–10aa–PGIS or cells coexpressing COX-1 and PGIS In the sample containing only platelet-rich plasma, the majority of the [14C]AA was converted into [14C]thromboxane B2 (Fig 7A), indicating the presence of endogenous COX-1 and TXAS in the plasma However, when cells expressing COX-1–10aa– PGIS were added to the plasma, the major product shifted to [14C]6-keto-PGF1a(degraded PGI2; Fig 7B), which demonstrated that COX-1–10aa–PGIS could effectively compete with endogenous COX-1 and TXAS for the substrate, [14C]AA On the other hand,
A
B
400
300
250
200
150
100
50
0
300
200
100
0
Time (s)
[ 14 C]-AA added (µ M )
Fig 4 Comparison of the time course (A) and dose-dependent
response (B) of HEK293 cells expressing Trip-cat enzyme-1 (closed
circles) and coexpressing its parent enzymes, COX-1 and PGIS
(tri-angles) The assay and HPLC analysis conditions used are
described in the caption for Fig 3.
0
100
200
300
400
Time (days)
Fig 5 Time course experiment for HEK293 cells expressing the
recombinant Trip-cat enzymes The cells transfected with the cDNA
of Trip-cat enzyme-1 (black squares) or Trip-cat enzyme-2 (white
squares) were selected by the G418 screening approach as
described in Experimental procedures, and then taken for assay
analysis at different days following the transfection The assay
con-ditions for the Trip-cat enzymes are described in the caption for
Fig 3 [13].
0
20
40
60
80
100
4
3
2
1
100
Time (min)
80 60 40 20
0 –20
AA added
B
C
D A
A
B
Fig 6 (A) Effects of Trip-cat enzyme-1 on antiplatelet aggregation The platelet-rich plasma was incubated with 100 lM AA at 37 C in the presence of NaCl ⁄ P i (a), HEK293 cells expressing Trip-cat enzyme-1 (b), HEK293 cells coexpressing individual COX-1 and PGIS (c), and nontransfected HEK293 cells (d) The number of HEK293 cells used for the experiments was approximately 0.2 · 10 6 per assay The addition of AA to the platelets is indicated
by an arrow (B) Comparison of the effects of HEK293 cells expressing Trip-cat enzyme-1 on platelet aggregation stimulated by collagen and AA The platelet-rich plasma, prepared from fresh human blood, was incubated with 100 lM of collagen (bars 1 and 2) or AA (bars 3 and 4) at 37 C in the presence of NaCl ⁄ P i (bars 1 and 3) or HEK293 cells (0.5 · 10 6
) expressing Trip-cat enzyme-1 (bars 2 and 4) Five minutes after the initiation of the experiment, the levels of platelet aggregation were recorded and plotted; n = 3.
Trang 7addition of cells coexpressing COX-1 and PGIS led to
only partial conversion of [14C]AA to [14C]PGI2
(Fig 7C) These results are consistent with the
obser-vations from the platelet aggregation assay described
above
Discussion
COX-1 is a housekeeping enzyme that is constantly
expressed in tissues to maintain the physiological
functions of the organs However, COX-2 is an induc-ible enzyme and is related to the pathological processes
of cancer cells and inflammation [6,7] From the point
of view of therapeutic potential, it should be safer to use the Trip-cat enzyme constructed with COX-1 rather than with COX-2 for upregulating PGI2 biosyn-thesis in vivo Thus, the successful engineering of the active COX-1–10aa–PGIS represents an advance in our COX-based enzyme engineering, and provides a basis for developing a novel therapeutic approach against thrombosis and ischemic diseases It should also be noted that the COX-2-based Trip-cat enzyme could not be simply replaced by the COX-1-based Trip-cat enzyme, because it is known that the mecha-nisms for the upregulation of COX-1 and COX-2 activities in vivo are different For instance, in a situa-tion where PGI2 is only required for a short time in the circulation, the COX-2 based Trip-cat enzyme could be preferable
It is known that the C-terminal amino acid sequence
of human COX-1 is different from that of human COX-2 [16] The crystal structures of the COX-1 C-ter-minal domain are not available yet Therefore, it remains a challenge to clearly define its orientation with respect to the ER membrane, which may affect
ER retention and anchoring, as well as enzyme cata-lytic functions Active Trip-cat enzyme-1 was prepared
by linking the human COX-1 C-terminus to the PGIS N-terminus through a 10 residue TM linker The fact that this linkage did not affect COX-1 catalytic func-tion is consistent with earlier studies, in which COX-1 activity was not affected by elongation of the C-termi-nal segment [16] The linkage also configured the COX-1 C-terminus on the membrane of the ER lumen
in Trip-cat enzyme-1 (Fig 1) Our data (Fig 2B) clearly indicate that catalytic activity and ER anchor-ing were not affected by this configuration This implies that the C-terminus is likely to be located close
to the ER membrane in native COX-1 Whether the C-terminal structure is related to COX-1 stable expres-sion in cells remains a challenging question to be explored
Recently, Smith’s group reported that the recombi-nant COX-1 (t1⁄ 2> 24 h), expressed in HEK293 cells, happens to be more stable than COX-2 (t1⁄ 2 approxi-mately 5 h), and found that a unique 19 amino acid cassette in the C-terminal region of COX-2 facilitates degradation of the expressed COX-2 in the cells [17] Without the 19 amino acid cassette in the COX-1 sequence, the expressed COX-1 maintains a higher expression level and activity level in the cells than COX-2 This finding has provided a partial explana-tion for the improved activity and stable expression of
50
A
B
C
40
30
20
10
0
50
40
30
20
10
0
500
400
300
200
100
0
30 20
10 0
30 20
10 0
30 20
10
Time (min) [ 14 C]-TXB 2
[ 14 C]-6-keto PGF 1αα
[ 14 C]-6-keto PGF 1α
[ 14 C]-TXB 2
0
Fig 7 HPLC analysis of the profiles of [ 14 C]AA metabolized by
platelets in blood in the absence and presence of HEK293 cells.
[ 14 C]AA (10 lM) was incubated with 100 lL of fresh blood in the
absence (A) and the presence (B) of HEK293 cells (0.1 · 10 6 )
expressing COX-1–10aa–PGIS, or coexpressing individual COX-1
and PGIS (C), for 5 min The metabolized [14C]eicosanoids produced
from the [ 14 C]AA in the supernatant were analyzed by the HPLC
system as described in the caption for Fig 3.
Trang 8the Trip-cat enzyme-1 derived from COX-1 In
addi-tion, 26S proteosome inhibitors retarded COX-2
degra-dation but not that of COX-1 in cells [16] This
indicated that COX-2 is easily degraded in cells, which
could be another key factor that would lead to more
difficulty in achieving stable expression of COX-2–
10aa–PGIS in cells However, the exact involvement of
gene regulation in the different expression levels of the
COX-1- and COX-2-derived Trip-cat enzymes remains
to be further characterized
One of the major difficulties in using membrane
pro-tein as a therapeutic agent is the limited number of
options currently available for solubilizing and
purify-ing the protein Nonionic detergent is commonly used
for solubilizing and purifying the membrane proteins,
but is not suitable for experiments requiring admission
of the membrane protein in vivo One way to deliver
the membrane protein in vivo is to introduce
engi-neered cells that specifically overexpress the target
pro-tein (COX-1–10aa–PGIS, Trip-cat enzyme-1) The
successful establishment of an HEK293 cell line that
can stably overexpress Trip-cat enzyme-1 and
con-stantly produce active PGI2, while demonstrating
strong antiplatelet aggregation properties, has
pro-vided a model for generating a therapeutic cell line for
potential therapeutic use of Trip-cat enzyme-1 in vivo
Antiplatelet aggregation assays provide an important
method for confirmation of the antithrombotic benefits
of the newly engineered Trip-cat enzyme-1 Human
platelet cells are rich in COX-1 and TXAS Following
the release of AA from the cell membrane (via stimuli
on the platelets), the majority of the AA is converted
to TXA2 by the coupling reaction of COX-1 and
TXAS The resultant TXA2 binds to its receptor on
the surface of the platelet and causes platelet
aggre-gation The inhibition of platelet aggregation by
HEK293 cells stably expressing Trip-cat enzyme-1
(Fig 6A) strongly indicates that: (a) expressed Trip-cat
enzyme-1 was able to compete with endogenous
COX-1 and use AA as a substrate; (v) PGH2 produced by
Trip-cat enzyme-1 was readily available to the PGIS
active site, even in the presence of TXAS, which
com-petitively binds to PGH2; and (c) the immediate
increase in PGI2 production by Trip-cat enzyme-1
reduced the amount of PGH2 available for TXAS to
produce TXA2 (Fig 7), which further prevented
plate-let aggregation These factors led Trip-cat enzyme-1 to
possess dual functions: increasing PGI2 biosynthesis
and reducing TXA2 biosynthesis, which could be a
unique and novel antithrombosis and anti-ischemic
approach that has not yet been available thus far In
addition, Trip-cat enzyme-1 also showed significant
activity in inhibiting non-AA-induced aggregation
(Fig 7B), such as that of collagen This indicates that HEK293 cells stably expressing Trip-cat enzyme-1 could use endogenous AA in the plasma, released from the platelets, to produce PGI2 which then acts against platelet aggregation Furthermore, this suggests the therapeutic potential of Trip-cat enzyme-1 in antiplat-elet aggregation through cell delivery
Experimental procedures
Materials
The HEK293 cell line was purchased from ATCC (Manas-sas, VA, USA) Medium for culturing the cell lines was purchased from Invitrogen (Carlsbad, CA, USA) [14C]AA was purchased from Amersham (Piscataway, NJ, USA) Goat anti-(rabbit IgG)–fluorescein isothiocyanate (FITC) conjugate, saponin, SLO, Triton X-100 and triethylenedi-amine were purchased from Sigma (St Louis, MO, USA) Mowiol 4-88 was purchased from Calbiochem (San Diego,
CA, USA)
Cell culture
HEK293 cells were cultured in a 100 mm cell culture dish with high-glucose DMEM (containing 10% fetal bovine serum and antibiotic and antimycotic), and were grown at
37C in a humidified 5% CO2incubator
Engineered cDNA plasmids with single genes encoding the human COX-1 and PGIS sequences
The sequence of COX-1 linked to PGIS through a 10 amino acid linker (COX-1–10aa–PGIS, Trip-cat enzyme-1) was generated by a PCR approach and subcloning proce-dures provided by the vector company (Invitrogen) The procedures have been previously described [13]
Transient and stable expression of the Trip-cat enzymes in cells
Recombinant Trip-cat enzyme-1 and Trip-cat enzyme-2 were expressed in HEK293 cells using the pcDNA3.1 vector Briefly, the cells were grown and transfected with the purified cDNA of the recombinant protein by the Lipofecta-mine 2000 method [13], following the manufacturer’s instruc-tions (Invitrogen) For transient expression, the cells were harvested approximately 48 h after transfection for further enzyme assays and western blot analysis For stable expres-sion, the transfected cells were cultured in the presence of geneticin (G418 screening) for several weeks, following the manufacturer’s instructions (Invitrogen) The cells stably expressing Trip-cat enzyme-1 and Trip-cat enzyme-2 were identified by enzyme assay and western blot analysis
Trang 9Enzyme activity determination for the Trip-cat
enzymes using the HPLC method
To determine the activities of the synthases that converted
AA to PGI2through the Trip-cat functions, different
con-centrations of [14C]AA (0.2–17.5 lm) were added to
HEK293 cells either expressing Trip-cat enzyme-1 or
coex-pressing COX-1 and PGIS, or to the nontransfected cells,
in a total reaction volume of 100 lL After 10 s to 15 min
of incubation, the reactions were terminated by adding
200 lL of the solvent containing 0.1% acetic acid and 35%
acetonitrile (solvent A) After centrifugation (8000 g for
5 min), the supernatant was injected into a C18 column
(Varian Microsorb-MV 100-5, 4.6· 250 mm), using
sol-vent A with a gradient from 35% to 100% of acetonitrile
for 45 min at a flow rate of 1.0 mLÆmin)1 The 14C-labeled
AA metabolites, including 6-keto-PGF1a (degraded PGI2),
were monitored directly with a flow scintillation analyzer
(Packard 150TR)
Immunofluorescence staining
The stable⁄ transiently transfected HEK293 cells either
expressing Trip-cat enzyme-1, coexpressing COX-1 and
PGIS, or expressing the vector (pcDNA 3.1) alone,
were cultured on a coverglass The cells were then
washed with NaCl⁄ Pi, and incubated either with 0.5 mU
of SLO for 10 min or with 1% saponin for 20 min
The cells were then incubated with the primary
anti-body (10 lgÆmL)1, affinity-purified antibody against
human PGIS or antibody against mouse COX-1) for 1 h
After being washed with NaCl⁄ Pi, the cells were
incu-bated with the FITC- or rhodamine-labeled
second-ary antibodies [13,18] and viewed under a fluorescence
microscope
Antiplatelet aggregation assays
A sample of fresh blood was collected using a collection
tube with 3.2% sodium citrate for anticoagulation, and
then centrifuged (150 g for 10 min) to separate the
plasma from the red blood cells A total of 450 lL of
this platelet-rich plasma was placed inside the 37C
incu-bator of an aggregometer (Chrono-Log) for 3 min The
nontransfected HEK293 cells, as well as those transfected
with the recombinant cDNAs of COX-1–10aa–PGIS
(Trip-cat enzyme-1) or coexpressed COX-1 and PGIS,
were added to different tubes containing platelet-rich
plasma The sample was then treated with 500 lgÆmL)1
AA, while inside the platelet aggregometer’s incubator,
to initiate the aggregation process Readings by the
anti-coagulant analyzer were obtained, indicating the amount
of platelet aggregation inhibited by each of the treated
samples
Acknowledgements
This work was supported by NIH Grants HL56712 and HL79389 (to Ke-He Ruan) In addition, we thank
R Kulmacz and Lee-Ho Wang for providing the origi-nal wild-type cDNAs of human COX-1 and PGIS, respectively We also thank Anita Mohite for her assis-tance with the anti platelet aggregation assays
References
1 Majerus PW (1983) Arachidonate metabolism in vascu-lar disorders J Clin Invest 72, 1521–1525
2 Pace-Asciak CR & Smith WL (1983) Enzymes in the biosynthesis and catabolism of the eicosanoids: prosta-glandins, thromboxanes, leukotrienes and hydroxy fatty acids Enzymes 16, 544–604
3 Samuelsson B, Goldyne M, Granstrom E, Hamberg M, Hammarstrom S & Malmsten C (1978) Prostaglandins and thromboxanes Annu Rev Biochem 47, 994–1030
4 Smith WL (1986) Prostaglandin biosynthesis and its compartmentation in vascular smooth muscle and endo-thelial cells Annu Rev Physiol 48, 251–262
5 Funk CD (2001) Prostaglandins and leukotrienes: advances in eicosanoid biology Science 294, 1871–1875
6 Cheng Y, Austin SC, Rocca B, Koller BH, Coffman
TM, Grosser T, Lawson JA & FitzGerald GA (2002) Role of prostacyclin in the cardiovascular response to thromboxane A2 Science 296, 539–541
7 Vane JR (2002) Biomedicine Back to an aspirin a day? Science 296, 474–475
8 Smith WL & Song I (2002) The enzymology of prosta-glandin endoperoxide H synthases-1 and -2 Prostaglan-dins Other Lipid Mediat 68-69: 115–128
9 Needleman P, Turk J, Jackschik BA, Morrison AR & Lefkowith JB (1986) Arachidonic acid metabolism Annu Rev Biochem 55, 69–102
10 Bunting S, Gryglewski R, Moncada S & Vane JR (1976) Arterial walls generate from prostaglandin endoperoxides a substance (prostaglandin X) which relaxes strips of mesenteric and coeliac arteries and inhibits platelet aggregation Prostaglandins 12, 897–913
11 Moncada S, Herman AG, Higgs EA & Vane JR (1977) Differential formation of prostacyclin (PGX or PGI2)
by layers of the arterial wall An explanation for the anti-thrombotic properties of vascular endothelium Thromb Res 11, 323–344
12 Weksler BB, Ley CW & Jaffe EA (1978) Stimulation of endothelial cell prostacyclin production by thrombin, trypsin, and the ionophore A 23187 J Clin Invest 62, 923–930
13 Ruan KH, Deng H & So SP (2006) Engineering of a protein with cyclooxygenase and prostacyclin synthase
Trang 10activities that converts arachidonic acid to prostacyclin.
Biochemistry 45, 14003–14011
14 Ruan KH (2007) Hybrid protein that converts
arachi-donic acid into prostacyclin WO Patent,
WO⁄ 2007 ⁄ 104000
15 Okada T & Palczewski K (2001) Crystal structure of
rhodopsin: implications for vision and beyond Curr
Opin Struct Biol 11, 420–426
16 Guo Q & Kulmacz RJ (2000) Distinct influences of
car-boxyl terminal segment structure on function in the two
isoforms of prostaglandin H synthase Arch Biochem
Biophys 384, 269–279
17 Mbonye UR, Wada M, Rieke CJ, Tang HY, Dewitt
DL & Smith WL (2006) The 19-amino acid cassette of cyclooxygenase-2 mediates entry of the protein into the endoplasmic reticulum-associated degradation system
J Biol Chem 281, 35770–35778
18 Lin YZ, Deng H & Ruan KH (2000) Topology of catalytic portion of prostaglandin I(2) synthase: identification by molecular modeling-guided site-specific antibodies Arch Biochem Biophys 379, 188– 197