(Received for publication, April 25, 1997, and in revised form, May 19, 1997)
From the Blood Research Institute, The Blood Center of Southeastern Wisconsin, Milwaukee, Wisconsin 53201-2178
The rapid movement of phospholipids (PL) between plasma membrane leaflets in response to increased intracellular Ca2+ is thought to play a key role in expression of platelet procoagulant activity and in clearance of injured or apoptotic cells. We recently reported isolation of a ~37-kDa protein in erythrocyte membrane that mediates Ca2+-dependent movement of PL between membrane leaflets, similar to that observed upon elevation of Ca2+ in the cytosol (Bassé, F., Stout, J. G., Sims, P. J., and Wiedmer, T. (1996) J. Biol. Chem. 271, 17205-17210). Based on internal peptide sequence obtained from this protein, a 1,445-base pair cDNA was cloned from a K-562 cDNA library. The deduced "PL scramblase" protein is a proline-rich, type II plasma membrane protein with a single transmembrane segment near the C terminus. Antibody against the deduced C-terminal peptide was found to precipitate the ~37-kDa red blood cell protein and absorb PL scramblase activity, confirming the identity of the cloned cDNA to erythrocyte PL scramblase. Ca2+-dependent PL scramblase activity was also demonstrated in recombinant protein expressed from plasmid containing the cDNA. Quantitative immunoblotting revealed an approximately 10-fold higher abundance of PL scramblase in platelet (~104 molecules/cell) than in erythrocyte (~103 molecules/cell), consistent with apparent increased PL scramblase activity of the platelet plasma membrane. PL scramblase mRNA was found in a variety of hematologic and nonhematologic cells and tissues, suggesting that this protein functions in all cells.
The plasma membrane phospholipids (PL)1 are normally asymmetrically distributed, with phosphatidylcholine (PC) and sphingomyelin located primarily in the outer leaflet, and the aminophospholipids, phosphatidylserine (PS) and phosphatidylethanolamine restricted to the cytoplasmic leaflet (1, 2). An increase in intracellular Ca2+ due to either cell activation, cell injury, or apoptosis causes a rapid bidirectional movement of the plasma membrane PL between leaflets, resulting in exposure of PS and phosphatidylethanolamine at the cell surface (1, 3-5). This exposure of the plasma membrane aminophospholipids has been shown to promote assembly and activation of several key enzymes of the coagulation and complement systems, as well as to accelerate the clearance of injured or apoptotic cells by the reticuloendothelial system, suggesting that Ca2+-induced remodeling of plasma membrane PL is central to both vascular hemostatic and cellular clearance mechanisms (1, 6-9).
We recently reported isolation of a ~37-kDa integral membrane protein from human erythrocytes that when reconstituted into liposomes mediated a Ca2+-dependent, bidirectional scrambling of PL between membrane leaflets mimicking the action of Ca2+ at the endofacial surface of the erythrocyte membrane (10, 11). Evidence for protein(s) in platelet that mediates a similar "PL scramblase" function when incorporated into liposomes has also been reported (12). Here we report the cDNA cloning and deduced structure of the PL scramblase from human erythrocyte and show evidence that this same protein is expressed in human platelet and various other cell lines and tissues where plasma membrane PL scramblase activity has been observed.
Egg yolk PC, brain PS, and
1-oleoyl-2-[6(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]caproyl-sn-glycero-3-phosphocholine
(NBD-PC) were obtained from Avanti Polar Lipids. Expressed sequence tag (EST) clone with GenBankTM accession number gb AA143025 was obtained through American Type Culture Collection (ATCC 962235). All restriction enzymes and amylose resin were from New England BioLabs, Inc. KlenTaq
polymerase was from CLONTECH Laboratories, wheat
germ agglutinin Sepharose was from Sigma,
isopropyl--D-thiogalactopyranoside was from Eastman
Kodak, factor Xa was from Hematologic Technologies, and Bio-Beads SM-2
were from Bio-Rad. N-Octyl-
-D-glucopyranoside (OG) and Glu-Gly-Arg chloromethyl ketone were from Calbiochem. Sodium
dithionite (Na2S2O4, Sigma) was
freshly dissolved in 1 M Tris, pH 10, at a concentration of
1 M.
PL scramblase was purified as described previously (10, 11), with the following modifications. The active fraction eluting from Mono S was concentrated and exchanged into 150 mM NaCl, 20 mM Tris, 0.1 mM EGTA, 0.1% Nonidet P-40, pH 7.4, and absorbed against 5 ml of wheat germ agglutinin-Sepharose to remove trace contaminating glycophorins. The breakthrough material was concentrated and exchanged into 20 mM Tris, 0.1 mM EGTA, 0.02% Nonidet P-40, pH 7.4, and subjected to SDS-PAGE under reducing conditions in a 10% NuPAGE gel (Novex, San Diego, CA). The band at ~37 kDa was visualized with 0.1% Brilliant Blue R-250 and excised for amino acid analysis and sequencing (University of Michigan Protein and Carbohydrate Structure Facility). 450 pmol of this protein was subjected to in situ cleavage with 10 mg/ml CNBr in 70% formic acid, the cleaved peptides were extracted into 60% acetonitrile, 10% trifluoroacetic acid, dried in a speed vacuum, and resolved by SDS-PAGE and electroblotted onto sequencing grade polyvinylidene difluoride. Peptides were observed by staining with Coomassie Blue, excised, and subjected to microsequencing using Edman chemistry on a model 494 Applied Biosystems sequencer run with standard cycles, yielding the sequence PAPQPPLNCPPGLEYLSQIDQILIHQQIELLE. This same sequence was partially confirmed in a second preparation of PL scramblase purified from erythrocyte ghosts and subjected to internal peptide sequencing as above. A BLAST homology search revealed 100% identity to the translation product of EST clone gb AA143025 with no significant sequence homology to any protein in available data bases.
Isolation of PL Scramblase cDNA by Plaque HybridizationThe 568-bp insert of EST clone gb AA143025 was
labeled with [-32P]dCTP by Random Primed DNA Labeling
Kit (Boehringer Mannheim) and used to screen a cDNA library derived
from human erythroleukemic cell line K-562 in
gt11
(CLONTECH). Escherichia coli strain
Y1090r was transformed by K-562 cDNA library (4.86 × 105 pfu) and poured on 27 agarose plates (15-cm diameter,
18,000 plaque-forming unit/plate). Plaques were lifted onto Hybond-N Nylon membranes (Amersham Corp.). After UV-cross-linking and
prehybridization in a solution composed of 5 × Denhardt, 5 × SSC, 1% SDS, and 200 µg/ml herring sperm DNA for 3 h at
68 °C, the membranes were hybridized in the same solution containing
5 ng/ml 32P-labeled probe for 16 h at 68 °C. The
membranes were washed once with 1 × SSC, 0.1% SDS, then three
times with 0.2 × SSC, 0.1% SDS for 20 min at 65 °C, and
exposed to x-ray film. Secondary plaque lifts and hybridization were
carried out on 32 strongly positive plaques at a density of 50-100
plaques/plate. Single positive and well isolated plaques were picked
and amplified. The length of cDNA insert was examined by PCR with
gt11 reverse and forward primers. Six clones with cDNA inserts
of >1.4 kilobase pairs were selected for DNA sequencing.
DNA was sequenced on an Applied Biosystems DNA Sequencer model 373 Stretch using PRISM Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit (Perkin-Elmer) and combinations of vector and insert sequence primers.
Cloning of PL Scramblase into pMAL-C2 Expression VectorTo
express PL scramblase as a fusion protein with maltose binding protein
(MBP), cDNA encoding PL scramblase was cloned into pMAL-C2 (New
England BioLabs). PCR was performed on a full-length clone using
the primers 5-TCAGAATTCGGATCCATGGACAAACAAAACTCACAGATG-3
with an
EcoRI site before the ATG start codon and
5
-GCTTGCCTGCAGGTCGACCTACCACACTCCTGATTTTTGTTCC-3
with a
SalI site after the stop codon. KlenTaq polymerase
(CLONTECH) was used to ensure high fidelity
amplification. The PCR product was digested with EcoRI and
SalI and isolated by electrophoresis on 1% low melting
agarose gel and purification with Wizard kit (Promega). The amplified
cDNA was cloned into pMAL-C2 vector digested with EcoRI
and SalI, immediately 3
of MBP. This construct was amplified in E. coli strain TB1, and the sequence of the
cDNA insert of plasmids from single colonies was confirmed.
10 ml of E. coli TB1 transformed with scramblase
cDNA-pMAL-C2 were used to inoculate 1 liter of rich LB containing 2 mg/ml glucose, 100 µg/ml ampicillin, and the bacteria were allowed to grow for about 4 h at 37 °C. When A600
reached ~0.5, isopropyl--D-thiogalactopyranoside was
added to a final concentration of 0.3 mM. After 2 h of
incubation at 37 °C, the cells were centrifuged at 4000 × g for 20 min. The cell pellet was suspended in 50 ml of 20 mM Tris, 200 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl
fluoride (column buffer), and subjected to a freeze/thaw cycle. After
sonication (3 × 30 s on ice) and centrifugation at
43,000 × g for 1 h, the supernatant was applied
to 10 ml of amylose resin. The column was washed with 20 volumes of
column buffer, and the scramblase-MBP fusion protein eluted with the
same buffer containing 10 mM maltose. Digestion of MBP-PL
scramblase protein with factor Xa was routinely performed at 1/100
(w/w) ratio of enzyme and monitored by SDS-PAGE. In addition to MBP,
the product of this digest is the PL scramblase translation product
containing the N-terminal extension Ile-Ser-Glu-Phe-Gly-Phe (codons
6
to
1).
Reconstitution into proteoliposomes was performed essentially as described previously (10, 11). Briefly, a mixture of PC and PS (9:1 molar ratio) was dried under a stream of nitrogen and resuspended in 100 mM Tris, 100 mM KCl, 0.1 mM EGTA, pH 7.4 (Tris buffer). Protein samples to be reconstituted were added to the liposomes at a final lipid concentration of 4 mg/ml in the presence of 60 mM OG and dialyzed overnight at 4 °C against 200 volumes of Tris buffer containing 1 g/liter Bio-Beads SM-2. To liberate PL scramblase from MBP, the proteoliposomes were incubated for 3 h at room temperature in the presence of 1/40 (w/w) factor Xa. The digestion was terminated by the addition of 100 µM Glu-Gly-Arg chloromethyl ketone. Completeness of the digest was monitored by SDS-PAGE. Following dialysis, the proteoliposomes were labeled in the outer leaflet by the addition of 0.25 mol % fluorescent NBD-PC (in dimethyl sulfoxide; final solvent concentration, 0.25%).
PL Scramblase ActivityPL Scramblase activity was measured as described previously (10, 11). Routinely, proteoliposomes labeled with NBD-PC were incubated for 2 h at 37 °C in Tris buffer in the presence or the absence of 2 mM CaCl2. Proteoliposomes were diluted 25-fold in Tris buffer containing 4 mM EGTA and transferred to a stirred fluorescence cuvette at 23 °C. Initial fluorescence was recorded (SLM Aminco 8000 spectrofluorimeter; excitation, 470 nm; emission, 532 nm), 20 mM dithionite was added, and the fluorescence was continuously monitored for a total of 120 s. The difference in nonquenchable fluorescence observed in the presence versus the absence of CaCl2 was attributed to Ca2+-induced change in NBD-PC located in the outer leaflet (10, 11, 13). Ionized [Ca2+] was calculated using FreeCal version 4.0 software (generously provided by Dr. Lawrence F. Brass, University of Pennsylvania, Philadelphia, PA).
Antibody against PL Scramblase C-terminal PeptideThe peptide CESTGSQEQKSGVW, corresponding to amino acids 306-318 of the predicted open reading frame of PL scramblase with an added N-terminal cysteine, was synthesized and conjugated to keyhole limpet hemocyanin (Protein Core Facility, Blood Research Institute). Antiserum to this protein was raised in rabbit (Cocalico Biologicals, Inc.), and the IgG fraction was isolated on protein A-Sepharose-CL4B (Sigma). Peptide-specific antibody was isolated by affinity chromatography on UltraLink Iodoacetyl beads (Pierce) to which peptide CESTGSQEQKSGVW was conjugated. This affinity-purified antibody (anti-306-318) was used for immunoprecipitation and Western blotting of PL scramblase (see "Results and Discussion").
Immunoprecipitation of PL ScramblasePL scramblase purified from human erythrocytes was 125I-labeled with Iodogen (Pierce), free iodide removed by gel filtration, and the protein was incubated (4 °C, overnight) with either anti-306-318,or an identical quantity of preimmune rabbit IgG (1 mg/ml in 150 mM NaCl, 10 mM MOPS, 50 mM OG, pH 7.4) or no IgG as control. The IgG was precipitated with protein A-Sepharose and washed exhaustively, and protein bands were resolved by 8-25% SDS-PAGE (Phast System, Pharmacia Biotech Inc.) under reducing conditions. Radioactive bands were visualized by autoradiography. To determine whether antibody to this peptide specifically removed the functional activity associated with the purified erythrocyte PL scramblase protein, the supernatant fractions remaining after immunoprecipitation were reconstituted in liposomes for activity measurements, performed as described above. For these experiments, unlabeled erythrocyte PL scramblase substituted for the 125I-labeled protein.
Western Blot Analysis2 × 108 washed
platelets, 2 × 108 erythrocyte ghost membranes, 0.9 pmol of purified recombinant PL scramblase (obtained by factor Xa
digest of the PL scramblase-MBP fusion protein), and 0.3 pmol of PL
scramblase purified from human erythrocyte were each denatured by
boiling in 40 µl of sample buffer containing 10% SDS,
4%-mercaptoethanol, and 1 mM EDTA, and protein bands
were resolved by SDS-PAGE. After transfer to nitrocellulose, the
blocked membrane was incubated with 1 µg/ml of anti-306-318, and the
blot was developed with horseradish preoxidase-conjugated goat
anti-rabbit IgG (Sigma) using Chemiluminescence Reagent (DuPont).
Protein concentrations were
estimated based upon optical density at 280 nm, using extinction
coefficients (M1 cm
1) of 39,000 (PL scramblase), 64,500 (MBP), and 105,000 (PL scramblase-MBP fusion).
PL scramblase contained in human platelet and erythrocyte membranes was
estimated by quantitative immunoblotting of the detergent extracts,
with reference to known quantities of purified MBP-PL scramblase fusion
protein.
Human multiple tissue Northern blot
and human cancer cell line multiple tissue Northern blot membranes were
obtained from CLONTECH. The blots were
prehybridized with ExpressHyb (CLONTECH) at
68 °C for 30 min and hybridized with ExpressHyb containing 5 ng/ml
32P-labeled PL scramblase cDNA probe at 68 °C for
1 h, then washed, and exposed to x-ray film. After development,
the blots were stripped and hybridized with 32P-labeled
-actin cDNA probe using identical conditions.
PL scramblase was purified
from human erythrocyte membranes and cleaved with cyanogen bromide, and
Edman degradation was performed on a 12-kDa peptide fragment to obtain
32 residues of peptide sequence (Fig. 1,
underlined sequence). This peptide sequence, plus the
anticipated methionine residue N-terminal to the predicted site of
cyanogen bromide cleavage, was identified in the translation product of
a 568-bp EST clone deposited in GenBank by the I.M.A.G.E. Consortium
(clone identification number 505141). No other significant matches to
this sequence were identified in any protein data base. The EST clone
was used to screen a human K-562 leukemic cell cDNA library. Of 32 positive clones identified by plaque hybridization, six clones were
sequenced yielding 1445 bp of cDNA (Fig. 1). The open reading frame
encodes a protein without a signal sequence that contains 318 residues
with a calculated mass of 35.1 kDa, a theoretical pI of 4.8, and a
single predicted transmembrane helix near the C terminus (residues
Ala291-Gly309), in good agreement with the
physical properties observed for the ~37-kDa protein band we
tentatively identified as PL scramblase in human erythrocyte membrane
(10). Whereas the deduced protein sequence is notable for its high
proline (12%) content, homology searching failed to reveal significant
concensus to identifiable protein family or domain structures, with the
exception of a single potential protein kinase C phosphorylation site
(Thr161). The possibility that PL scramblase function is
mediated by a phosphoprotein has previously been suggested based on an
observed decrease in PL scrambling activity in erythrocytes depleted of ATP (14).
Analysis of the cDNA-derived protein sequence (Tmpred program, ISREC server, University of Lausanne, Epalinges, Switzerland) revealed a strongly preferred (p < 0.01) inside-to-outside orientation of the predicted 19-residue transmembrane helix, consistent with a type II plasma membrane protein. Most of the polypeptide (residues 1-290) thereby extends from the cytoplasmic membrane leaflet, leaving a short exoplasmic tail (residues 310-318). The predicted orientation of this protein is consistent with the anticipated topology of PL scramblase in the erythrocyte membrane, where lipid-mobilizing function is responsive to [Ca2+] only at the endofacial surface of the membrane (3, 5, 10, 11, 15, 16).
To confirm that the cDNA we cloned from the K-562 cDNA library
actually encodes the same protein purified as PL scramblase from human
erythrocyte membrane, we raised a rabbit antibody against the deduced C
terminus predicted from the open reading frame of the cloned cDNA
(codons 306-318). As shown in Fig. 2, this antibody precipitated the ~37-kDa red cell protein we tentatively identified as PL scramblase and also absorbed the functional activity detected in
this isolated erythrocyte membrane protein fraction. As also evident
from Fig. 2 (inset), we often observed the partial
proteolysis of 37-kDa PL scramblase to a polypeptide of ~30 kDa. The
apparent susceptibility of this protein to proteolytic degradation may account for the reported rapid loss of activity observed in earlier attempts to purify PL scramblase from platelet (12).
Expression and Membrane Reconstitution of Recombinant PL Scramblase
Recombinant PL scramblase was expressed in E. coli as fusion protein with MBP, purified by amylose affinity
chromatography, and incorporated into PC/PS liposomes for assay of PL
scramblase activity. When incorporated into liposomes, the recombinant
protein mediated a Ca2+-dependent transbilayer
movement of NBD-PC mimicking the activity of PL scramblase isolated
from erythrocyte. PL scramblase activity was observed both for the
chimeric MBP-PL scramblase fusion protein (not shown) and for
recombinant PL scramblase liberated from MBP through proteolytic
digestion with factor Xa (Fig. 3). By contrast, no such
activity was observed for control protein consisting of the pMAL-C2
translation product MBP lacking the PL scramblase cDNA insert. The
specific PL mobilizing activity of recombinant PL scramblase expressed
and purified from E. coli was approximately 50% of that
observed for the endogenous protein purified from the erythrocyte
membrane, which is likely due to incomplete folding of the recombinant
protein. Half-maximal [Ca2+] required for activation was
approximately 100-200 µM for recombinant protein
purified from E. coli versus ~40 µM for the
erythrocyte-derived protein, raising the possibility that altered
folding or an unknown post-translational modification in mammalian
cells affects the putative Ca2+ binding site (10, 11). In
addition to activation by Ca2+, the transbilayer migration
of PL in erythrocytes is accelerated upon acidification of the inside
leaflet to pH < 6.0 (in absence of Ca2+), a response
that is also observed in proteoliposomes containing PL scramblase
purified from erythrocyte membranes (11). A similar acid-dependent activation of PL mobilizing function was
also exhibited by proteoliposomes incorporating recombinant PL
scramblase purified from E. coli (not shown).
Platelet PL Scramblase
In addition to the presumed role of PL
scramblase in PS exposure following cell injury and upon repeated
sickling of SS hemoglobin red cells, the capacity of activated
platelets to rapidly mobilize aminophospholipids across the plasma
membrane is thought to play a central role in the initiation of
thrombin generation required for plasma clotting (17). Whereas
incubation with Ca2+ ionophore causes a marked acceleration
in transbilayer movement of plasma membrane PL in both platelets and
erythrocytes, the apparent rate of transbilayer PL migration in
platelet exceeds that in erythrocyte by approximately 10-fold, implying
either a higher abundance of PL scramblase or the action of another
component in platelet with enhanced PL scrambling function (18, 19). Zwaal and associates recently reported evidence for the existence of
protein(s) in platelet with functional properties similar to that of PL
scramblase we isolated from erythrocyte (10-12). To determine whether
the protein we now identify in the erythrocyte membrane is also found
in platelets, we probed platelets with antibody against PL scramblase
residues 306-318. As shown in Fig. 4, this antibody
blotted a single protein in platelet with similar mobility to the
~37-kDa PL scramblase in erythrocyte. Based on quantitative
immunoblotting with anti-306-318, we estimate approximately 104 molecules/cell in platelet versus
103 molecules/cell in erythrocyte, consistent with the
increased PL scramblase activity and procoagulant function observed for human platelets versus erythrocytes.
Tissue Distribution
In addition to platelet and red blood
cell, PL scramblase activity has been observed in many other cells, and
this Ca2+-induced response is thought to be central to the
rapid movement of PS and phosphatidylethanolamine from inner plasma
membrane leaflet to the surface of perturbed endothelium and a variety of injured and apoptotic cells (17). The resulting exposure of PS
at the cell surface is thought to play a key role in removal of such
cells by the reticuloendothelial system, in addition to activation of
both the plasma complement and coagulation systems (8, 9, 17). Whereas
the molecular mechanism(s) in each circumstance remains
unresolved, evidence for a specific platelet membrane protein
functioning to accelerate migration of PL between membrane leaflets at
increased cytosolic [Ca2+] has been reported (12),
similar to the proposed role of PL scramblase in red blood cells (10,
11). It was thus of interest to determine whether mRNA for this
protein is expressed in nucleated cells where PL scramblase-like
activity has been observed. As shown by Fig. 5, Northern
blotting with PL scramblase cDNA revealed transcripts of ~1.6 and
~2.6 kilobases in all tissues and cell lines tested. Some
tissue-to-tissue and cell line variability in the relative abundance of
these two transcripts is apparent, the significance of which remains to
be determined. Also notable was markedly reduced expression in HL-60
and the lymphoma lines Raji and MOLT-4, whereas abundant message was
detected in spleen, thymus, and peripheral leukocytes. In addition to
the transformed cell lines shown, mRNA for PL scramblase was also
confirmed in human umbilical vein endothelial cells (not shown).
Whereas these data imply that the same protein identified as mediating
accelerated transbilayer flip-flop of the erythrocyte membrane PL also
plays a similar role in the plasma membrane of platelets, leukocytes, and other cells, actual confirmation for this role of PL scramblase awaits analysis of a cell line that is selectively deficient in this
protein. In Scott syndrome, a bleeding disorder related to an inherited
deficiency of plasma membrane PL scramblase function, erythrocytes and
other cells deficient in PL scramblase activity were found to contain
normal amounts of the PL scramblase protein (11).2 Furthermore, despite the apparent
deficiency in Scott syndrome cells of endogenous PL scramblase
function, when PL scramblase protein from these cells was purified and
reconstituted in proteoliposomes containing exogenous PL, it exhibited
normal Ca2+-dependent PL mobilizing activity
(11). This suggests that in addition to the known regulation by
intracellular [Ca2+], the activity of PL scramblase in
the plasma membrane is regulated by other as yet unidentified
membrane or cytoplasmic component(s).
The excellent technical assistance of Mary J. Blonski and Timothy T. O'Bryan is gratefully acknowledged. The assistance of Dr. Philip C. Andrews in protein sequencing and Trudy Holyst in peptide sequencing is also gratefully acknowledged.