(Received for publication, October 1, 1996, and in revised form, January 29, 1997)
From the Faculty of Pharmaceutical Sciences, Toyama Medical and Pharmaceutical University, 2630 Sugitani, Toyama 930-01, Japan
We found that isolated gastric vesicles contain a novel Mg2+-ATP-dependent phospholipid translocation (flippase) activity. Fluorescence analogue of phosphatidylcholine, 2-(12-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)dodecanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine, was ATP-dependently translocated from the outer (cytosolic) to inner (luminal) leaflet of the lipid membrane bilayer of hog gastric vesicles. The translocation was saturable and depended on time and the ATP concentration (Km = 3.1 µM). The basal Mg2+-ATPase activity of gastric vesicles in the absence of K+ showed high (Km = 1.6 µM) and low (Km = 80 µM) affinities for ATP, indicating that the present flippase activity is driven mostly by the high affinity Mg2+-ATPase activity. It required Mg2+ but not K+. Verapamil, which is an inhibitor of mouse mdr2 phosphatidylcholine flippase, did not inhibit the present flippase activity. Isolated sarcoplasmic reticulum vesicles that contain Ca2+-ATPase did not show any flippase activity. Fluorescence analogues of phosphatidylserine and phosphatidylethanolamine were similarly translocated by the gastric flippase. These phospholipid flippase activities were inhibited by 2-methyl-8-(phenylmethoxy)imidazo[1,2-a]pyridine-3-acetonitrile (SCH 28080) (IC50 = 0.14-0.25 µM), a specific K+-ATPase inhibitor of gastric H+,K+-ATPase rich in gastric vesicles. IC50 value for the SCH 28080-inhibitable Mg2+-ATPase activity was about 0.13 µM, indicating that the phospholipid translocation was driven mostly by the SCH 28080-sensitive Mg2+-ATPase activity. Possible physiological roles of flippases were discussed in relation with the gastric acid secretory and cytoprotective mechanisms.
Dramatic morphological changes occur in the parietal cell during the gastric acid secretory cycle: that is (a) the fusion of intracellular tubulovesicles that contain proton pump (gastric H+,K+-ATPase) with the apical membranes upon stimulation of acid secretion and (b) subsequent membrane retrieval and recycling (1, 2). Cytoskeletal proteins such as actin, spectrin, and ankyrin have been shown to be involved in the fusion process (3, 4). In isolated gastric vesicles that were tightly sealed and originated from the tubulovesicles (5), 82% of phosphatidylcholine, 77% of phosphatidylethanolamine, and 60% of phosphatidylserine were located in the outer (cytosolic) leaflet of the gastric vesicle membrane bilayer (6). However, the dynamic aspect of phospholipid translocation between the inner and outer leaflets has not been studied.
Phospholipid translocation (flippase) activities so far known are ATP-dependent or independent (7). For example, aminophospholipid translocases in the erythrocyte and chromaffin granule membranes ATP-dependently flip-flop phosphatidylserine and phosphatidylethanolamine (8, 9), and a mouse mdr2 gene product ATP-dependently translocates phosphatidylcholine (10) and secretes this phospholipid into bile (11). In the endoplasmic reticulum of eukaryotic cells, glycerophospholipids such as phosphatidylcholine, phosphatidylethanolamine, and phosphatidylserine are ATP-independently flip-flopped between the outer and inner leaflets of the membrane bilayer (12, 13).
We studied whether fluorescent analogues of phospholipids were translocated from the outer (cytosolic) to inner (luminal) leaflet of the lipid bilayer of isolated hog gastric vesicles in the presence and the absence of ATP, using the same experimental protocol employed for measuring the phosphatidylcholine translocation activity of mouse mdr2 gene product that was expressed in secretory vesicles from a yeast mutant (10). We found Mg2+-ATP-dependent phospholipid translocation in gastric vesicles. Interestingly, this flippase activity was inhibited by a low concentration of SCH 28080,1 which is a specific K+-competitive gastric H+,K+-ATPase inhibitor. This is the first evidence for the presence of the phospholipid flippase activity in gastric vesicles.
Fluorescent analogues of phosphatidylcholine (NBD-C12-HPC), phosphatidylethanolamine (NBD-PE), and phosphatidylserine (NBD-PS) were obtained from Molecular Probes Inc. (Eugene, OR) or Avanti Polar Lipids Inc. (Alabaster, AL). SCH 28080 was obtained from Schering-Plough Co. (Bloomfield, NJ). Pyruvate kinase (500 units/mg at 37 °C, solution in 50% glycerol) and lactate dehydrogenase (850 units/mg at 37 °C, solution in 50% glycerol) were obtained from Boehringer Mannheim-Yamanouchi Co. (Tokyo, Japan); ATP from Oriental Yeast Co. (Tokyo, Japan), DIDS, phosphoenolpyruvate, Ap5A, and NADH were from Sigma. Other chemicals used were of the highest purity available.
Preparation of Donor Liposomes Containing NBD-labeled PhospholipidDonor small unilamellar liposomes (DLs) were used to transfer NBD-labeled phospholipid into the gastric vesicle membrane. DLs containing 40 mol % of NBD-labeled phospholipid and 60 mol % of phosphatidylcholine were prepared by an ethanol injection method (10, 14). Ethanol solution of phospholipid mixture (1.3 µmol of NBD-labeled phospholipid and 2.0 µmol of phosphatidylcholine in 250 µl of ethanol) was injected at room temperature with a microsyringe into 3.1 ml of a stirred buffer solution containing 100 mM NaCl and 10 mM Tris/HCl (pH 7.40) (its final ethanol concentration was 7.5%). Then DLs were dialyzed at 4 °C for more than 16 h against the same buffer solution. The final liposome concentration was about 0.4 µmol of NBD-labeled phospholipid/ml.
Preparation of Gastric Vesicles Containing NBD-PhospholipidA gastric microsomal fraction prepared from hog
gastric mucosa was used (15). When this microsomal fraction is one-step further purified by discontinuous density gradient ultracentrifugation, so-called gastric vesicles are obtained (15). As found for the case of
DLs and secretory vesicles (10), NBD-labeled phospholipids were
transferred from DLs into the mostly outer leaflet of the bilayer of
the microsomes by simple incubation of a mixture of the microsomal
fraction and DLs for 30 min on ice, where the concentration of
NBD-phospholipid in the mixture was adjusted to about 5 nmol NBD-phospholipid/mg protein. Then this mixture was applied on the top
layer of discontinuous density gradient solutions consisting of a 250 mM sucrose solution (top layer), a mixture of 3.1% Ficoll 400 and 250 mM sucrose (middle layer), and a mixture of 7%
Ficoll 400 and 250 mM sucrose (bottom layer), and
ultracentrifuged using a vertical rotor (Hitachi RPV50T, 132,000 × g for 1 h at 4 °C). This centrifugation separated
the gastric microsomal fraction from DLs, accompanying further
purification of the microsomes and resulting in production of gastric
vesicles. Gastric vesicles (their diameter was about 0.2 µm) were
recovered from the boundary between the middle and bottom layers. Small
DLs (their diameter was about 0.03 µm) were recovered from the
boundary between the top and middle layers. The tightness of the
vesicle membrane was determined by the increase in K+
stimulation of ATPase activity in the presence of KCl due to the
ionophore, nigericin (16). The K+-ATPase activity in the
presence of 1 µg/ml nigericin, 20 mM KCl, 130 mM NaCl, 4 mM MgCl2, 0.2 mM ATP, and 40 mM Tris/HCl (pH 7.40) in the
presence of an ATP regeneration system (see the following section of
"Assay of Mg2+-ATPase Activity"), was 50.5 µmol
Pi/mg·h at 37 °C, and in the absence of nigericin, 3.7 µmol Pi/mg·h. These results indicated that about 93%
of vesicles were tightly sealed. Similar experiments at 25 °C gave
97%, indicating that almost all gastric vesicles were tightly sealed.
The tightness of the vesicle membrane was also evaluated by their
ability to accumulate proton into the vesicles; the rate of active
H+ uptake of NBD-labeled vesicles was almost the same as
that of unlabeled control vesicles. Gastric vesicles containing
NBD-labeled phospholipid were stored at 85 °C. Protein
concentration was determined by the method of Lowry et al.
(17) with bovine serum albumin as standard.
When indicated, isolated sarcoplasmic reticulum vesicles containing NBD-C12-HPC were used in place of gastric vesicles. Sarcoplasmic reticulum vesicles were prepared from rabbit leg and back muscles (18) and were incubated with DLs and separated by ultracentrifugation. The labeling efficiency of sarcoplasmic reticulum vesicles with NBD-C12-HPC was the same as that of gastric vesicles, and the integrity of sarcoplasmic reticulum membrane bilayer was not damaged by this labeling.
Assay of Phospholipid TranslocationGastric vesicles
(typically, 20 µg/ml) labeled with NBD-phospholipid were preincubated
at 37 °C for 10 min in 2 ml of solution containing 130 mM NaCl, 4 mM MgCl2, and 40 mM Tris/HCl (pH 7.40) in a fluorescence cuvette. Unless
indicated, K+ was absent in this solution. Fluorescence
emission from NBD-labeled phospholipid associated with the outer and
inner leaflets of the membranes bilayer of tight gastric vesicles was
monitored using a Shimadzu Spectrofluorophotometer (RF-5000) equipped
with a temperature-controlled water jacket and a magnetic stirrer.
Fluorescence intensity was recorded at a wavelength of 530 nm with a
10-nm slit width using an excitation wavelength of 470 nm with a 5-nm
slit width. After another incubation for the predetermined periods
(10-480 min) in the absence and the presence of ATP and chemicals
indicated, the fluorescence intensity was recorded (FT).
Then 20 µl of 1 M sodium dithionite in 1 M
Tris (pH 10) was added to reduce the fluorescent aryl-nitro group of
NBD-labeled phospholipid associated with the outer leaflet of the
vesicle membrane bilayer, and the fluorescence intensity was recorded
after 10 min (FD). Membrane-impermeable sodium dithionite
reduced only the NBD-labeled phospholipid present in the outer leaflet
(10). Then NBD-phospholipid fluorescence associated with the inner
leaflet of the vesicle membrane bilayer was quenched by sodium
dithionite after disrupting the vesicles by the addition of 80 µl of
a 25% (w/v) Triton X-100, and the fluorescence intensity was recorded
(F0). The fraction of NBD-labeled phospholipid in the outer
leaflet was calculated by the following equation (10): percentage of
outer leaflet = 100 [(FT FD)/(FT
F0)], and that of inner
leaflet was calculated by: percentage of inner leaflet = 100 [(FD
F0)/(FT
F0)].
When indicated, the translocation (flippase) activity, as evidenced by increase in "% of inner leaflet" was measured as a function of ATP concentrations (2-1,000 µM), where the ATP concentration was kept constant during the incubation period of 20 min by an ATP regeneration system (see the following section of "Assay of Mg2+-ATPase Activity").
Assay of H+ Uptake RateThe rate of H+ uptake into gastric vesicles was measured from the quench of fluorescence of acridine orange (19). The reaction mixture contained 10 µg/ml of gastric vesicles, 4 mM MgCl2, 130 mM KCl, 40 mM Tris/HCl (pH 7.40), 5 µM acridine orange, and 10 µg/ml of valinomycin. The rate of H+ uptake was determined as the initial slope of fluorescence quench after the addition of 2 mM ATP at 37 °C. The fluorescence intensity was recorded at a wavelength of 525 nm with a 3-nm slit width using an excitation wavelength of 495 nm with a 3-nm slit width.
Assay of Mg2+-ATPase ActivityWhen Mg2+-ATPase activity was measured as a function of ATP concentrations from 2 to 1,000 µM, the ATP concentrations were kept constant during the experiments by use of a pyruvate kinase-lactate dehydrogenase-linked reaction where hydrolysis of ATP is coupled with oxidation of NADH (19, 20). Although pyruvate kinase requires K+ under normal conditions, we employed the condition that did not require K+, taking the fact into consideration that high concentrations of Na+ activate pyruvate kinase without K+ (21). The reaction mixture contained 20 µg/ml of gastric vesicles, 40 mM Tris/HCl (pH 7.40), 130 mM NaCl, 4 mM MgCl2, 100 µM NADH, 0.8 mM phosphoenolpyruvate, 10 units/ml pyruvate kinase, 15.5 units/ml lactate dehydrogenase, and ATP at the indicated concentrations. The decrease in the amount of NADH was measured at 37 °C with a Beckman DU7500 photodiode array spectrophotometer in a dual wavelength mode at 340 and 500 nm. ATP hydrolyzing activity was calibrated by the addition of a known amount of ADP to a reaction mixture. Mg2+-activated ATP hydrolyzing activity was expressed as µmol of liberated inorganic phosphate/mg protein·h (µmol Pi/mg·h).
When indicated, Mg2+-ATPase activity was measured by the Yoda-Hokin method (22), where the reaction was performed in the presence of 2 mM ATP without the ATP regeneration system.
Fig.
1 shows effects of a 120-min incubation in the presence
and the absence of ATP on the fluorescence that comes from both the
inner and outer leaflets of the membrane bilayer of gastric vesicles.
The absolute fluorescence level of gastric vesicles incubated in the
presence of ATP did not differ from that incubated in the absence of
ATP. Then 10 mM sodium dithionite was added, which induced
the decrease in the fluorescence. The fluorescence level depended on
the presence and the absence of ATP (Fig. 1). This decrease was
attributed to the reduction of the aryl-nitro group of
NBD-C12-HPC to the aryl-amino group in the outer leaflet. The subsequent addition of Triton X-100 (its final concentration was
1%) destroyed the vesicle membrane, and sodium dithionite reduced
NBD-C12-HPC in the inner leaflet (maybe, also in the
intravesicular space). The final level (F0) did not depend
on the presence and the absence of ATP. From this particular experiment
shown in Fig. 1, the fraction of NBD-C12-HPC in the inner
leaflet after a 130-min incubation was calculated to be 11.5 and 20.9%
in the absence and the presence of ATP, respectively. Incubation with 2 mM ATP, therefore, increased the fraction of inner
NBD-C12-HPC by about 82%. Fig. 2 shows the
time-dependent increase in the fraction of inner
NBD-C12-HPC. This result indicates that
NBD-C12-HPC in the outer leaflet of the vesicle lipid
bilayer was significantly translocated into the inner leaflet and/or
the inner space of vesicles. The translocation was saturable.
When AMP or AMPPNP was used in place of ATP, the flippase activity was
not found (Fig. 3). In the case of ADP, an apparent flippase activity was observed. However, this activity was completely inhibited by Ap5A, a specific inhibitor of adenylate kinase (23), indicating that part of ADP was changed into ATP and AMP by adenylate kinase which was reported to be present in the gastric vesicle preparation (24). Without Mg2+, no
ATP-dependent flippase activity was observed, indicating that this NBD-C12-HPC flippase activity requests not the
binding but hydrolysis of ATP.
Fig. 4 shows that transbilayer motion of
glycerophospholipids occurs identically for phosphatidylethanolamine
and phosphatidylserine in the presence of ATP and Mg2+.
AMPPNP was not effective.
We could not find the phosphatidylcholine flippase activity in isolated sarcoplasmic reticulum vesicles that contain Ca2+-ATPase (data not shown), indicating that the flippase activity found in gastric vesicles, therefore, is not commonly associated with P-type ATPases.
Effects of K+ on the Phospholipid Flippase ActivityEffects of K+ (0.3-65 mM) on the phospholipid flippase activity were measured in the solution containing 20 µg/ml NBD-C12-HPC-labeled gastric vesicles, 4 mM MgCl2, 2 mM ATP, 40 mM Tris/HCl (pH 7.40), and various concentrations of KCl and NaCl, where the sum of the K+ and Na+ concentrations was constant (130 mM). Under the conditions, K+ gives no effect on the phospholipid flippase activity (data not shown).
Effects of Inhibitors on the Phospholipid Flippase ActivityThe concentration-dependent inhibition of
the flippase activity by orthovanadate is shown in Fig.
5. Orthovanadate at 10 and 100 µM inhibits
the flippase activity by 10 and 25%, respectively. At the same
concentrations of orthovanadate, the rate of H+ uptake into
gastric vesicles was inhibited by 20 and 88%, respectively. Table
I shows that 100 µM
N,N-dicyclohexylcarbodiimide, an inhibitor of
mitochondrial F0F1-ATPase, does not inhibit the
flippase activity. Verapamil, which inhibits both activities of the
mouse mdr2 gene product (phosphatidylcholine flippase) and
the mouse mdr3 gene product (multidrug resistance efflux
pump) (10), did not inhibit the present flippase activity (Table I).
DIDS (10 µM), which is used as an anion channel
inhibitor, significantly inhibited the flippase activity, although
diphenylamine-2-carboxylic acid (100 µM), which is also
an anion channel inhibitor, did not inhibit the flippase activity
(Table I). 1 mM N-ethylmaleimide, which at 2 mM inhibited the ATP-dependent
phosphatidylserine translocation from the luminal to cytosolic leaflet
of the bilayer of chromaffin granule membrane (9), did not inhibit the
present gastric flippase activity. Progesterone, vinblastine, and
tamoxifen, which are substrates of multidrug resistance pump and
inhibit multidrug resistance pump, did not inhibit the flippase
activity (Table I).
|
Next, we studied effects of
ATP concentrations on activities of phosphatidylcholine translocation
and Mg2+-ATPase under the same conditions, that is, both
solutions contained 20 µg/ml of gastric vesicles, 40 mM
Tris/HCl (pH 7.40), 130 mM NaCl, 4 mM
MgCl2, 100 µM NADH, 0.8 mM
phosphoenolpyruvate, 10 units/ml pyruvate kinase, 15.5 units/ml lactate
dehydrogenase, and ATP at the indicated concentrations, which were kept
constant during the incubation period. Fig.
6A demonstrates that the
Mg2+-ATPase activity shows high and low affinities for ATP
(Km = 1.6 and 80 µM determined from
the double reciprocal plot), whereas the flippase activity shows one
affinity for ATP (Km = 3.1 µM). Fig.
6B shows detailed comparison of the ATP dependences between
the flippase and Mg2+-ATPase activities in the range of low
ATP concentrations (0-50 µM ATP). In this range, both
the flippase activity (the amount of NBD-C12-HPC
translocated in 20 min) and the Mg2+-ATPase activity
followed single Michaelis-Menten enzyme reaction equations. These
results indicate that NBD-C12-HPC was translocated as a
substrate by the putative Mg2+-ATP-dependent
flippase.
Effects of SCH 28080 on the Flippase and Mg2+-ATPase Activities
SCH 28080 is a reversible K+-competitive
inhibitor of gastric H+,K+-ATPase (25). Effects
of SCH 28080 on the flippase and Mg2+-ATPase activities
were measured under the same conditions: that is, the both experimental
solutions contained 20 µg/ml of gastric vesicles, 130 mM
NaCl, 4 mM MgCl2, and 40 mM
Tris/HCl (pH 7.40). After a 10-min incubation with various
concentration of SCH 28080, 2 mM ATP was added to start the
flippase reaction. Here, the amount of NBD-C12-HPC
translocated from the outer to inner leaflet of the bilayer for 1 h in the presence of 2 mM ATP was measured. The amount of
liberated inorganic phosphate for 1 h was measured by the
Yoda-Hokin method (22). Fig. 7 shows that SCH 28080 concentration-dependently inhibits the flippase activity, its
IC50 value for NBD-C12-HPC being about 0.14 µM. Furthermore, we measured the inhibition of NBD-PS and
NBD-PE translocations by SCH 28080 and found the IC50 values of 0.25 and 0.20 µM, respectively. These
IC50 values were not significantly different.
The K+-ATPase activity of gastric vesicles measured in the presence of 15 mM KCl and 10 µg/ml of valinomycin was inhibited by SCH 28080 (IC50 = 0.2 µM) (26). Furthermore, it has been reported that SCH 28080 partially inhibited the Mg2+-ATPase activity of gastric vesicles (27). Fig. 7 shows that about 56% of the Mg2+-ATPase activity is inhibited by 10 µM SCH 28080, whereas 96% of the flippase activity is inhibited by 10 µM SCH 28080, suggesting that only part of the Mg2+-ATPase activity contributed to the flippase activity. From Fig. 7, IC50 value for the SCH 28080-inhibitable fraction of the Mg2+-ATPase activity was obtained to be 0.13 µM.
We have found that the Mg2+-ATP-dependent phospholipid flippase is present in hog gastric vesicles. Fluorescent analogues of phosphatidylcholine, phosphatidylserine, and phosphatidylethanolamine, which are distributed more densely in the outer (cytosolic) leaflet of the vesicle membrane bilayer than in the inner (luminal) leaflet (6), were translocated from the outer to inner leaflet of the bilayer in a same manner. It does not require the presence of K+ and is relatively insensitive to orthovanadate, discriminating this flippase from H+,K+-ATPase. The translocation (flippase) activity depends on the ATP concentration. The Km value of 3.1 µM was obtained from the experiments of a 20-min incubation (Fig. 6). When the incubation period was increased to 1 h, the value of Km increased to 9 µM (data not shown), indicating that the Km value would be smaller than 3.1 µM if the initial rate of the flippase activity were plotted against the ATP concentration. Therefore, the Km value of 3.1 µM may be comparable with that of the high affinity Mg2+-ATPase activity of gastric vesicles (1.6 µM). The low affinity Mg2+-ATPase activity (Km = 80 µM) seems to have no contribution to the flippase activity (Fig. 6A). From these results, we suggest that the phospholipid flippase activity may be attributed mostly to the high affinity Mg2+-ATPase activity of gastric vesicles.
Furthermore, the IC50 value obtained from the inhibition of the flippase activity by SCH 28080 (IC50 = 0.14-0.25 µM; Fig. 7, 1-h incubation) was comparable with the value obtained from the SCH 28080-sensitive Mg2+-ATPase activity (0.13 µM). This fact may indicate that the flippase activity mostly owes to the SCH 28080-sensitive Mg2+-ATPase activity.
Verapamil is known to inhibit the flippase activity of mouse mdr2 gene product (10). As shown in Table I, verapamil did not inhibit the present flippase activity in gastric vesicles. Furthermore, substrates of multidrug resistance pump (P-glycoprotein) such as progesterone, tamoxifen, and vinblastine did not inhibit this flippase activity (Table I).
Tamoxifen is also known to inhibit the volume-activated
Cl channel (28, 29). Similarly, a Cl
channel inhibitor diphenylamine-2-carboxylic acid did not inhibit this
phospholipid flippase, although DIDS, which binds to amino residue and
inhibits anion channels, inhibits the flippase.
The direction of the ATP-dependent phospholipid transporting activity of the present gastric flippase and mdr2 gene product is from the cytosolic to luminal leaflet of the lipid bilayer, whereas that of ATP-dependent aminophospholipid translocase in erythrocyte and chromaffin granules is opposite. Mouse mdr2 gene product plays an essential role in the secretion of phosphatidylcholine across the canalicular membrane into bile (10). Aminophospholipids such as phosphatidylserine and phosphatidylethanolamine are confined to the inner leaflet of the plasma membrane by aminophospholipid translocase (8), and the appearance of phosphatidylserine on the surface of animal cells triggers phagocytosis (30) and blood coagulation (31). ATPase II of chromaffin granule membrane was assigned to be this aminophospholipid translocase (32), and the gene that codes for this ATPase was cloned (33). The ATPase II is a member of an ancient subfamily of P-type ATPases.
The role of the present flippase is unknown. Several types of flippase functions are considered to be involved in the acid secretory and cytoprotective mechanisms. First, stimulation of the parietal cell by histamine and gastrin causes great morphological changes. One such event is the fusion of tubulovesicles (gastric vesicles) with the apical membrane accompanying increase in the area of the external plasma membrane as evidenced by increase in the external membrane capacitance (34). The fusion process will request the presence of flippase to readjust the balance of phospholipids between the outer and inner leaflets of the lipid bilayer, for example, enabling a curvature change in the membrane.
Second, there are several lines of evidence that gastric mucus is protected against the back diffusion of acid by a hydrophobic barrier comprising of surface-active phospholipids (35). Electron microscopic observations indicated that deeper mucus-free epithelial surfaces of oxyntic ducts and parietal cells were also covered by oligolamellar lining of phospholipids mainly phosphatidylcholine (36), which may be secreted by flippase. Other roles of flippases are also considered. The role of the present flippase in gastric parietal cells is necessary to be clarified in future studies.
The present new gastric phospholipid flippase is an interesting object because only several ATP-dependent flippases are known, and the properties of the present flippase are different from previously known phospholipid flippases.