Stimulation of Gs and Inhibition of Gi Protein Functions by Minimally Oxidized LDL
Abstract We have previously shown that treatment of aortic endothelial cells with minimally oxidized LDL (MM-LDL) induces their interaction with monocytes but not neutrophils and that these induced responses are associated with increased cAMP levels. Here we studied the mechanism by which MM-LDL elevates cAMP levels. Treatment of human aortic endothelial cells with MM-LDL resulted in a saturable dose-dependent increase in cAMP levels. Studies using a combination of pertussis toxin and MM-LDL suggested that part of the cAMP increase was due to the stimulation of Gs complexes. Studies with pertussis toxin–treated membranes in which Gi was completely inhibited were used to directly address the effect of MM-LDL on the Gs pathway. MM-LDL and an oxidized lipid (palmitoyl arachidonyl phosphatidylcholine), the effects of which mimic those of MM-LDL, caused a 40% to 100% increase in cAMP levels in these isolated membranes that was augmented by GTP, thus showing Gs stimulation. These results also show that MM-LDL increases cAMP levels by inhibiting Gi. MM-LDL inhibited ADP ribosylation of Gi by about 30% and completely abolished the ability of serotonin to interact with Gi complexes, whereas direct activation of Gi by mastoparan was not inhibited. This observation suggests that MM-LDL interferes with the interaction of Gi molecules with inhibitory receptors. There was no direct effect of MM-LDL on adenylate cyclase. Overall, these studies show that MM-LDL increases cAMP levels both by stimulating Gs and inhibiting Gi complexes.
- Received January 11, 1995.
- Accepted September 1, 1995.
Oxidized lipoproteins may play an important role in the pathogenesis of atherosclerosis, as shown by their presence in human lesions,1 as well as the ability of antioxidants to reduce lipoprotein oxidation and lesion formation.2 An important early event in the initiation of atherosclerosis is the increased interaction of monocytes with ECs lining the vessel wall.3 MM-LDL prepared by mild oxidation of LDL has substantial atherogenic properties, including its ability to induce endothelium-monocyte interactions.4 5 The biological activity of MM-LDL is highly associated with the elevation of cAMP levels in aortic ECs.6 In fact, cholera toxin, an activator of Gs; PT, an inhibitor of Gi; IP, a β-adrenergic receptor agonist; and dibutyryl cAMP, all of which elevate intracellular cAMP levels, elicit responses similar to those induced by MM-LDL.6 In the present study we have characterized the mechanism by which MM-LDL elevates cAMP levels.
Many review articles have described the regulatory mechanisms for the activity of AC, the family of enzymes responsible for cAMP formation in cells.7 8 Briefly, AC activity is regulated by the Gs-mediated stimulatory and Gi-mediated inhibitory arms of the AC system.7 The activation of Gs complexes occurs as the result of interaction with a ligand-bound stimulatory receptor, followed by the dissociation of the α subunit from the β and γ subunits and the activation of AC by these subunits. The dissociation of the subunits involves the exchange of a GDP for a GTP molecule, hence the GTP dependence of the activation process. Inhibition of AC occurs via a similar pathway mediated by the dissociation of the α subunit of Gi complexes from their β and γ subunits. In several systems AC activity is under tonic inhibition by Gi complexes.9 10 Inhibition of these Gi complexes by PT-catalyzed ADP ribosylation of their α subunit leads to the potentiation of the effects of the stimulatory arm of the AC system as well as the elevation of basal cAMP levels.9 10 The G protein–independent alteration of AC activity due to changes in the physical characteristics of cellular plasma membrane after lipid incorporation has also been reported.11 12 The present study addresses the following questions. Does MM-LDL cause the receptor-mediated activation of Gs complexes, thereby causing the stimulation of AC activity? Does MM-LDL impair the function of Gi complexes, thus potentiating the stimulation of AC? Does MM-LDL affect the activity of AC independent of the effects on G proteins?
Cell and Lipoprotein Preparation and Modification
HAECs at passages 5 through 8 were prepared.13 Human LDL was isolated by density-gradient centrifugation of serum and stored in phosphate-buffered 0.15 mol/L NaCl containing 0.01% EDTA.
MM-LDL was made by enzymatic modification by using soybean lipoxygenase and phospholipase A26 or by mild oxidation with iron.6 The preparations of MM-LDL used for these studies contained between 20 and 50 pg lipopolysaccharide/mL medium. Both methods of MM-LDL preparation caused minimal oxidation, giving 2 to 3 nmol of thiobarbituric acid–reactive substances per milligram of cholesterol after dialysis. The concentrations of lipoproteins used in this study are reported in micrograms of protein. The two methods of lipoprotein modification gave preparations of MM-LDL with similar activity. Each preparation of modified lipoprotein was used at a concentration determined to elicit maximal cellular response.
HAECs were cultured in six-well dishes, and cells were treated with or without agonists in medium 199 containing 5% fetal calf serum and 82 μmol/L IBMX (Calbiochem Corp), a phosphodiesterase inhibitor. After incubation, the cells were rinsed twice with phosphate-buffered saline containing 4 mmol/L EDTA (also an inhibitor of cAMP phosphodiesterase) and scraped into the same buffer. The cell pellet after centrifugation was resuspended in 200 μL boiling assay buffer (cAMP RIA kit, Amersham Corp). The suspension was sonicated, heated 5 minutes in a boiling-water bath, and microcentrifuged for 3 minutes at high speed to spin out the coagulated proteins. The supernatant was used to determine cAMP levels normalized to cell number.
HAECs were cultured in 100-mm dishes, and cells were left untreated or treated with agonist in medium 199 containing 5% fetal calf serum. Membrane was prepared by one of the following methods, both of which gave similar results when used in assays. Crude membrane preparation was made by rinsing the cells twice with cold phosphate-buffered saline containing 0.1 mmol/L PMSF (a protease inhibitor; Sigma), after which the cells were scraped into the same buffer. After centrifugation the cell pellet was resuspended in 250 μL of 10 mmol/L Tris-HCl buffer, pH 7.5, containing 1 mmol/L dithiothreitol, 0.1 mmol/L PMSF, 1 μmol/L each pepstatin and leupeptin (protease inhibitors; Sigma), 5 mmol/L MgCl2, and 1 mmol/L EDTA. The resuspended cell pellet was then subjected to three rounds of freeze-thawing in dry ice–ethanol to disrupt the cells. The suspension was microcentrifuged at 4°C for 10 minutes at 14 000g to obtain the pellet designated as the membrane fraction. The pellet was resuspended in 100 mmol/L potassium phosphate buffer, pH 7.5, containing 5 mmol/L MgCl2 and PMSF, pepstatin, and leupeptin at the same concentrations as before. The suspension was then sonicated for 10 seconds, and aliquots were used for protein determination. As an alternate method, partially purified membrane fractions were prepared by collecting the cell pellet as described above and freeze/thawing to disrupt the cells. The cells were centrifuged for 5 minutes at 500g to separate cellular organelles, and the supernatant from the first spin was further centrifuged for 10 minutes at 4°C at 14 000g to obtain the partially purified membrane pellet.
HAEC membranes were prepared as described above. Assay buffer consisting of 50 mmol/L Tris-HCl, pH 7.5, 1 mmol/L IBMX, 1.5 mmol/L EDTA, 1 mmol/L dithiothreitol, 5 mmol/L MgCl2, 0.25% bovine serum albumin, and 0.1 mmol/L PMSF was put in all assay tubes for a total reaction volume of 100 μL. ATP (Sigma) was added to all tubes (final concentration, 1 mmol/L). GTP was also added to each tube (final concentration, 0.5 mmol/L). GTP was also added to each tube (final concentration, 0.5 unless stated otherwise). If testing the ability of MM-LDL, IP, or PAPC to activate AC, these agonists were also added to the appropriate tubes. Finally, 5 μg membrane preparation was placed in each assay tube (except in the control tube that lacked membrane). Tubes were incubated for 15 minutes at 30°C, after which the reaction was stopped by adding 800 μL boiling cAMP assay buffer and heating for 5 minutes in a boiling-water bath. The tubes were then microcentrifuged for 3 minutes at high speed to spin out the coagulated proteins, and the supernatant was used to determine cAMP content as described above. This method measures the amount of cAMP formed as determined by RIA, in contrast to the formation of labeled cAMP derived from labeled ATP as used by other investigators. In addition, the present method does not require the separation of labeled cAMP from ATP due to the specificity of the RIA system.
HAEC membranes were prepared as described above. Membrane preparation (15 μg) was placed in each assay tube, except in the control tube without membrane. Assay buffer was added that consisted of 100 mmol/L potassium phosphate, 5 mmol/L MgCl2, 0.1 mmol/L PMSF, 1 μmol/L each leupeptin and pepstatin, 10 mmol/L dithiothreitol, 4 mmol/L thymidine, 1 mmol/L ATP, 1 mmol/L GTP, and 20 μg ovalbumin for a total volume of 200 μL. To each tube 8 μCi (0.2 μmol/L) nicotinamide adenine[α-32P]dinucleotide (ICN Radiochemicals) and 2.5 μg/mL activated PT (Calbiochem) were added. PT was activated by incubation at 25 μg/mL in 20 mmol/L dithiothreitol at 30°C for 15 minutes. Tubes were incubated for 30 minutes at 30°C, and the reaction was stopped by the addition of 800 μL ice-cold 13% TCA. TCA-precipitated pellets were obtained by microcentrifugation at 14 000g for 10 minutes at 4°C; the pellets were rinsed once with 13% TCA and twice with cold acetone. The acetone was discarded, and the pellets were dissolved in SDS loading buffer, heated for 10 minutes in a boiling-water bath, and analyzed in an 11% gel by using SDS-PAGE followed by autoradiography. The signals were analyzed by densitometric scanning and are reported in relative densitometric units that describe the extent of ADP ribosylation in each sample.
Membrane fractions were prepared as described above, and 15 μg was resuspended in SDS buffer containing β-mercaptoethanol. Proteins were separated by using PAGE and transferred to nitrocellulose membrane. The membrane was then incubated with antibody to Giα1,2 (Biodesign International), and antibody was detected by chemiluminescence by using the Amersham ECL reaction system. A band at 41 Kd was detected that corresponds to the position of Giα2.
Elevation of cAMP Levels by MM-LDL
Incubation of HAECs for 4 hours with MM-LDL resulted in a saturable dose-dependent increase in cAMP levels (Fig 1⇓). This saturation point was below the level produced with cholera toxin, suggesting a saturable process other than the saturation of the AC machinery.
MM-LDL Does Not Directly Activate AC
Because one mechanism by which MM-LDL may elevate cAMP levels is by a G protein–independent activation of AC, we examined the activity of the enzyme in membrane preparations from untreated cells and cells treated with MM-LDL in the absence of exogenously added GTP or stimulatory agonists. No increase in AC activity was detected in membrane preparations from MM-LDL–treated versus untreated control cells (4.1 versus 4.2 fmol cAMP·mg−1·min−1, respectively; n=6, P=.447). These data suggest that MM-LDL does not directly activate AC in these cells.
Activation of Gs by MM-LDL
To examine the role of Gs in activation of AC by MM-LDL, several approaches were taken. To detect Gs-specific increases in AC activity, the possible contribution from MM-LDL–induced Gi inhibition must be eliminated. This was achieved by using PT, an enzyme that has been shown to inhibit receptor–Gi protein interaction by ADP-ribosylating the Gi α subunits.14 Incubation of HAECs for 4 hours with 1 μg/mL PT or 125 μg/mL MM-LDL but not native LDL caused significant elevation of cAMP levels (Fig 2⇓). Preincubation with 1 μg/mL PT for 4 hours caused the complete inhibition of ADP ribosylation by PT (Fig 2⇓), an indication of maximal inhibition of all possible receptor-Gi interactions.15 When cells were treated for 4 hours with both 1 μg/mL PT and 125 μg/mL MM-LDL, an additional increase of 32% in cAMP levels was observed compared with that induced by PT alone (P<.005 for PT plus MM-LDL versus PT); native LDL did not have an effect similar to MM-LDL (Fig 2⇓). IP, a β-adrenergic receptor agonist that activates AC via activation of Gs, also caused an increase in cAMP levels when added in combination with PT, although the increase was much greater than that seen for MM-LDL (Fig 2⇓).
Activation of AC by MM-LDL Is GTP Dependent
To further determine whether MM-LDL increased cAMP by a Gs-dependent activation of AC activity, membranes were isolated from HAECs pretreated with 20 ng/mL PT for 4 hours to inhibit Gi complexes. This two-step isolation (described in “Methods”) resulted in partially purified membranes that showed no further ADP ribosylation by PT, indicating the complete inhibition of Gi. GTP (0.5 mmol/L), MM-LDL (500 μg/mL), oxidized PAPC, or IP were added separately or in combination with the membrane preparations for 15 minutes, and cAMP synthesis was measured by RIA (Fig 3⇓). Oxidized PAPC was used because in a separate study we have shown that the major activity of MM-LDL is derived from oxidized PAPC.24 GTP caused an ≈40% increase in cAMP synthesis in isolated membranes. Both MM-LDL and oxidized PAPC in the absence of GTP caused increases in cAMP levels formed. The addition of GTP to PAPC or MM-LDL caused additional 85% and 40% increases in AC activity, respectively, compared with those induced in the absence of exogenous GTP. IP increased AC activity twofold in the absence and sevenfold in the presence of GTP compared with control membranes (data not shown). Overall, these data show that MM-LDL and oxidized PAPC stimulate cAMP synthesis by a Gs-mediated pathway.
Effects of MM-LDL on Gi Molecules
The effect of MM-LDL treatment of ECs was also examined. Western blot analysis of HAEC membranes using an antibody to Giα2 detected a band at 41 Kd (Fig 4⇓). In two separate experiments, incubation of cells for 4 hours with 125 μg/mL MM-LDL had no effect on the levels of membrane-associated Giα2; however, in those same experiments ADP ribosylation of Giα2 by PT was inhibited by 54% (Fig 5⇓). MM-LDL–induced inhibition ranged between 40% and 60% in five different experiments. Similarly, incubation of cells with 1 μg/mL PT caused a 95% inhibition of ADP ribosylation by PT (Fig 5⇓). In addition, inhibition of PT-catalyzed ADP ribosylation of an 80-Kd protein was observed following MM-LDL treatment. The identity of this molecule, which was not immunoreactive with the antibody to Giα2, is presently unknown to us.
MM-LDL Pretreatment Inhibits Receptor-Mediated Activation of Gi by Serotonin
To further characterize the effect of MM-LDL on Gi activity, the ability of serotonin and mastoparan to activate Gi molecules in membrane preparations from untreated versus MM-LDL–treated cells was tested. PT only ADP ribosylates Gi molecules in their trimeric form, and receptor-mediated activation of Gi by agonists such as serotonin or activation by direct Gi activators such as mastoparan results in dissociation of the α subunit from the β and γ subunits, thereby inhibiting ADP ribosylation.14 Previous reports have examined the interaction of membrane receptors with Gi molecules by assessing the inhibition of PT-catalyzed ADP ribosylation of Gi.16 Incubation of membranes from untreated HAECs with serotonin resulted in a 60% inhibition of ADP ribosylation by PT (Fig 6⇓). However, serotonin did not inhibit ADP ribosylation in membrane preparations from MM-LDL–treated cells (Fig 6⇓). Although in several experiments the degree of MM-LDL effect varied depending on the activity of the MM-LDL preparation used, there was always an inhibitory effect on the ability of serotonin to activate Gi molecules. Mastoparan caused a 95% inhibition of PT-catalyzed ADP ribosylation in membrane preparations of untreated cells. In contrast to serotonin, the ability of mastoparan to activate Gi molecules and inhibit ADP ribosylation was not inhibited in membrane preparations from MM-LDL–treated cells (Fig 7⇓).
Exposure of aortic ECs to MM-LDL induced a saturable dose-dependent elevation of intracellular cAMP levels (Fig 1⇑). Several lines of evidence suggest that MM-LDL activates the stimulatory arm of the AC system mediated by Gs molecules. Under experimental conditions in which PT caused maximal inhibition of Gi protein activity and in the presence of the phosphodiesterase inhibitor IBMX, MM-LDL had an additive effect with PT in causing the elevation of cAMP levels, whereas native LDL was ineffective (Fig 2⇑). Further evidence for the involvement of Gs in mediating the MM-LDL–induced AC stimulation came from the observation that treatment of EC membrane preparations (in which Gi complexes were completely inhibited) caused a significant elevation of AC activity (Fig 3⇑). This increase was further augmented by the presence of GTP. There was no direct effect of MM-LDL on AC. The similarity of the action of MM-LDL to IP, in addition to the saturability of the MM-LDL effect, may indicate the presence of a Gs-linked, membrane-bound receptor for MM-LDL on aortic ECs. It must be noted that direct evidence for the existence of a receptor for MM-LDL requires confirmation. The putative MM-LDL receptor is not the same as the LDL receptor, since this receptor is not Gs coupled. The MM-LDL receptor may be one of the identified Gs-coupled lipid receptors or a related receptor, since lipids from MM-LDL can cause effects similar to the intact lipoprotein,4 and in this study we showed that an oxidized lipid that mimics the effect of MM-LDL caused a Gs-mediated increase in AC activity (Fig 3⇑).
The activation of Gs molecules appears to be a partial explanation of the mechanism by which MM-LDL elevates cAMP levels in aortic ECs. Our present study indicates that MM-LDL causes several effects on endothelial Gi molecules. Incubation of ECs with MM-LDL but not native LDL significantly inhibited PT-catalyzed ADP ribosylation of Gi (Fig 5⇑). The mechanism by which MM-LDL may cause such an effect is not known; however, posttranslational modification of the Gi molecules may be involved. One possibility may be the activation of a cellular ADP ribosyltransferase17 that would ADP ribosylate the sites normally targeted by PT. Activation of cellular ADP ribosyltransferases has been reported for other EC activators, such as l-arginine,18 the precursor of endothelium-derived nitric oxide, a potent vasodilator. Alternatively, MM-LDL may alter the PT binding sites, thereby blocking their interaction with PT. The effect of MM-LDL is not due to a decrease in the total amount of membrane-associated Giα2 (Fig 4⇑).
Our observations also indicate that MM-LDL interferes with receptor activation of Gi molecules, as suggested by the complete inhibition of serotonin-induced Gi activation and PT-catalyzed ADP ribosylation in membranes from MM-LDL–treated cells (Fig 6⇑). This effect appears to be specifically at the level of Gi-receptor interaction, since the direct activation of Gi by mastoparan was not affected (Fig 4⇑). MM-LDL may inhibit the site of interaction with inhibitory receptors, which may be close if not similar to the site of ADP ribosylation by PT. Impaired Gi-mediated function of the endothelium from atherosclerotic vessels has been reported.19 Similar results have been reported for the ability of lysophosphatidylcholine, a component of oxidized LDL, to inhibit receptor-mediated activation of Gi molecules in ECs.20 Lysophosphatidylcholine is unlikely to be responsible for inhibition of ADP ribosylation by MM-LDL, as by itself it does not increase cAMP levels in HAECs, and very low levels of lysophosphatidylcholine are present in the iron-oxidized LDL used for several of the ADP ribosylation studies (less than 1 μmol/L when MM-LDL is added to the cells at 125 μg/mL). In addition, Liao21 has shown the inhibition of bradykinin-stimulated, Gi-mediated release of nitric oxide in bovine aortic ECs treated for 72 hours with LDL. This effect may have been caused as a result of cell-mediated oxidation of LDL. Liao and Clark22 have further reported that such prolonged treatment of cells with LDL oxidized to a greater degree than the LDL used in the present studies causes a decrease in the level of expression of Giα2. Thus, the results of previous studies and the present one show that Gi inhibition in response to products of lipid oxidation may only involve alterations in the interaction of receptors with Gi complexes or may, under more severe conditions, involve actual loss of the Gi protein. An extensive review by Flavahan23 describes the potential mechanisms underlying the impaired Gi-mediated endothelial signaling associated with atherosclerosis.
We have identified an oxidized phospholipid in MM-LDL that is responsible for its ability to induce monocyte adherence to ECs and that is a target for inactivation by the platelet-activating factor acetylhydrolase.24 Our preliminary observations suggest that the phospholipids isolated from MM-LDL elevate cAMP levels in aortic ECs. We are currently further characterizing the active lipid components in MM-LDL that are responsible for stimulating AC activity. Here we have presented evidence that MM-LDL activates aortic ECs by interacting with their Gs as well as their Gi molecules. If the present in vitro study indicates the events occurring in vivo, our observations suggest that several pathways would need to be targeted to control endothelial activation in atherosclerosis.
Selected Abbreviations and Acronyms
|HAEC||=||human aortic endothelial cell|
|MM-LDL||=||minimally oxidized LDL|
|PAGE||=||polyacrylamide gel electrophoresis|
|PAPC||=||palmitoyl arachidonyl phosphatidylcholine|
|SDS||=||sodium dodecyl sulfate|
This research was supported by National Institutes of Health grants HL-30568, MO1 RR00865, and TRDRP RT372 and the Laubisch Fund. The authors thank Dr Nicholas A. Flavahan for his expert advice.
Yla-Herttuala S, Palinski W, Rosenfeld ME, Parthasarathy S, Carew TE, Butler S, Witztum JL, Steinberg D. Evidence for the presence of oxidatively modified low density lipoprotein in atherosclerotic lesions of rabbit and man. J Clin Invest. 1989;84:1086-1095.
Kita T, Nagano Y, Yokode M, Ishii K, Kume N, Ooshima A, Yoshida H, Kawai C. Probucol prevents the progression of atherosclerosis in Watanabe heritable hyperlipidemic rabbit, an animal model for familial hyperlipidemia. Proc Natl Acad Sci U S A. 1987;84:5928-5931.
Berliner JA, Territo MC, Sevanian A, Ramin S, Kim JA, Bamshad B, Esterson M, Fogelman AM. Minimally modified low density lipoprotein stimulates monocyte endothelial interactions. J Clin Invest. 1990;85:1260-1266.
Cushing SD, Berliner JA, Valentine AJ, Territo MC, Navab M, Parhami F, Gerrity R, Schwarz CJ, Fogelman AM. Minimally modified low density lipoprotein induces monocyte chemotactic protein 1 in human endothelial cells and smooth muscle cells. Proc Natl Acad Sci U S A. 1990;87:5134-5138.
Parhami F, Fang ZT, Fogelman AM, Andalibi A, Territo MC, Berliner JA. Minimally modified low density lipoprotein-induced inflammatory responses in endothelial cells are mediated by cyclic adenosine monophosphate. J Clin Invest. 1993;92:471-478.
Olansky L, Myers GA, Pohl SL, Hewlett EL. Promotion of lipolysis in rat adipocytes by pertussis toxin, reversal of endogenous inhibition. Proc Natl Acad Sci U S A. 1983;80:6547-6551.
Engelhard VH, Esko JD, Storm DR, Glaser M. Modification of adenylate cyclase activity in LM cells by manipulation of membrane phospholipid composition in vivo. Proc Natl Acad Sci U S A. 1976;73:4482-4486.
Berliner JA, Territo MC, Almada L, Carter A, Shafonsky E, Fogelman AM. Monocyte chemotactic factor produced by large-vessel endothelial cells in vitro. Arteriosclerosis. 1986;6:254-258.
Gold MR, Jakway JP, DeFranco AL. Involvement of a guanine-nucleotide-binding component in membrane IgM-stimulated phosphoinositide breakdown. J Immunol. 1987;139:3604-3613.
Shimokawa H, Flavahan NA, Vanhoutte PM. Loss of endothelial pertussis toxin–sensitive G protein function in atherosclerotic porcine coronary arteries. Circulation. 1991;83:652-660.
Flavahan NA. Lysophosphatidylcholine modifies G protein-dependent signaling in porcine endothelial cells. Am J Physiol. 1993;264:H722-H727.
Liao JK. Inhibition of Gi proteins by low density lipoprotein attenuates bradykinin-stimulated release of endothelial-derived nitric oxide. J Biol Chem. 1994;269:12987-12992.
Liao JK, Clark SL. Regulation of G-protein ai2 subunit expression by oxidized low-density lipoprotein. J Clin Invest. 1995;95:1457-1463.
Flavahan NA. Atherosclerosis or lipoprotein-induced endothelial dysfunction: potential mechanisms underlying reduction in EDRF/nitric oxide activity. Circulation. 1992;85:1927-1938.
Watson AD, Navab M, Hama SY, Sevanian A, Prescott SM, Stafforini DM, McIntyre TM, La Du BN, Fogelman AM, Berliner JA. Effect of platelet activating factor-acetylhydrolase on the formation and action of minimally oxidized low density lipoprotein. J Clin Invest. 1995;95:774-782.