- Email: [email protected]
The central type Y amphipathic α-helices of apolipoprotein AI are involved in the mobilization of intracellular cholesterol depots
Available online at www.sciencedirect.com
ABB Archives of Biochemistry and Biophysics 473 (2008) 34–41 www.elsevier.com/locate/yabbi
The central type Y amphipathic a-helices of apolipoprotein AI are involved in the mobilization of intracellular cholesterol depots Marina C. Gonzalez, Juan D. Toledo, M. Alejandra Tricerri, Horacio A. Garda * Instituto de Investigaciones Bioquı´micas de La Plata (INIBIOLP), CONICET/UNLP, Facultad de Ciencias Me´dicas, Calles 60 y 120, La Plata 1900, Buenos Aires, Argentina Received 16 January 2008, and in revised form 14 February 2008 Available online 23 February 2008
Abstract We studied the role of a central domain of human apolipoprotein AI (apoAI) in cholesterol mobilization and removal from cells. In order to check different protein conformations, we tested different sized and cholesterol-content reconstituted apoAI particles (rHDL). Meanwhile cholesterol-free discs were active to induce mobilization, only small cholesterol-containing rHDL were active. To test the influence of a central domain in such events, we used two apoAI variants: one, with its central Y helix pair replaced by the C-terminal domain, and the other having a lysine deleted in central region. The helix-swapping variant decrease the cholesterol pool available to acyl-CoA cholesterol acyl transferase and increase mobilization of newly synthesized cholesterol. Instead, the deletion mutant had no effect on both events. We conclude that the central domain of apoAI is involved in cholesterol cell traffic and solubilization, and that a Y-type charge distribution in polar face may be required, as well as a correct helices-polar face orientation. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Apolipoprotein AI; Cholesterol efflux; Chinese hamster ovary cells; Intracellular cholesterol mobilization; Mutants; Reverse cholesterol transport; Reconstituted high density lipoprotein
Apolipoprotein A-I (apoAI)1 is the major protein component of high density lipoproteins (HDL) which play a key role in reverse cholesterol transport (RCT), a process of great antiatherogenic importance that results in the removal of cholesterol (Chol) excess from peripheral cells, and its transport toward Chol metabolizing organs as liver or steroidogenic tissues [1,2]. Chol efflux mediated by lipidfree apoAI or other apolipoproteins can be considered as
*
Corresponding author. Fax: +54 221 425 8988. E-mail address: [email protected] (H.A. Garda). 1 Abbreviations used: ApoAI, apolipoprotein AI; rHDL, reconstituted high density lipoprotein; RCT, reverse cholesterol transport; CHO, chinese hamster ovary; MEM, minimal essential medium; FBS, fetal bovine serum; Chol, cholesterol; cholne, 4-cholesten-3-one; BSA, bovine serum albumin; POPC, 1-palmitoyl-2-oleyl sn-glycero-3-phosphocholine; PBS, phosphate-buffered saline; ACAT, acyl-CoA cholesterol acyl transferase; IPTG, isopropyl-beta-D-thiogalactopyranoside; ABCAI/GI, ATP binding cassette transporter A-I/G-I; FPLC, fast performance liquid chromatography. 0003-9861/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2008.02.021
the consequence of at least three partial events: (1) Initial apolipoprotein lipidation or loading with phospholipids at the cell membrane, resulting in the formation of discoidal HDL particles, (2) Stimulation of mobilization of intracellular Chol depots toward the cell membrane, and (3) Chol uptake from the cell membrane by discoidal HDL or other lipoprotein particles. There is a large body of evidence that the first process is catalyzed by ATP Binding Cassette A1, (ABCA1) [3]. It was proposed that ABCA1 catalyzes the concurrent loading of apolipoproteins with phospholipids and Chol [4], but there are also different pieces of evidence indicating that these events occur separately in time and in different specialized domains of the cell membrane [5]. Further Chol efflux toward the apoAI acceptors is proposed to occur mediated by the ABCG1 transporter, which should act sequentially to the ABCA1 [6]. Chol transfer from the cell membrane toward HDL can be produced by simple aqueous diffusion, in a process facilitated by Scavenger Receptor Type B1 (SR-B1) [7] and,
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
as we proposed for cholesterol-poor discoidal HDL [8–11], by the insertion in Chol-rich membrane domains of the central Y-helix of apoAI which facilitates its desorption. The molecular mechanisms of intracellular Chol pools mobilization toward the cell membrane induced by apoAI are poorly understood. It was proposed that apoAI induces the transport of intracellular Chol to cell-surface caveolae, possibly in part through the stimulation of caveolin expression [12]. Mobilization of intracellular pools also seems to be dependent on the expression of ABCA1, but can be separated from ABCA1/apoAI interaction or ABCA1-mediated phospholipid efflux under certain circumstances [13]. Activation of protein kinase C (PKC) by apoAI was reported to be involved in the mobilization of intracellular cholesterol pools, as the one available to be esterified by ACAT [13–15]. ApoAI contains several repeats which are predicted to form amphipathic a -helices [16,17] (see Fig. 3). Most of them (helices 1, 2, and 5–8) are 22mer repeats forming type A helices, a very common motif in all exchangeable apolipoproteins, characterized by two clusters of positively charged residues at the hydrophobic/hydrophilic interface and a cluster of negative charges at the center of the hydrophilic face. Besides, apoAI contains a G* helix at its N-terminus which is similar but not identical to the helices present in globular proteins, and two Y type helix pairs: one in the central region (residues 88–120, helices 3 and 4) and the other located at the C-terminus (residues 209– 241, helices 9 and 10). Type Y helices are similarly organized to type A helices, but they have positive residues disrupting the cluster of negative charges at the center of the hydrophilic face. The relative exposition of the different regions of apoAI to the aqueous/and or lipid environments is probable to play key roles in mediating cellular signals prone to redistribute the Chol among domains with different accessibility to be exported. Previously we have shown, using monoclonal antibodies, that the conformation of the central region, is strongly dependent on lipid environment in discoidal, reconstituted particles. This domain is highly exposed in small-cholesterol-free particles, and becomes masked as particles get larger and-or enriched in Chol [11]. Later, our group showed that the central region spanned residues 87–112 deeply inserted into the lipid bilayer of phospholipid membranes [10]; moreover, a synthetic peptide having the sequence of apoAI comprised from residues 77 to 120 (AI 77–120) promoted Chol efflux from Chinese hamster ovary (CHO) cells with high efficiency [8]. Thus, it is conceivable that a sequence spanning this peptide, could actively participate also in cell events destined to promote cholesterol efflux. Based on these previous results, we tested here the ability of different reconstituted particles to induce mobilization of the internal Chol pool from cells. We hypothesize that the exposition and/ or conformation of the central domain could play a role for inducing cellular events involved in cholesterol exportation. We also explored the
35
structural requirements of this central region by using two mutants. One of them (DK107) presents a deletion in helix 4, which changes by about 100° the helix registry and the orientation of the hydrophilic and hydrophobic faces of this amphipathic helix at both sides of the deletion point. In the second mutant (H9–10@H3–4) the central Yhelix pair 3–4 was replaced by the C end Y-helix pair. Altogether, our results suggest that the central type Y helix pair, which may be highly exposed in lipid-poor particles, contributes to the mobilization of intracellular Chol pools toward the cell membrane. Its functionality would not require a specific sequence, rather a type Y charge distribution in the polar helix face and a correct polar/hydrophobic face orientation throughout helix 4. Materials and methods Preparation of apoAI ApoAI was obtained from human serum (donated by Banco de Sangre, Instituto de Hemoterapia de la Provincia de Buenos Aires, La Plata, Argentina) as previously described [11]. It shows more than 95% purity as estimated by SDS polycrylamide gel electrophoresis (SDS–PAGE) using Coomasie blue staining.
Reconstitution of high density lipoprotein particles (rHDL) The cholate dialysis method was used to obtain rHDL containing ˚ Stokes diameter, with or without apoAI. Discoidal rHDL of 96 and 78 A cholesterol, were obtained starting from 95/1/150 and 40/1/65 POPC/ apoAI/sodium cholate, or 95/24/1/150 and 40/10/1/65 POPC/chol/apo˚ ) were purified AI/sodium cholate molar ratios, respectively. rHDL (120 A ˚ reconstitution mixture. Particles were purified by molecular from the 96 A size exclusion chromatography by FPLC as described before [10] and the molecular weight determined by polyacrylamide gel gradient electrophoresis in non-denaturing conditions (PAGGE). The composition was determined by the usual analytical methods.
Construction and expression of proapoAI variants Wild type proapoAI and DK107 DNA sequences were a generous gift from Dr. A. Jonas, University of Illinois at Urbana-Champaign, IL. The respective inserts, ligated to pET30a vectors were transformed into BL21 (DE) Escherichia coli cells (Novagen, Madison, WI), then expressed by induction with IPTG and purified as described elsewhere before [18]. Proteins were reconstituted into lipid particles and dialyzed against 10 mM Tris buffer, pH 8.0 (TNB). The fusion tag was removed by digestion with enterokinase followed by further elution through Ni-chelate columns (Novagen). Finally, proteins were delipidated as described previously [18]. The helix-swapping H9–10@H3–4 mutant inserted on a pET30 vector was a gift from Dr. W.S. Davidson (University of Cincinnati, OH). As explained above, in this construction amino acids spanning residues 88– 120 were replaced by residues 209–241, in such a way that the original central region of the protein was absent, but the variant kept the molecular weight preserved. A region with the sequence for an Igase protease cleavage site was designed directly upstream of the pro segment of the apoAI construction in order to specifically eliminate the histidine tag. BL21 (DE) E. coli cells were transformed with the vector containing the insert and plated in the presence of Kanamicyn. A colony was grown in Luria Bertani (LB) media at 37° C to an absorbance of 0.2 at 600 nm. Temperature was lowered to 25° C, and induced for 3 h with 0.4 mM IPTG. Proteins were purified through Ni-chelate columns in the presence
36
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
of 0.5 M Guanidine HCl, and the His tag eliminated with the Igase protease. Purity in the final preparations was confirmed by SDS–PAGE.
Cell culture About 400,000 Chinese hamster ovary (CHO-K1) cells were seeded in 6-well plates and incubated until confluence (3–4 days) with 2 ml of minimal essential medium (MEM) supplemented with penicillin/streptomycin solution (100 units/ml) and 10% FBS at 37 °C in a 5% CO2 atmosphere.
Cholesterol efflux and evaluation of cellular free and esterified cholesterol CHO-K1 cells were grown until confluence. Then, cell monolayers were washed with PBS and incubated for 41 h with serum-free medium containing 50 lg/ml cholesterol, 0.5 lCi/ml [3H] cholesterol, and 2 mg/ml BSA. After 8 h equilibration with MEM containing 1 mg/ml of BSA, monolayers were washed twice with PBS and treated for 12 h with serumfree medium containing 0.5 lM of apoAI, either as the lipid-free protein variant or as rHDL. Background efflux was determined from control cells incubated with serum-free medium. Then, media were collected and cells scrapped and suspended in methanol for lipid analysis. Radioactivity was measured by scintillation counting in aliquots of centrifuged culture medium and scrapped cells to calculate the percentage of cholesterol efflux. Cellular lipids were extracted by the method of Bligh and Dyer [19]. Sterol species were separated by thin layer chromatography on silica gel G plates developed in hexane:ethyl ether:acetic acid (80:20:1, v:v:v). Lipid spots corresponding to cholesteryl esters and unesterified cholesterol were identified by staining with I2 vapor and co-migration with authentic standards. Appropriate spots were scrapped for scintillation counting.
[14C] Mevalonate labeling of newly synthesized cellular cholesterol CHO-K1 cell monolayers, grown until confluence, were washed twice with PBS-BSA (2 mg/ml) and incubated for 48 h at 37° C with MEM, BSA (2 mg/ml) and cholesterol (50 lg/ml). Cells were washed three times with PBS-BSA and incubated at 37° C for 3 h in bicarbonate-free MEM containing 25 mM HEPES, 1 mg/ml BSA, 0.5 lCi/ml [14C] mevalonic acid lactone, 0.5 mM mevalonolactone, and 2 lg/ml of Sandoz ACAT inhibitor [compound 58-035] to prevent esterification of newly synthesized sterols [20]. After incubation, cell layers were rinsed twice with PBS-BSA and incubated in the same media without label for 1 h at 37° C to allow the conversion of labeled biosynthetic intermediates into [14C] sterols. Then, cells were washed and incubated with serum-free medium containing 0.2 lM apoAI, the protein variant or the rHDL to be tested for 90 min. After incubation at 37° C during the indicated time, cell layers were rinsed twice with 1 ml of PBS-BSA, fixed in 1% glutaraldehyde in PBS for 15 min on ice, and then treated with Chol oxidase to measure intracellular Chol translocation. Alternatively, fixed cells were stored frozen at 20° C before Chol oxidase treatment.
Treatment of cells with cholesterol oxidase After fixation, cells were washed two or three times with PBS, incubated at 37° C for 30 min with Chol oxidase (4 units/ml) in 5 mM NaPO4, pH 7.2, washed twice with PBS and stored frozen until lipids were extracted. The fraction of [14C] sterols oxidized to [14C] cholestenone was assumed to represent the plasma membrane associated sterol pool, whereas [14C] sterols resistant to enzymatic oxidation were assumed to be the intracellular pool.
Cholesterol pool available to ACAT The cellular cholesterol pool available to be esterified by ACAT was measured according to Francis et al. [21] after 12 h incubation with serum-
free medium containing 0.5 lM apoAI or the corresponding variant. Cells were then washed once with PBS before adding 0.25 lCi [14C] oleic acid and 10 lM of unlabeled oleate. After 1 h at 37 °C, cells were collected and lipid extracted as mentioned above. ACAT activity was estimated by measuring the incorporation of [14C] oleic acid into the cholesteryl ester fraction [15,21].
Other analytical methods Quantitative measurement of protein was performed by the method of Lowry el al. [22]. Student´s t-test was used to compare pooled data.
Other materials Minimal essential medium (MEM) was obtained from Invitrogen Corporation. Penicillin–streptomycin was from Parafarm (Bs. As., Argentina) and fetal bovine serum (FBS) was purchased from Bioser (Bs. As., Argentina). [3H] cholesterol, [14C] oleic acid and [14C] mevalonic acid lactone were obtained from Amersham Biosciences (Pittsburgh, PA). ACAT inhibitor [compound 58-035] was a generous gift from Sandoz Pharmaceuticals. Igase (endoproteinase), bovine serum albumin (BSA), cholesterol, mevalonolactone, glutaraldehyde, sodium cholate, Cholne and Chol oxidase were obtained from Sigma Co (St. Louis, MI) and enterokinase was from Boehringer-Mannheim. IPTG was from Promega (Buenos Aires, Argentina). POPC was purchased from Avanti Polar Lipids (Alabaster, AL).
Results and discussion Effect of apoAI lipidation state on cellular cholesterol mobilization and efflux from cells ˚ diamHomogeneous, POPC-apoAI rHDL (78 and 96 A ˚ eter), and POPC-Chol-apoAI rHDL (78, 96 and 120 A diameter) were incubated with cells for 90 min and their ability to induce de novo synthesized Chol mobilization to the plasma membrane tested in comparison with lipid-free 0 ˚ rHDL containing apoAI. Fig. 1A shows that 78 and 96 A POPC resulted as active as apoAI for increasing the cholestenone/cholesterol ratio which means, to mobilize internal Chol pools. Instead, when rHDL contain Chol, only the 0 ˚ particles were as effective as the lipid-free prosmall 78 A tein, while larger cholesterol-rich rHDL have completely lost that activity (Fig. 1B). These results indicate that Chol mobilization is dependent on the conformation of apoAI, which in fact is determined by the interaction with different lipid environments in the rHDL particles. As it was previously determined [11], both the increase in size and Chol content of rHDL decrease the exposure degree of a central epitope and the ability of these lipoprotein particles to interact with artificial phospholipid membranes. Thus, it is possible that in large and Chol-rich rHDL, the central region is not available to interact with the cell membrane sites or proteins responsible for triggering the signaling pathways conducing to the mobilization of the internal Chol pools. Following, we analyzed the ability of the same particles to remove Chol. Interestingly, the effect of size and sterol content does not seem to influence the final effectiveness of the tested particles when incubated with cells for 12 h. All of them removed about 60% of the amount
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
Fig. 1. Mobilization of recently synthesized cholesterol mediated by POPC-apoAI (A), or POPC-Chol apoAI rHDL (B). After 3 h incubation with [14C] mevalonic acid lactone, cells were treated with lipid-free apoAI or rHDL for 90 min and incubated with cholesterol oxidase. Results are expressed as the ratio of radioactivity associated to plasma membrane (cholestenone) and intracellular cholesterol pools (Chol oxidase resistant). * P 6 0.05, **P 6 0.01 (levels of significance with respect to control); ° P 6 0.05, °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
removed by lipid-free apoAI under the same conditions (Fig. 2A and B). Mobilization and efflux of intracellular lipid depots is not a simple consequence of plasma membrane Chol depletion since it is not induced by unspecific acceptors like albumin or cyclodextrin [23]. It is well-known, and it is confirmed here, that apoAI participates in triggering these processes, though the steps involved are not completely clear. Chol loading of different cells leads to the formation of oxysterols and the subsequent activation of the liver X receptor (LXR) [24]. A number of ATP binding cassette transporters (ABCs) involved in lipid transport are regulated through activation by LXR, including ABCA1 [24] and ABCG1. By using original experimental designs, Gelissen et al. [25] and Vaughan and Oram [6] showed that particles released from ABCA1-expressing cells incubated with apoAI, act as cholesterol acceptors from ABCG1-expressing cells, and thus they suggested that both ABC transporters work sequentially; the incubation of cells with lipid-free apoAI would lead to the efflux of Chol and phospholipid and the subsequent formation of nascent HDL particles,
37
Fig. 2. Cholesterol removal from CHO-K1 cells mediated by rHDL. ApoAI (either in the lipid-free state or as rHDL) was incubated with cells for 12 h as described in Methods. Results are shown as the % of radioactivity removed by cholesterol-free (A) or cholesterol containing (B) rHDL from the total radioactivity incorporated. **P 6 0.01 (levels of significance with respect to control); °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
which could be further enriched, via ABCG1, in Chol content [6]. Even though we use here a different cell system, it is plausible that, when loading cells with Chol, both transporters are being activated. Our experiments show that lipid-free apoAI, and small particles are effective to mobilize intracellular Chol, (Fig. 1) and that this ability is lost when particles are larger and Chol-enriched. It is probable then, that as the protein becomes lipidated, the domains required for triggering Chol mobilization become involved in contacts among the lipids of the particle, less exposed and unable to start cell signaling. Another observation is that Chol solubilization is more efficient when apoAI is in the lipid-free form than in rHDL, and independent, under our experimental conditions, on the size and Chol content of the rHDL particles (Fig. 2). We believe that, in our system, lipid-free apoAI is allowed to go through both steps: inducing ABCA1-dependent lipid solubilization, and giving rise to small particles, that are sequentially substrate for the ABCG1, which can suffer further lipidation. Instead, when incubating the same amount of protein
38
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
as discoidal particles, they can go through the second step only, thus is lipidation via ABCG1, and then lipid solubilization is limited. In Oram’s experiments, linear lipid solubilization is observed at shorter incubation times (less than 7 h), and thus, the fact that all particles showed a similar efficiency for Chol removal, could be representing a final condition in which all particles acquire phospholipids and cholesterol and reach a final, completely saturated and more stable apoAI conformation that could be further substrate of lecithin cholesterol-acyl transferase. This fact is in agreement with the observation that the product of apoAI lipidation through ABCA1 and ABCG1 overexpression results in a final, large and cholesterol rich lipoprotein [6]. This hypothesis can be also supported by the recent report from Sanchez et al [26]. By using Fluorescence Correlation Spectroscopy, they observed a decrease in the diffusion coefficient of rHDL in a single-molecule scale. They explain this observation as an increase in the size of every single particle as long as they capture lipids, but rather not due to a fusion of particles into a large aggregate. Structural requirements of the central apoAI Y helical region to promote cellular cholesterol efflux The second aim of this work, was to define the main role of the central helix of apoAI in the mentioned processes. The recently reported crystalline structure of full-length lipid-free apoAI [27] shows two well-separated domains: a 4-helix bundle formed by helices A to D, which includes predicted G*-helix and repeats 1 to 7; and a 2-helix bundle (helices E and F) which includes repeats 8 to 10 (Fig. 3). The N-terminal portion of the protein plays a key role in stabilizing the lipid-free state as it is indicated by this structure and also by the fact that its deletion results in totally different structure, which is extended, highly helicoidal, and thought to be similar to the lipid-bound state conformation [28]. While the long Y repeats (4 and 10) are helical in both reported crystalline structures, the short Y repeats (3 and 9) correspond to loops exposed to the surface in the lipid-free full length apoAI structure [27]. Next, we have studied the influence of the structural arrangement of the central region to evoke the previous events. Previous results from our group showed that the peptide spanning the apoAI central Y helix pair AI 77120 is active in promoting cell Chol efflux [8]. The physiological role of this region was investigated by using apoAI mutants with a disruption (DK107), or replacement of the domain spanning the central Y helix pair (H9–10@H3–4). Following exposition of [3H] -cholesterol loaded CHO-K1 cells to MEM alone (control) or to MEM containing the different apoAI variants, radioactivity was counted in the medium. In agreement with reported results in adipocytes and macrophages [29], deletion of the K107 residue did not significantly affect the capacity of the lipid-free protein to promote Chol removal from CHO cells (Fig. 4), or as shown previously by others as reconstituted HDL [29,30]. On the other hand, AI H9–10@H3–4 mutant showed lower
Fig. 3. Amphipathic a-helices in mature human apolipoprotein AI. Top: Amino acid sequence showing the correspondence between predicted amphipathic a-helices [16] and those found in the crystalline structure [41]. Predicted G*-helix is shown in black while class-A and class-Y repeats are shown in dark and light gray, respectively. The six a-helices found in the crystal structure are indicated with capital letters from A to F. The sequence of the peptide AI 77–120 is underlined. Lysine residue lacking in DK107 mutant is shown (as a K) in white. In H9–10@H3–4 mutant, the sequence spanning repeats 3–4 (residues 88–120) was replaced with that spanning repeats 9–10 (residues 209–241). Bottom: Wheel diagrams showing the typical charged residue distribution for class-A and class-Y helical repeats. Non-polar residues are shown in black, uncharged polar residues in white, and positive and negative residues in dark and light gray, respectively.
Fig. 4. Influence of the central helix on the activity of apoAI to promote cell cholesterol efflux. Confluent CHO-K1 cell monolayers grown and loaded with [3H]–cholesterol as explained in Materials and methods were treated with serum free medium (control) or medium containing 0.5 lM of either wild type apoAI or the apoAI mutants DK107 and H9–10@H3–4 for 12 h. Efflux was expressed as % of the cellular radioactivity released to the medium. *P 6 0.05, **P 6 0.01 (levels of significance with respect to control); °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
efficiency for Chol removal under the same conditions. This fact is not surprising since several experiments with trypsin-
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
39
ized or chemically modified HDL apolipoproteins have shown that intact apolipoproteins were required to translocate intracellular cholesterol to the plasma membrane, but not to deliver cholesterol from the cell membrane into the medium [31–34]. Thus, these results support the idea that intracellular Chol mobilization and efflux may be controlled by different cellular processes. Role of the central apoAI Y helical region on intracellular cholesterol esterification and mobilization toward the plasma membrane It is well-known that in skin fibroblasts, apoAI induces a decrease in the cellular Chol pool available to be esterified by ACAT [13]. We were interested in investigating whether changes at the central apoAI Y-helical repeats play a role in the induction of such cellular responses. Fig. 5 shows that the incubation of cells with apoAI or H9–10@H3–4 resulted in a similar decrease of the esterified/unesterified cholesterol ratio. Instead, DK107 variant was not efficient for modifying that ratio. The pool of cholesterol available to ACAT was directly measured by using the strategy of Francis et al. [21]. A short pulse of radiolabeled oleic acid was given to the cells after the treatment with proteins and radioactivity in the cellular cholesteryl ester fraction was measured. As shown in Fig. 6, the helix-swapping mutant induced a decrease in the amount of [14C]–oleic acid incorporated into the cholesteryl ester fraction which was similar to that evoked by apoAI. Instead, deletion of lysine 107 totally impaired the ability of apoAI to decrease the incorporation of [14C]–oleic acid in the cellular cholesteryl ester fraction.
Fig. 5. Activity of apoAI variants for decreasing the cellular cholesteryl ester pool in CHO-K1 cells. Cells loaded with [3H]–cholesterol and incubated with apoAI variants as explained above were scrapped, and radioactivity counted in cellular cholesterol and cholesteryl ester fractions. Data are the mean and standard deviation of three independent measurements. *P 6 0.05 (levels of significance with respect to control); ° P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
Fig. 6. Comparative influence of the central apoAI domain on the cellular cholesterol pool available for esterification. Confluent CHO-K1 cell monolayers loaded with cholesterol as explained in Materials and Methods were treated with serum free medium (control) or medium containing 0.5 lM of apoAI variants at 37 °C for 12 h. Then, cells were pulse-labeled with [14C] oleic acid for 1 h, and the amount of oleic acid incorporated in cholesteryl ester was estimated as described in Materials and methods. Data are the mean and standard deviation of three independent measurements. **P 6 0.01 (levels of significance with respect to control); °P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
In addition, we investigated the role of the central Yhelical domain of apoAI for mediating recently synthesized Chol mobilization. For this experiment, we labeled cells with [14C] mevalonolactone and incubated them with the lipid-free apoAI variants. The accessible fraction was assumed to correspond to Chol residing in the cell membrane. Fig. 7 shows, as expected, that apoAI induced an increase in the proportion of radioactive cholestenone, which was mimicked by H9–10@H3–4, indicating that this mutant retained the full activity of apoAI in evoking intracellular Chol mobilization toward the cell membrane. The DK107 mutant is instead, totally inactive for increasing the newly synthesized Chol fraction accessible to be oxidized by exogenous Chol oxidase. Thus, these results indicate that deletion of DK107 residue results in a severe impairment to trigger the mobilization of intracellular Chol pools. Instead H9–10@H3–4 retains the whole activity of apoAI to decrease the cholesteryl esters cell content and the cellular Chol pool available to be esterified by ACAT, as well as to increase the transport of newly synthesized Chol toward the cell membrane in CHO cells. As mentioned above, in this mutant occurs the replacement of the central 3–4 Yhelix pair by the 9–10 helix, also containing a Y-type charge distribution. Thus, a specific sequence in the central helix does not seem to be important for this function, rather, a class Y character of these helices would be required. The fact that deletion of lysine 107 residue in apoAI results in a total loss of its ability to induce intracellular Chol mobilization in CHO cells and an impaired esterification of intracellular pools could be attributed to the loss of a positive charge in the central apoAI region. However,
40
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
zation of intracellular cholesterol pools, could be relevant for its atherogenicity and it would merit further intensive studies. In short, these results suggest that the central apoAI Yhelix pair is involved in triggering intracellular Chol pool mobilization, and that its activity requires an appropriate orientation of the hydrophobic/hydrophilic helix faces as well as a type Y charge distribution at the hydrophilic face. Acknowledgments
Fig. 7. Comparative influence of apoAI variants on the mobilization of newly synthesized cholesterol toward the cell membrane. Confluent CHOK1 cell monolayers loaded with cholesterol as explained in Materials and Methods, were labeled with [14C] mevalonic acid lactone for 3 h at 37 °C, and chased with serum free medium (control) or medium containing 0.2 lM apoAI, DK107 and H9–10@H3–4 for 1.5 h. Then, cells were fixed and treated with cholesterol oxidase to determine oxidase sensitive (plasma membrane) and resistant (intracellular) cell cholesterol fractions. Data are the means and standard deviation of three independent measurements. *P 6 0.05, **P 6 0.01 (levels of significance with respect to control); °P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
this is very unlikely since a positive charge is also absent at this position in H9–10@H3–4 which is totally active in this process. As previously noted [29,35,36], a more drastic structural change in DK107 is that the deletion of a single amino acid residue would change the orientation of subsequent amino acids within this a-helix by about 100°. The orientation of the hydrophobic face of this presumed helix to the lipids or other protein domains may be critical to some functional property. Then it is possible that the deletion of lysine 107 may cause a disruption of the protein conformational arrangement that could affect the interaction with lipids as well as with cell membrane proteins involved in triggering the signaling pathways leading to the mobilization of intracellular cholesterol pools [37,38]. The structural change produced by K107 deletion in the central Y helix pair also influences the size distribution of reconstituted discoidal HDL complexes [30,36] as well as the capacity of these complexes to be rearranged in the presence of LDL [36]. We do not know if the inability to suffer size rearrangement is related with the incapacity of DK107 to evoke intracellular cholesterol mobilization. The DK107 mutation has been associated with premature atherosclerosis [37], but the molecular mechanisms leading to this disorder are poorly understood. An enhanced amyloidogenicity was proposed for this mutant [39]. The DK107 carriers present low cholesterol and high triacylglycerol levels in HDL [37], and increased apoAI catabolic rate [38]. Early works suggested a decreased capacity of DK107 to activate lecithin cholesterol acyl transferase [35], but other reports [40] contradicted that conclusion. Our finding showing that DK107 is inactive in the mobili-
We are grateful to Dr. W.S. Davidson for his helpful discussion and L.E. Herna´ndez for her technical assistance. ApoAI was donated by Banco de Sangre, Instituto de Hemoterapia de la Provincia de Buenos Aires, La Plata. This work was supported by grants from the Ministerio de Salud de la Nacio´n, and from Agencia Nacional de Promocio´n Cientı´fica y Tecnolo´gica and Consejo de Investigaciones Cientı´ficas y Tecnolo´gicas, (CONICET). HAG, MCG and MAT are members of the Carrera del Investigador Cientı´fico (CONICET), Argentina. References [1] W.M. Nofer Jr., G. Assmann, Expert Rev. Cardiovasc. Ther. 3 (6) (2005) 1071–1086. [2] A.K. Groen, R.P. Oude Elferink, H.J. Verkade, F. Kuipers, Ann. Med. 36 (2004) 135–145. [3] B.L. Knight, Biochem. Soc. Trans. 32 (2004) 124–127. [4] J.D. Smith, W. Le Goff, M. Settle, G. Brubaker, C. Waelde, A. Horwitz, M.N. Oda, J. Lipid Res. 45 (2004) 635–644. [5] P.E. Fielding, K. Nagao, H. Hakamata, G. Chimini, C.J. Fielding, Biochemistry 39 (2000) 14113–14120. [6] O.J. Vaughan AM, J Lipid Res. 47 (2006) 2433–2443. [7] S.T. Thuahnai, S. Lund-Katz, P. Dhanasekaran, M. de la LleraMoya, M.A. Connelly, D.L. Williams, G.H. Rothblat, M.C. Phillips, J. Biol. Chem. 279 (2004) 12448–12455. [8] J.D. Toledo, E.D. Prieto, M.C. Gonzalez, J.L. Soulages, H.A. Garda, Arch. Biochem. Biophys. 428 (2004) 188–197. [9] J.D. Toledo, M.A. Tricerri, B. Corsico, H.A. Garda, Arch. Biochem. Biophys. 380 (2000) 63–70. [10] B. Corsico, J.D. Toledo, H.A. Garda, J. Biol. Chem. 276 (2001) 16978–16985. [11] A. Tricerri, B. Corsico, J.D. Toledo, H.A. Garda, R.R. Brenner, Biochim. Biophys. Acta 1391 (1998) 67–78. [12] D. Sviridov, N. Fidge, G. Beaumier-Gallon, C. Fielding, Biochem. J. 358 (2001) 79–86. [13] Y. Yamauchi, C.C. Chang, M. Hayashi, S. Abe-Dohmae, P.C. Reid, T.Y. Chang, S. Yokoyama, J. Lipid Res. 45 (2004) 1943–1951. [14] Q. Li, S. Yokoyama, J. Biol. Chem. 270 (1995) 26216–26223. [15] Q. Li, M. Tsujita, S. Yokoyama, Biochemistry 36 (1997) 12045– 12052. [16] J.P. Segrest, M.K. Jones, H. De Loof, C.G. Brouillette, Y.V. Venkatachalapathi, G.M. Anantharamaiah, J. Lipid Res. 33 (1992) 141–166. [17] C.G. Brouillette, G.M. Anantharamaiah, J.A. Engler, D.W. Borhani, Biochim. Biophys. Acta 1531 (2001) 4–46. [18] A.K. Behling Agree, M.A. Tricerri, K. Arnvig McGuire, S.M. Tian, A. Jonas, Biochim. Biophys. Acta 1594 (2002) 286–296. [19] E.G. Bligh, W.J. Dyer, Can. J. Biochem. Physiol. 37 (1959) 911– 917. [20] A.J. Mendez, J.F. Oram, E.L. Bierman, J. Biol. Chem. 266 (1991) 10104–10111.
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41 [21] G.A. Francis, A.J. Mendez, E.L. Bierman, J.W. Heinecke, Proc. Natl. Acad. Sci. USA 90 (1993) 6631–6635. [22] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951) 265–275. [23] D. Sviridov, L.E. Pyle, N. Fidge, J. Biol. Chem. 271 (1996) 33277– 33283. [24] Y. Uehara, S. Miura, A. von Eckardstein, S. Abe, A. Fujii, Y. Matsuo, S. Rust, S. Lorkowski, G. Assmann, T. Yamada, K. Saku, Atherosclerosis 191 (2007) 11–21. [25] I.C. Gelissen, M. Harris, K.A. Rye, C. Quinn, A.J. Brown, M. Kockx, S. Cartland, M. Packianathan, L. Kritharides, W. Jessup, Arterioscler. Thromb. Vasc. Biol. 26 (2006) 534–540. [26] S.A. Sanchez, M.A. Tricerri, E.J. Gratton, Lipid Res. 48 (2007) 1689– 1700. [27] J.W. Burgess, P.G. Frank, V. Franklin, P. Liang, D.C. McManus, M. Desforges, E. Rassart, Y.L. Marcel, Biochemistry 38 (1999) 14524– 14533. [28] D.W. Borhani, D.P. Rogers, J.A. Engler, C.G. Brouillette, Proc. Natl. Acad. Sci. USA 94 (1997) 12291–12296. [29] A. von Eckardstein, G. Castro, I. Wybranska, N. Theret, P. Duchateau, N. Duverger, J.C. Fruchart, G. Ailhaud, G. Assmann, J. Biol. Chem. 268 (1993) 2616–2622. [30] W. Huang, A. Matsunaga, W. Li, H. Han, A. Hoang, M. Kugi, T. Koga, D. Sviridov, N. Fidge, J. Sasaki, Atherosclerosis 159 (2001) 85– 91.
41
[31] I. Tabas, A.R. Tall, J. Biol. Chem. 259 (1984) 13897–13905. [32] J.B. Karlin, W.J. Johnson, C.R. Benedict, G.K. Chacko, M.C. Phillips, G.H. Rothblat, J. Biol. Chem. 262 (1987) 12557– 12564. [33] C.M. Mendel, S.T. Kunitake, J. Lipid Res. 29 (1988) 1171–1178. [34] W.J. Johnson, F.H. Mahlberg, G.K. Chacko, M.C. Phillips, G.H. Rothblat, J. Biol. Chem. 263 (1988) 14099–14106. [35] S.C. Rall Jr., K.H. Weisgraber, R.W. Mahley, Y. Ogawa, C.J. Fielding, G. Utermann, J. Haas, A. Steinmetz, H.J. Menzel, G. Assmann, J. Biol. Chem. 259 (1984) 10063–10070. [36] A. Jonas, A. von Eckardstein, K.E. Kezdy, A. Steinmetz, G. Assmann, J. Lipid Res. 32 (1991) 97–106. [37] M. Tilly-Kiesi, Q. Zhang, S. Ehnholm, J. Kahri, S. Lahdenpera, C. Ehnholm, M.R. Taskinen, Arterioscler. Thromb. Vasc. Biol. 15 (1995) 1294–1306. [38] M. Tilly-Kiesi, A.H. Lichtenstein, J.M. Ordovas, G. Dolnikowski, R. Malmstrom, M.R. Taskinen, E.J. Schaefer, Arterioscler. Thromb. Vasc. Biol. 17 (1997) 873–880. [39] M. Amarzguioui, G. Mucchiano, B. Haggqvist, P. Westermark, A. Kavlie, K. Sletten, H. Prydz, Biochem. Biophys. Res. Commun. 242 (1998) 534–539. [40] A. Jonas, A. von Eckardstein, L. Churgay, W.W. Mantulin, G. Assmann, Biochim. Biophys. Acta 1166 (1993) 202–210. [41] A.A. Ajees, G.M. Anantharamaiah, V.K. Mishra, M.M. Hussain, H.M. Murthy, Proc. Natl. Acad. Sci. USA 103 (2006) 2126–2131.
ABB Archives of Biochemistry and Biophysics 473 (2008) 34–41 www.elsevier.com/locate/yabbi
The central type Y amphipathic a-helices of apolipoprotein AI are involved in the mobilization of intracellular cholesterol depots Marina C. Gonzalez, Juan D. Toledo, M. Alejandra Tricerri, Horacio A. Garda * Instituto de Investigaciones Bioquı´micas de La Plata (INIBIOLP), CONICET/UNLP, Facultad de Ciencias Me´dicas, Calles 60 y 120, La Plata 1900, Buenos Aires, Argentina Received 16 January 2008, and in revised form 14 February 2008 Available online 23 February 2008
Abstract We studied the role of a central domain of human apolipoprotein AI (apoAI) in cholesterol mobilization and removal from cells. In order to check different protein conformations, we tested different sized and cholesterol-content reconstituted apoAI particles (rHDL). Meanwhile cholesterol-free discs were active to induce mobilization, only small cholesterol-containing rHDL were active. To test the influence of a central domain in such events, we used two apoAI variants: one, with its central Y helix pair replaced by the C-terminal domain, and the other having a lysine deleted in central region. The helix-swapping variant decrease the cholesterol pool available to acyl-CoA cholesterol acyl transferase and increase mobilization of newly synthesized cholesterol. Instead, the deletion mutant had no effect on both events. We conclude that the central domain of apoAI is involved in cholesterol cell traffic and solubilization, and that a Y-type charge distribution in polar face may be required, as well as a correct helices-polar face orientation. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Apolipoprotein AI; Cholesterol efflux; Chinese hamster ovary cells; Intracellular cholesterol mobilization; Mutants; Reverse cholesterol transport; Reconstituted high density lipoprotein
Apolipoprotein A-I (apoAI)1 is the major protein component of high density lipoproteins (HDL) which play a key role in reverse cholesterol transport (RCT), a process of great antiatherogenic importance that results in the removal of cholesterol (Chol) excess from peripheral cells, and its transport toward Chol metabolizing organs as liver or steroidogenic tissues [1,2]. Chol efflux mediated by lipidfree apoAI or other apolipoproteins can be considered as
*
Corresponding author. Fax: +54 221 425 8988. E-mail address: [email protected] (H.A. Garda). 1 Abbreviations used: ApoAI, apolipoprotein AI; rHDL, reconstituted high density lipoprotein; RCT, reverse cholesterol transport; CHO, chinese hamster ovary; MEM, minimal essential medium; FBS, fetal bovine serum; Chol, cholesterol; cholne, 4-cholesten-3-one; BSA, bovine serum albumin; POPC, 1-palmitoyl-2-oleyl sn-glycero-3-phosphocholine; PBS, phosphate-buffered saline; ACAT, acyl-CoA cholesterol acyl transferase; IPTG, isopropyl-beta-D-thiogalactopyranoside; ABCAI/GI, ATP binding cassette transporter A-I/G-I; FPLC, fast performance liquid chromatography. 0003-9861/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2008.02.021
the consequence of at least three partial events: (1) Initial apolipoprotein lipidation or loading with phospholipids at the cell membrane, resulting in the formation of discoidal HDL particles, (2) Stimulation of mobilization of intracellular Chol depots toward the cell membrane, and (3) Chol uptake from the cell membrane by discoidal HDL or other lipoprotein particles. There is a large body of evidence that the first process is catalyzed by ATP Binding Cassette A1, (ABCA1) [3]. It was proposed that ABCA1 catalyzes the concurrent loading of apolipoproteins with phospholipids and Chol [4], but there are also different pieces of evidence indicating that these events occur separately in time and in different specialized domains of the cell membrane [5]. Further Chol efflux toward the apoAI acceptors is proposed to occur mediated by the ABCG1 transporter, which should act sequentially to the ABCA1 [6]. Chol transfer from the cell membrane toward HDL can be produced by simple aqueous diffusion, in a process facilitated by Scavenger Receptor Type B1 (SR-B1) [7] and,
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
as we proposed for cholesterol-poor discoidal HDL [8–11], by the insertion in Chol-rich membrane domains of the central Y-helix of apoAI which facilitates its desorption. The molecular mechanisms of intracellular Chol pools mobilization toward the cell membrane induced by apoAI are poorly understood. It was proposed that apoAI induces the transport of intracellular Chol to cell-surface caveolae, possibly in part through the stimulation of caveolin expression [12]. Mobilization of intracellular pools also seems to be dependent on the expression of ABCA1, but can be separated from ABCA1/apoAI interaction or ABCA1-mediated phospholipid efflux under certain circumstances [13]. Activation of protein kinase C (PKC) by apoAI was reported to be involved in the mobilization of intracellular cholesterol pools, as the one available to be esterified by ACAT [13–15]. ApoAI contains several repeats which are predicted to form amphipathic a -helices [16,17] (see Fig. 3). Most of them (helices 1, 2, and 5–8) are 22mer repeats forming type A helices, a very common motif in all exchangeable apolipoproteins, characterized by two clusters of positively charged residues at the hydrophobic/hydrophilic interface and a cluster of negative charges at the center of the hydrophilic face. Besides, apoAI contains a G* helix at its N-terminus which is similar but not identical to the helices present in globular proteins, and two Y type helix pairs: one in the central region (residues 88–120, helices 3 and 4) and the other located at the C-terminus (residues 209– 241, helices 9 and 10). Type Y helices are similarly organized to type A helices, but they have positive residues disrupting the cluster of negative charges at the center of the hydrophilic face. The relative exposition of the different regions of apoAI to the aqueous/and or lipid environments is probable to play key roles in mediating cellular signals prone to redistribute the Chol among domains with different accessibility to be exported. Previously we have shown, using monoclonal antibodies, that the conformation of the central region, is strongly dependent on lipid environment in discoidal, reconstituted particles. This domain is highly exposed in small-cholesterol-free particles, and becomes masked as particles get larger and-or enriched in Chol [11]. Later, our group showed that the central region spanned residues 87–112 deeply inserted into the lipid bilayer of phospholipid membranes [10]; moreover, a synthetic peptide having the sequence of apoAI comprised from residues 77 to 120 (AI 77–120) promoted Chol efflux from Chinese hamster ovary (CHO) cells with high efficiency [8]. Thus, it is conceivable that a sequence spanning this peptide, could actively participate also in cell events destined to promote cholesterol efflux. Based on these previous results, we tested here the ability of different reconstituted particles to induce mobilization of the internal Chol pool from cells. We hypothesize that the exposition and/ or conformation of the central domain could play a role for inducing cellular events involved in cholesterol exportation. We also explored the
35
structural requirements of this central region by using two mutants. One of them (DK107) presents a deletion in helix 4, which changes by about 100° the helix registry and the orientation of the hydrophilic and hydrophobic faces of this amphipathic helix at both sides of the deletion point. In the second mutant (H9–10@H3–4) the central Yhelix pair 3–4 was replaced by the C end Y-helix pair. Altogether, our results suggest that the central type Y helix pair, which may be highly exposed in lipid-poor particles, contributes to the mobilization of intracellular Chol pools toward the cell membrane. Its functionality would not require a specific sequence, rather a type Y charge distribution in the polar helix face and a correct polar/hydrophobic face orientation throughout helix 4. Materials and methods Preparation of apoAI ApoAI was obtained from human serum (donated by Banco de Sangre, Instituto de Hemoterapia de la Provincia de Buenos Aires, La Plata, Argentina) as previously described [11]. It shows more than 95% purity as estimated by SDS polycrylamide gel electrophoresis (SDS–PAGE) using Coomasie blue staining.
Reconstitution of high density lipoprotein particles (rHDL) The cholate dialysis method was used to obtain rHDL containing ˚ Stokes diameter, with or without apoAI. Discoidal rHDL of 96 and 78 A cholesterol, were obtained starting from 95/1/150 and 40/1/65 POPC/ apoAI/sodium cholate, or 95/24/1/150 and 40/10/1/65 POPC/chol/apo˚ ) were purified AI/sodium cholate molar ratios, respectively. rHDL (120 A ˚ reconstitution mixture. Particles were purified by molecular from the 96 A size exclusion chromatography by FPLC as described before [10] and the molecular weight determined by polyacrylamide gel gradient electrophoresis in non-denaturing conditions (PAGGE). The composition was determined by the usual analytical methods.
Construction and expression of proapoAI variants Wild type proapoAI and DK107 DNA sequences were a generous gift from Dr. A. Jonas, University of Illinois at Urbana-Champaign, IL. The respective inserts, ligated to pET30a vectors were transformed into BL21 (DE) Escherichia coli cells (Novagen, Madison, WI), then expressed by induction with IPTG and purified as described elsewhere before [18]. Proteins were reconstituted into lipid particles and dialyzed against 10 mM Tris buffer, pH 8.0 (TNB). The fusion tag was removed by digestion with enterokinase followed by further elution through Ni-chelate columns (Novagen). Finally, proteins were delipidated as described previously [18]. The helix-swapping H9–10@H3–4 mutant inserted on a pET30 vector was a gift from Dr. W.S. Davidson (University of Cincinnati, OH). As explained above, in this construction amino acids spanning residues 88– 120 were replaced by residues 209–241, in such a way that the original central region of the protein was absent, but the variant kept the molecular weight preserved. A region with the sequence for an Igase protease cleavage site was designed directly upstream of the pro segment of the apoAI construction in order to specifically eliminate the histidine tag. BL21 (DE) E. coli cells were transformed with the vector containing the insert and plated in the presence of Kanamicyn. A colony was grown in Luria Bertani (LB) media at 37° C to an absorbance of 0.2 at 600 nm. Temperature was lowered to 25° C, and induced for 3 h with 0.4 mM IPTG. Proteins were purified through Ni-chelate columns in the presence
36
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
of 0.5 M Guanidine HCl, and the His tag eliminated with the Igase protease. Purity in the final preparations was confirmed by SDS–PAGE.
Cell culture About 400,000 Chinese hamster ovary (CHO-K1) cells were seeded in 6-well plates and incubated until confluence (3–4 days) with 2 ml of minimal essential medium (MEM) supplemented with penicillin/streptomycin solution (100 units/ml) and 10% FBS at 37 °C in a 5% CO2 atmosphere.
Cholesterol efflux and evaluation of cellular free and esterified cholesterol CHO-K1 cells were grown until confluence. Then, cell monolayers were washed with PBS and incubated for 41 h with serum-free medium containing 50 lg/ml cholesterol, 0.5 lCi/ml [3H] cholesterol, and 2 mg/ml BSA. After 8 h equilibration with MEM containing 1 mg/ml of BSA, monolayers were washed twice with PBS and treated for 12 h with serumfree medium containing 0.5 lM of apoAI, either as the lipid-free protein variant or as rHDL. Background efflux was determined from control cells incubated with serum-free medium. Then, media were collected and cells scrapped and suspended in methanol for lipid analysis. Radioactivity was measured by scintillation counting in aliquots of centrifuged culture medium and scrapped cells to calculate the percentage of cholesterol efflux. Cellular lipids were extracted by the method of Bligh and Dyer [19]. Sterol species were separated by thin layer chromatography on silica gel G plates developed in hexane:ethyl ether:acetic acid (80:20:1, v:v:v). Lipid spots corresponding to cholesteryl esters and unesterified cholesterol were identified by staining with I2 vapor and co-migration with authentic standards. Appropriate spots were scrapped for scintillation counting.
[14C] Mevalonate labeling of newly synthesized cellular cholesterol CHO-K1 cell monolayers, grown until confluence, were washed twice with PBS-BSA (2 mg/ml) and incubated for 48 h at 37° C with MEM, BSA (2 mg/ml) and cholesterol (50 lg/ml). Cells were washed three times with PBS-BSA and incubated at 37° C for 3 h in bicarbonate-free MEM containing 25 mM HEPES, 1 mg/ml BSA, 0.5 lCi/ml [14C] mevalonic acid lactone, 0.5 mM mevalonolactone, and 2 lg/ml of Sandoz ACAT inhibitor [compound 58-035] to prevent esterification of newly synthesized sterols [20]. After incubation, cell layers were rinsed twice with PBS-BSA and incubated in the same media without label for 1 h at 37° C to allow the conversion of labeled biosynthetic intermediates into [14C] sterols. Then, cells were washed and incubated with serum-free medium containing 0.2 lM apoAI, the protein variant or the rHDL to be tested for 90 min. After incubation at 37° C during the indicated time, cell layers were rinsed twice with 1 ml of PBS-BSA, fixed in 1% glutaraldehyde in PBS for 15 min on ice, and then treated with Chol oxidase to measure intracellular Chol translocation. Alternatively, fixed cells were stored frozen at 20° C before Chol oxidase treatment.
Treatment of cells with cholesterol oxidase After fixation, cells were washed two or three times with PBS, incubated at 37° C for 30 min with Chol oxidase (4 units/ml) in 5 mM NaPO4, pH 7.2, washed twice with PBS and stored frozen until lipids were extracted. The fraction of [14C] sterols oxidized to [14C] cholestenone was assumed to represent the plasma membrane associated sterol pool, whereas [14C] sterols resistant to enzymatic oxidation were assumed to be the intracellular pool.
Cholesterol pool available to ACAT The cellular cholesterol pool available to be esterified by ACAT was measured according to Francis et al. [21] after 12 h incubation with serum-
free medium containing 0.5 lM apoAI or the corresponding variant. Cells were then washed once with PBS before adding 0.25 lCi [14C] oleic acid and 10 lM of unlabeled oleate. After 1 h at 37 °C, cells were collected and lipid extracted as mentioned above. ACAT activity was estimated by measuring the incorporation of [14C] oleic acid into the cholesteryl ester fraction [15,21].
Other analytical methods Quantitative measurement of protein was performed by the method of Lowry el al. [22]. Student´s t-test was used to compare pooled data.
Other materials Minimal essential medium (MEM) was obtained from Invitrogen Corporation. Penicillin–streptomycin was from Parafarm (Bs. As., Argentina) and fetal bovine serum (FBS) was purchased from Bioser (Bs. As., Argentina). [3H] cholesterol, [14C] oleic acid and [14C] mevalonic acid lactone were obtained from Amersham Biosciences (Pittsburgh, PA). ACAT inhibitor [compound 58-035] was a generous gift from Sandoz Pharmaceuticals. Igase (endoproteinase), bovine serum albumin (BSA), cholesterol, mevalonolactone, glutaraldehyde, sodium cholate, Cholne and Chol oxidase were obtained from Sigma Co (St. Louis, MI) and enterokinase was from Boehringer-Mannheim. IPTG was from Promega (Buenos Aires, Argentina). POPC was purchased from Avanti Polar Lipids (Alabaster, AL).
Results and discussion Effect of apoAI lipidation state on cellular cholesterol mobilization and efflux from cells ˚ diamHomogeneous, POPC-apoAI rHDL (78 and 96 A ˚ eter), and POPC-Chol-apoAI rHDL (78, 96 and 120 A diameter) were incubated with cells for 90 min and their ability to induce de novo synthesized Chol mobilization to the plasma membrane tested in comparison with lipid-free 0 ˚ rHDL containing apoAI. Fig. 1A shows that 78 and 96 A POPC resulted as active as apoAI for increasing the cholestenone/cholesterol ratio which means, to mobilize internal Chol pools. Instead, when rHDL contain Chol, only the 0 ˚ particles were as effective as the lipid-free prosmall 78 A tein, while larger cholesterol-rich rHDL have completely lost that activity (Fig. 1B). These results indicate that Chol mobilization is dependent on the conformation of apoAI, which in fact is determined by the interaction with different lipid environments in the rHDL particles. As it was previously determined [11], both the increase in size and Chol content of rHDL decrease the exposure degree of a central epitope and the ability of these lipoprotein particles to interact with artificial phospholipid membranes. Thus, it is possible that in large and Chol-rich rHDL, the central region is not available to interact with the cell membrane sites or proteins responsible for triggering the signaling pathways conducing to the mobilization of the internal Chol pools. Following, we analyzed the ability of the same particles to remove Chol. Interestingly, the effect of size and sterol content does not seem to influence the final effectiveness of the tested particles when incubated with cells for 12 h. All of them removed about 60% of the amount
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
Fig. 1. Mobilization of recently synthesized cholesterol mediated by POPC-apoAI (A), or POPC-Chol apoAI rHDL (B). After 3 h incubation with [14C] mevalonic acid lactone, cells were treated with lipid-free apoAI or rHDL for 90 min and incubated with cholesterol oxidase. Results are expressed as the ratio of radioactivity associated to plasma membrane (cholestenone) and intracellular cholesterol pools (Chol oxidase resistant). * P 6 0.05, **P 6 0.01 (levels of significance with respect to control); ° P 6 0.05, °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
removed by lipid-free apoAI under the same conditions (Fig. 2A and B). Mobilization and efflux of intracellular lipid depots is not a simple consequence of plasma membrane Chol depletion since it is not induced by unspecific acceptors like albumin or cyclodextrin [23]. It is well-known, and it is confirmed here, that apoAI participates in triggering these processes, though the steps involved are not completely clear. Chol loading of different cells leads to the formation of oxysterols and the subsequent activation of the liver X receptor (LXR) [24]. A number of ATP binding cassette transporters (ABCs) involved in lipid transport are regulated through activation by LXR, including ABCA1 [24] and ABCG1. By using original experimental designs, Gelissen et al. [25] and Vaughan and Oram [6] showed that particles released from ABCA1-expressing cells incubated with apoAI, act as cholesterol acceptors from ABCG1-expressing cells, and thus they suggested that both ABC transporters work sequentially; the incubation of cells with lipid-free apoAI would lead to the efflux of Chol and phospholipid and the subsequent formation of nascent HDL particles,
37
Fig. 2. Cholesterol removal from CHO-K1 cells mediated by rHDL. ApoAI (either in the lipid-free state or as rHDL) was incubated with cells for 12 h as described in Methods. Results are shown as the % of radioactivity removed by cholesterol-free (A) or cholesterol containing (B) rHDL from the total radioactivity incorporated. **P 6 0.01 (levels of significance with respect to control); °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
which could be further enriched, via ABCG1, in Chol content [6]. Even though we use here a different cell system, it is plausible that, when loading cells with Chol, both transporters are being activated. Our experiments show that lipid-free apoAI, and small particles are effective to mobilize intracellular Chol, (Fig. 1) and that this ability is lost when particles are larger and Chol-enriched. It is probable then, that as the protein becomes lipidated, the domains required for triggering Chol mobilization become involved in contacts among the lipids of the particle, less exposed and unable to start cell signaling. Another observation is that Chol solubilization is more efficient when apoAI is in the lipid-free form than in rHDL, and independent, under our experimental conditions, on the size and Chol content of the rHDL particles (Fig. 2). We believe that, in our system, lipid-free apoAI is allowed to go through both steps: inducing ABCA1-dependent lipid solubilization, and giving rise to small particles, that are sequentially substrate for the ABCG1, which can suffer further lipidation. Instead, when incubating the same amount of protein
38
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
as discoidal particles, they can go through the second step only, thus is lipidation via ABCG1, and then lipid solubilization is limited. In Oram’s experiments, linear lipid solubilization is observed at shorter incubation times (less than 7 h), and thus, the fact that all particles showed a similar efficiency for Chol removal, could be representing a final condition in which all particles acquire phospholipids and cholesterol and reach a final, completely saturated and more stable apoAI conformation that could be further substrate of lecithin cholesterol-acyl transferase. This fact is in agreement with the observation that the product of apoAI lipidation through ABCA1 and ABCG1 overexpression results in a final, large and cholesterol rich lipoprotein [6]. This hypothesis can be also supported by the recent report from Sanchez et al [26]. By using Fluorescence Correlation Spectroscopy, they observed a decrease in the diffusion coefficient of rHDL in a single-molecule scale. They explain this observation as an increase in the size of every single particle as long as they capture lipids, but rather not due to a fusion of particles into a large aggregate. Structural requirements of the central apoAI Y helical region to promote cellular cholesterol efflux The second aim of this work, was to define the main role of the central helix of apoAI in the mentioned processes. The recently reported crystalline structure of full-length lipid-free apoAI [27] shows two well-separated domains: a 4-helix bundle formed by helices A to D, which includes predicted G*-helix and repeats 1 to 7; and a 2-helix bundle (helices E and F) which includes repeats 8 to 10 (Fig. 3). The N-terminal portion of the protein plays a key role in stabilizing the lipid-free state as it is indicated by this structure and also by the fact that its deletion results in totally different structure, which is extended, highly helicoidal, and thought to be similar to the lipid-bound state conformation [28]. While the long Y repeats (4 and 10) are helical in both reported crystalline structures, the short Y repeats (3 and 9) correspond to loops exposed to the surface in the lipid-free full length apoAI structure [27]. Next, we have studied the influence of the structural arrangement of the central region to evoke the previous events. Previous results from our group showed that the peptide spanning the apoAI central Y helix pair AI 77120 is active in promoting cell Chol efflux [8]. The physiological role of this region was investigated by using apoAI mutants with a disruption (DK107), or replacement of the domain spanning the central Y helix pair (H9–10@H3–4). Following exposition of [3H] -cholesterol loaded CHO-K1 cells to MEM alone (control) or to MEM containing the different apoAI variants, radioactivity was counted in the medium. In agreement with reported results in adipocytes and macrophages [29], deletion of the K107 residue did not significantly affect the capacity of the lipid-free protein to promote Chol removal from CHO cells (Fig. 4), or as shown previously by others as reconstituted HDL [29,30]. On the other hand, AI H9–10@H3–4 mutant showed lower
Fig. 3. Amphipathic a-helices in mature human apolipoprotein AI. Top: Amino acid sequence showing the correspondence between predicted amphipathic a-helices [16] and those found in the crystalline structure [41]. Predicted G*-helix is shown in black while class-A and class-Y repeats are shown in dark and light gray, respectively. The six a-helices found in the crystal structure are indicated with capital letters from A to F. The sequence of the peptide AI 77–120 is underlined. Lysine residue lacking in DK107 mutant is shown (as a K) in white. In H9–10@H3–4 mutant, the sequence spanning repeats 3–4 (residues 88–120) was replaced with that spanning repeats 9–10 (residues 209–241). Bottom: Wheel diagrams showing the typical charged residue distribution for class-A and class-Y helical repeats. Non-polar residues are shown in black, uncharged polar residues in white, and positive and negative residues in dark and light gray, respectively.
Fig. 4. Influence of the central helix on the activity of apoAI to promote cell cholesterol efflux. Confluent CHO-K1 cell monolayers grown and loaded with [3H]–cholesterol as explained in Materials and methods were treated with serum free medium (control) or medium containing 0.5 lM of either wild type apoAI or the apoAI mutants DK107 and H9–10@H3–4 for 12 h. Efflux was expressed as % of the cellular radioactivity released to the medium. *P 6 0.05, **P 6 0.01 (levels of significance with respect to control); °°P 6 0.01 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
efficiency for Chol removal under the same conditions. This fact is not surprising since several experiments with trypsin-
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
39
ized or chemically modified HDL apolipoproteins have shown that intact apolipoproteins were required to translocate intracellular cholesterol to the plasma membrane, but not to deliver cholesterol from the cell membrane into the medium [31–34]. Thus, these results support the idea that intracellular Chol mobilization and efflux may be controlled by different cellular processes. Role of the central apoAI Y helical region on intracellular cholesterol esterification and mobilization toward the plasma membrane It is well-known that in skin fibroblasts, apoAI induces a decrease in the cellular Chol pool available to be esterified by ACAT [13]. We were interested in investigating whether changes at the central apoAI Y-helical repeats play a role in the induction of such cellular responses. Fig. 5 shows that the incubation of cells with apoAI or H9–10@H3–4 resulted in a similar decrease of the esterified/unesterified cholesterol ratio. Instead, DK107 variant was not efficient for modifying that ratio. The pool of cholesterol available to ACAT was directly measured by using the strategy of Francis et al. [21]. A short pulse of radiolabeled oleic acid was given to the cells after the treatment with proteins and radioactivity in the cellular cholesteryl ester fraction was measured. As shown in Fig. 6, the helix-swapping mutant induced a decrease in the amount of [14C]–oleic acid incorporated into the cholesteryl ester fraction which was similar to that evoked by apoAI. Instead, deletion of lysine 107 totally impaired the ability of apoAI to decrease the incorporation of [14C]–oleic acid in the cellular cholesteryl ester fraction.
Fig. 5. Activity of apoAI variants for decreasing the cellular cholesteryl ester pool in CHO-K1 cells. Cells loaded with [3H]–cholesterol and incubated with apoAI variants as explained above were scrapped, and radioactivity counted in cellular cholesterol and cholesteryl ester fractions. Data are the mean and standard deviation of three independent measurements. *P 6 0.05 (levels of significance with respect to control); ° P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
Fig. 6. Comparative influence of the central apoAI domain on the cellular cholesterol pool available for esterification. Confluent CHO-K1 cell monolayers loaded with cholesterol as explained in Materials and Methods were treated with serum free medium (control) or medium containing 0.5 lM of apoAI variants at 37 °C for 12 h. Then, cells were pulse-labeled with [14C] oleic acid for 1 h, and the amount of oleic acid incorporated in cholesteryl ester was estimated as described in Materials and methods. Data are the mean and standard deviation of three independent measurements. **P 6 0.01 (levels of significance with respect to control); °P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
In addition, we investigated the role of the central Yhelical domain of apoAI for mediating recently synthesized Chol mobilization. For this experiment, we labeled cells with [14C] mevalonolactone and incubated them with the lipid-free apoAI variants. The accessible fraction was assumed to correspond to Chol residing in the cell membrane. Fig. 7 shows, as expected, that apoAI induced an increase in the proportion of radioactive cholestenone, which was mimicked by H9–10@H3–4, indicating that this mutant retained the full activity of apoAI in evoking intracellular Chol mobilization toward the cell membrane. The DK107 mutant is instead, totally inactive for increasing the newly synthesized Chol fraction accessible to be oxidized by exogenous Chol oxidase. Thus, these results indicate that deletion of DK107 residue results in a severe impairment to trigger the mobilization of intracellular Chol pools. Instead H9–10@H3–4 retains the whole activity of apoAI to decrease the cholesteryl esters cell content and the cellular Chol pool available to be esterified by ACAT, as well as to increase the transport of newly synthesized Chol toward the cell membrane in CHO cells. As mentioned above, in this mutant occurs the replacement of the central 3–4 Yhelix pair by the 9–10 helix, also containing a Y-type charge distribution. Thus, a specific sequence in the central helix does not seem to be important for this function, rather, a class Y character of these helices would be required. The fact that deletion of lysine 107 residue in apoAI results in a total loss of its ability to induce intracellular Chol mobilization in CHO cells and an impaired esterification of intracellular pools could be attributed to the loss of a positive charge in the central apoAI region. However,
40
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41
zation of intracellular cholesterol pools, could be relevant for its atherogenicity and it would merit further intensive studies. In short, these results suggest that the central apoAI Yhelix pair is involved in triggering intracellular Chol pool mobilization, and that its activity requires an appropriate orientation of the hydrophobic/hydrophilic helix faces as well as a type Y charge distribution at the hydrophilic face. Acknowledgments
Fig. 7. Comparative influence of apoAI variants on the mobilization of newly synthesized cholesterol toward the cell membrane. Confluent CHOK1 cell monolayers loaded with cholesterol as explained in Materials and Methods, were labeled with [14C] mevalonic acid lactone for 3 h at 37 °C, and chased with serum free medium (control) or medium containing 0.2 lM apoAI, DK107 and H9–10@H3–4 for 1.5 h. Then, cells were fixed and treated with cholesterol oxidase to determine oxidase sensitive (plasma membrane) and resistant (intracellular) cell cholesterol fractions. Data are the means and standard deviation of three independent measurements. *P 6 0.05, **P 6 0.01 (levels of significance with respect to control); °P 6 0.05 (levels of significance with respect to lipid free apoAI). When P was > 0.05, no symbol was used.
this is very unlikely since a positive charge is also absent at this position in H9–10@H3–4 which is totally active in this process. As previously noted [29,35,36], a more drastic structural change in DK107 is that the deletion of a single amino acid residue would change the orientation of subsequent amino acids within this a-helix by about 100°. The orientation of the hydrophobic face of this presumed helix to the lipids or other protein domains may be critical to some functional property. Then it is possible that the deletion of lysine 107 may cause a disruption of the protein conformational arrangement that could affect the interaction with lipids as well as with cell membrane proteins involved in triggering the signaling pathways leading to the mobilization of intracellular cholesterol pools [37,38]. The structural change produced by K107 deletion in the central Y helix pair also influences the size distribution of reconstituted discoidal HDL complexes [30,36] as well as the capacity of these complexes to be rearranged in the presence of LDL [36]. We do not know if the inability to suffer size rearrangement is related with the incapacity of DK107 to evoke intracellular cholesterol mobilization. The DK107 mutation has been associated with premature atherosclerosis [37], but the molecular mechanisms leading to this disorder are poorly understood. An enhanced amyloidogenicity was proposed for this mutant [39]. The DK107 carriers present low cholesterol and high triacylglycerol levels in HDL [37], and increased apoAI catabolic rate [38]. Early works suggested a decreased capacity of DK107 to activate lecithin cholesterol acyl transferase [35], but other reports [40] contradicted that conclusion. Our finding showing that DK107 is inactive in the mobili-
We are grateful to Dr. W.S. Davidson for his helpful discussion and L.E. Herna´ndez for her technical assistance. ApoAI was donated by Banco de Sangre, Instituto de Hemoterapia de la Provincia de Buenos Aires, La Plata. This work was supported by grants from the Ministerio de Salud de la Nacio´n, and from Agencia Nacional de Promocio´n Cientı´fica y Tecnolo´gica and Consejo de Investigaciones Cientı´ficas y Tecnolo´gicas, (CONICET). HAG, MCG and MAT are members of the Carrera del Investigador Cientı´fico (CONICET), Argentina. References [1] W.M. Nofer Jr., G. Assmann, Expert Rev. Cardiovasc. Ther. 3 (6) (2005) 1071–1086. [2] A.K. Groen, R.P. Oude Elferink, H.J. Verkade, F. Kuipers, Ann. Med. 36 (2004) 135–145. [3] B.L. Knight, Biochem. Soc. Trans. 32 (2004) 124–127. [4] J.D. Smith, W. Le Goff, M. Settle, G. Brubaker, C. Waelde, A. Horwitz, M.N. Oda, J. Lipid Res. 45 (2004) 635–644. [5] P.E. Fielding, K. Nagao, H. Hakamata, G. Chimini, C.J. Fielding, Biochemistry 39 (2000) 14113–14120. [6] O.J. Vaughan AM, J Lipid Res. 47 (2006) 2433–2443. [7] S.T. Thuahnai, S. Lund-Katz, P. Dhanasekaran, M. de la LleraMoya, M.A. Connelly, D.L. Williams, G.H. Rothblat, M.C. Phillips, J. Biol. Chem. 279 (2004) 12448–12455. [8] J.D. Toledo, E.D. Prieto, M.C. Gonzalez, J.L. Soulages, H.A. Garda, Arch. Biochem. Biophys. 428 (2004) 188–197. [9] J.D. Toledo, M.A. Tricerri, B. Corsico, H.A. Garda, Arch. Biochem. Biophys. 380 (2000) 63–70. [10] B. Corsico, J.D. Toledo, H.A. Garda, J. Biol. Chem. 276 (2001) 16978–16985. [11] A. Tricerri, B. Corsico, J.D. Toledo, H.A. Garda, R.R. Brenner, Biochim. Biophys. Acta 1391 (1998) 67–78. [12] D. Sviridov, N. Fidge, G. Beaumier-Gallon, C. Fielding, Biochem. J. 358 (2001) 79–86. [13] Y. Yamauchi, C.C. Chang, M. Hayashi, S. Abe-Dohmae, P.C. Reid, T.Y. Chang, S. Yokoyama, J. Lipid Res. 45 (2004) 1943–1951. [14] Q. Li, S. Yokoyama, J. Biol. Chem. 270 (1995) 26216–26223. [15] Q. Li, M. Tsujita, S. Yokoyama, Biochemistry 36 (1997) 12045– 12052. [16] J.P. Segrest, M.K. Jones, H. De Loof, C.G. Brouillette, Y.V. Venkatachalapathi, G.M. Anantharamaiah, J. Lipid Res. 33 (1992) 141–166. [17] C.G. Brouillette, G.M. Anantharamaiah, J.A. Engler, D.W. Borhani, Biochim. Biophys. Acta 1531 (2001) 4–46. [18] A.K. Behling Agree, M.A. Tricerri, K. Arnvig McGuire, S.M. Tian, A. Jonas, Biochim. Biophys. Acta 1594 (2002) 286–296. [19] E.G. Bligh, W.J. Dyer, Can. J. Biochem. Physiol. 37 (1959) 911– 917. [20] A.J. Mendez, J.F. Oram, E.L. Bierman, J. Biol. Chem. 266 (1991) 10104–10111.
M.C. Gonzalez et al. / Archives of Biochemistry and Biophysics 473 (2008) 34–41 [21] G.A. Francis, A.J. Mendez, E.L. Bierman, J.W. Heinecke, Proc. Natl. Acad. Sci. USA 90 (1993) 6631–6635. [22] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, J. Biol. Chem. 193 (1951) 265–275. [23] D. Sviridov, L.E. Pyle, N. Fidge, J. Biol. Chem. 271 (1996) 33277– 33283. [24] Y. Uehara, S. Miura, A. von Eckardstein, S. Abe, A. Fujii, Y. Matsuo, S. Rust, S. Lorkowski, G. Assmann, T. Yamada, K. Saku, Atherosclerosis 191 (2007) 11–21. [25] I.C. Gelissen, M. Harris, K.A. Rye, C. Quinn, A.J. Brown, M. Kockx, S. Cartland, M. Packianathan, L. Kritharides, W. Jessup, Arterioscler. Thromb. Vasc. Biol. 26 (2006) 534–540. [26] S.A. Sanchez, M.A. Tricerri, E.J. Gratton, Lipid Res. 48 (2007) 1689– 1700. [27] J.W. Burgess, P.G. Frank, V. Franklin, P. Liang, D.C. McManus, M. Desforges, E. Rassart, Y.L. Marcel, Biochemistry 38 (1999) 14524– 14533. [28] D.W. Borhani, D.P. Rogers, J.A. Engler, C.G. Brouillette, Proc. Natl. Acad. Sci. USA 94 (1997) 12291–12296. [29] A. von Eckardstein, G. Castro, I. Wybranska, N. Theret, P. Duchateau, N. Duverger, J.C. Fruchart, G. Ailhaud, G. Assmann, J. Biol. Chem. 268 (1993) 2616–2622. [30] W. Huang, A. Matsunaga, W. Li, H. Han, A. Hoang, M. Kugi, T. Koga, D. Sviridov, N. Fidge, J. Sasaki, Atherosclerosis 159 (2001) 85– 91.
41
[31] I. Tabas, A.R. Tall, J. Biol. Chem. 259 (1984) 13897–13905. [32] J.B. Karlin, W.J. Johnson, C.R. Benedict, G.K. Chacko, M.C. Phillips, G.H. Rothblat, J. Biol. Chem. 262 (1987) 12557– 12564. [33] C.M. Mendel, S.T. Kunitake, J. Lipid Res. 29 (1988) 1171–1178. [34] W.J. Johnson, F.H. Mahlberg, G.K. Chacko, M.C. Phillips, G.H. Rothblat, J. Biol. Chem. 263 (1988) 14099–14106. [35] S.C. Rall Jr., K.H. Weisgraber, R.W. Mahley, Y. Ogawa, C.J. Fielding, G. Utermann, J. Haas, A. Steinmetz, H.J. Menzel, G. Assmann, J. Biol. Chem. 259 (1984) 10063–10070. [36] A. Jonas, A. von Eckardstein, K.E. Kezdy, A. Steinmetz, G. Assmann, J. Lipid Res. 32 (1991) 97–106. [37] M. Tilly-Kiesi, Q. Zhang, S. Ehnholm, J. Kahri, S. Lahdenpera, C. Ehnholm, M.R. Taskinen, Arterioscler. Thromb. Vasc. Biol. 15 (1995) 1294–1306. [38] M. Tilly-Kiesi, A.H. Lichtenstein, J.M. Ordovas, G. Dolnikowski, R. Malmstrom, M.R. Taskinen, E.J. Schaefer, Arterioscler. Thromb. Vasc. Biol. 17 (1997) 873–880. [39] M. Amarzguioui, G. Mucchiano, B. Haggqvist, P. Westermark, A. Kavlie, K. Sletten, H. Prydz, Biochem. Biophys. Res. Commun. 242 (1998) 534–539. [40] A. Jonas, A. von Eckardstein, L. Churgay, W.W. Mantulin, G. Assmann, Biochim. Biophys. Acta 1166 (1993) 202–210. [41] A.A. Ajees, G.M. Anantharamaiah, V.K. Mishra, M.M. Hussain, H.M. Murthy, Proc. Natl. Acad. Sci. USA 103 (2006) 2126–2131.