Sphingomyelin inhibits the lecithin-cholesterol acyltransferase reaction with reconstituted high density lipoproteins by decreasing enzyme binding.

Lecithin-cholesterol acyltransferase (LCAT) catalyzes the formation of cholesterol esters on high density lipoproteins (HDL) and plays a critical role in reverse cholesterol transport. Sphingomyelin, an important constituent of HDL, may regulate the activity of LCAT at any of the key steps of the enzymatic reaction: binding of LCAT to the interface, activation by apo A-I, or inhibition at the catalytic site. In order to clarify the role of sphingomyelin in the regulation of the LCAT reaction and its effects on the structure of apolipoprotein A-I, we prepared reconstituted HDL (rHDL) containing egg phosphatidylcholine, cholesterol, apolipoprotein A-I, and up to 22 mol % sphingomyelin. Because the interfacial properties of substrate particles can dramatically affect LCAT binding and kinetics, we also prepared and analyzed proteoliposome substrates having the same components as the rHDL, except for a 4-fold higher ratio of phospholipid to apolipoprotein A-I. The reaction kinetics of LCAT with the rHDL particles revealed no significant change in the apparent Vmax but showed a concentration-dependent increase in slope of the reciprocal plots and in the apparent Km values with sphingomyelin content. The dissociation constant (Kd) for LCAT with these particles increased linearly with sphingomyelin content up to 22 mol %, changing in parallel with the apparent Km values. No structural changes of apolipoprotein A-I were detected in the particles with increasing content of sphingomyelin, but fluorescence results with lipophilic probes revealed that significant changes in the acyl chain, backbone, and head group regions of the lipid bilayer of the particles are introduced by the addition of sphingomyelin. On the other hand, the proteoliposome substrates also had increased Kd values for LCAT at high sphingomyelin contents but compared with the rHDL particles had a 6-10-fold lower affinity for LCAT binding and exhibited kinetics consistent with competitive inhibition by sphingomyelin at the active site. These results show conclusively that the dominant mechanism for the inhibition of LCAT activity with rHDL particles by sphingomyelin is the impaired binding of the enzyme to the interface. The results also underscore the significant differences in the enzyme reaction kinetics with different substrate particles.

Lecithin-cholesterol acyltransferase (LCAT) 1 plays a critical role in the maintenance of cholesterol homeostasis. LCAT participates in the reverse cholesterol transport pathway, maintaining a gradient for the diffusion of free cholesterol from peripheral tissues into high density lipoproteins (HDL) by catalyzing cholesterol ester (CE) formation from HDL surface phosphatidylcholine (PC) and cholesterol. As a result of LCAT activity, cholesterol is removed from peripheral cell membranes and, as CE, is ultimately removed and metabolized by the liver.
Several factors contribute to the maintenance of cellular cholesterol homeostasis, including the activity of HMG-CoA reductase and acyl-CoA acyltransferase, which are responsible for production and esterification of intracellular cholesterol, respectively. In addition, the sphingomyelin (SPM) content of the cell membrane is thought to contribute to the maintenance of cellular cholesterol homeostasis. The cholesterol content of the cell membranes is positively correlated with SPM content (1). Sphingomyelin is thought to bind cholesterol with high affinity and inhibit its efflux from the plasma membrane by preventing cholesterol desorption (2); SPM also prevents the exchange of cholesterol between the plasma membrane and intracellular pools (3). Although its role in the regulation of cellular cholesterol homeostasis has become more clearly established, the role of SPM in the function of circulating lipoproteins remains unclear.
has an altered epitope expression compared with discoidal rHDL particles, suggesting that the altered lipid composition of lymphatic HDL promotes an apoA-I conformation that may render it incapable of activating LCAT (11). Clearly, structural and compositional differences of lymphatic HDL contribute to their lower reactivity with LCAT. The phospholipid content of HDL could dramatically alter LCAT activity by interacting with and changing the conformation of apoA-I, by changing the nature of the lipid interface, by inhibiting LCAT through competition for substrate binding at the active site, or by a combination of the above. It is possible that the presence of SPM in HDL alters LCAT reactivity by several of these mechanisms.
In addition to its effects in native lipoproteins, SPM has been shown to be a poor matrix for the LCAT reaction with synthetic substrates. In fact, LCAT activity is lower with PC regardless of the acyl chain composition when the PC is presented in an SPM matrix compared with a fluid ether PC matrix (12). Research by Subbaiah and Liu (13) using proteoliposome substrates and native lipoproteins suggests that SPM competes with PC for binding to the active site of LCAT and thus participates in the regulation of the LCAT reaction. Although proteoliposomes are useful as substrates for LCAT in vitro, they lack the defined apoA-I structures characteristic of discoidal reconstituted HDL (rHDL) and have very different structures from native HDL. In order to clarify the role of SPM in the regulation of the LCAT reaction and its effects on the structure of apoA-I, we have prepared discoidal rHDL with egg PC, cholesterol, and apoA-I containing up to 22 mol % SPM. We have characterized these particles in terms of protein structure and properties of the lipid components and the lipid-water interface. We examined the structure of apoA-I and the binding affinity and reaction kinetics of LCAT with these particles. Our results suggest that in discoidal rHDL particles SPM introduces changes in the structure of the lipids, decreases the binding of LCAT to the substrate particles, and thus regulates the LCAT reaction. In contrast, proteoliposome substrates bind more weakly to LCAT and experience the regulatory effect of SPM not only at the binding step but also at the catalytic step.

EXPERIMENTAL PROCEDURES
Materials and Preparations-Human LCAT was purified by methods described previously (14,15). Its average specific activity, using standard rHDL substrates, was around 100 nmol CE/hr/g LCAT, and it remained fully active over an 8-month period. Human apoA-I was prepared using a modification of the method of Nichols et al. (16). Egg PC, egg SPM, cholesterol, and sodium cholate were obtained from Sigma. Radiolabeled [4-14 C]cholesterol and 3 H-labeled [2-palmitoyl-9,10-3 H]dipalmitoylphosphatidylcholine ( 3 H-DPPC) were purchased from DuPont NEN.
The rHDL were prepared using the sodium cholate dialysis method (17)  . Radiolabeled cholesterol (5,000 cpm/nmol) was incorporated only into the particle preparations that were used for the determination of reaction kinetics with LCAT. Radiolabeled 3 H-DPPC (20,000 cpm/ nmol of cholesterol) was included in the preparation of the standard substrate for the activity inhibition measurements (18). Cholate was removed by exhaustive dialysis against 0.1 M Tris-HCl, 0.0005% EDTA, 0.15 NaCl, 1 mM NaN 3 , pH 8.0 buffer at 4°C. Diameters of rHDL were determined by nondenaturing 8 -25% polyacrylamide gradient gel electrophoresis (Pharmacia PHAST gradient gel electrophoresis). Phosphatidylcholine was separated from SPM and cholesterol by TLC (Analtech Analytical) using chloroform/methanol/ammonia (65:25:4, v/v/v) and quantified using the method of Chen et al. (19). Protein content was determined from absorbance at 280 nm using the percentage extinction coefficient for apoA-I, 11.5 ϫ 10 2 g Ϫ1 cm 2 (20) and by the method of Lowry et al. (21).
Egg PC/C or SPM/C vesicles were prepared in ratios of 10:1, phos-pholipid/cholesterol. Preparations were dried down and dispersed in 10 ml of standard buffer. The samples were sonicated on ice (egg PC) or at 50°C (SPM) until they cleared, alternating 3 min of sonication with 1-min rest periods. The samples were centrifuged at 15,000 rpm at 15°C for 1 h. The phospholipid content of the supernatant (ϳ5 mg/ml) was assessed by the method of Chen (19) and with a standard phospholipid assay kit (Wako Phospholipids B). The vesicles were used immediately following the preparation. Activity Inhibition Assay for Determining the Binding of LCAT-LCAT affinity for the rHDL and proteoliposome particles was assessed by the activity inhibition assay previously described (18). Briefly, the activity inhibition assay uses rHDL containing 3 H-DPPC as substrates for the LCAT reaction. When unlabeled rHDL or proteoliposomes are present in the reaction mixture, LCAT equilibrates between labeled and unlabeled rHDL, and the total radiolabeled CE production is decreased. Lineweaver-Burke plots of reciprocal velocity versus reciprocal substrate apolipoprotein concentration give a family of lines that are consistent with the pattern expected for competitive inhibition. Plots of the slopes of these lines versus the concentration of apolipoprotein in the competing rHDL give a straight line from which the K i (K d , the dissociation constant) for the competing rHDL can be obtained. The substrates used were the egg PC/cholesterol/apoA-I rHDL prepared as described above including 3 H-DPPC. The LCAT reaction mixture consisted of rHDL or proteoliposomes with substrate apoA-I contents ranging from 2.5 to 43 g, 2 mg of defatted bovine serum albumin, 4 mM 2-mercaptoethanol, unlabeled test rHDL or proteoliposomes with apolipoprotein contents ranging from 0 to 150 g, and standard buffer to 0.45 ml of total volume. When vesicles were analyzed for LCAT binding, 9 g of substrate apoA-I were included, and vesicle phospholipid contents from 0 up to 0.95 mg (egg PC) or 1.56 mg (SPM) were used as inhibitors; the remainder of the reaction mixture was unchanged. The reaction mixtures were incubated at 37°C for 5 min, and 50 l of a suitable LCAT dilution were added to start the reaction. The reaction proceeded 20 min and was stopped by the addition of 5 ml of chloroform/ methanol (2:1, v/v). Labeled CE were separated from cholesterol and phospholipids by thin layer chromatography and were quantitated by scintillation counting as described previously (14,17). All experiments were performed in duplicate on two separate particle preparations.
Enzymatic Reactions-LCAT reactions with rHDL or proteoliposomes containing [4-14 C]cholesterol were performed in standard buffer as reported previously (22). Reaction mixtures for kinetic analysis contained substrate concentrations ranging from 1 ϫ 10 Ϫ7 M (2.8 g) to 3 ϫ 10 Ϫ6 M (84 g) apoA-I, 2 mg of defatted bovine serum albumin, 4 mM ␤-mercaptoethanol, and 30 -50 ng of pure LCAT. Apparent kinetic constants were obtained from Lineweaver-Burke analysis of the data. Experiments on two separate preparations were performed in duplicate, each giving similar results. A more detailed analysis of the effects of an interfacial inhibitor on the enzyme kinetics was based on the following expression derived by Verger et al. (23) and applied by Jonas et al. (24) to the inhibition of LCAT by ether PC: In this expression, v o is the initial velocity; K m * is the intrinsic Michaelis-Menten constant (in molecules/surface); I is the inhibitor concentration (in molecules/surface); K i * is the intrinsic inhibition constant; S is the interfacial substrate concentration (in molecules/surface); K cat is the catalytic rate constant, E o is the total enzyme concentration (in molecules/volume); K d is the dissociation constant of the enzyme from the interface; and S o is the bulk substrate concentration (in molecules/volume). According to this equation, in the presence of an interfacial inhibitor, reciprocal plots of v o versus bulk PC concentration should give a family of straight lines with an increasing 1/v o intercept, i.e. decreasing apparent V max values. The slope of the lines should be constant if K d remains the same. However, changes in K d would give increasing or decreasing slopes.
Fluorescence Characterization and Circular Dichroism-The lipid dynamics and hydration of the rHDL containing SPM were examined using fluorescent probes. The motions and polarity of the environment of the acyl chain, glycerol backbone, and head group regions were assessed using 1,3,5 diphenylhexatriene (DPH), trimethylammonium-DPH (TMA-DPH), and 6-propionyl-2-dimethylaminonaphthalene (PRODAN), respectively. All fluorescent lipophilic probes were obtained from Molecular Probes (Eugene, OR). Fluorescence measurements and analysis of the data were performed as described previously (25). Circular dichroism spectra were measured with a Jasco J-720 spectropo-larimeter at 24°C between 200 and 250 nm using 0.1 mg/ml sample solutions and a 1-mm path length cuvette. The ␣-helical content of apo A-I in the rHDL particles containing SPM was estimated from the molar ellipticities at 222 nm using the method of Chen et al. (26) as previously reported (27). Two separate fluorescence experiments were performed, each giving similar results.

RESULTS
ApoA-I has been shown to combine with a variety of phospholipids, including SPM, to form stable discoidal rHDL particles (12, 28 -30). Table I summarizes the properties of rHDL particles prepared in this study; their size distribution is shown in Fig. 1. The composition and size of the rHDL particles are consistent with a discoidal morphology (31). The moderate content of egg SPM, with saturated acyl chains (86% palmitoyl) (32), does not appear to alter significantly the structure of apoA-I. Circular dichroism spectra for all of the rHDL in this series were quite similar (data not shown) indicating that the ␣-helical content of apoA-I changes very little, as shown on Table I. It is clear that rHDL particles with similar size and total lipid contents can be prepared with apoA-I and mixtures of egg PC and SPM. The proteoliposome preparations had protein and lipid compositions very similar to those of the initial reaction mixtures and migrated on nondenaturing gradient gel electrophoresis as heterogeneous populations of particles most having diameters greater than 180 Å (data not shown).
LCAT reactivity with rHDL is highly dependent upon the phospholipid composition of the interface of the substrate particle. SPM provides a poor matrix for the LCAT reaction when it is present as 89% of the interfacial phospholipid (12). To determine the effect of SPM incorporated into discoidal rHDL or into proteoliposomes on the LCAT reaction, we used the two series of substrates containing [4-14 C]cholesterol. Lineweaver-Burke plots for the enzymatic reactions are shown in Fig. 2. The resulting kinetic parameters are summarized in Table II. We found (Fig. 2B) that increasing SPM content in the rHDL had a minimal effect on the apparent V max (appV max ) of the LCAT reaction. The insignificant change in appV max in this rHDL series suggests that SPM competition with PC for the LCAT active site is minimal in these substrates. However, the increasing slopes of the Lineweaver-Burke plots with SPM content clearly indicate, by reference to the analysis of interfacial enzyme inhibition of Verger et al. (23) and the work of Jonas et al. (24), that the major effect is on the K m *K d /K cat E o parameters. Because the K m * and the K cat E o parameters of the catalytic step are in effect constants, then K d is the likely variable that affects the slope of the Lineweaver-Burke plots shown in Fig. 2.
The proteoliposome substrates had increasing slopes for the highest SPM contents ( Fig. 2A) consistent with increasing K d . However, the appV max values decreased with added SPM, suggesting competition of SPM for PC at the active site. These results are in complete agreement with those reported by Subbaiah and Liu (13). The apparent kinetic constants are summarized in Table II. In the absence of the SPM inhibitor, the appV max is 37% higher for an rHDL than for a proteoliposome substrate, and the appK m (in terms of PC concentration) is 4.4-fold lower for rHDL, giving an overall 5.2-fold greater catalytic efficiency (appV max /appK m ) for the rHDL substrates. This is the first quantitative comparison of these two widely used synthetic substrates for LCAT.
To confirm that SPM increases the K d for the interaction of rHDL and proteoliposomes with LCAT, we measured directly the LCAT binding affinity of a series of rHDL and proteoliposome particles without 14 C-cholesterol using the activity inhibition assay previously described (18). We found that the K d for rHDL increased about 5-fold in a linear manner with increasing SPM, and the K d for proteoliposomes increased 2-fold (see Table II). Clearly, the presence of SPM in the particles decreases LCAT affinity for the phospholipid interface. To determine if the effect of SPM on LCAT binding affinity could be observed independently of apoA-I, we prepared vesicles with cholesterol and either egg PC or SPM in a 1:10 molar ratio of cholesterol to phospholipid. The vesicles were used as test particles in competition with the standard LCAT substrate rHDL under the same conditions as the activity inhibition assay. As the amount of vesicle phospholipid increases, LCAT binds to the vesicle surface, and as a result, net production of radiolabeled cholesterol ester at the standard substrate surface decreases. Fig. 3 shows that the amount of SPM vesicle phos- a PC was separated from SPM by TLC (Analtech Analytical) using chloroform/methanol/ammonia (65:25:4, v/v/v) and quantified using the method of Chen et al. (19). Cholesterol content from the initial phospholipid/cholesterol ratios. Protein content from absorbance at 280 nm and extinction coefficient. The errors of measurement are approximately Ϯ5%. A-I, apolipoprotein A-I; C, cholesterol; PC, egg phosphatidylcholine; SPM, egg sphingomyelin.
b From nondenaturing gradient gel electrophoresis, relative to protein standards: bovine serum albumin, lactate dehydrogenase, thyroglobulin, and ferritin. Errors are Ϯ2 Å.
c Estimated from ellipticity at 222 nm from circular dichroism spectra using the method of Chen et al. (26). The errors of measurement are approximately Ϯ5%.

FIG. 1. Size distribution of rHDL prepared with apoA-I, cholesterol, and mixtures of egg PC and egg SPM. rHDL preparations
were separated by 8 -25% polyacrylamide gradient gel electrophoresis using Pharmacia Phast System. Bands were visualized by Coomassie staining. Gel bands were scanned with LKB UltroScan XL laser densitometer. Absorbance scales are arbitrary, and peak height has been adjusted to give peaks of comparable height. Stokes diameter was calculated using protein markers with known Stokes radii as standards: bovine serum albumin, lactate dehydrogenase, horse ferritin, and thyroglobulin. pholipid (0.25 mg/ml) required to inhibit the LCAT reaction by 50% was nearly 5-fold higher than the amount of egg PC vesicle phospholipid (0.048 mg/ml) necessary for a similar inhibition. This result indicates that LCAT binds to the SPM/cholesterol vesicle surface in the absence of apoA-I with less affinity than it does to the egg PC/cholesterol vesicle surface. Because the K d for LCAT binding to the egg PC rHDL particles is 2.3 ϫ 10 Ϫ5 M or 0.018 mg/ml (Table II), it follows that the affinity of LCAT for the rHDL is about 3-fold greater than for the egg PC/ cholesterol vesicles. Table II also shows that the affinity of LCAT for rHDL in the absence of SPM is 10-fold greater than for comparable proteoliposomes.
To investigate the effects of the addition of SPM on the properties of the rHDL phospholipid phase, we examined the lipid dynamics and hydration of the surface of the rHDL particles using lipophilic fluorescent probes. DPH fluorescence polarization reports on the fluidity of the acyl chain region of the rHDL particles. Fig. 4 shows the temperature dependence of the polarization of DPH for the rHDL containing 22% SPM, 11% SPM, and 0% SPM. DPH polarization increases with increasing SPM content in the rHDL. As shown in Fig. 5, changes in TMA-DPH polarization with temperature indicate similar effects in the phospholipid backbone region, the region between the hydrophobic acyl chains, and the hydrophilic head group region. The higher polarization values observed with both probes indicate that the mobility of the lipids is restricted and order is increased.
The fluorescence of PRODAN was used to probe the polarity of the phospholipid head group region. The fluorescence spectra of PRODAN are quite sensitive to the polarity of the probe environment (33,34). Fig. 6 shows the fluorescence intensity ratio for PRODAN at 440/490 nm. The probe is in a more polar environment in the egg PC control rHDL compared with the rHDL containing SPM as indicated by the blue-shift that occurs with increased SPM content in the rHDL. This suggests that the presence of SPM shields PRODAN from water molecules in the head group region. These changes are consistent with decreased hydration of egg PC/SPM interfaces as a result of altered phospholipid packing or hydrogen bonding of SPM to cholesterol or to PC, which displaces water (35) and/or allows the probe to penetrate more deeply into the head group region. DISCUSSION Much of what is known about the regulation of LCAT activity with lipoprotein substrates has come from kinetic studies of LCAT with substrate analogs, such as proteoliposomes or rHDL. Many factors have been shown to modulate LCAT activity in vitro with these substrates. Among these are substrate size and morphology (disc versus sphere), phospholipid composition (head group, acyl chain, and unsaturation), and apolipoprotein composition and conformation (36). It is clear that these factors are closely interrelated. As a consequence, exam-  c rHDL particles described in Table I. ining the effect of one of these parameters on LCAT activity while holding the others constant is a difficult task; however, rHDL prepared by the sodium cholate dialysis method have made possible detailed studies of the LCAT reaction with particulate substrates of defined apolipoprotein and phospholipid compositions and similar morphologies. The conformation of apoA-I, the principal physiological activator of LCAT, is one of the key determinants of LCAT activity. ApoA-I conformation is related to the size of the particle and the phospholipid composition and content. In terms of the number of ␣-helices/molecule, apoA-I may adopt a conformation with six, seven, or eight ␣-helices, depending upon the amount of lipid complexed with the protein. But not all confor-

FIG. 3. Inhibition of LCAT activity by phospholipid vesicles.
Vesicles of egg PC/cholesterol (å) or SPM/cholesterol (f) were prepared by sonication in ratios of 10:1, phospholipid/cholesterol. Reaction conditions were identical to those of the activity inhibition method. Concentrations of vesicle phospholipid in the assay mixture ranged from 1.3 ϫ 10 Ϫ3 to 3.2 mg/ml.  6. PRODAN fluorescence intensity ratio at 440/490 nm in SPM containing rHDL as a function of temperature. f, 22 mol % SPM rHDL; q, 11 mol % SPM rHDL; å, 0 mol % SPM rHDL. PRODAN was added to rHDL samples (0.1 mg/ml apoA-I) in the ratio 300:1, phospholipid/probe (mol/mol). Experiments were performed as described previously (26). Two separate experiments were performed giving similar results. mations of apoA-I are equally effective in activating LCAT; the form of apoA-I with six helical segments is a poor activator of LCAT (37). rHDL prepared with DPPC and cholesterol are good substrates for LCAT if the particles are 97 Å in diameter. The same lipids, however, are 20-fold less reactive in particles that are 186 Å in diameter. Clearly, the amount of phospholipid complexed with apoA-I can profoundly influence LCAT activity through changes in apoA-I conformation and interfacial lipid properties.
Gross changes in the conformation of apoA-I can be avoided by preparing rHDL particles of nearly identical size and total phospholipid content. We have previously prepared a series of 96 Å diameter rHDL with apoA-I, cholesterol, and various mixtures of phospholipids to examine the effect of the interfacial lipid mixture on LCAT reactivity and binding affinity (18). In this study we prepared particles with apoA-I, cholesterol, and mixtures of egg PC and SPM, which are very similar in size. CD spectra for this rHDL series indicate that all have similar apoA-I ␣-helical contents. It is possible that subtle changes not detected by CD are introduced by the increased SPM content in the structure of apoA-I, which could alter its ability to activate LCAT, but within the resolution of our spectral methods the apoA-I structure appears identical for the series of rHDL particles used in this study.
We examined the reactivity of these rHDL particles with LCAT to investigate the effect of SPM content in discoidal substrates with similar apoA-I structure on LCAT reaction kinetics. The reaction of LCAT with aggregated lipid substrates involves multiple steps: the association of LCAT with the lipid interface, followed by activation by apolipoproteins, binding of lipid substrate(s) to the active site, and subsequent catalytic events (36). Sphingomyelin could influence the initial binding step through altered phospholipid packing as a consequence of its backbone structure or acyl chain composition, or SPM could influence subsequent steps in the LCAT reaction by acting as a competitor with PC in the LCAT phospholipase reaction or by sequestering cholesterol from LCAT on the rHDL surface. The role(s) of SPM are reflected in the kinetic parameters of the LCAT reaction.
If SPM were competing with PC in the active site of LCAT, significant changes would be expected in the appV max parameter according to the Verger et al. analysis (23). Such changes were observed by Subbaiah and Liu (13) with SPM and palmitoyl oleoyl-PC diether inclusion in proteoliposome substrates. In this study we confirmed their results with SPM (see Fig. 2A). Previously, we (24) had shown that DPPC diether incorporated into rHDL particles acts as an interfacial competitive inhibitor of LCAT with a K i * comparable with K m * and little change in K d . Similarly, Massey et al. (38) prepared rHDL substrates with various ratios of dimyristoyl PC (DMPC) and DMPC diether and observed that although the appK m did not change as a function of DMPC content, the appV max increased linearly with increasing DMPC. Pownall et al. (12) demonstrated that the appV max increases with increasing sterol content. We observe no change in appV max with increasing SPM content in the rHDL particles ( Fig. 2B and Table II), suggesting that SPM is not competing with PC at the active site of LCAT nor is it sequestering cholesterol in this rHDL series as was observed for SPM in the proteoliposomes studied by Subbaiah and Liu (13). The qualitative difference between our results for rHDL and those of Subbaiah and Liu (13) for proteoliposomes can be attributed to the different structure of the substrate particles. The properties of phospholipid bilayers in liposomes depend upon liposome composition and curvature (39). Differences in the physical properties of liposomes arise from variations in phospholipid molecular packing with curvature of the liposome (39,40). It is possible that the highly curved surface of the proteoliposomes decreases the hydrogen bonding of SPM molecules to other surface components and facilitates diffusion and binding to the active site, whereas the planar surface of the rHDL discs maximizes the intermolecular interactions of SPM. In addition, lipid phase separation of SPM or interactions with apoA-I are likely to be very different in the planar, protein-rich rHDL compared with the curved, relatively protein-poor proteoliposomes.
We observe changes in the slopes of the Lineweaver-Burke plots (Fig. 2) and appK m with increasing SPM content, suggesting that SPM content alters the initial binding of LCAT to the rHDL surface. The changes in appK m parallel the relationship observed between SPM content and dissociation constant for LCAT determined with these particles. Thus, it appears that SPM effects on the initial binding of LCAT are more important in discoidal substrates than SPM's role in sequestration or competition with molecular substrates. In the case of proteoliposomes, we also demonstrate an effect of SPM on the binding of LCAT to the interface. The effects of SPM on the binding of LCAT to proteoliposomes are also apparent in the results of Subbaiah and Liu (13); however, these authors do not address the change in slope of their 1/v o versus 1/PC plot. Rather, they interpret the reversal of the effect of SPM on the activity of LCAT by sphingomyelinase treatment of the substrates, as an indication that the physical state of the lipid does not affect the enzymatic reaction. Because ceramide remains in the bilayer following the removal of the phosphocholine group, treatment of the substrates by sphingomyelinase may not affect the lipid order and motions in the acyl chain region. However, the effects of ceramide on the hydration and packing at the interface may be profound and opposite to those of SPM, considering the major differences that are known to exist between diacylglycerols and corresponding phosphatidylcholines (41).
We provide strong evidence that the interfacial lipid structure of the rHDL changes as a result of the addition of SPM. We observe significant changes in the fluorescence properties of probes in the acyl chain, backbone, and head group regions of the rHDL phospholipids containing increasing SPM. The change in DPH polarization is most likely due to the high content of saturated palmitoyl acyl chains (86%) in egg SPM (32). Saturated acyl chains restrict the mobility and increase the order of the lipids and result in higher polarization values, as has been previously observed with DPH in rHDL prepared with DPPC or palmitoyl oleoyl-PC (29).
The higher polarization values for TMA-DPH in rHDL with increasing SPM content reflect the effects of SPM in the backbone region of the bilayer. In PC, the backbone region includes carbons 1, 2, and 3 of the glycerol backbone and the two ester bonds linking the acyl chains; PC has no capacity as a hydrogen bond donor. In SPM, however, this region includes the amide bond linking the palmitoyl acyl chain and the amino group on carbon 2, the hydroxyl group on carbon 3, and possibly the trans double bond found between carbons 4 and 5 of sphingosine (42). Both the amino and hydroxyl groups of sphingosine have been suggested to participate in intra-and intermolecular hydrogen bonding, which impart important hydrogen bonding potential to SPM that is not found in PC (1,42). The structure of SPM in the backbone region affects the packing of the acyl chains below and the orientation of the PC head group above (43), facilitating close lipid packing and a more condensed lipid organization (44). Inter-and intramolecular hydrogen bonds, as well as a compact lipid organization, would increase the order and restrict the mobility of TMA-DPH and explain the higher fluorescence polarization observed when SPM is present.
The effects of SPM on the head group region are also significant. The presence of SPM facilitates close lipid packing and condensed organization of the head group region due to its hydrogen bonding capacity. The results of 31 P NMR in egg PC/SPM vesicles suggest that intramolecular hydrogen bonds of SPM and close lipid packing may partially exclude water molecules from hydrating the phosphate group of PC, resulting in decreased hydration of the head group region (35). This would explain the relative blue shift of PRODAN fluorescence in rHDL with increased SPM content.
During the revision of this paper a study by Rye et al. (45) was published on the effects of SPM on the structure and function of spherical and discoidal rHDLs. Rye and co-workers demonstrated that SPM affects the lipid order and packing in these particles, in close agreement with our observations. They reported that SPM does not influence neutral lipid transfers involving spherical rHDL and cholesterol ester transfer protein; they also showed that SPM inhibits LCAT reaction with rHDL substrates. However, Rye et al. did not address the mechanism of LCAT inhibition by SPM, which is the main topic of this report.
In summary, we report that a SPM content up to 22 mol % does not alter the size of rHDL prepared with bulk egg PC, cholesterol, and apoA-I. LCAT binding affinity decreases as rHDL SPM content increases. The inhibition of LCAT by SPM at the active site has a minimal effect in the modulation of enzyme activity with these substrates. The results of our studies with lipophilic probes suggest that SPM significantly changes the properties of the phospholipid interface at the surface and also in the backbone and acyl chain regions. Furthermore, these changes in the surface properties of the rHDL correlate with decreased LCAT binding and reactivity.