Small high density lipoprotein subclasses: some of their physico-chemical properties and stability in solution

Small high density lipoproteins (SHDL) contribute to the protection from atherosclerosis, but detailed information about their properties is not available yet. We isolated four of the smallest HDL subclasses that contain apoA-I alone, the small lipoprotein A-I (SLpAI), by their separation on gradient polyacrylamide gel followed by electroelution. Their physico-chemical properties were calculated from their displacement in non-denaturing gradient polyacrylamide gel under the effect of electrical potential. The properties are: Stokes' radii 2.96-3.56 nm; molecular masses 42-70 kDa; net negative charge 7.2-13.5; surface charge densities 3139-4069 -esu.cm(-2); surface potentials 10.6-15.7 -mV; coefficients of friction 5.74-6.90 x 10(-8) g.s(-1); and diffusion coefficients 5.76-6.94 x 10(-7) cm(2).s(-1). We found that these particles were of low stability as they underwent molecular transformation into larger particles on storage. The estimated dimensions of these particles do not support ellipsoidal shape, therefore, the most probable shape is spherical; consequently, their hydrated characteristics were estimated. We conclude that these particles have high values of negative surface charge and diffusion coefficients, and are of low stability. Their small Stokes' radii were similar to each other and they are spherical and highly hydrated.

Small high density lipoprotein (SHDL) particles are important because they act as acceptors of excess cholesterol from the peripheral body cells and deliver it to the liver in a process known as "reverse cholesterol transport" (Glomset, 1968).This process contributes to the prevention and regression of atherosclerosis where such particles should be of small size (Barter & Connor, 1976) and contain ApoA-I, an activator of the enzyme lecithin : cholesterol acyltransferase (LCAT; Fielding & Fielding, 1980;1981); moreover, their charge characteristics are expected to affect their role (Braschi et al., 1999) as well as their clearance rate from circulation (Desrumaux et al., 1998).Recently, five discrete types of SHDL of molecular mass of 43-70 kDa were detected and quantitated in fresh plasma (Atmeh, 1990;Atmeh & Robenek, 1996).The smallest three of these subclasses were shown to be of physiological R.F.Atmeh and A.O. Issa be determined.This will also aid in the design of protocols for their individual isolation and purification; a target that has not been achieved yet.
In this study we report the isolation, as a group, of four native discrete SHDL subclasses that contain apoA-I as the sole protein, the small lipoprotein A-I (SLpAI) group.To determine the net charge and other physicochemical properties of these particles we utilized their time-dependent migration in non-denaturing gradient polyacrylamide gel (gPAG) under the effect of electrical field.Moreover, we report here the low stability of these particles in vitro and present a prediction of the shape of the smallest particle.

MATERIALS AND METHODS
Subjects.Apparently healthy non-smoker subjects were studied.Due to the relatively low level of the SHDLs in the blood of fasting healthy subjects who do not drink alcohol, and in order to get reasonable amounts of the SHDLs for the study, blood was withdrawn 4 h aĞer a breakfast meal rich in fat (Atmeh & Robenek, 1996).As not all of the SHDL subclasses appear in the plasma of every subject, we analyzed individual samples from many subjects and choose 8 samples for each subclass and their average and S.E.M. were calculated.Venous blood was withdrawn into EDTA-containing tubes and plasma was isolated promptly by low-speed centrifugation at 4 o C and gentamycin (0.08 mg • mL -1 ), sodium azide (0.1 mg • mL -1 ), and chloramphenicol (0.08 mg • mL -1 ) were added immediately.
Nomenclature.In previous work (Atmeh, 1990;Atmeh & Robenek, 1996) five small HDL types of particles were reported with the molecular mass range of 43-70 kDa, and were shown to be present in both fresh plasma and in the isolated HDL fraction.The smallest three of them with molecular masses of 43-50 kDa were designated as SHDL (Atmeh & Robenek, 1996).The fourth and fiĞh types had the same molecular mass of 70 kDa, and apoA-I was the sole apolipoprotein in one type, and both apoA-I and apoA-II were present in the other type with a molar ratio of 1:1 (Atmeh, 1990).For convenience, it would be reasonable to consider these types of particles as the small lipoprotein A (SLpA) group, and the particles that contain apoA-I alone as SLpAI and give the abbreviations SLp(AI+AII), SLpAI 1 , SLpAI 2 , SLpAI 3 , SLpAI 4 for the individual types as detailed in Table 1.
Preparation of HDL and lipoprotein fraction of d < 1.22 g • mL -1 .HDL (d = 1.063-1.22g • mL -1 ) was prepared by sequential ultracentrifugation at 10 o C according to Havel et al. (1955) in a Beckman model L8-70M ultracentrifuge using a Beckman 50.3 Ti rotor (Beckman, Palo Alto, CA, USA).Similarly, lipoprotein fraction of d < 1.22 g • mL -1 was prepared by a single spin of plasma overnight.Both fractions were dialyzed against 0.15 M NaCl solution containing 0.01% EDTA for several hours.Both HDL and the lipoprotein fraction of d < 1.22 g • mL -1 were passed through an affinity chromatography column containing anti-albumin to remove any traces of albumin.
Nondenaturing gradient polyacrylamide gel electrophoresis (gPAGE).A 4-25% gPAG was prepared and run as was described previously (Atmeh, 1990;Atmeh & Abuharfeil, 1993).In brief, the gPAG was cast and run in an LKB 2001 vertical electrophoresis system (LKB, Sweden).An appropriate amount of lipoprotein was applied to each well of the gPAG and run at 15 o C in Tris (14 mM) and glycine (110 mM) buffer, pH 8.3.The voltage gradient applied was 2542 V • m -1 .Stokes' radius (R s ) and relative molecular mass (M r ) were calculated in comparison with molecular weight markers (Sigma, St. Louis, MO, USA) of known Stokes' radius and molecular mass which were run in the same gel.After the run was complete, either the gel was stained and destained or treated as in the following section.
Electroelution of SLpA from the gradient gel.Freshly isolated lipoprotein fraction of d < 1.22 g • mL -1 or HDL (300 μg protein) was separated on 4-25% gPAG and run as described in the above section.AĞer the run was complete, the gel was cut longitudinally and the lane that contained the marker proteins was stained with Coomassie Brilliant Blue R-250 for a short time and the rest of the lanes were compared with the stained one in order to locate the position of individual lipoproteins.The gel portion that contained the required bands of molecular mass equal to or less than 70 kDa (SLpA) was cut and placed in an electroelution apparatus developed by Atmeh (1989) and electroeluted as described elsewhere (Atmeh et al., 1991).The eluted SLpA fraction was tested by the double immunodiffusion technique using specific antibodies against albumin, apoC-I, apoC-II, and apoC-III.Since the subclasses SLp(AI-AII) and SLpAI 1 have the same molecular mass (Table 1) and migrate together (Atmeh, 1990), therefore, from now on we will confine our analysis to the SLpAI 1 particle in addition to the other three.

Determination of migration distance (D) and migration velocity (V).
The migration distance of each SLpAI type of particles in the 4-25% gPAG was measured at various time intervals (1, 2, 4, 8, 12, and 16 h).Samples were applied to gradient gels and run, as described above, for the specified time interval followed by staining and destaining.The migration distance in millimetres (D; mm) and migration velocity (V; mm • s -1 ) of each SLpAI particle in each sample were measured.The migration behaviour of each lipoprotein was utilized to determine its maximum migration distance (D max ) and maximum gel concentration (T max ).The following parameters were calculated: coefficient of friction (f), diffusion coefficient (D c ), apparent electrophoretic mobility (μ), free electrophoretic mobility (U), net negative charge (Z), density of surface charge (C d ), and surface potential (S).The mathematical treatment reported by Rothe (1991;1994) that described some of those parameters was used.Following are the net equations that were used.
Determination of maximum migration distance (D max ).Migration distances (D) of globular proteins in linear gPAG are not linearly correlated with time (t) of electrophoresis nor with the reciprocal of time (t -1 ), therefore, this relation does not follow a simple hyperbolic function.Moreover, such distance and time are not related to the other by a simple exponential function.A linear relationship was found (Rothe, 1991) by ploĴing the double logarithm of D, [ln(lnD)] versus the square root of t -1 , therefore, a straight line equation can be wriĴen as where a and b are the slope and intercept of the line, respectively, and can be determined from the plot of ln(lnD) vs. t -1/2 .When the migration time (t) approaches infinity, t -1/2 approaches zero and the migration distance will approach the maximum migration distance (D max ) of the lipoprotein, thus ln(lnD max ) = b, consequently, D max can be determined by extrapolating the straight line of Eqn 1 and measuring the intercept b.Determination of gel concentration (T) and maximum gel concentration (T max ).The gel concentration (T; %) and the migration distance (D; mm) are linearly related by the following Eqn (Rothe, 1991): where α is the slope of the line and β is the starting (the lowest) concentration of the gradient gel.The value of the constant α was obtained from the scan at 405 nm of a 4-25% linear gradient PAG contain-ing p-nitrophenol (Rothe, 1991).These gels were prepared by adding the yellow-coloured p-nitrophenol (0.075 mg • mL -1 ) to the concentrated (25%) acrylamide solution during the preparation of the gel.The formed gel was scanned at 405 nm from top to bottom with correction for the absorbance of the supporting glass plate.The y-axis of the scan (absorbance) was related to the gel concentration where the top of the gel corresponds to the lower concentration of the gradient (4%), and the boĴom of the gel corresponds to the higher concentration of the gradient (25%).The interval of the absorbance axis was divided into percentage concentration scale and for each point on the gel length (distance; x-axis), the corresponding T(%) value was obtained from the straight line of the scan.The slope of the line equals α and the intercept equals β, which is the lower gel concentration (4%).
From Eqn (2) D max can be wriĴen as Determination of the apparent free electrophoretic mobility (μ).At polyacrylamide gel concentration of zero (T = 0), migration of a macromolecule is unhindered by the gel matrix and can be considered as migration in solution, i.e., the migration velocity is equivalent to the apparent free electrophoretic mobility (μ; mm • s -1 ).The migration velocity is the rate of migration and equals: The relation between V and D is curvilinear and can be described by the following exponential equation (Rothe, 1991): where ε and δ are constants.Converting this equation into a logarithmic form we get a straight line equation: PloĴing lnV (m • m-1 ) vs. ln(D max -D), both δ and ε can be determined.Substituting the values of D and D max with the corresponding values of T and T max , respectively, from equations ( 2) and ( 3) and rearrangement, we get: where h = ε α -δ .At T = 0 we can write: thus μ can be calculated.
Determination of the free electrophoretic mobility (U).The free electrophoretic mobility is defined as the apparent free electrophoretic mobility (in meters per second) relative to the electrophoretic potential (E, voltage in volts/gel length in meters) applied, i.e.,

Calculation of the coefficient of friction (f) and diffusion coefficient (D c
).The coefficients of friction of SLpAI molecules were determined from the Stokes' equation: and their diffusion coefficients were determined from the Stokes-Einstein equation: where k B is Boltzmann constant, T absolute temperature, and η is the dynamic viscosity of the medium.
Calculation of the net negative charge (Z).The number of unit charge was estimated from Debye-Hückel equation: where e is the unit charge and equals 1.602 × 10 -19 Coulomb; F is a retardation factor which results from the electrophoretic effect where its value depends on R s and the composition and strength of the small ions of the buffer solution.The value of F can be calculated from the equation: where κ is related to the thickness of the ion layer around the macromolecule and equals the reciprocal of the radius of the ion cloud; X(κR s ) is the Henry's function.
Calculation of the density of surface charge (C d ) and surface potential (S).The density of surface charge in electrostatic units was calculated from the following relation (Sparks & Phillips, 1992): The surface potential was calculated from Henry's equation: where D is the solvent dielectric constant.
Calculation of the partial specific volume [v(bar)] of SLpAI particles.The partial specific volume of the SLpAI particles can be calculated from the partial specific volumes of their components.Unfortunately, we are not able yet to isolate the individual SLpAI particles in a pure form and sufficient amount necessary for compositional analysis, therefore, we estimated their v(bar) as follows.The molecular masses are 42, 45, 50, and 70 kDa and each particle comprises one apoA-I molecule (v(bar) 0.736 cm 3 • g -1 ; Vitello & Scanu, 1976) and the rest of the mass is expected to be phospholipids and small amount of cholesterol with traces of triglycerides, if present.However, we assumed that the lipid moiety is mainly the phospholipid palmitooleoylphosphatidylcholine (POPC).Our assumption was based on the following findings: 1) Dergunov et al. (2001) found that the dimensions of reconstituted HDL (rHDL) particles did not change significantly in the presence or absence of cholesterol; 2) the size of the phospholipid bilayers does not change significantly in the presence or absence of cholesterol (Schwartz & Paltauf, 1976;Tu et al., 1998) or if saturated or unsaturated faĴy acids are present in the phospholipid (Armen et al., 1998;Chiu et al., 1999); 3) in similar spherical small HDL particles, apoA-I and phospholipid components represent more than 90% of the particle volume (Chen et al., 1984).The POPC v(bar) value of 1.006 cm 3 • g -1 was calculated from its molecular volume (1263 nm 3 ) reported by Chiu et al (1990) using the formula reported by Tristram-Nagle et al. (1998).
Calculation of the lengths of the axes of SL-pAI particles as ellipsoids.The values of the long axis (2a) and the short axis (2b) as well as the axial ratio (a/b) of the ellipsoidal particles were predicted by the UltraScan soĞware (Borries Demeler, PhD, University of Texas Health Science Center at San Antonio, USA).The soĞware utilizes the linear Lamm equation (Lamm, 1929) model for the analysis of ultracentrifugation and other molecular parameters (Demeler & Saber, 1998).
Calculation of the radius of unhydrated spherical SLpAI particles (R o ).The volume of one mole of a spherical particle equals the product of its molecular mass by its partial specific volume, and the volume of one molecule is obtained on dividing the product by Avogadro's number (N), or calculated from its radius (R o ), i. Calculation of the hydration properties of SLpAI particles.The radius and size of hydrated SLpAI particles are larger than those of unhydrated ones due to the presence of the hydration layer, therefore the hydration factor (δ; g of water/g of lipoprotein) can be introduced into Eqn 17 which will be wriĴen as: where v o (bar) is the partial specific volume of water and approximately equals 1.
The difference between the radius of a hydrated (R s ) and that of an unhydrated spherical particle (R o ) was used to calculate the thickness of hydration water around each SLpAI particle.Dividing the resulted value by 0.28 nm (thickness of one water layer) gave the number of hydration layers around the particle.The number of bound water molecules around each particle was calculated from the product of δM r divided by 18 (the molar mass of water).
In vitro stability of SLpAI particles.To investigate the effect of storage on the stability of the isolated SLpA fraction, it was incubated with and without 5,5-dithiobis(2-nitrobenzoic acid) (an LCAT inhibitor), at 4 o C for up to ten days followed by separation on 4-25% gPAGE.

RESULTS
The density of the Tris/glycine buffer and its viscosity at 20 o C were 1.002 g • cm -3 and 0.01028 g • cm -1 • s -1 (Poise), as measured by a pycnometer and Ostwald viscometer, respectively.The electroeluted SLpAI fraction was free from albumin and apoC-I, apoC-II, and apoC-III when tested against the corresponding specific antibody by the double immunodiffusion method.The SLpAI subclasses are shown in Fig. 1 before and aĞer elution from gPAG.

Migration distance (D) and maximum migration distance (D max )
The freshly isolated SLpA fraction from each sample (average n = 8) was run in duplicate on 4-25% gPAG for various time intervals.The values of ln(lnD) were ploĴed against t -1/2 according to Eqn 1 and the value of D max for each type of particles was calculated from the intercept (Table 2).

Gel concentration (T), and maximum gel concentration (T max )
The slope (α) of the straight line of Eqn 2 was 0.181 as calculated from the scanned gPAG con-taining p-nitrophenol.The value of the constant (β) equals 4.0%, which is the lower concentration of the gradient gel.Thus Eqn 2 will be: The values of T max (%) for each SLpAI particle were calculated from the corresponding D max value (Table 2).

Migration velocity (V), apparent free electrophoretic mobility (μ) and free electrophoretic mobility (U)
The migration velocity for each type of particles at each time point was calculated from equation 4, and ploĴing the values of lnV versus ln(D max -D) (Eqn 6) enabled the calculation of the constants ε and δ which were used to calculate the value of the constant h.The values of μ and U were calculated from Eqns 8 and 9, respectively (Table 2).Lane 1: molecular mass standards; lane 2: total HDL; lane 3: SLpAI subclasses; HDL was run on 4-25% gPAGE, the region containing the SLpAI was cut out, eluted and rerun on another 4-25% gPAGE.

Table 2. Calculated physico-chemical properties of the SLpAI particles from their measured electrophoretic migration behaviour.
Electrophoresis was done in 4-25% gPAG for various time intervals and the voltage gradient used was 2542 V • m -1 .

Stokes' radius (R s ), molecular mass diffusion coefficient (D c ), and charge characteristics
The plot of lnR s vs lnD max for the marker proteins was linear, and from this plot the R s of the SLpAI subclasses were obtained and found to be in the range 2.96-3.56nm.The molecular mass was determined in a similar manner and was found in the range 42-70 kDa.The value of the diffusion coefficient (D c ) was calculated from Eqn 11 and was found to be in the range 5.76-6.94× 10 -7 cm 2 • s -1 .The quantities Z, Cd, and S were calculated from Eqns 12, 14, and 15, respectively.All these values are shown in Table 2.

Radius and volume of SLpAI particles
The partial specific volumes of SLpAI 1-4 were calculated as 0.840, 0.855, 0.838, and 0.826 cm 3 • g -1 , respectively (Table 3).These values are less than that reported by Patsch et al. (1980) for a high density lipoprotein fraction with molecular mass of 152 kDa designated as HDL i (0.961 cm 3 • g -1 ).From these values and the relative molecular masses of SLpAI 1-4 we predicted the lengths of their axes by the soĞware UltraScan (Table 4).The length of the shorter axis (2b) of the particles was 1.74, 1.54, 1.52, and 1.50 nm as oblate ellipsoids (discs).These values are much shorter than the thickness of a phospholipid bilayer (3.5 nm; Tristram-Nagle et al., 1998;BrouilleĴe et al., 2001;Rog et al., 2003) which excludes the possibility of the particles being ellipsoids.This renders the sphere as the probable shape of the particles.The ellipsoidal shape can be further excluded by calculating the circumferences of the assumed discoidal particles from their phospholipid content.The calculated circumferences of the SLpAI 1-4 particles were 14.83, 10.69, 9.49, and 8.55 nm, respectively (Table 5).These circumference lengths are much shorter than the length of one apoA-I helical molecule; this means that apoA-I will make more than one turn around these discs.Being spherical particles, their unhydrated radii were calculated from Eqn 17 from their corresponding M r and v(bar) values and were found to be 2.85, 2.56, 2.46, and 2.39 nm (Table 6).The unhydrated and hydrated volumes were calculated from Eqn 16 using the corresponding R o and R s values, respectively (Table 6).

Hydration properties of the SLpAI particles
The calculated thickness of the hydration shell around the SLpAI particles was in the range 0.71-0.57nm equivalent to 2.5-2.0 layers of water comprising 3100-1700 water molecules (Table 6).

In vitro stability of SLpA particles
When the isolated SLpA subclasses were incubated at 4 o C, in the presence or absence of an LCAT  inhibitor, for 1, 2, and 3 days followed by separation on a gradient gel, no detectable change regarding the number and position of bands was observed after their separation on the gradient gel.On the other hand, incubation at 4 o C, in the presence or absence of an LCAT inhibitor, for 10 days showed the presence of at least twelve bands of molecular masses ranging from 41 to 504 kDa (Table 7).Some samples showed a band of molecular mass around 28 kDa which is expected to be the lipid free apoA-I.

DISCUSSION
SLpAI particles are important because they are the initiators of the reverse cholesterol transport process (Miller, 1990;Chiesa et al., 1998;GilloĴe et al., 1998), but not much is known about their physico-chemical and metabolic properties.Therefore, further investigation of their properties contributes to the beĴer understanding of their individual metabolism and physiological function (Davidson et al., 1995;Colvin & Parks, 1999).Such physiological role is a function of small size (Oram et al., 1981;Dory et al., 1986), high negative charge (Desrumaux et al., 1989;Dory et al., 1986;Sparks et al., 1995;Saku et al., 1999), high diffusion (Castro & Fielding, 1988), and low thermodynamic stability (Sparks et al., 1999) of the particles.The physico-chemical properties of macromolecules are commonly determined from the study of their displacement in solution by the effect of an external gradient of a potential caused by either electrical (electrophoresis) or gravitational (sedimentation) forces (Giddings, 1991).Isolation of the individual SLpAI particles in reasonable purity and significant amounts has not been achieved yet, due to the minor differences in their size and molecular mass, which has hindered the determination of their physico-chemical properties.In order to overcome this problem we isolated the four SLpAI particles as a group and studied their displacement in gradient polyacrylamide gel by the effect of electrical potential, and the migration behaviour of each type of these particles was followed.The migration distances at various time intervals were measured and used for the determination of the maximum migration distance (D max ) that each particle can migrate in the gradient gel at time infinity.The D max value of the SLpAI particles increased with the decrease in particle size and ranged from 154.2 to 216.2 mm (Table 2).Since the migration is in a gradient gel, then D max depends on the particle dimensions that correspond to the gel pore size.From the knowledge of the gel concentration (T%) at each point a Lipid mass = molecular mass -28 (mass of one apoA-I molecule); b Mass of one POPC molecule = 0.76 kDa.The number of POPC molecules in each particle is obtained by dividing the mass of the particle by the mass of POPC molecule.c Number of POPC molecules in each phospholipid layer = ½ the number of POPC molecules in the particle.d Surface area of one POPC = 0.65 nm 2 (Murzyn et al., 2001;Koln et al., 2002).The surface area of the circular leaflet face of the particle = number of POPC molecules × 0.65.e Surface area of the circular leaflet face of the particle = πr 2 , therefore, r can be calculated.f Circumference of the discoidal particle = 2πr.along the gradient gel, a certain gel concentration value (T max ) was obtained for each particle (Table 2) at its corresponding D max value, where beyond this value it cannot migrate.The migration velocity for each type of particles was calculated from the corresponding value of T max (Eqn 7), and this velocity at zero gel concentration represents the apparent free electrophoretic mobility (μ) which was used to calculate the free electrophoretic mobility (U) (Table 2).
Comparing our measured U values for SLpA-I 1-4 with high density lipoproteins of larger size (HDL 3 and VHDL as reported by Sparks & Phillips, 1992) of Stokes' radii of 4.5 and 3.8 nm, respectively, we found that the free electrophoretic mobility of SLpA-I 1 (7.21 m 2 • s -1 • V -1 ) is larger than that of both HDL 3 and VHDL (5.9 and 6.2 m 2 • s -1 • V -1 , respectively), whereas the U values for SLpA-I 2-4 are smaller.This is in accordance with the values of the net charge (Z) of the particles which was -13.5 for SLpA-I 1 and -11.5 and -9.0 for HDL 3 and VHDL, respectively, and -9.0 to -7.2 for SLpA-I 2-4 .Similarly, the surface potential values of SLpA-I 1 (-13.5 mV) and SLpA-I 2-4 (-11.7 to -10.6 mV) were higher and lower, respectively, than those of HDL 3 and VHDL (-11.5 and -11.9 mV, respectively).However, the reported values of the surface charge of HDL 3 and VHDL (2220 and 1800 -esu • cm -2 ; Sparks & Phillips, 1992) were lower than our values for SLpA-I 1-4 (4069-3139 -esu • cm -2 ).Moreover, a comparison of the charge characteristics of SLpA-I 1-4 with those of pure apoA-I (U: 4.3 m 2 • s -1 • V -1 , Z: -3.7, S: -8.3 mV, and C d : 1800 -esu • cm -2 ; Sparks & Phillips, 1992) shows that these values of apoA-I are less than those of SLpA-I 1-4 (Table 2).This increase in the charge characteristics of SLpA-I 1-4 more than those of pure apoA-I can be explained by changes in apoA-I conformation that result from its binding with lipids to form lipoproteins, where different domains are exposed, and this exposure is a function of the size and lipid content of the particles (Petit et al., 1987;Pio et al., 1988;Marcovina et al., 1990;Collet et al., 1991;Fielding et al., 1994;Tricerri et al., 1998).In the light of these findings, the presence of several SLpA-I particles with varying physico-chemical properties can be easily understood.In accordance with that, Sparks et al. (1992a;1992b) showed that different HDL subspecies maintain particular apoA-I conformations that confer unique charge and structural characteristics on the particles.The presence of several discrete particles of SLpA in blood with varying physico-chemi-cal properties may imply different roles both in lipid transport and in other physiological processes, and we can speculate that they may have different origins, functions, and/or fates.Recently, similar small particles were proposed to have a role in the regulation of human placental lactogen during pregnancy (Handwerger et al., 1999).Moreover, the differences in their properties may imply an involvement as cholesterol acceptors in various efflux mechanisms or preferential removal of cholesterol from different types of tissue cells (Rothblat et al., 1999;Haynes et al., 2000).The relative importance of each efflux mechanism would be determined by the cell type, metabolic status, tissue location, and the nature of the extracellular cholesterol acceptor (Rothblat et al., 1999).The diffusion coefficients of the SLpAI particles were high, which is a requirement for these particles to approach the aqueous environment of the peripheral cells and come in close contact with them in order to initiate the process of cholesterol removal from these cells (Castro & Fielding, 1988).
HDL particles are either spherical or ellipsoidal in shape; and in order to predict the shape of the SLpAI particles we estimated their axial lengths as oblate ellipsoids.The estimated values did not produce reasonable dimensions for the ellipsoids, and the calculated circumferences were much shorter than the length of one apoA-I helix.This result looks logical in the light of the small number of lipid molecules in the particles.The spherical shape of such small particles is more probable and may be more stable especially if neutral lipids are present.Indeed, Jonas et al. (1990) reported that apoA-I in spherical rHDL is more stable than in a discoidal one, suggesting a more stable interaction of the αhelical segment of apoA-I with lipid in the spherical particles.The presence of neutral lipids in small HDL particles was reported by many workers.Atmeh (1990) reported the presence of neutral lipids in the 70 kDa SHDL subclass, and Chen et al. (1984) described a spherical particle of about 83 kDa that contained 4.5 mole % non-polar lipids.Moreover, Physico-chemical properties of small high density lipoprotein subclasses Nanjee and Brinton (2000) reported a spherical shape of small HDL particles of 40-60 kDa, and Jaspard et al. (1996) reported a spherical shape for the small particle pre-β 1 HDL that contained 9% of its total mass as non-polar lipids.Considering the spherical shape of SLpAI particles, we calculated their hydration properties and found that their frictional ratios were 1.27-1.24and the molecules were surrounded by 2.0-2.5 layers of hydration water (Table 6).A frictional ratio of more than one indicates either a non-spherical particle or a hydrated spherical particle.The first possibility was excluded as discussed above, therefore, the second possibility supports the spherical shape.Lipidation of apoA-I to form small recombinant HDL particles is associated with decreased thermodynamic stability that may contribute to the metabolic reactivity of similar particles.This enhances their ability to accept cholesterol from cells and subsequent conversion into more stable larger particles (Sparks et al., 1999).In agreement with that we found that the pure isolated SLpA particles were stable at 4 o C for three days, but they were not stable on storage for longer periods of time (10 days) and a large number of particles of increasing molecular mass were detected (Table 7).This interconversion occurred in the absence of the enzyme LCAT because the solution contained pure electroeluted SLpA particles; with that, the incubation was done in the presence or absence of 5,5'-dithiobis(2-nitrobenzoic acid), an LCAT inhibitor.In support of that, Chen et al. (1984) reported low stability of small HDL particles of 2.99 nm radius where they decomposed and formed larger and smaller particles.The large particles may have been formed by either self-association or fusion of different SLpA particles.In accordance with that, Lusa et al. (1996) reported HDL fusion that involved the formation of unstable intermediate HDL particles.Castro and Fielding (1988) explained the presence of a large HDL subclass (LpA-I pre-β2 ), of molecular mass of 325 kDa in an incubation mixture, as a result of fusion of smaller particle (LpA-I pre-β1 ) of molecular mass of 71 kDa.
For a beĴer understanding of the physiological function of apoA-I-containing lipoproteins, it is important to study the conformation of apoA-I in the native lipid-bound state rather than in rHDL particles.This requires the isolation of individual native discrete apoA-I-containing lipoprotein particles, a target that has not been achieved yet.This problem was addressed by Phillips et al. (1997) where they stated: "So far, the structure of apoA-I in native lipoproteins has not been amenable to study, as it exists in very heterogeneous particles and, possibly, in diverse conformations.Reconstituted HDL, of defined compositions and sizes, has provided the best opportunities to learn about the structure-function relationships of apoA-I".We here present a study of native discrete small HDL particles, and we are in the process of aĴempting their individual isolation.This will offer a good opportunity for the study of the structure and conformations of apoA-I in native, not reconstituted, lipoproteins.
e., volume of one molecule = M r v (bar) / N = 4π R o 3 /3 (16) Thus the value of R o can be wriĴen as: R o = {3 M r v (bar)/4πN} ⅓ (17) Calculation of the coefficient of friction of spherical unhydrated SLpAI particles (f o ), their frictional ratio (ff o -1 ), and their sedimentation coefficient (s).These values were calculated by the above mentioned UltraScan soĞware.

Table 7 . Molecular masses of SLpA transformation products.
SLpA were isolated by electroelution from 4-25% gradient polyacrylamide gel on which HDL was separated.The isolated SLpA fraction was incubated at 4 o C for 10 days followed by separation on similar gradient polyacrylamide gel.