Formation of phospholipid-rich HDL: a model for square-packing lipoprotein particles found in interstitial fluid and in abetalipoproteinemic plasma

386

~j~)chjmicoet Bjo~~v.~icuArfu 834 (1985) 386-395 Elseviel

BBA 53907

Fo~ation of phospholipid-rich HDL: a model for square-packing lipoprotein particles found in interstitial fluid and in abetalipoproteinemic plasma Trudy M. Forte a-*, C. Lurning Ren b, Robert W. N~rdhausen a and Alex V. Nichols a aDonner Laboratory, l-213, Lawrence Berkelq laboratory, University of California, Berkeiqv, CA 94720, and ’ Prirzker School of Medicine, University of Chicago, Chicago, IL 60615 ( U. .S.A. 1 (Received (Revised manuscript

Key words:

Abetaiipoproteinemic

HDL;

October 16th. 1984) received January 22nd, 1985)

Dimyristoylphosphatidyicholine; structure

HDL:

Electron

microscopy:

Lipoprotein

The major bovine HDL subfraction, fraction I-HDL, was incubated with increasing amounts of dimy~stoyI~osp~tidylcboline (DMPC). HDL size, as determined by gradient gel electropboresis and electrou microscopy, increased with increasing HDtphospholipid to DMPC mole ratios. Control fraction I-HDL were spherical, hexagonally-packing particles with a peak on gradient gel electrophoresis at 12.3 rt 0.1 nm; at a ratio of 1: 0.5, larger, mainly spherical particles with a peak at 12.9 t_ 0.08 nm were formed. At a ratio of 1: 1, occasional square-shaped particles were seen by electron microscopy; by gradient gel analysis, the mean diameter of the HDL-product increased to 13.7 t 0.1 nm. At the 1: 2 ratio, extensive domains of squarepacking particles were noted; the major size peak of this product was 14.6 it 0.08 nm. In ah incubations with DMPC, a smai19.4 f 0.08 nm product was formcxl; it was most pronounced at the 1: 2 ratio. The large, less dense particles generated by incubation contained apolipoprotein A-I and small molecular weight proteins. The 9.4 nm product contained only apolipoprotein A-I. The less dense product formed during incubation at the 1: 2 ratio had a decreased protein-to-lipid ratio relative to control HDL and a tfold increase in percent phospholipid. At a 1: 2 ratio, ineo~ra~on of DMPC into fraction I-I-IDL results in the loss of one molecule of apohpoprotein A-I; the resultant particle is a stable phospholipid-rich and protein-poor HDL which has a square-packing geometry. These phospholipid-laden HDL are morphologicaily similar to lipoproteins isolated from interstitial fluid or from plasma of a~~i~proteinemic patients. Our data suggest that the unusual morphologicaI properties of the latter bioIo~~lly formed particles may be due to increases in the polar lipid contents, and concomitant decreases in surface protein.

Introduction The morphology of lipoproteins as seen by negative staining electron microscopy frequently depends on the lipid composition of the particle. Plasma high-density lipoproteins (HDL) normally are spherical particles rich in phospholipid and * To whom correspondence O#S-27~~/85/$03.30

should

be addressed.

0 1985 Elsevier Science Publishers

cholesteryl esters [ 11; isolated subfractions are highly uniform in size and thus have a propensity to hexagonally close-pack upon negative staining [2]. In the absence of the cholesteryl ester core, HDL have been shown to assume a discoidal shape [3,4]. These particles were first isolated from patients with lecithin : cholesterol acyltransferase deficiency [S] where the particles contained only polar lipids, phospholipid and free cholesterol, in

B.V. (Biomedical

Division)

387

addition to protein. In vitro studies with HDL apolipoproteins and their constitutive lipids, however, indicated that the discoidal morphology of HDL is determined by phospholipid [6-81. In abetalipoproteine~a, yet another unusual lipoprotein structure, variously described as square-packing or rectangular, has been reported [1,9]. These particles appear in the d I 1.063 g/ml fraction of the patients’ plasma and contain apolipoproteins A-I, A-II, and E; hence, they are lipid-enriched HDL. Square-packing lipoproteins containing apolipoproteins A-I and E have also recently been isolated from the d I 1.063 g/ml fraction of sheep lung lymph [lo] and the d 1.063-1.085 g/ml fraction of human interstitial fluid [ll]; hence, these particles may be physiologically significant entities generated during interaction of HDL with cells and tissues. The percent phospholipid and unesterified cholesterol in the d I 1.063 g/ml fraction from abetalipoproteine~c plasma is elevated relative to normal plasma HDL [12]; the sheep lung lymph fraction of d s 1.063 g/ml also shows increased polar lipid content when compared with plasma [lo]. Previous studies in our laboratory have shown that human HDL,, incubated with phospholipid generated some particles which assumed rectangular shapes [ 131; biologically occurring rectangular HDL structures have also been found in human umbilical cord blood HDL,, [14]. These observations suggest that increased phospholipid uptake by HDL may account for the unusual rectangular and/or square-packing particles seen in abetalipoproteinemic plasma and in interstitial fluid. To test this hypothesis we have developed an in vitro incubation system consisting of bovine HDL and dimyristoylphosphatidylcholine (DMPC) which produces the unusual square-packing HDL found in lymph and plasma. The characteristics of these particles are described in the present report. Methods isolation of i~~o~ro~ein~.Fresh steer blood in EDTA (1 mg/ml) was collected from a local slaughterhouse and was immediately placed on ice. Plasma was separated from cells by centrifuging at 1000 x g for 20 min, and gentamicin sulfate (0.1 mg/ml) was added. HDL were isolated by adjust-

ing the plasma to d 1.063 g/ml, centrifuging at 100000 x g (40.3 rotor) for 24 h at 15°C and removing the d I 1.063 g/ml material by pipetting. The infranatant was adjusted to d 1.21 g/ml and centrifuged for 48 h, and the top 1 ml containing the HDL was removed. Steer HDL was then subjected to density-gradient centrifugation, according to the method of Anderson et al. [2] in order to obtain the HDL,, fraction which is the major HDL subpopulation in the steer. Vesicle preparation. Dimyristoylphosphatidylcholine (DMPC) (Sigma, St. Louis, MO) vesicles were prepared by sonication of 5 mg DMPC/ml in 0.13 M ammonium acetate buffer (pH 7.4) containing EDTA (345 PM) and gentamicin sulfate (0.1 mg/ml). Sonication for 4-5 cycles of 5 min each were carried out on a Sonifier Cell Disurptor 350 (Branson, Sonic Power Co., Danbury, CT). The sonicator was tuned for maximum energy output using a microtip. Samples were kept under nitrogen in a 20-23°C water bath. The sonicated dispersion was centrifuged (16000 x g, 40 min) at 4°C to remove titanium particles and aggregated phospholipid. The resultant DMPC vesicles were unilamellar structures (38.2 + 4.6 nm diameter) as determined by electron microscopy. ~ncabatia~ procedure. HDL fractions were dialyzed against 0.13 M ammonium acetate (pH 7.4) containing 345 PM EDTA and 0.1 mg/ml gentamicin sulfate. DMPC was added to HDL to achieve HDL-phospholipid to DMPC molar ratios of 1: 0,5,1: 1, and 1: 2 (a molecular weight of 750 was used to calculate HDL-phospholipid). Incubations were carried out in a total volume of 2 ml at 37°C for 12-16 h in a shaking-water bath. Following incubation, samples were placed on ice and examined immediately by electron microscopy and gradient gel el~trophoresis. ~~ectrap~aresis. HDL fractions were analyzed by gradient gel electrophoresis to determine size distribution of the incubated products. Precast 4-30% polyacrylamide slab gels (Pharmacia Fine Chemicals, Piscataway, NJ) were used to analyze HDL fractions according to the procedure described by Blanche et al. [15]. Reference proteins used to determine particle diameters and molecular weights consisted of thyroglobulin, apoferritin, catalase, lactate dehydrogenase and bovine serum albumin. Gels were stained with

388

Coomassie G250 to identify protein bands. 10% polyacrylamide gel electrophoresis in the presence of 0.1% sodium dodecyl sulfate was used to determine molecular weights of apolipoproteins; the procedures used were essentially those of Weber and Osborn (161. Protein standards consisting of bovine serum albumin (M, 66000) apolipoprotein A-I (M, 28 000), apolipoprotein E (M, 34 000), and egg lysozyme (M, 14 000) were co-electrophoresed for determinations of molecular weights. All SDS polyacrylamide gels were stained with Coomassie R250. Densitometric scans of gradient gels and polyacrylamide tube gels were obtained with a Transidyne RFT densitometer. Gels were scanned at 603 nm wavelength. Chemical analyses. Protein and phospholipid concentrations were obtained by the method of Markwell et al. [17] and Bartlett [18], respectively. Unesterified cholesterol and cholesteryl ester were determined by gas-liquid chromatography according to Hindriks et al. [19]. Triacylglycerol was measured with the enzymatic reagent kit from Gilford Diagnostics (Oberlin, OH). Electron microscopy. Samples were diluted to a final concentration of approx. 300 pg protein/ml for negative staining. This concentration favors close packing of particles for evaluation of lipoprotein morphological characteristics. The procedures used for negative staining and sizing lipoprotein particles were those previously described [lo]. A minimum of 200 free-standing particles per sample was measured. Results Gradient gel electrophoresis of steer HDL (Fig. 1A) shows two distinct classes of particles: a major class with a peak at 11.9 nm and a minor one with a peak at 9.4 nm. These two classes were separated by equilibrium density gradient centrifugation as demonstrated in Fig. 1B and C. The major HDL class, termed fraction I-HDL, has a density of d 5 1.097 g/ml and corresponds with the 11.9 nm peak in the parent sample. The minor class, termed fraction II-HDL, was isolated at d 2 1.131 g/ml and corresponds with the 9.4 nm peak seen in the parent sample. These two fractions were incubated with DMPC at varying concentrations, and the

Fig. 1. Size distribution of bovine HDL on 4-30s gradtent gels. Inset shows gradient gel stained with Coomassie G250; protein standards used to calibrate the gel include: thyroglobulin, Thy; apoferritin, Fer; catalase, Cat; lactic dehydrogenase, Ldh; albumin, Alb. (A) Parent HDL (d 1.063-1.21 g/ml) shows two distinct peaks: one of large diameter, 11.9 nm, and one of smaller diameter, 9.4 nm. These two components were separated into two fractions by density gradient centrifugation: (B) fraction I-HDL which represents the large-sized particle seen in total HDL, (C) fraction II-HDL which represents the small-sized particles. The numbers over the peaks represent particle diameters in nm. 8 pg HDL protein were electrophoresed per lane.

products formed were evaluated by gradient gel electrophoresis and electron microscopy. Fraction I-HDL incubated with increasing amounts of DMPC showed a concentration-dependent shift in HDL size (Fig. 2 and Table I); particle diameter of fraction I-HDL increased from 12.5 nm (estimated M, 425000) in the control sample to 14.6 nm at the HDL-phospholipid/

389

12.5

A

Fer

Cat

’ Ldh

! Alb 0

0.2

0.4

0.6

0.6

1.0

Rf Fig. 2. Gradient gel (Coomassie stained gel in inset) and corresponding scans of fraction I-HDL incubated with increasing amounts of DMPC. (A) HDL-control; (B) HDL-phospholipid/DMPC, 1: 0.5; (C) HDL-phospholipid/DMPC, 1 : 1; (D) HDL-phospholipid/DMPC, 1: 2. In the presence of increasing DMPC, fraction I-HDL peak (12.5 nm for control) shifts to larger diameters; concomitantly, a product of smaller diameter is also generated. The numbers over the peaks indicate particle diameter in mn. Standards used to calibrate the gels are the same as those in Fig. 1. 8 pg HDL protein were applied to each lane.

DMPC ratio of 1 : 2 (estimated M, 567000). The shift of fraction I-HDL to larger particles was paralleled by the appearance of a minor component with a diameter of 9.4 nm (Fig. 2B-D). As indicated in Fig. 2 and Table I, the mean diameter of this smaller particle remained constant with increasing DMPC concentration, although the area under this peak increased with increasing DMPC ratio. The large- and small-sized products generated by DMPC incubation were separated by den-

A

B

C

Fig. 3. Gradient gel electrophoresis patterns of the two major incubation products isolated from fraction I-HDL by density gradient centrifugation following incubation with DMPC at a ratio of 1 : 2. (A) Large molecular weight particle with d s 1.094 g/ml. (B) Small molecular weight fraction, d > 1.137 g/ml. (C) High molecular weight standards which are the same as those described in Fig. 1.

sity gradient centrifugation, similar to that used in the initial separation of fraction I- and fraction II-HDL. The gradient gel patterns of the two isolated products are shown in Fig. 3, A and B. The large diameter product (14.6 nm) accounted for 65% of the total protein and had a density of d I 1.094 g/ml. The more dense (d L 1.137 g/ml), small-sized product (9.4 nm) accounted for 19% of the total protein. The remaining protein was

Fig. 4. Electron micrographs of negatively stained fractions. (A) Fraction I-HDL incubated without DMPC. (B) Product of fraction I-HDL incubated with DMPC (ratio 1 :OS); note occasional rectangular or square particles (arrow). (C) Product of fraction I-HDL incubated with DMPC (ratio 1: 1); a few square-packing particles are apparent (arrow). (D) Reisolated large molecular weight product of fraction I-HDL incubated with DMPC (ratio 1: 2); note characteristic square-packing particles. (E) HDL particles of d s 1.063 g/ml isolated from the plasma of a patient with abetalipoproteinemia; particle size 12.8kO.7 nm on a side. Note similarity of square-packing geometry to&at produced in D. (F) Square-packing particles, d 1.057-1.063 g/ml fraction, isolated from sheep lung lymph; particte size 14.9 i 2.1 nm on a side. Bars represent 100 nm.

TABLE

I

EFFECT OF DMPC INCUBATION ON BOVINE FRACTION I-HDL PARTICLE SIZE DISTRIBUTION AS DETERMINED BY GRADIENT GEL ELECTROPHORESIS. Fraction I-HDL was incubated at 37°C for 12-16 h, with increasing amounts of DMPC. Values represent meanrtS.D. for four determinations. Mole ratio HDL-phospholipid/DMPC

Large particle

Small particle

(nm)

(nm)

0 1 : 0.5 I:1 1:2

12.3kO.l 12.9rtO.08 13.7rf:O.l 14.6 I 0.08

9.5 i 0.1 9.4 + 0.08 9.4kO.98

located primarily within the d 1.098-1.102 g/ml fraction. Electron micrographs of the large diameter particles resulting from incubation of fraction I-HDL with increasing amounts of DMPC are seen in Fig. 4. Control particles (Fig. 4A) consist of spherical structures with a mean diameter of 12.7 f 1.0 nm which have a tendency to hexagonally pack. At an HDL-phospholipid/DMPC ratio of 1 : 0.5, the particles are mostly spherical (mean diameter 13.1 : 2.5 nm), but an occasional square or rectan-

Fig. 5. Electron micrograph of the negatively stained 9.4 nm product produced by incubation of fraction I-HDL with DMPC. Small discoidal structures are indicated by arrow. The fine granular structures in the background probably represent apolipoprotein A-I oligomers. Bar marker represents 100 nm.

gular profile is in evidence (Fig. 4B). At a ratio of 1 : 1, the particles are still predominantly round (Fig. 4C), although, occasionally, square-packing structures are present. At a ratio of 1: 2, large areas of square-pacing particles can be seen (Fig. 4D); the size of the square profiles is 13.5 & 1.2 nm on a side. Square-packing particles are the hallmark of this incubation product, but spherical particles are also present; the latter are 14.1 & 1.0 nm in diameter. For comparison, Figs. 4E and 4F show negatively stained, d 5 1.063 g/ml HDL particles from a patient with abetalipoproteinemia, and sheep lung lymph particles of d i 1.063 g/ml, respectively. These particles formed in vivo have packing characteristics very similar to those of fraction I-HDL incubated with DMPC at a ratio of 1 : 2. The 9.4 nm product generated during incubation is morphologically heterogeneous (Fig. 5) and contains small-disc-shaped structures (10.8 f 1.9 nm long axis and 4.4 f 0.3 nm short axis) and what appears to be aggregated protein; the latter may be associated with small amounts of lipid. Fraction II-HDL isolated from the parent HDL was also incubated with increasing amounts of DMPC; however, at a ratio of 1: 2, the major peak on gradient gel electrophoresis shifted only slightly (from 9.3 to 9.7 nm). Exa~nation of the incubation products by electron microscopy showed that no square-packing particles were formed, even when the ratio was increased to 1: 6; therefore, fraction II - HDL was not used in subsequent studies. The protein and lipid compositions of fraction I-HDL incubated without DMPC and the two major products isolated after incubation with DMPC at a 1 : 2 ratio were determined in order to ascertain whether the changes in HDL electron microscopic characteristics were accompanied by compositional changes. The compositional data is summarized in Table II. Fraction I-HDL incubated without DMPC consists of 29% protein and 71% lipid; on the other hand, the square-packing product generated by incubation of fraction I-HDL with DMPC contained 15% protein and 85% lipid. The most distinctive compositional change in this DMPC-enriched product is the large percentage of phospholipid (54% vs. 24% in fraction I-HDL). The more dense, 9.4 nm product generated during

392 TABLE

II

COMPOSITION

OF HDL

PRODUCTS

FOLLOWING

INCUBATION

The composition represents the mean and SD. of four animals. animal. Results are expressed as % total wt.

Fraction I - HDL control Fraction I-HDL square-packing Fraction I-HDL

product

I-HDL WITH DMPC (I: 2 RATIO)

I more dense product.

which is from a single

Protein

Phospholipid

Unesterified cholesterol

Cholesteryl ester

Triacylglycerol

28.9 + 2.2

26.2 + 3.0

5.3 IO.4

38.9+3.1

0.9 + 0.1

15.3*

54.4 + 2.3

3.4 & 0.4

25.9 + 2.6

2.3 f 0.7

1.7

87.0

more dense, 9.4 nm product Fraction II-HDL control

OF FRACTION

except for the fraction

8.1

54.7 & 4.2

15Xi+o.2

incubation is extremely rich in protein (87%) and contains phospholipid as its chief lipid moiety. Although this product has the same size as frac-

____._.

Alb aP0 E aP0 Al

LYS

0.2 1.4+0X

0.5 X.1+5.1

4.2 2.3 f 0.7

tion II-HDL, it is wholly dissimilar from the latter, which contains a cholesteryl ester-rich core. The unusual composition of the 9.4 nm product, however, is consistent with its electron microscopic image, which showed aggregated protein and small discoidal structures. Fig. 6 shows the distribution of apolipoproteins in fraction I-EDL incubated in the presence and absence of the 1 : 2 ratio of DMPC. Based on densitometry and assuming equal chromogenicity, the major apolipoproteins in the bovine fraction I-HDL without DMPC were apolipoprotein A-I (67%) and small molecular weight proteins (33%). The square-packing product formed during incubation with the 1 : 2 ratio of DMPC showed a redistribution of apolipoproteins where apolipoprotein A-I decreased to 56% and small molecular weight proteins increased to 44%. This decrease in apohpoprotein A-I can be accounted for by the loss of apolipoprotein A-I into the more dense, 9.4 nm product. The protein moiety of the latter incubation product (Fig. 6C) contains primarily apolipoprotein A-I. Discussion

A

B

C

D

Fig. 6. Apolipoprotein distribution on 10% SDS-polyacrylamide gel electrophoresis (A) Control. (B) Square-packing, less dense d d 1.094 g/ml product formed during incubation of fraction I-HDL with DMPC, ratio 1: 2. (C) More dense, d 2 1.137 g/ml product formed during phosphoiipid incubation. (D) Standards: albu~n, alb; apo~poprotein E, apo E, apotipoprotein A-I, apo AI; lysozyme, lys, HDL protein (30 pg) was applied to each gel.

It has been documented by several laboratories that the plasmas of abetalipoproteinemic patients possess a chemically and morphologically unique, less dense, HDL particle ]1,9,12,20]. Upon negative staining electron microscopy, such particles in these patients form square-packing arrays and, occasionally, rectangular structures. Kostner et al. f12] have isolated less dense HDL in the d 2 1.063

393

g/ml region of abetalipoproteine~c patients and have shown that the particles are enriched in phospholipid and unesterified cholesterol, as compared with normal HDL. The abetahpoproteinemic HDL of d s 1.063 g/ml are known to contain the major apolipoproteins (apolipoproteins A-I, A-II, E and C’s) normally associated with plasma HDL of d 1.063-1.21 g/ml [9,12,20]. It has recently been reported that human interstitial fluid contains a less dense HDL (d 1.063-1.085 g/ml) subclass which forms extensive square-packing arrays similar to those of abetalipoproteinemia and contains apolipoprotein AI and apolipoprotein E [ll]. A lipoprotein particle of d I 1.063 g/ml isolated from sheep lung lymph has also been shown to possess square-packing particles upon electron microscopy [lo]. The latter density fraction is enriched in phospholipid and unesterified cholesterol and also contains apolipoproteins A-I and E. In all instances where square-packing HDL have been noted, the particles characteristically are less dense than normal plasma HDL and frequently fall within the density range associated with lowdensity lipoproteins. The association of apolipoprotein A-I with such particles is an indicator that the particles constitute a less dense subclass of the HDL family. The increase in polar lipids in such particles and a decrease in protein content can also be considered characteristic. Abetalipoproteinemic particles in the d I 1.063 g/ml range are known to have, in addition to an increased phospholipid content, elevated sphingomyelin relative to phosphatidylcholine levels. No information is available, however, on the phospholipid composition of lung lymph or interstitial fluid. The presence of elevated sp~ngomyelin levels, however, is not a necessary prerequisite for the unusual electron microscopic properties of square-packing particles, since in our present study we can demonstrate the same phenomenon with bovine HDL enriched in DMPC. The present investigation shows that bovine HDL can readily incorporate DMPC into its structure. At low ratios of HDL-phospholipid to DMPC, phospholipid incorporation is associated with an increase in particle diameter, but particle morphology does not change; hence, the particle remains spherical. At a mole ratio of approx, 1: 2, there is a distinct change in the morphological

behavior of the particles, so that they have a tendency to form squ~e-pacing arrays. Based on gradient gel electrophoresis data, control fraction I-HDL has an estimated molecular weight of 425 000; according to its apolipoprotein composition, such a particle would contain three molecules of apolipoprotein A-I. The square-packing, phospholipid-enriched product, however, is considerably larger and has an estimated molecular weight of 567000. Based on apolipoprotein content, the product particle would contain only two apolipoprotein A4 molecules. The apparent reduction of apolipoprotein A-I per particle suggests that, during incubation, DMPC is incorporated into the particle until it reaches a critical phospholipid to protein ratio; when this ratio is exceeded, apolipoprotein A-I is displaced from the surface. The resulting phospholipid-rich, protein-poor particle, however, is a stable entity. Two factors may be responsible for the square-packing arrays of phospholipid-enriched HDL during negative staining: (I) The particles may be spherical, but the distribution and conformation of surface molecules, especially protein, are altered so that they are favorably disposed to foster inter-particle interactions during negative staining. Because of the nature of these particle-to-particle interactions, the particles are organized into square lattices; however, the nature of the interactions are as yet unknown. (2) Free-standing particles may be square and, at appropriate concentrations during negative staining, aggregate into square-packing arrays. Although free-standing square profiles were not seen in the present study, such structures have recently been noted in our laboratory on a specimen of human interstitial fluid HDL (work in progress); hence, one cannot completely rule out this possibility. The displacement of apolipoprotein A-I from the bovine fraction I-HDL during DMPC uptake is not unique to this species, for Nichols et al. [21] have reported that human HDL,, incubated with DMPC responds in the same manner. Based on our electron microscopic analysis, the apolipoprotein A-I dissociated from the surface of bovine fraction I-HDL forms discoidal structures and aggregated protein. It is reasonable to assume that the former structures are phospholipidapolipoprotein A-I complexes such as those de-

394

scribed previously [21,22]. The appearance of aggregated apolipoprotein A-I is not surprising, since Jonas and Krajnovich [23] found that bovine apolipoprotein A-I readily self-associates and forms ohgomers. The ability of HDL to incorporate large amounts of phospholipid into its surface suggests a for the formation of possible mechanism square-packing, apolipoprotein A-I- and apolipoprotein E-containing particles found in sheep lung lymph and human interstitial fluid. A fraction of normal HDL filters from the plasma into the interstitial space and bathes the surrounding cells and tissues. It has been proposed that HDL has the capacity for removing phospholipid and cholesterol from cell membranes [24-263. Although no direct evidence is available at present, it is conceivable that HDL in the interstitial fluid can incorporate enough polar lipids into its surface so that the electron microscopic structure is similar to that of our in vitro system of DMPC-enriched particles. It is also possible that apolipoprotein A-I is released from the HDL surface during uptake of phospholipid in the interstitial space and that the liberated apolipoprotein A-I can interact with phospholipids to produce discoidal structures. This could, in part, explain the appearance of smalf numbers of discoidal structures in sheep lung lymph and the more prevalent appearance of dislymph of coidal particles in the peripheral cholesterol-fed dogs [10,26,27]. Although wholly speculative, the square-packing, phospholipid-enriched and protein-poor HDL in the d < 1.063 g/ml fraction of abetalipoproteinemic plasma may represent lymph particles which have filtered back into the plasma compartment. Under normal conditions, these phospholipid-rich HDL are either rapidly removed through an apolipoprotein E receptor mechanism or are immediately converted to spherical, plasma HDL; however. in the case of abetalipoproteinemia, the altered metabolism of lipoproteins precludes catabolism or transformation, and these unusual particles persist in the plasma.

Mary Lou Olbrich for preparation of the manuscript. This work was supported by Program Project Grant HL-18574 from the National Heart, Lung, and Blood Institute, National Institutes of Health. C. Luming Ren was supported by a scholarship from the Association of Western Universities. References

4 5 6 7

a 9 10 11 12 13 14

19 20 21 22

Acknowledgements We wish to thank Janet Selmek-Halsey for her excellent technical assistance, and Linda Abe and

23 24

Forte, T.M. and Nichols, A.V. (1972) Adv. Lipid Res. 10. l-40 Anderson. D.W., Nichols, A.V.. Forte, T.M. and Lindgren, F.T. (1977) Biochim. Biophys. Acta 493, 55-68 Forte, T.M.. Nichols, A.V.. Glomset. J. and Norum. K. (1974) Stand. J. Clin. Lab. Invest. 33, suppl. 137, 123-132 Iiamilton. R.L., Williams, M.C., Fielding, C.J. and Havel, R.J. (1976) J. Clin. Invest. 58.667-680 Forte, T.M., Norum. K.R., Glomset, J.A. and Nichols. A.V. (1971) J. Clin. Invest. 50, 1141-1148 Forte, T.M., Nichols, A.V., Gong, EL. and Levy. R.I. (1971) Biochim. Biophys. Acta 248. 381-386 Forte, T.M., Gong, EL. and Nichols, A.V. (1974) Biochim. Biophys. Acta 337, 169-183 Kruski, A.W. and Scanu, A.M. (1974) Chem. Phys. Lipids 13, 27-48 Scanu, A.M.. Aggerbeck, L.P.. Kruski. A.W., Lim. C.T. and Kayden, H.J. (1974) J. Clin. Invest. 53, 440-453 Forte, T.M., Cross, C.E., Gunther. R.A. and Kramer. G.C. (1983) 1. Lipid Res. 24, 13581367 Forte. T.M., Reichi, D.R., Hong, J.L. and Ruda, D.N. f 1984) Arteriosclerosis 4, 564a Kostner,G., Holasek, A., Bohlmann, H.G. and Thiede. H. (1974) Clin. Sci. Mol. Med. 46, 457-468 Nichols, A.V., Gong. EL.. Blanche, P.J. and Forte. T.M. (1980) Biochim. Biophys. Acts 617. 480-488 Forte, T.M.. Davis, P.A., Nordhausen. R.W. and Giueck. C.J. (1982) Artery 10. 223-236 Blanche, P.J.. Gong, EL., Forte. T.M. and Nichols. A.V. (1981) Biochim. Biophys. Acta 665, 408-419 Weber, K. and @born, M. (1969) J. Biol. Chem. 244, 440664412 Markwell, M.A.K., Haas, SM., Bieber, L.L. and Tolbert, N.E. (1978) Anal. Biochem. 87. 206-210 Bartlett, G.R. (1959) J. Biol. Chem. 234. 466-468 Hindriks, F.R., Wolthers, B.C. and Groen. A. (1977) Ciin. Chim. Acta 74, 207-215 Blum, C.B., Deckelbaum, R.J., Witte, L.D., Tall, A.R. and Cornicelli. J. (1982) J. Clin. Invest. 70. 115771169 Nichols, A.V., Gong, E.L., Forte, T.M. and Blanche. P.J. (19781 Lipids 13.943-950 Tall, A.R., Small, D.M.. Deckelbaum, R.J. and Shipley. G.G. (1977) J. Biol. Chem. 252.4701-4711 Jonas, A. and Krajnovich, D.J. (1977) J. Biol. Chem. 252. 2194-2199 Glomset. J.A. and Norum. K.R. (1973) Adv. Lipid Res. II. l-65

25 Stein, 0. and Stein, Y. (1973) B&him. Biophys. Acta 326, 232-244 26 Sloop, C.H., Dory, L.. Hamilton, R., Krause, B.R. and Roheim. P.S. (1983) J. Lipid Res. 24. 1429-1440

27 Sloop, C.H., Dory, L., Krause, B.R., Castle, C. and Roheim, P.S. (1983) Atherosclerosis 49, 9-21