Lipoprotein B37, a Naturally Occurring Lipoprotein Containing the Arnino-terminal Portion of Apolipoprotein B100, Does Not Bind to the Apolipoprot~in B,E(Low Density Lipoprotein) Receptor”

In 1979, Steinberg and colleagues described a unique kindred with familial hypobetalipoproteinemia (Steinberg, D., Grundy, S. M., Mok, H. Y. I., Turner, J. D., Weinstein, D. B., Brown, W. V., and Albers, J. J. (1979) J. Clin. Invest. 64, 292-301). Recently, we demonstrated the existence of an abnormal species of apolipoprotein (apo-) B, apo-B37 (Mr = 203,000) in nine members of that kindred (Young, S. G., Bertics, S. J., Curtiss, L. K., and Witztum, J. L. (1987) J. Clin. Invest. 79, 1831-1841; Young, S. G., Bertics, S. J., Curtiss, L. K., Dubois, B. W., and Witztum, J. L. (1987) J. Clin. Invest. 79, 1842-1851). Apolipoprotein B37 contains only the amino-terminal portion of apo-B100. In affected individuals most of the apo-B37 is contained in the high density lipoprotein (HDL) fraction (d = 1.063-1.21 g/ml), where it is the principal apolipoprotein in a unique lipoprotein (Lp) particle, Lp-B37, which contains little, if any, apo-A-I. However, the most abundant lipoprotein in the HDL density fraction is a smaller particle, which contains apo-A-I, but no apo-B. The Lp-B37 particles were isolated from the HDL of affected individuals by immunoabsorption of apo-B37. Selected affinity antibodies specific for apo-B37 were used to prepare an anti-apo-B37-Sepharose 4B column. Lipoproteins not bound by the column (unbound HDL fraction) contained apo-A-I, but no apo-B. The Lp-B37, which was eluted from the column with 3 M KI, contained apo-B37 and trace amounts of apo-A-I, but no apo-B100. Over a 4-h period, normal human fibroblasts degraded 10-fold more 125I-low density lipoprotein (LDL) than 125I-Lp-B37. Also, whereas addition of excess unlabeled LDL markedly reduced degradation of 125I-LDL, it did not significantly reduce the degradation of 125I-Lp-B37. Unlabeled Lp-B37 did not inhibit uptake and degradation of 125I-LDL by fibroblasts. These data suggest that the amino-terminal portion of apo-B100, when expressed on a naturally occurring lipoprotein particle, does not contain a functional apo-B,E(LDL) receptor binding domain.

J . Clin. Invest. 79, 1842-1851). Apolipoprotein B37 contains only the amino-terminal portion of apo-Bl00. In affected individuals most of the apo-B37 is contained in the high density lipoprotein (HDL) fraction (d = 1.063-1.21 g/mI), where it is the principal apolipoprotein in a unique lipoprotein (Lp) particle, Lp-B37, which contains little, if any, apo-A-I. However, the most abundant lipoprotein in the HDL density fraction is a smaller particle, which contains apo-A-I, but no apo-B.
The Lp-B37 particles were isolated from the HDL of affected individuals by irnmunoabsorption of apo-B37. Selected affinity antibodies specific for apo-B37 were used to prepare an anti-apo-B37-Sepharose 4B column. Lipoproteins not bound by the column (unbound HDL fraction) contained apo-A-I, but no apo-B. The Lp-B37, which was eluted from the column with 3 M KI, contained apo-B37 and trace amounts of apo-A-I, but no apo-BlOO. Over a 4-h period, normal human fibroblasts degraded 10-fold more *261-low density lipoprotein (LDL) than '251-Lp-B37. Also, whereas addition of excess unlabeled LDL markedly reduced degradation of lZ51-LDL, it did not significantly reduce the degradation of I2'1-Lp-B37. Unlabeled Lp-B37 did not inhibit uptake and degradation of '"I-LDL by fibro-* This work was supported by Grant HL-14197 (Arteriosclerosis Specialized Center for Research in Atherosclerosis) from United States Public Health Service, a merit review grant award from the Veterans Administration, and by a grant-in-aid from the American Heart Association, California Affiliate. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a clinical investigator award from the National Heart, Lung, and Blood Institute. To whom reprint requests should be addressed.
11 Recipient of a summer student fellowship award from the American Heart Association, California Affiliate.
$$ Established investigator of the American Heart Association. blasts. These data suggest that the amino-terminalportion of apo-B100, when expressed on a naturally occurring lipoprotein particle, does not contain a functional apo-B,E(LDL) receptor binding domain.
Apolipoprote~n (apo-)z Bl00 plays a central role in human lipoprotein metabolism. It is the predominant protein in human low density lipoproteins (LDL) and is the ligand responsible for the receptor-mediated uptake of LDL by cells (1,2). Because of the large size of apo-B100, its insolubility, and its propensity to break down into a number of smaller polypeptides, there has been, until recently, little information on its structure and functional domains. However, within the past year, the complete nucleotide sequence of the apo-B cDNA has been determined by four independent groups (3-6). The derived amino acid sequence of apo-B will be invaluable in understanding apo-B structure and function. An important issue will be to localize the apo-B,E(LDL) receptor binding domain(s) of apo-€3 as precisely as possible.
Several clues about the probable nature of the apo-B receptor binding domain already exist. The structure and functional domains of apo-E, the other ligand for the apo-B,E(LDL) receptor, have been extensively studied (2,7,8). The receptor binding domain of apo-E is centered near amino acid residues 140-150 and is rich in lysine and arginine residues. Chemical modification of the lysine and arginine residues of apo-E abolishes its ability to bind to the apo-B,E(LDL) receptor (9,lO).
Chemical modification of the lysine and arginine residues of apo-B also disrupts LDL binding, suggesting that these positively charged amino acids may play a critical role in the receptor binding domain of apo-Bl00 (10,11). Two basic amino acid sequences in the carboxyl-terminal region of apo-B (residues 3147-3157 and 3359-3367) have been proposed as candidates for the receptor binding domain of apo-B (3, 12). The second of these sequences has partial homology to the receptor binding domain sequence of apo-E.
Studies with apo-B-specific monoclona~ antibodies support the idea that these two carboxyl-terminal amino acid sequences, or at least sequences close by, may actually be important for receptor binding. Some apo-B-specific antibodies are capable of blocking the binding of LDL to the apo-B,E(LDL) receptor, whereas other antibodies specific for apo-B do not block binding (13)(14)(15)(16). The epitopes for several of the receptor-blocking antibodies have been localized to sequences flanking the two basic amino acid regions (3). For example, antibody MB47, which was developed and characterized in our laboratory (15), has been shown to bind to apo-B within residues 3350-3505 (3). The epitopes of two other receptor-blocking antibodies, 4G3 and 3F5 (13,14), bind within residues 3037-3132 (3). Recently, it has been shown that the two basic regions may be brought closer to one another as a result of a disulfide bridge between residues 3167 and 3297 (3).
In spite of strong immunochemical evidence localizing the receptor binding domain to the carboxyl-terminal region of apo-B100, uncertainty about t.he nature and location of the receptor binding domain has persisted. Protter et al. (17) have pointed out the existence of sequences rich in basic amino acids in the amino-terminal region of apo-Bl00; the functional significance of these sequences, if any, has not been determined. Recently, Hospattankar et a1. (18) proposed that 12 widely dispersed regions of apo-B100, including some aminoterminal sequences, may be involved in receptor binding.
Localization of the receptor binding domain of apo-E was greatly aided by a detailed m u~t i d i s c i p l i n~y examination of clinically important mutations in apo-E (2,7,8). Similarly, examination of mutations in apo-B will likely be rewarding in localizing the receptor binding domain of apo-B100. In this study, the binding characteristics of a lipoprotein containing a mutant apo-B were examined the results yield insights into the location of the receptor binding domain of apo-B100.
In 1979, Steinberg et al. (19) described a unique kindred with familial hypobetalipoproteinemia. In a further evaluation of this kindred, we found that the lipoproteins of nine family members with hypobetalipoproteinemia had an abnormal species of apo-B, apo-B37 (20, 21). Apolipoprotein B37 contains only the amino-terminal portion of apo-B100 (20). In affected individuals, apo-B37 is found in the high density lipoprotein (HDL) density fraction in a unique lipoprotein, Lp-B37.
These particles contain apo-B37, but no apo-B100, and little, if any, apo-A-I. They constitute only a small fraction of lipoproteins found in the HDL density range.
In this study, we successfully purified Lp-B37 particles from the HDL density fraction of several affected family members.
The ability of '"I-Lp-B37 particles t o be degraded by cultured human fibroblasts and the ability of unlabeled Lp-B37 particles to compete with intact lZ5I-LDL for cellular uptake and d e~a d a t i o n was tested. No significant uptake of Lp-B37 particles was observed. Thus, the amino-terminal portion of apo-B100, a t least as expressed in Lp-B37, a naturally occurring lipoprotein particle, does not appear to contain a physiologically important receptor binding domain.

Human Subjects
The H. J. B. kindred was originally described by Steinberg et at. (19). Recently, Young and co-workers (20,21) further examined the kindred and found evidence for two abnormal apo-B alleles, one encoding for a truncated apo-B species, apo-B37, and a second associated with low plasma concentrations of apo-B100. A family tree, in which each family member is identified by a number and characterized according to apo-B genotype, has been published (21).
Fresh plasma samples were obtained from selected family members whose lipoproteins contained apo-B37: subjects 1 (H. J. B.), 3, 13, and 14, as previously identified (21). Blood was collected into tubes containing EDTA (1.5 mg/ml of blood), and plasma was immediately isolated by centrifugation at 4 "C. Mult~ple proteolytic inhibitors were then added to the plasma, as previously described (20). Control plasma was obtained from normal laboratory personnel.

Lipoprotein Isolation and Characterization
Low density lipoproteins (d = 1.025-1.063 g/ml) and HDL (d = 1.070-1.21 g/ml) were isolated from plasma samples under sterile conditions by standard ultracentrifugation techniques (22). The lipoprotein fractions were then dialyzed extensively against phosphatebuffered saline (PBS (0.154 M NaCI, 21 mM NaZHPO,, 15 mM NaH2P04, 0.3 mM EDTA, pH 7.35)). The protein concentration of the lipoprotein fractions was determined by a modification of the Lowry technique using a bovine serum albumin standard (23).
Polyclonal antisera to apo-B37 were developed in rabbits (20). The apo-B37 used for immunization was purified from the HDL of subject 1 by preparative sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (29). On analytical SDS-polyacrylamide gel electrophoresis, the immunogen contained apo-B37, but no apo-E or apo-A-I, By Western blot analysis, the resultant antisera bound to apo-B37, apo-B100, and apo-B48, but not to apo-E or apo-A-I (20).

Product~on and Use of the An~i-lipoprotein B37
lmmunoaffinity Column The techniques for immunoabsorption of Lp-B37 from HDL are similar to those previously reported by McVicar et al. (30) for im-munoabso~tion of apo-A-I from plasma. Forty milligrams of LDL protein were covalently bound to 5 g of CNBr-activated Sepharose 4B (Pharmacia Biotechnology, Inc.) according to manufacturer's instructions, and a 6 X 2-cm LDL-Sepharose column was prepared. After the column was washed with PBS, 10 ml of anti-apo-B37 rabbit serum was circulated slowly (0.1 ml/min) over the column for 14 h at 4 "C; the column was then washed with PBS until no more protein could be removed, as determined by a UV monitor at 280 nm. Then 1 M acetic acid, pH 2.8, was passed over the column at 1.0 mI/min, and elution of antibody was monitored at 280 nm. The eluted antibody was immediately neutralized to pH 7.35 with 0.2 M sodium phosphate, dialyzed against PBS, and concentrated on an Amicon Centriflow membrane cone (type CF25; Amicon Corp., Danvers, MA). The average yield of antibody per 10 ml of antiserum was 1.5 mg.
Six milligrams of purified antibody was bound to 2 g of CNBractivated Sepharose 4B, according to manufacturer's instructions, and a 3 X 2-cm anti-apo-B37 column was prepared. The column was washed extensively with PBS.
To isolate Lp-B37 particles, 10 mg of HDL from an affected family member was circulated over the column for 14 h at 4 "C. Then PBS was pumped over the column at 1.0 ml/min, and lipoproteins not bound to the column (unbound HDL fraction) were collected. Phosphate-buffered saline was pumped over the column for 30-60 min after protein could no longer be detected by the UV monitor. The Lp-B37 particles then were eluted from the column with 3 M KI. The eluate was immediately transferred to a dialysis bag and dialyzed against multiple changes of PBS for 48 h.
Characterization of the LpB37 Preparation Chemical Composition-The protein concentration of Lp-B37 was determined by a modification of the Lowry technique (23). Phospholipid concentration was determined using a micromodification of the Bartlett procedure (31). Total cholesterol concentration and triglyceride concentration were determined enzymatically.
Apolipoprotein Content-The apolipoprotein content of Lp-B37 was assessed by SDS-polyacrylamide slab gels stained with Coomassie Brilliant Blue (24). Identity of individual apolipoproteins was con-Brmed by Western blot analysis with monospecific antibodies as previously described (24,27).
The apo-B content of Lp-B37 was assessed by solid-phase competitive radioimmunoassay (RIA) using apo-B-specific monoclonat antibodies or the polyclonal antibody to apo-B37.
This assay was performed exactly as previously described (15,26). Briefly, flexible 96-well plates were coated with PBS containing 10 pg of control LDL/ml. The remaining binding sites were blocked by coating with PBS containing 4% bovine serum albumin. Control LDL was utilized for the standard curve, which was included on each plate. The control LDL and other lipoprotein competitors (including Lp-B37) were diluted in PBS containing 3% bovine serum albumin, 0.02% sodium azide, and 0.05% Tween 20 to concentrations ranging from 0.23 to 123 pg/ml. The control LDL and other lipoprotein competitors (25 p l ) were added to the LDL-coated wells, followed by 25 ~1 of a fixed and limiting amount of antibody. After an 18-h incubation at 4 "C, the plates were washed and the amount of the first antibody bound to the immobilized LDL was quantitated by addition of the lZ5Ilabeled second antibody. Competition curves were plotted as BIB, uersus log of the protein concentration of competitor added, where B and Bo are specific counts bound in the presence and absence of competitor, respectively. The concentration of apo-A-I in the Lp-B37 preparations was determined with an immunoenzymometric assay utilizing monoclonal antibodies specific for apo-A-I.' Lipoprotein Sizing-The size distribution of Lp-B37 particles was determined by gradient polyacrylamide gel electrophoresis under nondenaturing conditions using commercially available gels (PAA 4/ 30 gels, Pharmacia Biotechnology, Inc.), as previously described (20,32). Gels were stained with silver or were electrophoretically blotted onto nitrocellulose sheets for Western blot analysis (20).
The size distribution of lipoprotein particles was also assessed by electron microscopy. Aliquots of lipoproteins (100 pg/ml) were allowed to adhere to a carbon film formed by evaporation of carbon onto a freshly cleaved mica surface. The floating film with adherent lipoproteins was picked up on a grid and rinsed with several drops of 1% sodium phosphotungstate (pH 7.0) to which 0.1% sucrose had been added to enhance the spreading of the stain. After removal of the excess stain and drying, the grids were viewed and photographed. Micrographs printed at 100,000 X magnification were used for particle sizing. The size of 65 free-standing lipoprotein particles was determined by two independent observers using a calibrated magnifying glass.
Normal human foreskin fibroblasts were grown in 22-or 35-mm wells in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum for 3-4 days. The cells were then washed, and the medium was changed to DMEM containing 10 mg of lipoproteindepleted serum (LPDS)/ml. After 48 h in DMEM-LPDS, when the cells had reached 85% confluency, the uptake and degradation experiments were performed.
Low density lipoproteins (d = 1.019-1.063) isolated from pooled normal plasma and Lp-B37 isolated from the HDL density fraction of subject 1 were iodinated using the IODO-GEN technique (36). The labeled lipoproteins were then dialyzed for 48 h against multiple changes of PBS containing 0.01% EDTA. The labeled lipoproteins, in addition to any unlabeled lipoprotein competitors, were adjusted to 0.5-1.5 ml with DMEM-LPDS and then dialyzed for 24 h against two changes of DMEM. The individually dialyzed media were then placed on the fibroblast monolayers at 37 "C for 4-7 h. Following the incubation, the media were removed and analyzed for lZ5I-labeled lipoprotein degradation products. This was performed essentially as described (15,(33)(34)(35). Trichloroacetic acid and silver nitrate were added to the media to concentrations of 10 and 1.7%, respectively. The trichloroacetic acid and silver nitrate were added to precipitate undegraded lipoproteins and free lZ5I, respectively. The media were then spun at 2500 X g, and an aliquot of the trichloroacetic acidsoluble silver nitrate-soluble supernatant fluid was counted in the y counter. At every concentration of labeled lipoprotein, the amount of trichloroacetic acid-soluble silver nitrate-soluble radioactivity present in the media after a parallel incubation in wells containing no cells was also determined, and this amount was subtracted from the amount present in the media of the fibroblast wells. The data were expressed as micrograms of LDL protein degraded/mg of cell protein.
Following removal of the media, the fibroblast wells were immediately placed on ice and were washed extensively, exactly as described *D. Hogle, R. S. Smith, and L. K. Curtiss . (35). The cells were then solubilized in 1 ml of 0.1 M NaOH. One aliquot was counted for radioactivity, and another aliquot was used for determination of total cell protein. The "cellassociated'. radioactivity represents the sum of bound and internalized lipoproteins present in the pellet at the end of the incubation period. The amount of cell-associated radioactivity was expressed as micrograms of cell-associated lipoprotein/mg of cell protein.

RESULTS
Isolation of Lp-B37 by Anti-apolipoprotein B37 Columns-In previous papers (20,21), we documented that HDL density fraction (d = 1.063-1.21 g/ml) of affected subjects contained two types of particles: those containing apo-B37, which we termed Lp-B37, and normal HDL particles containing apo-A-I. To purify Lp-B37 particles, an immunoaffinity column was prepared using antibodies directed against apo-B37. The apo-B37-specific antibodies were first selected for their ability to bind to and be readily eluted from an LDL-Sepharose column. The anti-apo-B37 antibodies were eluted from the LDL column with 1 M acetic acid, pH 2.8. These antibodies were then used to prepare an anti-apo-B37-Sepharose column.
T h e whole HDL fraction from one of the affected subjects was then circulated over the column. High density lipoprotein particles not bound by the column (the unbound HDL fraction) contained no apo-B37 (Fig. l), indicating that the column was effective in removing the Lp-B37 particles from the whole HDL preparation. Our initial experiments indicated that 1 M acetic acid, pH 2.8, was ineffective in eluting all Lp-B37 particles from the anti-apo-B37 column. When 3 M KI was used to elute the column, the yield of Lp-B37 particles was far greater. After dialysis and concentration of the sample o n a n Amicon filter, we obtained yields of 300 fig of Lp-B37 protein for each 10 mg of HDL loaded onto the column.
Characterization of Purified Lp-B37"The Lp-B37 preparations utilized in the cell culture experiments described below were intensively characterized. We determined the protein, cholesterol, and phospholipid concentrations of normal HDL, HDL of subject 1, and the unbound HDL fraction and Lp-B37 preparations isolated from the HDL fraction of subject 1. The phospholipid/protein and cholesterol/protein ratios of these lipoproteins are summarized in Table I. Triglycerides accounted for less than 5% of Lp-B37 mass. T h e chemical composition of Lp-B37 particles was similar to that of normal HDL.
When the purified Lp-B37 preparations were subjected to SDS-polyacrylamide gel electrophoresis and the gel was stained with Coomassie Brilliant Blue, a dense apo-B37 band a n d a fainter apo-A-I band were consistently observed. No apo-E was observed (data not shown). Western blots of the SDS-polyacrylamide gels, using specific monoclonal antibodies, were used to identify the apolipoproteins in the Lp-B37 preparations. The results of the Western blots of purified Lp-B37 from subject 14 are shown in Fig. 1. A Western blot using antibody MB19 demonstrated a dense apo-B37 band. A faint apo-B26 band was observed, but no other breakdown products were seen (Fig. 1, panel A ) . A Western blot using an apo-A-1specific antibody demonstrated a small amount of apo-A-I in the Lp-B37 preparations (Fig. 1, panel  B ) . No apo-E was detectable in the Lp-B37 preparation (Fig. 1, panel C). Western blots were also performed with the Lp-B37 preparations from subjects 1, 3, and 13, and similar results were obtained.
However, a small amount of apo-E was detectable in t h e Lp-B37 preparations isolated from the HDL of subjects 3 and 13.
Consequently, the Lp-B37 preparations from these two subjects were not used in the cell culture experiments described below.
The ability of the purified Lp-B37 preparation to compete

~~
with LDL for binding to apo-R-specific antibodies was assessed in solid-phase RIAs. Fig. 2 A illustrates the results obtained with the RIA using antibody MB11. In this assay the lipoprotein competitors were added on the basis of their protein content. Lp-B37 competed with the immobilized LDL for binding to the antibody MI311 (an amino-terminal-specific apo-B monoclonal antibody) slightly better than did native LDL. A likely reason for this finding is that, compared with LDL, there is a higher molar concentration of Lp-I337 particles when they are added to the assay at the same protein concentration. No competition was observed with the unbound HDL fraction, indicating that the anti-apo-R37 column had effectively removed nearly all the apo-B from the HDL. Assessment of the apo-A-I content of the Lp-B37 preparations in a sensitive immunoassay showed that apo-A-I constituted, at most, 13% of the Lp-B37 protein. Purified Lp-B37 did not bind to antibody MB47, indicating t,hat the purified Lp-B37 did not contain any apo-B100 (Fig. 2B).
The purified Lp-B37 particles had the same size as Lp-B37 particles contained within HDL, as determined by gradient polyacrylamide gel electrophoresis under nondenaturing conditions (Fig. 3). There was no evidence for aggregation of the Lp-B37 preparation on the nondenaturing gels. On 1% agarose gels, the purified Lp-B3'i particles had essentially the same pre-/3 mobility as did Lp-B37 particles in the HDL density fraction (data not shown).
Electron microscopy of the purified Lp-B37 particles demonstrated that they were larger than the apo-A-I-containing particles in the unbound HDL fraction. Representative elec- petitor was added to the assay on the hasis of its protein content. R and Bo, specific counts hound in the presence or absence of competitor, respectively. Similar results were ohtained with HDL, LpH37, and unbound HDL fraction preparations from subjects 3 and 14.
Also, similar results were obtained in other assays using other apo-R-specific antibodies (MB3, MR19) that hind to the amino-terminal portion of apo-€3100. =F' 1 2.

4.
FIG Ability of Lp-H37 to Hind to thc Apolipoprotein H,E(LIlIJ Receptor of Cultured Fibroblasts-In a preliminary experi-ment, we tested the ability of unlabeled control HDL and unlabeled H D L from subject 1 to compete with ""I-LDL for degradation by fibroblasts (Fig. 5 ) . Even at a very high concentration of competitor, HDL from subject 1 competed no better than control HDL. Our RIA results (Fig. 2) suggested that Lp-R3i constitutes -55 of the HDL mass. Thus, at the 1000 pg/ml point, there was -50 pg of Lp-RX/ml. However, whole HDL from an affected subject, even at 1000 pg/ml, failed to compete with ""I-LDI, any better than control HDL. In contrast, very low concentrations of unlabeled LDL competed quite well with "'I-LDL for degradation by fibroblasts.
Next, purified Lp-BX preparations were tested for their ability to compete with '""ILDL for uptake and degradation by cultured fibroblasts. Three different Lp-R,?i preparations were tested in t,hree different experiments (Fig.  6). Fig. 6 includes the results obtained by analyzing the ""I-labeled degradation products in the media (strippd bars) as well as the '""I-LDL cell-associated radioactivity (ushitc bars). Both the degradation products and cell-associated radioactivity yielded the same results. Even at high competitor concentrations, Lp-RXi particles failed to compete with ""I-LDL for uptake and degradation. In fact, they did not compete any better than whole HDL or the unbound HDL fraction, which contained very little, if any, apo-R. These data strongly support the idea that apo-RX'i does not contain an LDL receptor binding domain.
Further evidence that. Lp-RX'i did not bind to the LDI, receptor of fibroblasts was obtained wit.h an Lp-R3i preparation that was radioiodinated and tested directly for uptake and degradation by cultured fibroblasts. Fig. i shows the results of the degradation of '""I-Lp-R37 and '"'1-LDL by fibroblasts. Whereas degradation of LDL showed the usual high-affinity receptor-mediated curve, degradation of "'I-Lp-    14 were used as competitors. Control fibroblast wells contained no competitor and degraded 1.7 pg of lz5I-LDL/mg of cell protein over a 6.5-b incubation. In each of the three experiments, the amount of fibroblast protein per well was 80-120 pg. All determinations were performed in duplicate. Specific activity of lZ5I-LDL ranged from 130 to 515 cpm/ng. tion. Also, whereas addition of excess unlabeled LDL (500 pg/ ml) to the culture medium dramatically reduced the degradation of the lZ5I-LDL, addition of excess LDL to the culture medium reduced lZ51-Lp-B37 binding by only 10%. Thus, there was little, if any, specific degradation of Lp-B37 by the LDL receptor pathway. Analysis of the cell-associated radioactivity data yielded similar conclusions. At each concentration of labeled lipoprotein, the specific cell-associated radioactivity for lz5I-Lp-B37 was, at most, 15% that of lZ5I-LDL. It is important to note that the labeled lipoproteins were added to the cell culture media on the basis of their protein concentration. At comparable protein concentrations, the molar concentration of Lp-B37 particles was much higher than the concentration of LDL particles. In spite of the higher molar concentration of particles, there was much less uptake and degradation of Lp-B37.
The Lp-B37 used in these cell culture experiments had been eluted from an immunoaffinity column with a chaotropic agent, 3 M KI, whereas the control LDL had not been subjected to 3 M KI. However, it is unlikely that the brief exposure of the Lp-B37 particles to 3 M KI destroyed a physiologically important binding domain on the Lp-B37 particles because incubation of control LDL with 3 M KI for 2 h did not significantly affect its ability to be taken up and degraded by fibroblasts (data not shown).

DISCUSSION
In 1979, Steinberg and colleagues reported a unique kindred with familial hypobetalipoproteinemia (19). Further evalua-tion of that kindred provided evidence for two abnormal apo-B alleles associated with hypobetalipoproteinemia (20, 21).
One allele was associated with extremely low plasma concentrations of apo-B100. The other abnormal allele yielded an abnormal species of apo-B, apo-B37 (Mr = 203,000). Previously, we demonstrated that apo-B37 contains only the amino-terminal portion of apo-Bl00 (i.e. apo-B37 is a truncated version of apo-Bl00). Based on its apparent molecular weight, we estimate that apo-B37 must end near amino acid residue 1700. Recently, we have shown that polyclonal antibodies to a synthetic apo-B peptide (residues 2008-2024) bind to apo-B48 and apo-B100,3 but not to ap0-B37.~ In affected individuals, apo-B37 is the principal apolipoprotein in a unique lipoprotein particle, Lp-B37, which is present in the HDL fraction. In the present study, the Lp-B37 particles were purified from HDL by using an anti-apo-B37 immunoaffinity column. The approach and techniques were very  or lz5I-Lp-B37 (0) purified from the HDL of subject 1. When 500 pg of unlabeled LDL/ml was added to the medium at each concentration of "'I-LDL, there was a marked decrease in degradation of "'I-LDL (U). When 500 pg of unlabeled LDL/ml was added to the medium at each concentration of lz5I-Lp-B37, there was essentially no change in the degradation of "'I-Lp-B37 (0). All determinations were performed in duplicate. The amount of fibroblast protein/well was 80-100 pg. Results with cell-associated also indicated little specific uptake of lZ5I-Lp-B37, compared with "'1-LDL (see "Results"). naturing gradient polyacrylamide gels and agarose gels was identical to that of Lp-B37 particles in the whole HDL fraction. There was no evidence that the purification scheme caused aggregation of the particles. Apolipoprotein B37 was not degraded during the purification process (Fig. 1).
Each of the experiments with cultured fibroblasts suggested that Lp-B37 had little ability to be taken up by the apo-B,E(LDL) receptor. Lp-B37 competed with lz5I-LDL for uptake by fibroblasts no better than the unbound HDL fraction, which contained no apo-B. The very low levels of competition observed with the unbound HDL fraction and with Lp-B37 ( Fig. 6) may have been nonspecific or may have been due to the extremely low concentrations of apo-E in the unlabeled lipoprotein competitors.
Recently, Corsini et al. (38) prepared recombinant lipoprotein particles containing large thrombolytic fragments of apo-B100. Thrombin cleaves LDL-apo-B100 into three fragments: T4 (residues 1-1297), T3 (residues 1298-3249), and T2 (residues 3250-4536) (3, 12,20). Denatured thrombolytic fragments were purified from preparative SDS-polyacrylamide gels and used to form recombinant particles. Recombinant particles containing each of the three thrombin fragments bound specifically to the apo-B,E(LDL) receptor. Corsini et al. (38) recommended caution in interpreting these studies as demonstrating physiologically significant binding domains in multiple regions of apo-B100. They pointed out that the series of steps involved in making the recombinant particles (i.e. SDS denaturation of the thrombolytic fragments) could potentially unmask sequences capable of binding to the apo-B,E(LDL) receptor that were not normally exposed on native LDL particles. In our studies, we examined the binding of a naturally occurring lipoprotein particle containing the amino-terminal portion of apo-B100 and found no specific binding to the apo-B,E(LDL) receptor.
While providing no direct evidence localizing the apo-B,E(LDL) receptor binding domain to the carboxyl-terminal region of apo-B100, our data are quite consistent with immunochemical evidence suggesting that the receptor binding domain of apo-B100 is contained in the carboxyl-terminal region of apo-B100 near the T3-T2 cleavage point (3,12,20). Our data make the possibility of a receptor binding domain in the amino-terminal portion of apo-B100 seem unlikely.
Our previous studies with monoclonal antibodies have suggested that apo-B48 contains only the amino-terminal portion of apo-B100 (20). Studies with polyclonal antibodies to synthetic apo-B peptides have confirmed this finding.3 Hui et al.
(39) reported in vivo metabolic studies and in vitro cell culture experiments demonstrating that apo-B48 has no role in the receptor-mediated uptake of triglyceride-rich lipoproteins, suggesting that apo-B48 does not contain a receptor binding domain. The disadvantage of using the apo-B48-containing particles is that they are large triglyceride-rich lipoproteins with significant amounts of other apolipoproteins, including apo-E. In contrast, Lp-B37 particles are cholesterol-rich, triglyceride-poor particles with very low levels of apo-E. Both the Hui et al. study and this study are consistent in suggesting that there is no physiologically important LDL receptor binding domain in the amino-terminal portion of apo-B100.
We believe that studies of other apo-B mutations will yield new insights about the location and the structure of the receptor binding domain of apo-B100. Recently, Vega and Grundy (40) reported the existence of hypercholesterolemic patients whose LDL are cleared from plasma more slowly than LDL isolated from a normal subject. It seems likely that some of these patients may have a defect in the receptor binding domain of their apo-B100. Detailed study of such patients should yield additional insights into the receptor binding domain of apo-B100.