- Email: [email protected]
Chapter 12 Lipoprotein genetics and molecular biology
A . M . Gorto, J r . (Ed.) Plasma Lipoproreins
‘91987 Elsevier Science Publishers B.V. (Biomedical Division)
359 CHAPTER 12
Lipoprotein genetics and molecular biology JAN L. BRESLOW The Rockefeller University, New York, NY, USA
1. Perspectives and summary Apolipoproteins are important structural constituents of lipoprotein particles and have been shown to participate in lipoprotein synthesis, secretion, processing and catabolism. These subjects have been recently reviewed [ l , 21. In the last 15 years, advances in protein chemistry techniques have allowed the identification, isolation, and characterization of at least eight apolipoproteins (Table 1). Protein sequencing techniques have been used to derive the primary amino acid sequence of the plasma form of six of these polypeptides. In the last few years, cDNA and genomic clones have been derived for each of the apolipoproteins. The DNA sequences combined with cell-free synthesis and tissue and organ culture studies have revealed the presence of apolipoprotein precursors containing NH2-terminal extensions, including prepeptides and in some cases propeptides (Table 2). In addition, the apolipoprotein genes have been mapped in the human genome. Finally, human mutations in the apolipoprotein genes have been identified at both the amino acid and the DNA level. Some of these have profound effects on lipoprotein metabolism and are associated with premature atherosclerosis. This chapter will review current knowledge of human apolipoprotein gene structure and genetic variation. For additional background and references, several other recent reviews should be consulted
[3-51.
2. ApoA-1 ApoA-I is the major protein constituent of high density lipoproteins (HDL). HDL particles are about 50% protein and 50% lipid, and apoA-I is 70% of HDL protein. HDL levels are inversely correlated with susceptibility to coronary artery disease and recently the same association has been demonstrated for apoA-I [6, 71. ApoA-I is abundant in plasma with a concentration of 1 .O to 1.2 mg/ml. ApoA-I is thought to participate by two mechanisms in the reverse transport of cholesterol from tissues to the liver for excretion. ApoA-I can promote cholesterol efflux from tissues.
TABLE 1 Apolipoproteins and their association with human diseases
APOprotein
Plasma concentration (mg/ml)
Isoelectric point (PI)
A-I
1.0- 1.2
A-11
Function
Association with clinical disorders
5.85 - 5.40a 28K
Activates LCAT
Tangier disease; apoA-I - apoCIII deficiency; atherosclerosis
0.3 - 0.5
5.0
8.5K
-
-
A-IV
0.16
5.5
46K
-
-
B-100
0.7 - I .O
-
550K
Receptor-mediated catabolism of LDL
Abetalipoproteinemia; normotriglyceridemic abetalipoproteinemia (B-100 deficiency); atherosclerosis
-
275K
Chylomicron production
7.5
6.5K
Activates (moder-
B-48
M.W.
-
CI
0.04 - 0.06
CI1
0.03 - 0.05
4.9
9K
Activates lipoprotein lipase
Familial Type I hyperlipoproteinemia
CIII
0.12 - 0.14
4.7 - 5.0b
9K
lnhibits catabolism of triglyceride-rich lipoproteins
ApoA-I -apoCIlI deficiency
E
0.025 - 0.050
6.0- 5.7'
34.2K
Receptor-mediated catabolism of apoE-containing lipoproteins
Familial Type 111 hyperlipoproteinemia
-
ately) LCAT
a The isoelectric points of apoA-I isoproteins are: apoA-I, = 5.85; apoA-I, = 5.74; apoA-1, = 5.65; apoA-IS = 5.52; apoA-16 = 5.40. The major plasma isoprotein is apoA-I,. The isoelectric points of individual apo CIII isoproteins are: apoCIII-0 = 5.0; apoCII1-1 = 4.85; apoC1II-2 = 4.65. The isoelectric points of individual apoE3 isoproteins are: apoE3 = 6.02; apo E3,-, = 5.89; a p 0 E 3 , - ~ = 5.78; apoE3,_, = 5.68. The isoelectric points of the common apoE variants are apoE2 = 5.89; apoE4 = 6.18.
361 TABLE 2 Human apolipoprotein amino acids residues ~
ApoA-I ApoA-I1 APOA-1V ApoB-48 ApoB-100 ApoCI ApoCII ApoCI I I ApoE n.p.
-
not present; un.
Prepeptide
Propeptide
Mature protein
18 18 20
6 5 n.p. un . un. n.p. n.p. n.p. n.p.
243 77
un . un.
26 22 20 18 -
316 un. un.
57 79 79 299
unknown.
ApoA-I also displays cofactor activity for the lecithin cholesterol acyltransferase (LCAT) enzyme, which is responsible for almost all plasma cholesterol esterification. This reaction is thought to play a role in transforming nascent HDL to mature HDL particles. In mammals, apoA-I synthesis is approximately equally divided between liver and small intestine, whereas in avians other major sites of synthesis have been identified (references for background material on each of the apolipoproteins are in reviews listed as references 1 and 2, except where specifically provided).
(a) ApoA-I cDNA ApoA-I cDNA clones have been obtained and their DNA sequences derived [8 - 141. From this information, apoA-I mRNA is thought to be 893 bp in length and includes a 5' untranslated region of 35 bp, a region coding for 267 amino acids of 801 bp, a termination codon, TGA, and a 3' untranslated region of 54 bp followed by a poly A tail. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoA-I mRNA in cell-free synthesis experiments, indicates translation initiation at the methionine 24 amino acids upstream of the NH2-terminus of the mature protein. The 18 NH2-terminal amino acid polypeptide, which can be cotranslationally cleaved by microsomal membranes, represents the apoA-I prepeptide. The 6 amino acid polypeptide adjacent to the NH2-terminus of the mature apoA-I is the propeptide and has the rather unusual sequence Arg-His-Phe-Trp-GlnGln [15 - 171. The propeptide is not cleaved intracellularly, but rather is present in secreted apoA-I [16-201. Thus, it is necessary to postulate the existence of a previously unsuspected protease activity in lymph and/or plasma required to cleave this hexapeptide and generate mature apoA-I. This converting protease presumably
plays a role in apoA-I processing and may be an important determinant of apoA-I and, thereby, HDL metabolism. Protease activity, which is inhibited by EDTA, has been demonstrated in human serum and on the surface of human endothelial cells and hepatoma cells [21, 221. The reported cDNA sequences specify an amino acid sequence for mature apoA-I of 243 amino acids, which is very similar to that derived previously by protein sequencing methods [23]. The only difference is at residue 34 where the protein sequence specified Gln and the cDNA sequence indicates Glu. It had been previously noted that the apoA-I amino acid sequence from residues 99 to 230 was composed of six tandem 22 amino acid repeats, and five of the six repeats begin with proline [24, 251. Examination of the DNA sequence in this region confirms this and shows a tandemly repeated DNA structure 66 bp in length [12]. This finding suggests that this portion of the apoA-I gene arose by intragenic duplications. When the six 66 bp repeats are aligned and a consensus nucleotide at each position of the repeat derived, the consensus sequence is 64 to 80% homologous with each of the repeats [12]. Translation of the consensus sequence reveals an interesting underlying protein structure for this region of apoA-I. As noted from the protein structure, proline, an alpha-helix breaker, occurs every 22 amino acids. The intervening amino acids, when placed in an Edmundsen wheel diagram [4, 261, specify an alpha-helix with a nonpolar and a polar face. This is the general character of the amphipathic alpha-helical configuration which is a common feature of the apolipoproteins [27]. It is thought that the nonpolar face interacts with the hydrophobic lipid core of the lipoprotein particle, whereas the polar face interacts with the aqueous plasma environment. In addition, the positively charged residues tend to cluster between the nonpolar and polar faces. The latter has been shown to be important in stabilizing the lipid protein association [28]. A recent, more sophisticated, computer analysis derived to specifically look for DNA repeats in apoA-I reveals that the basic structure is a 33 bp (1 1 codon) repeat and a model based on gene duplication and unequal crossing-over events has been derived [29].
(6) ApoA-I gene The apoA-I gene has been isolated and its DNA sequence derived [12- 14, 301. From the transcription initiation to the polyadenylation site, the gene is 1863 bp in length. A comparison of the sequence of the apoA-I gene with the cDNA reveals three introns (IVS). IVS-1 is 197 bp long and occurs in the 5 ’ untranslated region between bases 20 and 21 upstream of the codon for Met that initiates translation. IVS-2 is 186 bp long and interrupts the codon specifying amino acid - 10, which is in the apoA-I prepeptide. IVS-3 is 588 bp long and interrupts the codon specifying amino acid 43 of the mature protein. The intron locations indicate that apoA-I exons may code for functionally distinct regions of apoA-I. For instance, exon 2 contains most of the apoA-I prepeptide, exon 3 contains the propeptide and the NH2-terminal sequences, whereas exon 4 contains codons for the 200 amino acids
363
which comprise the COOH-terminal portion of the molecule. The latter includes the 6 6 bp tandem DNA repeats. The apoA-I gene transcription initiation site has been designated based on the length of several apoA-I cDNA clones [14]. Upstream of the proposed apoA-I transcription initiation site is a 7 bp long AT-rich region which may be the apoA-I promoter, 'TATA box' [12- 14, 301. Another feature of the apoA-I gene is the presence of an Alu repetitive element approximately 1000 bp 3 ' to the gene [31].
(c) ApoA-I genetic variation ApoA-I is the principal structural protein in HDL. Because of the importance of HDL levels in predicting atherosclerosis susceptibility, extensive population screening for apoA-I structural variants, principally by isoelectric focusing, has been undertaken. Thus far, these studies have resulted in the discovery of at least 1 1 variants (Table 3). These variants have been shown in people who appear to be heterozygotes for one normal apoA-I structural allele and another allele that specifies a gene product that is either one charge unit more acidic or one or two charge units more basic than wild type. The acidic alleles have been designated A'Milano [32, 331, A-lMarburg [34~351, A-lMunster 2A and A-lMUnster 2B 1371. These mutations result from the following single amino acid substitutions: Arg17, Cys, LysIo7 0, Lysl,, 0, and AlaIs8 Glu, respectively. The basic alleles have been designated A-IGjessen 1341 A-lMunster 3.4, A-lMunster 3B1 A-lMunster 3C 1361, A-IMunster3D, A-IMunster4, and A-INorway[37]. These mutations result from the substitutions Prol4, Arg, Asp,,, Asn, Pro4 Arg, Pro3 His, Asp,,, Gly, GluI9, Lys, and Glu,,, Lys, respectively. A-IMiIano and A-IMarburg,
-
-
-
-
wl,
9
-
-
-
-
-
-
-
TABLE 3 ApoA-I genetic variants
Name
Charge difference
Defect
-1 -1 -1
Argl,3
___
A-lMilano A-lMarburg A-1Munster2A A-1Munster2B
A-IGiessen A-1Munsrer3A A'1Munsrer3B A-1Munster3C
A-1Munster3 D A-1Munsrer4 A-INorway
-1
+I
+I +I +I +I t 2 +2
- CYS
-0 -0 Ala,,, - Glu Prold3- Arg Asp,o3 - Asn Pro4 - Arg - His LYSIO, LYSIO,
Pro3
-GlY GluIYs -LYS
ASP213 GluI3,
- LYS
364 but not the other structural variants, have been associated with reduced HDL levels [32, 351. Normal apoA-I activates lecithin cholesterol acyltransferase and A-IGiessen,A-IMarburgand A-IMunster2A have been reported to be defective in this regard [38]. These variants have been discovered because they produce gene products with different net charge from wild type. Presumably, other apoA-I gene mutations exist that affect the protein coding region of the gene, but do not change the charge of apoA-I and are as yet undetected.
3. APOA-11 ApoA-I1 is the second most abundant protein in HDL, comprising approximately 20% of its protein. Plasma apoA-I1 concentrations are 0.3 to 0.5 mg/ml. In vitro, apoA-I1 has been shown to displace apoA-I from HDL particles, as well as both activate hepatic lipase and inhibit LCAT. However, the physiological role of apoA-I1 has not been determined, and primary qualitative or quantitative abnormalities of human apoA-I1 have not been reported. ApoA-I1 is made in the liver and intestine. (a) ApoA-II cDNA
ApoA-I1 cDNA clones have been isolated and sequenced [14, 39- 411. From this information, it has been deduced that apoA-I1 mRNA is 473 bp in length and includes a 5 ’ untranslated region of 58 bp, a region coding for 100 amino acids of 300 bp, a termination codon TGA, and a 3’ untranslated region of 112 bp followed by a poly A tail. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoA-I1 mRNA in cell-free synthesis experiments [42] indicates translation initiation at the methionine 23 amino acids upstream of the NH2terminus of the mature protein. Cotranslational cleavage of the primary translation product by microsomal membranes removes the 18 NH2-terminal amino acids which presumably represent the apoA-I1 prepeptide. The remaining 5 amino acid polypeptide adjacent to the NH2-terminus of mature apoA-I1 is a propeptide with the sequence Ala-Leu-Val-Arg-Arg. The occurrence of two basic amino acids in the propeptide adjacent to the NH2-terminus of the mature protein is rather typical of propeptides and quite different from the apoA-I propeptide [15 - 171. Therefore, apoA-I1 should be cleaved intracellularly, in contrast to apoA-I. However, studies with the human hepatoma cell line, HepG2 [43], indicate that only a fraction of proapoA-I1 is cleaved intracellularly with the rest cleaved extracellularly by cathepsin B. The significance of these extracellular processing events has yet’to be elucidated. The cDNA sequence specifies an amino acid sequence for mature apoA-I1 of 77
365 amino acids, very similar to the previously reported amino acid sequence determined by protein sequencing methods [44]. The difference was at residue 35, where the DNA sequence predicts Glu, and the protein-derived sequence specified Gln. In human plasma, apoA-I1 exists as a dimer, the monomers are connected by a disulfide bridge between the only cysteine in the mature protein which resides at amino acid residue 6 [44]. (b) ApoA-11 gene
The apoA-I1 gene has been isolated and sequenced [45 - 471. From the transcription initiation to the polyadenylation site, the gene is either 1330 or 1343 bp according t o two different reports. A comparison of the apoA-I1 gene and cDNA sequences reveals three introns. IVS-1 is 169 or 182 bp in length and occurs in the 5 ‘ untranslated region between bases 24 and 25 upstream of the codon for Met that initiates translation. IVS-2 is 293 bp long and interrupts the codon specifying amino acid - 6, which is the apoA-I1 prepeptide. IVS-3 is 395 bp long and interrupts the codon specifying amino acid 39 of the mature protein. The intron locations are remarkably similar to those for the apoA-I gene and suggest, as specified for apoAI , that the apoA-11 exons may code for functionally distinct regions. The apoA-I1 transcription initiation site has been designated by using HepG2 cell mRNA for primer extension and Sl nuclease protection experiments. Upstream of this site in the appropriate location there is a 6 bp long AT-rich region, which is presumably the ‘TATA box’ portion of the apoA-I1 gene promoter. Upstream of the ‘TATA box’ is a sequence resembling another regulatory element in some eukaryotic gene promoters, the ‘CAT box’. Finally, an Alu repetitive element has been identified between approximately 300 and 650 bp 3 ’ to the apoA-I1 gene.
4. APOA-IV ApoA-IV was originally discovered as a major constituent of rat HDL. In humans, it is a relatively minor HDL protein with most apoA-IV in the nonlipoprotein plasma fraction. Human plasma apoA-IV concentrations are approximately 0.16 mg/ml. In rats, apoA-IV synthesis occurs both in intestine and liver, and in this animal fat feeding doubles intestinal apoA-IV synthesis [48]. On this basis, it is hypothesized that apoA-IV plays a role in the synthesis and secretion of intestinal triglyceride-rich lipoproteins, but direct proof of this or any other functional role for apoA-IV is lacking.
(a) ApoA-IV cDNA Human apoA-IV cDNA sequence has been recently reported [49, 501. From this in-
formation, human apoA-IV mRNA is thought t o be 1469 bp in length, and includes a 5 ’ untranslated region of 113 bp, a region coding for 396 amino acids of 1188 bp, a termination codon, TGA, and a 3 ’ untranslated region of 165 bp followed by a poly A tail. In other experiments, utilizing intestinal mRNA, the NH2-terminal sequence of the apoA-IV primary translation product was analyzed by microsequencing techniques and compared to the NH2-terminal sequence of the mature protein [51]. In this manner, it was determined that human apoA-IV contains a prepeptide sequence 20 amino acids in length. Unlike apoA-I and apoA-11, apoA-IV does not contain a propeptide. These data combined with the cDNA sequence indicates that the mature apoA-IV protein is 376 amino acids in length. The deduced apoA-IV protein sequence was analyzed and multiple repetitions of 22 amino acid segments with amphipathic alpha-helical character were identified [49]. This is analogous to the previous findings for human apoA-I.
(b) ApoA-IV gene The human apoA-IV gene was recently isolated and characterized [49, 501. It is approximately 2600 bp in length and includes two introns [50]. IVS-1 is 359 bp long and interrupts the codon specifying amino acid - 4 in the apoA-IV prepeptide region. IVS-2 is 777 bp long and interrupts the codon for amino acid 39 of the mature protein. IVS-1 and IVS-2 of the apoA-IV gene are in strikingly similar locations to two of the introns in most of the other apolipoprotein genes, A-I, A-11, CI, CII, CIII and E. However, all of these genes also contain another intron, which interrupts the sequence coding for the 5 ’ untranslated region, which is evidently missing from the apoA-IV gene [50].
(c) Apo-A-IV genetic variation Genetic variation in human apoA-IV has been demonstrated [34, 361. In most people, isoelectric focusing of plasma apoA-IV results in a single major isoprotein of pH 5.50, other people have this isoprotein plus another, which is one charge unit more basic, and a few people have just the more basic isoprotein. In one large German study, the frequencies of these patterns in a normal population were 85.6070, 13.8% and O.6%, respectively [36]. Genetic studies were consistent with a single genetic locus two allele model [52]. From these data, the major allele frequency, specifying the more acidic gene product, was 92.5% and the minor allele frequency, specifying the more basic gene product, was 7.5%. Neither heterozygosity nor homozygosity for the minor allele have been associated with plasma lipoprotein abnormalities or atherosclerosis susceptibility.
367
5. ApoB ApoB is the major protein constituent of low density lipoproteins (LDL), but is also found in chylomicrons and VLDL. LDL particles are approximately 25% protein and 75% lipid and virtually all of the protein is apoB. LDL levels are directly correlated with coronary artery disease susceptibility and recently the same association has been demonstrated for apoB levels [6, 71. ApoB is abundant in plasma with a concentration of 0.7 to 1.0 mg/ml. ApoB synthesis appears to be required for the secretion into plasma of intestinal and hepatic triglyceride-rich lipoproteins. ApoB is also recognized by specific high affinity receptors that mediate clearance of LDL particles from plasma [53]. Human apoB is a glycoprotein which occurs in two forms, designated B-100 and B-48 [54]. B-100 is thought to be a single polypeptide of molecular weight approximately 550 000 produced primarily in the liver, whereas B-48 is approximately half that molecular weight and is produced primarily in the small intestine. Proteolytic degradation products of B-100 called B-74 and B-26 have been described, which are produced by a kallikrein-like activity present in plasma. Whether this happens in vivo remains to be determined [54a]. Classical protein chemistry techniques have provided very little information about the primary structure of apoB. The main problems were that the protein becomes quite insoluble after delipidation and standard methods of proteolytic digestion do not result in a high enough yield of unique peptides for structural studies. It has been suggested that the insolubility problem is due to an abnormal sensitivity of the protein to oxidation after delipidation, and recently, with proper precautions, soluble preparations of delipidated apoB have been obtained [55]. In addition, the use of bacterial proteases has improved the yield of unique apoB peptides [56]. A recent report provided the partial amino acid sequence of two apoB peptides [56]. The paucity of data on apoB structure, using standard protein chemistry techniques, has led to immunochemical studies to gain relevant information about this protein. In this regard, several groups have developed monoclonal antibodies to human apoB that recognize distinct epitopes, and some interesting information has been derived. In one study, the immunoreactivity of apoB in VLDL changed significantly after in vitro lipolysis, suggesting that apoB conformation might change at different stages of lipoprotein metabolism [57]. In another study, epitopes of apoB-100 have been mapped in a linear nonrepetitive array. For a subset of these epitopes, the monoclonal antibodies disrupt apoB binding to the LDL receptor, but d o not bind to B-48. For another subset of epitopes, the monoclonal antibodies bind B-48, but do not disrupt receptor binding [58]. Thus, B-48 and B-100 are antigenically related and it has been suggested that B-48 represents approximately one half of the B-100 protein and that this part of apoB is not involved in receptor binding [54, 581. Recently, a monoclonal antibody has been used to show parallel expression of an
368 epitope in both B-100 and B-48. This is very strong evidence that both B-100 and B-48 are products of the same gene [%a]. (a) ApoB cDNA
ApoB cDNA clones from human liver libraries have recently been isolated by several groups [59 - 64,64al. Two strategies were used. Oligonucleotides were synthesized based on limited apoB protein sequence data and used as probes to screen cDNA libraries. Alternatively, polyclonal antibodies to apoB were used to screen expression libraries and identify clones producing fusion proteins containing apoB epitopes. It appears that two types of cDNA clones have been identified. The first type corresponds to internal parts of the coding sequence and contains the regions specifying the previously reported apoB peptides. Deeb et al. reported a cDNA sequence coding for 197 amino acids of apoB, including the sequence of the previously reported peptide R3-1 [59]. Mehrabian et al. reported another cDNA sequence coding for 104 amino acids of apoB, which included the sequence of the peptide R2-5 [61]. These two sequences do not overlap. The second type of apoB cDNA clone corresponds to the COOH terminal amino acid part of the coding sequence and the 3 ' untranslated region. Knott et al. reported a cDNA sequence coding for the COOH terminal 1955 amino acids of apoB as well as the termination codon TAA and 266 bp of the 3 ' untranslated region [62]. The latter is imcomplete since it lacks the polyadenylation signal and tail. The sequence reported by Deeb et al. is not contained within this cDNA, but the sequence of Mehrabian corresponds to the 5 ' end of this sequence. Wei et al. reported a cDNA clone, coding for the COOH terminal 836 amino acids of apoB, derived from an expression library. This clone produces a protein that reacts with monoclonals that detect B-74 but not B-26 and establishes that B-74 corresponds to the COOH terminal end of apoB [64a]. Knott et al. identified within apoB an amino acid sequence that they propose might be the apoB/E receptor binding region [62]. This region has the sequence ThrThr-Arg-Leu-Thr-Arg-Lys-Arg-Gly-Leu-Lys (residues 276 - 286), which they note is structurally similar to the amino acid sequence of the apoE receptor binding domain (apoE residues 140- 150) [65 - 671. Both these apoB and apoE regions have similarly spaced, positively charged residues and would be ideal for interaction with the postulated ligand binding site of the apoB/E receptor, which is enriched in negatively charged amino acid residues [68]. Further proof that the proposed region of apoB is indeed the receptor binding domain will be required. Thus far, all the apoB cDNA clones described detect on Northern blot analysis of liver mRNA a species of 18 to 22 kb in length. mRNA of this length could contain enough coding information for B-100, which has an apparent molecular weight of 550 000 on SDS PAGE and may contain > 5000 amino acids. These same clones
369 have been used for Northern blot analysis with RNA from adult intestine, which makes B-48 but not B-100. Surprisingly, these studies also show an mRNA species of approximately 20 kb in length [61, 641. The explanation for this, as well as the structural relationship between B-100 and B-48 at the protein, mRNA and gene level, should be forthcoming in the near future.
(b) ApoB genetic variation Genetically determined variation in both the quality and quantity of apoB has been demonstrated. Antisera from multiply transfused patients have been used to define two series of allelic variants called Lp and Ag. Both systems appear to have marginal effects on plasma cholesterol levels and risk of atherosclerosis [69]. Recently, individual variation in apoB reactivity with an anti-apoB monoclonal antibody has been demonstrated [70, 711. Three phenotypes of strong, weak, and intermediate binding have been identified and family studies suggest a genetic basis for this consistent with a single genetic locus with two alleles specifying strong and weak binding forms of apoB. Thus, the intermediate binding pattern is the result of heterozygosity for the strong and weak binding alleles, whereas the other two patterns represent homozygosity for their respective alleles. In a study of the phenotype frequency in unrelated individuals the major allele was the one specifying weak binding, and its frequency was found to be 61% [71]. This apoB antigenic variation presumably results from an alteration in the amino acid sequence of the apoB polypeptide. This variation in apoB does not seem to affect LDL lipid composition or density [71], but whether it is associated with altered plasma lipoprotein levels has yet to be determined. The monoclonal antibody used to detect the apoB variation has been shown not to interfere with receptor binding [71]. Inherited disorders of lipoprotein metabolism associated with diminished plasma levels of apoB have been described. The most striking of these is abetalipoproteinemia [2]. In this disorder, individuals suffer from fat malabsorption and lack apoB-containing lipoproteins in their plasma, including chylomicrons, VLDL and LDL. Parents of affected individuals have normal lipoprotein and apoB levels, however, siblings with this disorder have been described and inheritance is assumed to be autosomal recessive. Immunologically detectable apoB is absent from both the plasma and tissues of these individuals, and the disorder is thought to be a genetic defect in apoB synthesis. A phenotypically similar disorder has been described, termed homozygous hypobetalipoproteinemia, in which parents of affected individuals have half-normal levels of LDL cholesterol and apoB [2]. This condition may also involve a genetic defect in apoB synthesis, but by a different mechanism. Finally, individuals have been described with normal fat absorption and the ability to produce chylomicrons, but low to absent levels of LDL cholesterol [72 - 741. Apparently, these individuals can produce the intestinal form of apoB, B-48, but not the hepatic form, B-100. The existence of this disorder suggests separate genetic con-
370 trol of B-48 and B-100 synthesis. However, at this time, it cannot be determined whether these two gene products are the result of distinct genetic loci or represent differential splicing of the transcript produced from a single genetic locus. Increased apoB levels are associated with atherosclerosis susceptibility [6, 71. In familial hypercholesterolemia (FH), an autosomal dominant disorder characterized by premature atherosclerosis, apoB and cholesterol in LDL are elevated and the genetic lesion is a defect in the apoB/E receptor [53]. A substantial fraction of nonFH individuals with coronary disease have been shown to have increased apoB, but normal cholesterol in LDL. This phenotype has been called hyperapobetalipoproteinemia [75]. A subset of these individuals may have the autosomal dominant disorder associated with premature atherosclerosis, familial combined hyperlipidemia (FCHL) [76]. Plasma from these individuals consistently contains elevated plasma apoB levels but only occasionally is the LDL cholesterol elevated [77]. Genetic abnormalities associated with apoB may be one possible explanation for the hyperapobetalipoproteinemia phenotype, but this has not been proven.
6. ApoCI ApoCI is a constituent of VLDL and HDL. VLDL particles are approximately 10% protein and 90% lipid and apoCI is 10% of VLDL protein. As previously noted, HDL particles are approximately 50% protein and 50% lipid and apoCI is 2% of HDL protein. Human plasma apoCI concentrations are in the range of 0.04 to 0.06 mg/ml. In vitro apoCI has been shown to activate LCAT, but not as efficiently as apoA-I. The physiological role(s) of apoCI has not been defined. Synthesis mainly occurs in liver and to a minor degree in intestine, but has not been evaluated in other organs. Primary qualitative or quantitative abnormalities of human apoCI have not been reported. (a) ApoCI cDNA and gene
ApoCI cDNA clones have been isolated [77], From this information, apoCI mRNA is thought to be approximately 419 bp in length, including a 5' untranslated region of 56 bp, a region coding for 83 amino acids of 249 bp, a termination coding, TGA, and a 3' untranslated region of 111 bp followed by a poly A tail. The DNA-derived amino acid sequence contains the 57 residues of mature apoCI and agrees entirely with the results derived by protein sequencing techniques [79, 801. The DNA sequence also specifies a 26 amino acid NH2-terminal extension [77, 781. This amino acid sequence is compatible with the entire 26 amino acids being the apoCI prepeptide. Recently, apoCI genomic clones have been isolated and partially characterized [81, 821. It appears that there are two copies of the apoCI gene [81]. One of these
37 1
has been sequenced [Sl] and found to contain four exons and three introns and is 4374 bp in length. The exons are 33, 78, 136 and 169 bp in length, and the introns are 138, 1190, and 2630 bp in length. The introns are in strikingly similar locations to several of the other apolipoprotein genes with 1VS-1 interrupting the 5‘ untranslated region, 1VS-2 interrupting the region coding for the prepeptide, and IVS-3 interrupting the region coding for the mature protein. The second apoCI gene has yet to be sequenced. It is not known whether one or both genes are expressed or whether one is a pseudogene. In this regard, it is of interest that Northern blotting of adult human liver RNA with an apoCI cDNA probe indicates two species of apoCI mRNA that differ in length by approximately 20 bp [77]. In addition, primer extension on a template of human liver mRNA indicates two different transcription initiation points, also approximately 20 bp different from each other [77]. It is tempting to speculate that both apoCl genes are transcriptionally active and produce two different size and perhaps different sequence mRNA’s. This remains to be proven.
7. ApoCII ApoCII is a constituent of VLDL and HDL and comprises 10% of VLDL protein and approximately 1% of HDL protein. Human plasma apoCII concentrations are in the range of 0.03 to 0.05 mg/ml. Purified apoCII has cofactor activity for the enzyme lipoprotein lipase, which catalyzes the hydrolysis of triglycerides in chylomicrons and VLDL. The physiological importance of apoCII in activating lipoprotein lipase has been established by the finding of patients with inherited apoCII deficiency, who are severely hypertriglyceridemic and have functional lipoprotein lipase deficiency (see [83]). In such cases, both the hypertriglyceridemia and the lipase deficiency can be relieved by an exogenous source of apoCII [84]. Studies using proteolytic fragments and synthetic peptides of apoCII [85 - 871 have shown that lipoprotein lipase interacts with the COOH terminal amino acids 56 to 79. This interaction is enhanced by residues 44 to 55, which possess phospholipid binding activity. Deletion of the last three residues, 77 t o 79, prevents the protein from activating lipoprotein lipase. Synthesis of apoCII is mainly in liver and, to a minor degree, in intestine. (a) ApoCII cDNA
ApoCII cDNA clones have been isolated and characterized [14, 88 - 901. From this information, apoCII mRNA appears to be 488 t o 494 bp in length, according to different reports. This includes a 5 ’ untranslated region of 38 bp, a coding region for 101 amino acids of 303 bp, a termination codon, TAA, and a 3 ’ untranslated region of from 144 to 150 bp followed by a poly A tail. The heterogeneity in the 3 ’ untranslated region is due t o a minor variation in the polyadenylation site used between different apoCII cDNA clones.
372 The DNA-derived amino acid sequence contains the 79 residues of mature apoCII. This is in agreement with a recent result derived by protein sequencing techniques [91] but differs somewhat from the previously derived amino acid sequence [92]. The DNA sequence also specifies a 22 amino acid NH2-terminal extension compatible with the existence of an apoCII prepeptide. There is a striking homology between the NH2-terminal regions of both apoA-I and apoCII at the amino acid level. ApoA-I amino acids - 2, - 1, + 1, + 2 are identical to apoCII amino acids 5 , 6, 7 and 8 and, in each case, specify Gln-Gln-AspGlu. This region spans the site of cleavage of the protease that converts proapoA-I to mature apoA-I. This suggests that apoCII should be a substrate for the apoA-I converting protease. However, in plasma, mature apoCII is found to be the full 79 amino acids in length and is fully active as a lipoprotein lipase activator in this form. Thus, apoCII is not a physiological substrate for the apoA-I converting protease, for reasons which remain to be elucidated. Preliminary analysis of the apoA-I and apoCII cDNA sequences reveals extensive homology at the DNA level extending between apoA-I bp 102 - 285 (amino acids - 2 to 59) and apoCII bp 79 - 261 (amino acids 5 to 65). Alignment of the DNA sequences in these regions by metric analysis indicates 56% overall matching with runs of matches 11, 8 and 10 bp in length (B. Erickson and J. L. Breslow, unpublished observations). These observations suggest a close phylogenetic relationship between apoA-I and apoCII.
(b) ApoCII gene structure The apoCII gene has been isolated and sequenced “3, 931. The gene is 3320 bp in length. It contains four exons and three introns. The exons are 25, 68, 160 and 241 bp, whereas the introns are 2356, 166 and 304 bp in length. IVS-1 occurs in the 5’ untranslated region between bases 13 and 14 upstream of the codon for Met that initiates translation. IVS-2 and IVS-3 interrupt the codons specifying amino acids - 4 and 50, respectively. There are introns in similar locations in the other apolipoprotein genes, where this has been studied. The two apoCII introns interrupting the protein coding region divide the DNA sequence coding for the signal peptide from that coding for the mature protein, and the DNA sequence coding for the NH, from that coding for the COOH terminal portions of the mature protein, respectively. As previously noted, the COOH terminal amino acids appear to be the functional domain involved in binding lipoprotein lipase. The apoCII gene was found to have four Alu-type repetitive sequences in IVS-1 and another in the 3’ flanking region of the gene. In addition, IVS-3 was composed largely of six copies of a 36 to 40 bp tandem repeat. The significance, if any, of both types of repeats in the apoCII gene is not known. Finally, a 6 bp AT-rich region is present in the appropriate location upstream of the transcription initiation site, which may be part of the apoClI promoter, the so-called ‘TATA box’.
373
(c) ApoCII genetic variation Several patients have been described who lack apoCII in their plasma [83]. They are functionally lipoprotein lipase deficient and severely hypertriglyceridemic. ApoCII deficiency appears to be a recessive genetic disorder with obligate heterozygotes having half-normal apoCII levels, but normal triglyceride concentrations. Southern blot analysis with a probe made from an apoCII cDNA clone reveals that at least a subset of these patients possess the apoCII gene and that it is grossly intact [94]. Two independent families of probands with this disorder have been studied using the TaqI polymorphism t o assess disease linkage with the apoCII gene locus. In each family, cosegregation was observed, which is consistent with the defect in this condition being within or near the apoCII gene [94]. Recently, several probands with apoCII deficiency have been found to have a small amount of a mutant form of the protein in their plasma [95, 961. Presumably the defect in the apoCII gene in these individuals results in the production of a n unstable nonfunctional protein. In most probands with apoCII deficiency, it is possible to fully activate lipoprotein lipase in their post-heparin plasma by adding exogenous pre-heparin plasma and/or purified apoCII [83]. A single family has been described in which several affected individuals are apoCII deficient and where exogenous apoCII fails to activate their lipoprotein lipase [97]. Perhaps in this family the apoCII gene defect is linked t o a defect in the lipoprotein lipase gene or another gene contributing to lipase activation. The gene for lipoprotein lipase is not yet cloned, and it is not known whether it localizes to the same region of the genome as the apoCII gene. A protein polymorphism of apoCII has also been described [98, 991. I n most individuals, plasma a p e 1 1 is a single spot on two-dimensional gels and a single band on one-dimensional size or charge separation gels. However, three hypertriglyceridemic individuals were found who had a normal apoCII component and a n apoCII isoprotein 1 charge unit more acidic than normal. It has been determined that this is due to a substitution of glutamine for lysine at residue 5 5 . The apoCII DNA sequence at this residue is AAA which specifies lysine. It is possible to explain the occurrence of glutaniine at this residue by a single base substitution in the first base of this codon, substituting C for A. Although this mutant form of apoCII was isolated from hypertriglyceridemic patients, it appears to activate lipoprotein lipase normally. Thus, the relationship between the amino acid subsitution and the hypertriglyceridemia probably is not cause and effect.
8. ApoCIII ApoCIII is a constituent of VLDL and HDL and comprises about 50% of VLDL protein and 2% of HDL protein. Human plasma apoCIII concentrations are in the range of 0.12 to 0.14 mg/ml. ApoCIII is a glycoprotein containing 1 mol each of
374 galactose, galactosamine and either 0, 1 or 2 mol of sialic acid. The three resultant isoproteins recognizable by isoelectric focusing are designated CIII-0, CIII- 1 and CIII-2 and comprise 14%, 59% and 27% of plasma apoCIII, respectively. In vitro apoCIII has been shown to inhibit the activities of both lipoprotein lipase and hepatic lipase. ApoCIII has also been shown to decrease the uptake of lymph chylomicrons by the perfused rat liver. These in vitro studies suggest that apoCIII might delay catabolism of triglyceride-rich particles. Recently, hypertriglyceridemic subjects were shown t o have circulation lipoprotein and non-lipoprotein inhibitors of lipoprotein lipase [ 1001. The lipoprotein-associated inhibition correlated best with apoCIII concentration. In the same study, apoCIII was shown to be a noncompetitive inhibitor of the activity of partially purified lipoprotein lipase. In addition, patients with combined apoA-I, apoCIII deficiency were shown to have low plasma triglyceride levels [ 1011, and in vivo studies showed that they rapidly convert VLDL to LDL [102]. In vitro lipolysis of their VLDL was inhibited by added apoCIII [ 1021. Thus, it appears that primary abnormalities in the quantity or quality of apoCIII may affect plasma triglyceride levels and the physiological role of apoCIII may be in the regulation of the catabolism of triglyceride-rich lipoproteins. Functional domains of apoCIII have been demonstrated. The NH2-terminal 40 amino acids do not bind phospholipid, whereas the COOH-terminal 39 amino acids do [103]. Synthesis of apoCIII is mainly in liver and to a lesser degree in intestine. (a) ApoCIII cDNA
ApoCIII cDNA clones have been isolated and sequenced [14, 104- 1061. Based on this information, the apoCIlI genomic sequence (see below), and knowledge of the consensus sequence for transcription initiation, the apoCIII 5 ’ untranslated region is 49 bp. This is followed by a coding region for 99 amino acids of 297 bp, a termination codon, TGA, and a 3 ’ untranslated region of from 183 to 189 bp followed by a poly A tail. The heterogeneity in the 3 ’ untranslated region is due to a minor variation in the polyadenylation site used between apoCIII cDNA clones. Thus, apoCIII mRNA is 532 to 538 bp in length. The DNA-derived apoCIll amino acid sequence differs from the previously reported protein-derived apoCllI amino acid sequence [lo71 at residues 32, 33, 37, and 39. At this location, the DNA sequence predicts Glu, Ser, Gln, Ala, respectively, whereas the previously reported protein-derived, sequence specified Ser, Gln, Ala, Gln, respectively. Several cDNA clones from different cDNA libraries [14, 105, 1061 all have shown the same DNA-derived amino acid sequence. In addition, the DNA sequence coding for residues 32 and 33, GAGTCC, includes a recognition site for the restriction endonuclease HinfI (GANTC), whereas the DNA sequence required to code for the corresponding residues in the protein-derived sequence could not possibly contain a HinfI recognition site [ 1051. Southern blotting analysis of genomic DNA from four normal and two hypertriglyceridemic individuals identifies
315 a HinfI site in this region, also compatible with the DNA-derived amino acid sequence [105]. Thus, the protein sequence, unless it was derived from a person homozygous for a rare apoCIII allele, is probably in error and should be revised. The DNA-derived amino acid sequence indicates a 20 amino acid NH2-terminal extension for the primary translation product of apoCIII [14, 105, 1061. The sequence is compatible with previously reported prepeptide sequences. Cell-free synthesis experiments using mRNA from rat liver and intestine indicates that rat apoCIII is made with a 20 amino acid NHz-terminal extension [108]. This can be co-translationally cleaved by signal peptidase t o yield a product with the same NH2-terminus as the mature protein. Therefore, apoCIII is made as a preprotein and does not contain a propeptide sequence.
(6) ApoCIII gene The apoCIII gene has been isolated and sequenced [31, 1051. The gene is approximately 3133 bp in length and contains four exons and three introns. The exons are 36, 68, 124 and 308 bp in length. IVS-1 is 625 bp long and occurs in the 5 ’ untranslated region between bases 13 and 14 upstream of the codon for Met that initiates translation. IVS-2 is 135 bp long and interrupts the codon specifying amino acid -2, which is in the apoCIII prepeptide. IVS-3 is 1837 bp long and interrupts the codon specifying amino acid 40 of the mature protein. The introns are in similar locations to those in most of the other apolipoprotein genes characterized. The intron locations indicate that apoCIIl exons may code for functionally distinct domains of the protein. For instance, intron 2 separates the prepeptide from the mature protein, and intron 3 seems to separate the phospholipid binding domain from the non-binding domain. Another feature of the apoCIII gene is the presence of an Alu repetitive element in the third intron and another approximately 1500 bp 3 ’ to the gene. Finally, an 8 bp AT-rich region is present in the appropriate location upstream of the proposed transcription initiation site. This may be the apoCIII promoter, ‘TATA box’. Upstream of the ‘TATA box’ is a sequence resembling another regulatory element in some eukaryotic gene promoters, the ‘CAT box’. (c) ApoCIII genetic variation
Variation in the apoCIII gene has been documented. Four apoCIII cDNA clones and a genomic clone have been sequenced [14, 31, 105, 1061. Three sites of variation were identified (Table 4). One site of variation is in the coding region affecting the third base of a codon specifying amino acid residue 14, but in each case, the codon specifies glycine. The other two sites of variation are 31 bp apart in the 3’ untranslated region. The first of these sites affects the first base of the recognition sequence for the restriction enzymes SstI and Sac1 [104, 1091. The sequence is
376 TABLE 4 Human apoCIII genetic variation Clonea
Base pairb 151
388
419
1 (cDNA) 2 (cDNA) 3 (cDNA) 4 (cDNA) 5 (genomic a
cDNA clones were sequenced in three different laboratories. Refers to bp location in apoCIII mRNA.
GAGCTC and those individuals with a G in position 388 would have a restriction enzyme cutting site revealed by Southern blot analysis (see section on RFLP’s). Whether the bp variations identified in the 3 ’ untranslated region are themselves clinically significant remains to be determined.
9. ApoE ApoE in normal plasma is equally divided between VLDL and HDL. It comprises about 10 to 20% of VLDL protein and 1 to 2% of HDL protein. ApoE occurs in a metabolically distinct subfraction of HDL particles where it is a larger fraction of the protein. Human plasma apoE concentrations are in the range of 0.025 to 0.050 mg/ml. Two-dimensional gel electrophoresis of human plasma apoE has shown it to consist of several isoproteins which differ in size and/or charge [ 110, 1111. This is the result of both common genetic variation of apoE in the population (discussed below) and post-translational modification of apoE with carbohydrate chains containing sialic acid [110- 1141. ApoE is synthesized and secreted as sialo apoE and subsequently desialated in plasma [19, 20, 1151, but the physiological significance of this process is unknown. ApoE can be recognized by high affinity receptors and can mediate the binding, internalization, and catabolism of lipoprotein particles. ApoE can serve as a ligand for the LDL (apoB/E) receptor present on hepatic as well as on extrahepatic tissues [ 116 - 1181. Hepatic tissues also possess a high affinity receptor that recognizes particles that contain apoE, but not apoB [ 119 - 1211. This receptor is genetically distinct from the LDL receptor and has been called the chylomicron remnant or apoE receptor. Mature apoE is a 299 amino acid polypeptide [122] and the receptor binding region has been localized to the middle
377 portion of the polypeptide chain between residues 140 and 150, with residue 158 being important for the conformation of the binding domain [65 - 671. Structural mutations in apoE affect receptor recognition and are believed t o underlie Type 111 hyperlipoproteinemia (HLP), a condition associated with increased plasma levels of cholesterol and triglycerides, xanthomas, and premature atherosclerosis (for review, see [123]). Glycosylation of apoE occurs via an 0-glycosidic linkage to a single site a t Thr,9,. The sugar chain contains GlcNAc and GalNAc and one or more sialic acid residues [ 1241. ApoE synthesis occurs in liver and to a minor extent in intestine. However, in contrast to the other apolipoproteins, synthesis has been documented in a wide variety of other tissues including brain, kidney, adrenal gland and reticuloendothelial cells [125, 1261.
(a) ApoE cDNA ApoE cDNA clones have been isolated and sequenced [115, 127- 1301. From the transcription initiation point to the polyadenylation site, apoE mRNA is 1163 bp in length and includes a 5 ' untranslated region of 67 bp, a region coding for 317 amino acids of 951 bp, a termination codon, TGA, and a 3' untranslated region of 142 bp. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoE mRNA in cell-free synthesis experiments indicates translation initiation at the methionine 18 amino acid upstream of the mature protein [115]. The NH2-terminal 18 amino acid polypeptide can be co-translationally cleaved by microsomal membranes and represents the apoE signal peptide. There is no propeptide. By analogy with both apoA-I and apoA-IV, human apoE contains eight tandem repetitions of exactly 22 amino acids from residues 62 t o 237 [131]. Only one of these repeats actually begins with proline. However, the sequence of charges of the amino acids in each repeat is strikingly similar. For example, two consecutive acidic amino acid residues occur in the same position in six of the eight repeats. When the DNA segments coding for these repeats are aligned and a consensus nucleotide at each position of the repeat derived, the consensus sequence is 51 to 75% homologous with each of the apoE repeats and 72% homologous to a similarly derived consensus sequence for the six human apoA-I DNA repeats that code for apoA-I residues 99 to 230. Thus, extreme similarity exists with respect to the 66 bp repeats in apoE, apoA-I and apoA-IV, suggestive of a common ancestral origin of this portion of these three apolipoprotein genes. As with the others, the consensus amino acid sequence for apoE, when placed in an Edmundson wheel diagram, specifies an amphipathic alpha-helix [4].
378
(6) ApoE gene structure The apoE gene has been isolated and sequenced by two different groups [131, 1321. In both reported sequences, the exons are 44, 66, 193 and 860 bp in length. There were minor differences between the intron sequences reported. The reported intron lengths were 760,1092 and 582 bp [132], and 757, 1093 and 580 bp [131], respectively. The apoE gene is, therefore, between 3593 and 3597 bp in length, and contains four exons and three introns. IVS-1 occurs in the 5’ untranslated region between bases 23 and 24 upstream of the codon for Met that initiates translation. IVS-2 interrupts the codon specifying amino acid - 4 which is in the apoE prepeptide. IVS-3 interrupts the codon specifying amino acid 61 of the mature protein. The intron locations are strikingly similar to those identified for most of the other apolipoprotein genes and, as previously suggested, may indicate that each exon codes for a functionally distinct region of apoE. The apoE transcription initiation site has been assigned to the A 44 bp upstream of the GT that begins the first intron based on S1 nuclease protection and primer extension experiments with human liver mRNA [ 131, 1321. However, both types of experiments indicated that the start site was heterogeneous and could be within a few bases on either side of the assigned cap site. In addition, the sequence TATAATT occurs beginning 33 bp upstream of the proposed transcription initiation site and is the putative apoE promoter. Furthermore, upstream of the TATA sequence there are four CCCGCC sequences, which have been shown t o play a major role in enhancing SV40 gene transcription. Finally, four Alu repetitive elements have been identified in association with the apoE gene. Two of these are located in the second intron, another is approximately 400 to 700 bp 5 ’ to the gene, and the fourth is approximately 130 to 400 bp 3 ’ to the gene.
(c) ApoE mutations One-dimensional isoelectric focusing of human plasma apoE reveals several bands whose relative concentrations vary between different individuals [112, 1141. Utilizing two-dimensional gel electrophoresis, it was possible to determine that some of these bands were due to sialo apoE isoproteins and others due to variations in the isoelectric point of the major asialo apoE isoprotein(s) [llO, 11I]. Studies of large numbers of individuals revealed six common apoE phenotypes in the population ((1 10, 1111, Fig. 1). Family studies showed that these phenotypes were the result of a single apoE gene locus with three common alleles [llO, 1111. The alleles have been designated €4, €3 and €2 and their gene products from basic to acidic are E4, E3 and E2, respectively. There are three homozygous phenotypes, E4/4, E3/3 and E2/2, and three heterozygous phenotypes, E4/3, E3/2 and E4/2 ([133], Fig. 1). In most of the large studies of apoE phenotype prevalence the range of allele frequencies were E4 14 to 15’70, €3 74 to 78% and €2 8 to 12% ([134- 1371, Table 5). These have
379 Apo E4
E4/4
E
alleles
63 -
E2
-
0 0
E 313 wl
aJ Q
21 c 0 C
E 212
8
E4/3
aJ Q
0
0
0
Fig. I . Schematic presentation of the three-allele model of apoE inheritance and nomenclature of the apoE alleles and phenotypes. The closed circles represent the major asialo apoE isoproteins.
TABLE 5 ApoE phenotype prevalence in population studies; derived apoE allele frequenciesa
Phenotype
Germany
Germany
USA
New Zealand
E4/4 E3/3 E2/2 E4/3 E4/2 E3/2
2.8 59.8 1 .o 22.9 1.5 12.0
2.2 62.2 0.9 19.9 2.9 11.1
1.1 61.3 2.4 20.8 4.2 10.1
51.4 1.4 25.0 1.2 20.0
58.3 0.5 24.8 2.8 12.8
6.3 54.0 0.3 31.9 0.5 6.1
No. of subjects
1031
1557
168
426
400
615
15 77 8
14 78 8
14 71
14 14 12
15 71 8
23 73 4
Allele €4 €3 €2
9
1 .o
Scotland 1 .o
Finland
The apoE phenotype frequencies are given as percentage of that phenotype occurring in the population under study. Allele frequencies were calculated and given as percentage of total (see [134- 1381).
a
been done in diverse geographical areas but primarily in Caucasians. Large studies assessing the frequencies of the apoE alleles in other racial groups have not been reported. A recent report from Finland indicates that this population may have a higher €4 and a lower €2 allele frequency than other Caucasian populations ([138], Table 5). This common apoE polymorphism has been found to play a role in Type I11 hyperlipoproteinemia (HLP) [ 110 - 112, 1391. This disorder is characterized by elevated cholesterol and triglyceride levels, as a result of delayed chylomicron remnant clearance, xanthomas and premature coronary as well as peripheral vascular disease [123]. Over 90% of individuals with Type I11 HLP have the E2/2 phenotype [140], whereas this occurs in only approximately 107'0 (Table 5) of normal individuals [134- 1371. In addition, when E2 is isolated and studied in vitro, it does not bind as well as E3 or E4 to high affinity lipoprotein receptors [141- 1431. It has been suggested that chylomicron remnants with E2 on their surface are recognized poorly by receptors and cleared slowly with the result being an accumulation of these particles in plasma. Chylomicron remnants are potent stimulators of macrophage cholesteryl ester accumulation in vitro and high plasma concentrations of these particles may be involved in the atherogenic process in vivo [123]. These data all suggest that homozygosity for the €2 allele may be the underlying cause of Type I11 HLP . However, the disease frequency is such that only 1 to 2% of people with the E2/2 phenotype actually express the disease. It is known that other hormonal and environmental factors are necessary for disease expression. However, the current belief is that Type I11 HLP is the result of two gene defects. One of these is in the apoE structural gene and the other in another gene that also influences chylomicron remnant synthesis and/or catabolism. The second gene product has yet to be defined. In addition t o the striking involvement of the E2/2 phenotype in Type 111 HLP, it appears that the apoE gene locus may be one of the factors influencing lipid levels in the general population [ 134 - 1371. The €2 allele appears to exert a stepwise gene dosage effect on lowering LDL levels as well as increasing VLDL cholesterol and triglyceride levels. The €4 allele also appears to exert a stepwise gene dosage effect on raising LDL levels. The increased frequency of the €4 allele and the decreased frequency of the €2 allele in the Finnish population may be an underlying genetic factor influencing the exceptionally high LDL cholesterol levels found in this population [ 1381. Amino acid sequence analysis established that the two common variants of apoE, E4 and E2, differ from E3 by single amino acid substitutions [ 1221. E4 differs from E3, at residue 112, because of an arginine for cysteine substitution, and E2 differs at residue 158, because of a cysteine for arginine substitution. Isoelectric focusing and amino acid and DNA sequencing have identified other rare apoE alleles. In all, 11 alleles are known and these are listed in Table 6 . In all but two of these alleles, E7 and E5 [144, 1451, the amino acid substitution underlying the variation has been
identified. Alleles E3** [146], E2 [122], E2* [143], E2** [147], E2*** [148] and E l [ 1491 all involve amino acid substitutions replacing positively charged with neutral amino acids in the region of the apoE receptor binding domain. Where information is available, these gene products have been shown to be defective in receptor binding and/or isolated from individuals with the Type 111 HLP phenotype. This emphasizes the importance of the positively charged amino acid residues in the receptor binding domain. Allele E3* was determined from the DNA sequence [130], but the substitution of proline for alanine at residue 152 might have a significant effect on the conformation of the receptor binding region. Functional studies of this gene product have not yet been reported. The alteration responsible for the E4 allele is not in the receptor binding region, and this gene product is fully functional in receptor binding studies [141- 1431. The E7 and E5 mutations have been found in Japanese, but not Caucasians. Although they are relatively rare alleles, it is said that they appear more frequently in individuals with hyperlipidemia and/or atherosclerosis [144, 1451. In one recent study, four of 58 lipid clinic patients and three of 69 coronary care unit patients were heterozygotes for one of these two alleles, whereas they were not present in 100 normal controls. The molecular basis for these mutations is not known, however, on SDS PAGE, E7 is of normal size, whereas E5 is smaller by 1500-2000 daltons [145]. TABLE 6 Human apoE protein polymorphism
Name"
Charge difference"
Defectb
El E5 E4 E3 E3* E3** E2 E2* E2** E2*** El
f 4 f 2 +I 0 0
? ?
0 -1 -1 -1 -1
-2
Ala,, - Thr, Ala,,, - Pro CYS,,,- Arg, - CYS - CYS Arg,,, - CYS Lys,,, - Gln Arg,,, - Ser C~Y,,,- Asp, Arg,,, - CYS CYSl,,
" Nomenclature for the apoE allele gene products recognized by the isoelectric focusing position of their
major asialo apoE isoprotein as specified by Zannis et al. [133]. The most common allele gene product E3 has an isoelectric point of p H 6.02. Alleles specifying gene products E4, E5, and E7 are 1, 2, and 4 charge units, respectively, more basic and E2 and El are 1 and 2 charge units, respectively, more acidic than wild type. *, **, *** indicate rare apoE variants recently discovered with the same isoelectric focusing pattern as E3 and E2. Amino acid sequence difference with reference t o the most common allele, E3 [122].
3 82
In addition to the apoE mutations just described, a rare form of Type I11 HLP has been reported associated with apoE deficiency [150]. A single family was described in which affected siblings had no immunodetectable apoE in their plasma. Southern blot analysis of genomic DNA from these patients showed that their apoE gene was present and grossly intact [ 15 11. However, monocyte-macrophage cultures from one of these patients was studied and showed a 50-fold decrease in apoE mRNA levels compared to similar cultures from normal people. In addition, two different apoE mRNA sizes were seen with one larger than normal. The apoE mutation in this case appears either to affect the transcription of the apoE gene or the processing of its primary transcript [ 1511.
10. Chromosomal location of apolipoprotein genes (a) Chromosome I : apoA-II Several different groups have used molecular probes and panels of somatic cell hybrids to map the apoA-I1 gene to chromosome 1 ([152- 1541, Fig. 2). Hybrids containing only a portion of chromosome 1, due to translocations, have provided subchromosomal localization. A hybrid containing chromosome lp2 1 to 1qter was positive for the human apoA-I1 gene [153]. Although the data are as yet unpublished, another group claims, in the discussion of one of their papers, that the apoA-I1 gene localizes to the region of lq21 to lq23 [155]. In the mouse, apoA-I1 is in a linkage group with peptidase C and renin found on mouse chromosome 1. In
E
CI
CI I
} A-IZ
I
2
Fig. 2. Chromosomal location of the apolipoprotein genes.
II
19
383 humans, such markers are found on the long arm of chromosome 1, whereas markers for the short arm of human chromosome 1 are found on mouse chromosomes 3 and 4 [153]. These data taken together indicate that the human apoA-I1 gene is on the long arm of chromosome 1, probably in the q21 and q23 region.
(b) Chromosome 2: apoB In recent studies, apoB cDNA probes have been used to map this gene to chromosome 2 ([62, 156, 157, 157a], Fig. 2). Both, in situ and somatic cell hybridization techniques were employed. Hybrids containing translocations involving chromosome 2 were used to provide information about subchromosomal localization. In one study, a hybrid containing 2pter to 2p23 was positive for apoB, whereas a hybrid containing 2p23 to 2qter was negative [157]. In another study, two hybrids containing only 2pter to cen were positive for apoB, whereas a hybrid with a complex rearrangement of the short arm of chromosome 2, due to both terminal and interstitial deletions was negative. The latter hybrid retained the MDH-1 and N-MYC regions [156]. These data taken together indicate that the apoB gene is distal to the MDH-1 and N-MYC loci, which have been mapped to 2p23. The somatic cell hybrid studies are compatible with the in situ hybridization results, which place apoB at the top of the short arm of chromosome 2, between 2pter to 2p24 [62, 1561.
(c) Chromosome 11: apoA-I, apoCIII, apoA-IV Somatic cell hybrids and DNA probes have been used to map the gene for human apoA-I [158 - 1601. In all studies, the apoA-I gene appears to be at the single locus and co-segregates with human chromosome 11 (Fig. 2). Some of the hybrids examined contained only a portion of chromosome 11, and apoA-I cosegregated with p l 1 t o qter [158], p l l to q13 [159], and q13 to qter [160] in three different studies. In the mouse, apoA-I has been mapped to chromosome 9 [161] and shown to be 1.3 & 0.7 centimorgans from uroporphyrinogen I synthetase [162]. The latter has been mapped in humans to chromosome region 1 lq23 t o qter [ 1631. If the linkage group present in the mouse is conserved in humans, it suggests that the apoA-I gene is more distal on the short arm of chromosome 11 than the q13 region. The translocation hybrid thought to contain pl 1 to q13, that was positive for apoA-I, may have to be re-evaluated. Altogether, these data suggest that the apoA-I gene resides on the long arm of chromosome 11 in the region of q23. Recently, the apoA-I gene has been mapped by in situ hybridization to the 1lq22-q23 region [163a]. In addition to the apoA-I gene, other apolipoprotein genes have been localized on chromosome 11 (Fig. 2). cDNA clones and somatic cell hybrids have been used to map the apoCIII gene to this chromosome [158]. A translocation hybrid further
localized the gene to the p l l to qter region. In a family study, the gene for apoA-IV was shown to be linked to the gene for apoA-I [164]. A large Norwegian kindred was identified in which apoA-IV and apoA-I alleles could be distinguished because of isoelectric point differences in their protein products. No recombination was seen in 1 1 observed meioses where this could have been detected. Linkage was suggested by a Lod score of 3.01 at a recombination fraction of 0.00. This study suggests that the apoA-IV gene resides on chromosome 1 1 in the region of the apoA-I gene, q23. The exact relationships of the apoA-I, apoCIII and apoA-IV genes have been determined (Fig. 3). ApoCIII cDNA clones were used to identify the apoCIII gene on human genomic DNA cloned in lambda phage which contained the apoA-I gene [104]. Mapping of the apoCIII gene reveals that it is about 2500 bp from the 3 ’ end of the apoA-I gene. Further mapping and DNA sequence analysis revealed that these genes are coded for by opposite DNA strands [ 1041. The 3 ’ end of the apoCIII gene is located closest to the 3 ‘ end of the apoA-I gene, and the 5 ’ end of the apoCIII gene, containing the apoCIII promoter, is furthest away from the apoA-I gene. Thus, these two genes are convergently transcribed. It is not known where their primary transcripts end, or whether there is any functional significance t o this unusual configuration. Recently, apoA-IV genomic sequences have been located beginning 4 kb 3 ’ to the apoA-I/apoCIII gene complex [49]. The apoA-IV gene is in the same orientation as the apoA-I gene but in the opposite orientation to the apoCIII gene. Thus, the apoCIII and apoA-IV genes share a common 5’ upstream region and are divergently transcribed. Thus, all three apolipoprotein genes at the chromosome 1 1 locus lie within a 14 kb DNA segment.
(d) Chromosome 19: apoE, apoCI, apoCII Family studies have shown that the inheritance of the apoE protein polymorphism cosegregated with the protein polymorphism for the third component of complement [165]. Since the latter had been mapped using DNA probes and somatic cell hybrids to human chromosome 19 [166], the apoE gene was also assigned to this chromosome [165]. ApoE cDNA probes and somatic cell hybrids have now been used to confirm this assignment ([131, 167, 1681, Fig. 2). Limited subchromosomal localization has also been achieved with somatic cell hybrids containing chromo-
* c -
ApoA-I
Apo CIII
___)
Ape A -
IZ
1 Kb
H
Fig. 3. The arrangement of the apolipoprotein genes on chromosome 1 1 . The arrows indicate the direction of transcription.
385 some 19 translocations. Two somatic cell hybrids with chromosome 19’s missing the tip of the short arm and long arm, respectively, were both positive for apoE. In addition to the apoE gene, other apolipoprotein genes have also been localized to chromosome 19 (Fig. 2). cDNA clones and somatic cell hybrids have been used to map the apoCI and apoCII genes to this chromosome [78, 88, 168 - 1701. To determine the relative orientation of the apoE, apoCI and apoCII genes, cosmid clones containing apoE genomic fragments were probed with apoCI cDNA. This has revealed the apoCI gene approximately4 kb 3 ’ to the apoE gene and in the same orientation ([81, 821, Fig. 4). Preliminary evidence indicates a second copy of the apoCI gene 5 to 10 kb 3 ’ to the first apoCI gene [81]. The exact relationship of the apoCII gene to the apoE/apoCI gene cluster has not yet been determined. Genomic clones for apoCII do not hybridize strongly with cDNA clones for apoE or apoCI and vice versa. In spite of this, other genetic evidence indicates a close association. For example, linkage disequilibrium has been demonstrated between the €2 apoE allele and alleles revealed by an apoCII gene-associated RFLP after digesting genomic DNA with the enzyme TaqI [171]. In addition, family studies done by two different groups show linkage of the apoE protein polymorphism and the apoCII TaqI polymorphism [ 171, 1721. In each case, the Lod score was greater than 4.0 at a recombination fraction of 0.00. The combined data from both studies revealed no recombinations in 53 observed meioses. Therefore, apoCII is probably no greater than 2 centimorgans from the apoE/apoCI gene complex. Besides apoE, apoCI and apoCII, other genes have been mapped to chromosome 19 [ 173 - 1751. The gene for familial hypercholesterolemia, the LDL receptor gene, appears to be at the tip of chromosome 19, pter to p13, approximately 20 cm distal to the gene for the third component of complement. The apolipoprotein gene locus is about the same distance from the complement gene, but unlinked to the familial hypercholesterolemia gene. Therefore, the apolipoprotein gene locus must be proximal to the complement gene locus. In addition, the apolipoprotein gene locus appears to be very near the locus for the Lutheran blood group and approximately 15 cm from the loci for myotonic dystrophy and neurofibromatosis.
1 Kb H
Fig. 4. The arrangement of the apolipoprotein genes on chromosome 19. The arrows indicate the direction of transcription. A second apoCI gene is 3 ‘ to the apoE/apoCI gene complex. The dashed lines show uncertainties of the scale in this part of the region. At this point, it is not known whether the apoCII gene is 5 ’ or 3’ to this complex nor is the exact distance from the complex known.
386
1 I . Apolipoprotein gene family In 1977, Barker and Dayhoff, using amino acid sequence data available only for apoA-I, apoA-11, apoCI and apoCIII proposed that the apolipoproteins were all derived from a common evolutionary precursor [24]. In modern genetic terminology, the apolipoprotein genes would be a multigene family, in other words, a group of functionally related genes evolved from an ancestral gene. The data thus far characterizing the genes for apoA-I, apoA-11, apoA-IV, apoCI, apoCII, apoCIII and apoE support this hypothesis. Strong evidence is provided by the existence of gene clusters. The apoA-I, apoCIII and apoA-IV genes all reside within a 15 kb region of chromosome 11. The apoE and apoCI genes reside next to each other, and the apoCII gene is close by on chromosome 19. Other supporting evidence lies in the structural similarities between the genes (Fig. 5). All of these genes but one have four exon, three intron structures with the introns in strikingly similar locations. The exception is the apoA-IV gene in which the IVS corresponding to the IVS-1 of the other genes has apparently been deleted. Other similarities between these apolipoprotein genes include extensive DNA homology in the 5 ’ coding region of apoA-I and CII exon 3 regions that contain a common block of 33 codons, and the existence of multiple 66 bp intragenic duplications in the 3 ’ coding region of the apoA-I, apoA-IV and apoE genes [3, 12, 49, 131, 175al. At the moment it is not certain whether the apoB gene is part of this multigene family. The location of the apoB gene t o a region of the genome quite separate from
-
100 bp
E
v v
v
Fig. 5 . The apolipoprotein mRNA structures are drawn to scale. The length of the protein-coding region is indicated by the rectangles. The thin lines indicate the lengths of the 5 ’ and 3 ‘ untranslated regions o n the left and right, respectively, of the protein-coding region. The locations of the introns are indicated by the triangles. To show the similarities between the genes, the beginnings of all the protein-coding regions are aligned.
387 the other apolipoprotein genes suggests that it may not be part of this gene family. Further work on the structure of the apoB gene will be necessary to establish this conclusion.
12. Apolipoprotein gene associated RFLP’s RFLP’s have been reported for the apolipoprotein genes, and some of these have been associated with lipoprotein abnormalities. RFLP’s have been identified in the apoA-I, apoCIII, apoA-IV gene cluster on chromosome 11 using the enzymes SstI (SacI), PstI and XmnI [104, 109, 1761. The polymorphic site for SstI (SacI) is approximately 2.7 kb 3 ’ to the apoA-I gene and has been shown to be due to a C to G transmutation in the 3 ’ untranslated region of the apoCIII gene, which establishes a n SstI (SacI) cutting site [104]. Utilizing a n apoA-I genomic probe, Southern blotting of genomic DNA after SstI digestion reveals bands of 4.5 kb (S1 allele), 3.2 kb (S2 allele), or both, in heterozygotes. The S2 allele is the minor one and has the additional SstI (SacI) site. In normolipidemic individuals, the frequency of the S2 allele varies in different racial groups. The reported frequency varies from 0 to 6 % in Caucasians, 18% in Indian-Asians, 15% in Africans, 19% in Japanese, and 48% in Chinese [176, 1771. Different clinical studies of Caucasian populations have indicated a n increased frequency of the S2 allele in hypertriglyceridemic individuals (25% and 19%) [109, 1771, Type V hyperlipoproteinemia (27%) [176], low HDL levels (16 YO) [ 1781, and survivors of myocardial infarction (13 Yo) [ 1791. Therefore, the evidence suggests an association between a DNA polymorhpism in the 3 ‘ untranslated region of the apoCIII gene and clinical lipoprotein abnormalities in Caucasians. It is uncertain whether the particular base change involved is causative or merely in linkage desequilibrium with another mutation in the apoA-I, apoCIII, apoA-IV gene cluster that is affecting lipoprotein metabolism. In one study, cDNA clones representing S1 and S2 alleles were sequenced and no associated mutations were found that altered the apoCIII primary amino acid sequence [105]. However, two other base differences were found. One affected the third position of the codon that specifies the glycine a t amino acid residue 14 and the other, affected a base 31 bp 3 ‘ to the SstI polymorphic site, also in the apoCIII 3’ untranslated region. Since the proposed physiological role of apoCIII is in regulating the metabolism of triglyceride-rich lipoproteins, it is attractive to speculate that mutations in and around this gene may be involved in causing hypertriglyceridemia. The polymorphic site for PstI is approximately0.3 kb 3 ’ to the apoA-I gene [176]. Utilizing an apoA-I probe, Southern blotting of genomic DNA after PstI digestion reveals bands of 2.2 kb (P1 allele), 3.3 kb (P2 allele), or both in heterozygotes. The P2 allele, which is missing the PstI site, is the minor one with a frequency of 12%. I n one clinical study, no association between the P2 allele and lipoprotein abnormalities was found [176].
388 The polymorphic site for XmnI is approximately 2.5 kb 5 ’ to the apoA-I gene [176]. Utilizing an apoA-I probe, Southern blotting of genomic DNA after XmnI digestion reveals bands of 8.4 kb (X1 allele), 6.3 kb (X2 allele), or both in heterozygotes. The X2 allele, which is missing the XmnI site, is the minor one with a frequency of 12%. In one clinical study, the X2 allele was associated with Type IIb (25%), Type I11 (25%), and Type V (22Vo), but not with Type IIa or Type IV hyperlipidemia. In the same study, the X2 allele showed an increased frequency in individuals in the highest triglyceride quintile (26%). There was a suggestion of an association with the lowest HDL quintile (19’70), but this did not reach significance [176]. The SstI, PstI, and XmnI RFLP’s in association with the apoA-I, apoCIII, apoAIV gene cluster were found to be in linkage equilibrium and could serve to establish a haplotype [1761. The combined polymorphism information content (PIC value) of these three RFLP’s is 0.5 and they, therefore, could be quite useful in family studies to establish linkage of this gene cluster to clinical phenotypes. In spite of the common protein polymorphism of the apoE gene, RFLP’s have not been reported for the apoE, apoCI gene cluster. Two RFLP’s have been reported for the closely linked apoCII gene using the enzymes TaqI and BglI [171, 172, 1801. The polymorphic site for TaqI is approximately 2 kb 3 ‘ to the apoCII gene. Utilizing an apoCII cDNA probe, Southern blotting after TaqI digestion reveals bands of 3.8 kb (T1 allele), 3.5 kb (T2allele), or both in heterozygotes. The T2 allele, which contains the polymorphic TaqI site, is the minor one with a frequency of 44%. The polymorphic site for BglI is approximately 9 kb 5 ’ to the apoCII gene. Utilizing a 400 bp genomic probe from a unique region approximately 1 to 1.4 kb upstream from the gene, Southern blotting after Bgl I digestion reveals bands of 9 kb (B1 allele), 12 kb (B2 allele), or both, in heterozygotes. The B2 allele, which is missing the polymorphic BglI site, is the minor one with a frequency of 47%. Neither the TaqI nor Bgl I polymorphism have been associated with any clinical lipoprotein abnormalities [171, 172, 1801. The PIC values for both of these polymorphisms separately are approximately 0.37. Utilizing the two together to establish a haplotype should give a PIC value of 0.68;however, due to a degree of linkage disequilibrium the PIC value is reduced to 0.51 [180]. These RFLP’s could be quite useful in family studies to establish linkage of the a p e 1 1 gene, and thereby the chromosome 19 apolipoprotein gene cluster, to clinical phenotypes. The TaqI RFLP alone has already been used to establish linkage to the apoCII deficiency phenotype in families with this form of Type I hyperlipoproteinemia [94]. At the apoA-I1 gene locus on chromosome 1, an RFLP has been reported using the enzyme MspI [155]. The polymorphic MspI site is approximately 0.2 kb 3 ’ to the gene in an Alu-repetitive element. Utilizing an apoA-I1 cDNA probe, Southern blotting after MspI digestion reveals bands of 3.0 kb (M1 allele), 3.7 kb (M2 allele), or both in heterozygotes. The M2 allele, which is missing the polymorphic MspI site,
3 89 is the minor one with a frequency of 19%. In one clinical study, homozygotes for the M2 allele had higher plasma apoA-I1 concentrations and an increased ratio of apoA-I1 to apoA-I when compared to controls or heterozygotes for the M2 allele. HDL cholesterol levels did not differ between the groups [155]. The clinical significance of this observation is as yet unknown. In the mouse, the apoA-I1 locus is also on chromosome 1 and is closely associated with genetic loci that affect HDL composition, the response of HDL cholesterol levels to a diet rich in saturated fat and cholesterol, and atherosclerosis susceptibility [161, 1811. More information will need to be gathered about the effect of the apoA-I1 gene locus or other closely linked loci on these phenomena in man. Several RFLP’s have been reported at the apoB gene locus on chromosome 2. cDNA probes corresponding to the 3 ‘ end of apoB mRNA reveal an EcoRI RFLP [156, 157a, 1821. The polymorphic EcoRI site is located in the coding region of the gene. The two alleles are 11 kb (Rl allele) and 13 kb (R2 allele). R2 is the minor allele and has a frequency of approximately 20%. The same probes also detect a polymorphic region just 3 ’ to the gene, which is revealed with many different enzymes. Presumably, this polymorphism is of the insertion-deletion type. There appears to be significant linkage disequilibrium between these two RFLP’s. The R2 allele is associated with the region of increased length 3 ‘ to the gene [183]. A third RFLP seen with the enzyme XbaI is revealed by more 5 ‘ cDNA probes [184]. The two alleles are 8.6 kb (X1 allele) and 5.0 kb (X2 allele). X2 is the minor allele and has a frequency of approximately 40%. Data on the clinical significance, if any, of these RFLP’s should be forthcoming in the near future.
13. Apolipoprotein gene rearrangements A family has been described in which two sisters had very low HDL but normal LDL levels, xanthomas, severe premature atherosclerosis, and absent plasma apoA-I and apoCIII. First degree relatives of these individuals had half-normal plasma levels of HDL, apoA-I, and apoCIII [loll. Southern blotting of genomic DNA from the probands, after digestion with EcoRI, with an apoA-I cDNA probe revealed a single band of 6.5 kb, whereas normals showed a single 13 kb band. First degree relatives, including the mother and father, of the probands showed one normal band and one abnormal band [185]. Therefore, they appeared to be carriers of a mutant allele associated with the apoA-I gene, and the probands appeared to be homozygous for this mutant allele. Southern blotting of probands’ DNA, after digestion with other restriction endonucleases with an apoA-I cDNA probe, consistently revealed differences from wild type [ 1861. This suggested that the genetic lesion was not a single bp substitution, but rather a more major DNA alteration. Southern blotting with other probes derived from the region of the apoA-I gene indicated that the lesion was a DNA insertion into the fourth exon of the apoA-I gene. This insertion inter-
rupts the normal coding region of the apoA-I gene at approximately the codonspecifying residue 80 of the mature protein and may explain the lack of apoA-I in the plasma of these patients. To define the nature of the DNA insertion, a genomic library was made from the DNA of one of the probands, and clones containing the apoA-I gene and the contiguous up and downstream portions of the insertion were isolated. Southern blotting of genomic DNA from normal individuals, after digestion with EcoRI, with a probe that included apoA-I sequences and insertion sequences, revealed the normal 13 kb apoA-I genomic fragment, as expected, plus another unique band. In the probands the same probe revealed only the 6.5 kb band. This suggested that in the probands a unique piece of DNA, normally present in the genome, has been deleted and was inserted into the apoA-I gene. Further experiments with a probe made to just the insert showed that it hybridized to a region 3 ’ to the apoA-I gene in normal individuals, which contains the apoCIII gene [187]. Thus, it appears that a portion of the apoCIII gene was inserted into the apoA-I gene in the probands and that this is the underlying molecular basis of their apoAI/apoCIII deficiency.
A ckn owledgernents This work was supported by grants from the National Institutes of Health (HL33714, HL32435, AG04727). Dr. Jan L. Breslow is an Established Investigator of the American Heart Association. I would like to express my sincere appreciation to Miss Lorraine Duda for her expert assistance in preparing this review.
References 1 2
3 4 5
6.
7. 8.
Zannis, V.I. and Breslow, J.L. (1985) Adv. Hum. Genet. 14, 125-215. Herbert, P.N., Assmann, G., Gotto, A.M., Jr. and Fredrickson, D.S. (1982) in: The Metabolic Basis of Inherited Disease (J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein and M.D. Brown, Eds.) pp. 589-651, McGraw-Hill, New York. Breslow, J . L . (1985) Annu. Rev. Biochem. 54, 699-727. Breslow, J . L . (1986) in: Biochemistry and Biology of Plasma Lipoproteins (A.M. Scanu and A. Spector, Eds.) pp. 85 - 143, Marcel Dekker, New York. Williams, D.L. (1985) Arteriosclerosis 5, 213 - 227. Heiss, G. and Tyroler, H.A. (1982) in: Proceedings of the Workshop on Apolipoprotein Quantification, pp. 7 - 24, Department of Health and Services, National Institutes of Health (NIH Publication No. 83 - 1266) Bethesda. Brunzell, J.D., Sniderman, A.D., Albers, J.J. and Kwiterovich, P.O., Jr. (1984) Arteriosclerosis 4, 79- 83. Breslow, J.L., Ross, D., McPherson, J., Williams, H . , Kurnit, D., Nussbaum, A., Karathanasis, S . K . and Zannis, V.1. (1982) Proc. Natl. Acad. Sci. USA 79, 6861 -6865.
391 9. 10 I1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
38 39
Shoulders, C.C. and Baralle, F.E. (1982) Nucl. Acids Res. 10, 4873 -4882. Cheung, P. and Chan, L. (1983) Nucl. Acids Res. 1 1 , 3703-3715. Law, S.W. and Brewer, H.B., Jr. (1984) Proc. Natl. Acad. Sci. USA 81, 66-70. Karathanasis, S.K., Zannis, V.1. and Breslow, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 6147-6151. Shoulders, C.C., Kornblihtt, A.R., Munro, B.S. and Baralle, F.E. (1983) Nucl. Acids Res. I I , 2827 - 2837. Sharpe, C.R., Sidoli, A., Shelley, C.S., Lucero, M.A., Shoulders, C.C. and Baralle, F.E. (1984) Nucl. Acids Res. 12, 3917 - 3932. Gordon, J.I., Smith, D.P., Andy, R., Alpers, D.H., Schonfeld, G. and Strauss, A.W. (1982) J. Biol. Chem. 257, 971 -978. Zannis, V.I., Karathansis, S.K., Keutmann, H., Goldberger, G. and Breslow, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 2574 - 2578. Gordon, J.I., Sims, H.F., Lentz, S.R., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1983) J. Biol. Chem. 258, 4037 -4044. Zannis, V.I., Breslow, J.L. and Katz, A.J. (1980) J. Biol. Chem. 255, 8612-8617. Zannis, V.I., Breslow, J.L., SanGiacomo, T.R., Aden, D.P. and Knowles, B.B. (1981) Biochemistry 20, 7089 - 7096. Zannis, V.I., Kurnit, D.M. and Breslow, J.L. (1982) J. Biol. Cheni. 257, 536-544. Edelstein, C., Gordon, J.I., Toscas, K., Sims, H.F., Strauss, A.W. and Scanu, A.M. (1983) J. Biol. Cheni. 258, 11430 - 11433. Kooistra, T., VanHinsbergh, V., Havekes, L. andKenipen, H.J. (1984)FEBS Lett. 170, 109- 113. Brewer, H.B., J r . , Fairwell, T., LaRue, A., Ronan, R., Houser, A. and Brouzert, T.J. (1978) Biochem. Biophys. Res. Commun. 80, 623 - 630. Barker, W.C. and Dayhoff, M.O. (1977) Conip. Biocheni. Physiol. 57B, 309-315. Fitch, W.M. (1977) Genetics 86, 623-644. Schiffer, M. and Edmundson, A.B. (1967) Biophys. J. 7, 121 - 135. Segrest, J.P., Jackson, R.L., Morrisett, J.D. and Gotto, A.M. (1974) FEBS Lett. 38, 247-253. Segrest, J.P., Chung, B.H., Brouillette, C.G., Kanellis, P . and McGahan, R. (1983) J. Biol. Chem. 258, 2290 -2295. Fitch, W.M., Smith, T. and Breslow, J.L. (1985) in: Methods in Enzymology, Chapter 99, Section VI, pp. 773 - 788, Academic Press, New York. Seilhanier, J.J., Protter, A.A., Forssard, P. and Levy-Wilson, B. (1984) DNA 3, 309-317. Protter, A.A., Levy-Wilson, B., Miller, J., Bencen, G., White, T. and Seilhamer, J.J. (1984) DNA 3, 449-456. Franceschini, G., Sirtori, C.R., Capurso, A., Weisgraber, K.H. and Mahley, R.W. (1980) J. Clin. Invest. 66, 892-900. Weisgraber, K.H., Rall, S.C., Bersot, T.P., Mahley, R.W., Franceschini, G. and Sirtori, C.R. (1983) J. Biol. Chem. 258, 2508-2513. Uterniann, G., Feussner, G., Franceschini, G., Haas, J . and Steinmetz, A. (1982) J. Biol. Chem. 257, 501 -507. Utermann, G., Steinmetz, A , , Paetzold, R., Wilk, J., Feussner, G., Kaffarnik, H., MuellerEckhardt, C., Seidel, D., Volgelbert, K.H. and Zimmer, F. (1982) Hum. Genet. 61, 329-337. Menzel, H.J., Kladetzky, L. and Assniann, C. (1982) J . Lipid Res. 23, 915 -922. Weisgraber, K.H., Rall, S.C., Jr., Mahley, R.W., Ogawa, Y., Fielding, C.J., Utermann, G., Haas, J., Steinmetz, A., Menzel, H.J. and Assmann, G. (1985) in: Proceedings of 7th International Syniposiuni on Atherosclerosis (P.J. Nestel, Ed.) pp. 113 - 114, International Atherosclerosis Society. Rall, S.C., Jr., Weisgraber, K.H., Mahley, R.W., Ogawa, Y., Fielding, C.J., Utermann, G., Haas, J., Steinmetz, A., Menzel, H.J. and Assniann, (1984) J . Biol. Chem. 259, 10063 - 10070. Knott, T.J., Priestley, L.M., Urdea, M. and Scott, J. (1984) Biochem. Biophys. Res. Commun. 120, 734 - 740.
Moore, M.N., Kao, F.T., Tsao, Y .K. and Chan, L. (1984) Biochem. Biophys. Res. Commun. 123, 1-7. 41 Lackner, K.J., Law, S.W. and Brewer, H.B., Jr. (1984) FEBS Lett. 175, 159- 164. 42 Gordon, J.I., Budelier, K.A., Sims, H.E., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1983) J . Biol. Chem. 258, 14054- 14059. 43 Gordon, J.I., Sims, H.F., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1985) J. Biol. Chem. 260, 14824- 14831. 44 Brewer, H.B., Jr., Lux, S.E., Ronan, R. and John, K.M. (1972) Proc. Natl. Acad. Sci. USA 69, 1304- 1308. 45 Knott, T.J., Wallis, S.C., Robertson, M.E., Priestley, L.M., Urdea, M., Rall, L.B. and Scott, J. (1985) Nucl. Acids Res. 13, 6387-6398. 46 Lackner, K.J., Law, S.W. and Brewer, H.B., Jr. (1985) Nucl. Acids. Res. 13, 4597-4608. 47 Tsao, Y.K., Wei, C.F., Robberson, D.L., Gotto, A.M., Jr. and Chan, L. (1985) J. Biol. Chem. 260, 15222 - 15231. 48 Gordon, J.I., Smith, D.P., Alpers, D.H. and Strauss, A.W. (1982) Biochemistry 21, 5424- 5430. 49 Karathanasis, S.K. (1985) Proc. Natl. Acad. Sci. USA 82, 6374-6378. 50 Elshourbagy, N.A., Boguski, M.S., Reardon, C.A., Levy-Wilson, B., Jefferson, L.S., Gordon, J.I. and Taylor, J.M. (1985) Arteriosclerosis 5 , 502a. 51 Gordon, J.I., Bisgaier, C.L., Sims, H.F., Sachdev, O.P., Glickman, R.M. and Strauss, A.W. (1984) J. Biol. Chem. 259, 468-474. 52 Menzel, H. J., Kovary, P.M. and Assmann, G. (1982) Hum. Genet. 62, 349 - 352. 53 Goldstein, J.L., Brown, M.S., Anderson, R.G.W., Russell, D.W. and Schneider, W.J. (1985) Annu. Rev. Cell Biol. 1, 1-39. 54 Kane, J.P. (1983) Annu. Rev. Physiol. 45, 637-650. 54a Cardin, A.D., Witt, K.R., Chao, J . Margolius, H.S., Donaldson, V.H. and Jackson, R.L. (1984) J. Biol. Chem. 259, 8522-8528. 55 Cardin, A.D., Witt, K.R., Barnhart, C.L. and Jackson, R.L. (1982) Biochemistry 21,4503-4511. 56 LeBoeuf, R.C., MIller, C., Shively, J.E., Schumaker, V.N., Balla, M.A. and Lusis, A.J. (1984) FEBS Lett. 170, 105 - 108. 57 Schonfeld, G., Patsch, W., Pfleger, B., Witztum, J.L. and Weidman, S.W. (1979) J. Clin. Invest. 64, 1288- 1297. 58 Marcel, Y.L., Hogue, M., Theolis, R. and Milne, R.W. (1982) J. Biol. Chem. 257, 13165- 13168. 58a Young, S.G., Bertics. S.J., Scott, T.M., Dubois, B.W., Curtis, L.K. and Witztum, J.L. (1986) J. Biol. Chem. 261, 2995-2998. 59 Deeb, S.S., Motulsky, A.G. and Albers, J.J. (1985) Proc. Natl. Acad. Sci. USA 82, 4983-4986. 60 Huang, L.S., Bock, S.C., Feinstein, S.I. and Breslow, J.L. (1985) Proc. Natl. Acad. Sci. USA 82, 6825 - 6829. 61 Mehrabian, M., Schumaker, V.N., Fareed, G.C., West, R., Johnson, D.F., Kirchgessner, T., Lin, H.C., Wang, X.,Ma, Y., Mendiaz, E. and Lusis, A.J. (1985) Nucl. Acids Res. 13, 6937-6953. 62 Knott, T.J., Rall, S.C., Jr., Innerarity, T.L., Jacobson, S.F., Urdea, M.S., Levy-Wilson;, B., Powell, L.M., Pease, R.J., Eddy, R., Nakai, H., Beyers, M., Priestley, L.M., Robertson, E., Rall, L.B., Bertsholtz, C, Shows, T.B., Mahley, R.W. and Scott, J. (1985) Science 230, 37-43. 63 Law, S.W., Lackner, K.J., Hospattankar, A.V., Anchors, J.M., Sakaguchi, A.Y., Naylor, S.L. and Brewer, H.B. (1985) Proc. Natl. Acad. Sci. USA 82, 8340-8344. 64 Shoulders, C.C., Myant, N.B., Sidoli, A., Rodriguez, J.C., Cortese, C., Cortese, R. and Baralle, F.E. (1985) Atherosclerosis , in press. 64a Wei, C.-F., Chen, S.H., Yang, C.-Y., Marcel, Y.L., Milne, R.W., Li, W.-H., Sparrow, J.T., Gotto, A.M., Jr. and Chan, L. (1985) Proc. Natl. Acad. Sci. USA 82, 7265-7269. 65 Innerarity, T.L., Friedlander, E.J., Rall, S.C., Jr., Weisgraber, K.H. and Mahley R.W. (1983) J. Biol. Chem. 258, 12341 - 12347. 40
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80
81 82 83
84 85 86 87 88 89 90 91
Weisgraber, K.H., Innerarity, T.L., Harder, K. J., Mahley, R.W., Milne, R.W., Marcel, Y.L. and Sparrow, J.T. (1983) J. Biol. Chem. 258, 12348 - 12354. Innerarity, T.L., Weisgraber, K.H., Arnold, K.S., Rall, S.C., Jr. andMahley, R.W. (1984) J. Biol. Chem. 259, 7261 -7267. Yamamoto, T., Davis, C.G., Brown, M.S., Schneider, W.J., Casey, M.L., Goldstein, J.L. and Russell, D.W. (1984) Cell 39, 27-38. Berg, K. (1983) in: Progress in Medical Genetics (A.G. Steinberg, A.G. Bearn, A.R. Motulsky and B. Childs, Eds.) pp. 35 -90, W.B. Saunders and Co., Philadelphia. Schumaker, V.N., Robinson, M.T., Curtiss, L.K., Butler, R. and Sparkes, R.S. (1984) J. Biol. Chem. 259, 6423 -6430. Young, S.G., Bertics, S.J., Curtiss, L.K., Casal, D.C. andwitztum, J.L. (1986)Proc. NatLAcad. Sci. USA 83, 1101 - 1105. Malloy, M.J., Kane, J.P., Hardman, D.A., Hamilton, R.L. andDolal, K.B. (1981)J. Clin. Invest. 67, 1441 - 1450. Herbert, P.N., Hoyams, J.S., Bernier, D.N., Berman, M.M., Saritelli, A.L., Lynch, K.M., Nichols, A.V. and Forte, T.M. (1985) J. Clin. Invest. 76, 403 -412. Takashima, Y., Kodama, T., Iida, H., Kawamura, M., Aouratini, H., Itakura, H., Akanuma, Y., Takaku, F. and Kawade, M. (1985) Pediatrics 75, 541 -546. Sniderman, A.D., Shapiro, S., Marpole, D., Skinner, B., Teng, B. and Kwiterovich, P.O. (1980) Proc. Natl. Acad. Sci. USA 77, 604-608. Goldstein, J.L., Schrott, H.G., Hazzard, W.R., Bierman, E.L. and Motulsky, A.R. (1973) J. Clin. Invest. 52, 1544- 1568. Knott, T.J., Robertson, M.E., Priestley, L.M., Urdea, M., Wallis, S. and Scott, J. (1984) Nucl. Acids Res. 12, 3909-3915. Tata, F., Henry, I., Markham, A.F., Wallis, S.C., Weil, D., Grzeschik, K.H., Junien, C., Williamson, R. and Humphries, S.E. (1985) Hum. Genet. 69, 345-349. Jackson, R.L., Sparrow, J.T., Baker, H.N., Morrisett, J., Taunton, O.D. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 5308-5313. Shulman, R.S., Herbert, P.N., Wehrly, K. and Fredrickson, D.S. (1975) J. Biol. Chem. 280, 182 - 190. Paik, Y.K., Lauer, S.J., Levy-Wilson, B., Chang, D.J., Elshourbagy, N.A., Mahley, R.W., Reardon, C.A. and Taylor, J.M. (1985) Arteriosclerosis 5 , 503a. Das, H.K. and Breslow, J.L. (1985) Arteriosclerosis 5 , 503a. Nikkila, E.A. (1983) in: The Metabolic Basis of Inherited Disease (J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein and M.D. Brown, Eds.) pp. 622 - 642, McGrawHill, New York. Breckenridge, W.C., Little, J.A., Steiner, G., Chow, A. and Poapst, M. (1978) N. Engl. J. Med. 298, 1265 - 1273. Musliner, T.A., Church, E.C., Herbert, P.N., Kingston, M.J. and Shulman, R.S. (1977) Proc. Natl. Acad. Sci, USA 74, 5358 - 5362. Kinnunen, P.K.J., Jackson, R.L., Smith, L.C., Gotto, A.M. andsparrow, J.T. (1977)Proc. Natl. Acad. Sci. USA 74, 4848 -4851. Vainio, P., Virtanen, J.A., Kinnunen, P.K.J., Gotto, A.M., Jr., Sparrow, J.T., Pattus, F., Bougis, P. and Verger, R. (1983) J . Biol. Chem. 258, 5477-5482. Jackson, C.J., Bruns, G.A.P. and Breslow, J.L. (1984) Proc. Natl. Acad. Sci. USA 81, 2945 - 2949. Myklebost, O., Williamson, B., Markham, A.F., Myklebost, S.R., Rogers, J., Woods, D.E. and Humphries, S.E. (1984) J. Biol. Chem. 259, 4401 -4404. Fojo, S.S., Law, S.W. andBrewer, H.B., Jr. (1984)Proc. Natl. Acad. Sci. USA 81, 6354-6357. Hospattankar, A.V., Fairwell, T., Ronan, R. and Brewer, H.B., Jr. (1984) J. Biol. Chem. 259, 3 18 - 322.
92 Jackson, R.L., Baker, H.N., Gilliam, E.B. and Gotto, A.M. (1977) Proc. Natl. Acad. Sci. USA 74, 1942- 1945. 93 Wei, C.F., Tsai, Y.K., Robberson, D.L., Gotto, A.M., Jr., Brown, K. andChan, L. (1985) J. Biol. Chem. 260, 15211 - 15221. 94 Humphries, S.E., Williams, L., Myklebost, O., Stalenhoef, A.F.H., Demacker, P.N.M., Baggio, G., Crepaldi, G., Galton, D.J. and Williamson, R. (1984) Hum. Genet. 67, 151 - 155. 95 Maguire, G.F., Little, J.A.. Kakis, G. and Breckenridge, W.C. (1984) Can. J. Biochem. Cell Biol. 62, 847-852. 96 Fojo, S.S., Law, S.W., Baggio, G., Sprecher, D.L., Taam, L. and Brewer, H.B., Jr. (1985) Clin. Res. 33, 569a. 97 Stalenhoef, A.F.H., Casparie, A.F., Demacker, P.N.M., Stouten, J.T.J., Lutterman, J.A. and Van’t Laar, A. (1981) Metabolism 30, 919-923. 98 Havel, R.J., Kotite, L. and Kane, J.P. (1979) Biochem. Med. 21, 121 - 128. 99 Menzel, H.J., Kane, J.P., Malloy, M.J. and Havel, R.J. (1986) J. Clin. Invest. 99, 595-601. 100 Wang, C.S., McConathy, W.J., Kloer, H.U. and Alaupovic, P. (1985) J. Clin. Invest. 75, 384 - 390. 101 Norum, R.A., Lakier, J.B., Goldstein, S., Angel, A., Goldberg, R.B., Block, W.D., Noffze, D.K., Dolphin, P.J., Edelglass, J., Bogorad, D.D. and Alaupovic, P. (1982) N. Engl. J. Med. 306, 1513- 1519. 102 Ginsberg, H.L., Le, N.A., Norum, R.A., Gibson, J. and Brown, W.V. (1984) Clin. Res. 32,396a. 103 Sparrow, J.T., Pownall, H.J., Hsu, F.J., Blumenthal, L.E., Culwell, A.R. andGotto, A.M. (1977) Biochemistry 16, 5427 - 5431. 104 Karathanasis, S.K., McPherson, J., Zannis, V.1. and Breslow, J.L. (1983) Nature 304, 371 -373. 105 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1985) J. Lipid Res. 26, 451 -456. 106 Levy-Wilson, B., Appleby, V., Protter, A., Auperin, D. and Seilhamer, J.J. (1984) DNA 3, 359 - 364. 107 Brewer, H.B., Shulman, R., Herbert, P., Ronan, R. and Wehrly, K. (1974) J. Biol. Chem. 249, 4975 - 4984. 108 Blaufuss, M.C., Gordon, J.I., Schonfeld, G., Straws, A.W. and Alpers, D.H. (1984) J. Biol. Chem. 259, 2452 - 2456. 109 Rees, A., Shoulders, C.C., Stocks, J., Galton, D.J. and Baralle, F.E. (1983) Lancet 1 #8322, 444 - 446. 110 Zannis, V.I., Just, P.W. and Breslow, J.L. (1981) Am. J. Hum. Genet. 33, I 1 -24. 111 Zannis, V.I. and Breslow, J.L. (1981) Biochemistry 21, 1033-1041. 112 Utermann, G., Jaeschke, M. and Mangel, J. (1975) FEBS Lett. 56, 352-355. 113 Jain, R.S. and Quarfordt, S.H. (1979) Life Sci. 25, 1315 - 1323. 114 Utermann, G., Langenback, U., Beisiegel, U. and Weber, W. (1980) Am. J. Hum. Genet. 32, 339- 347. 115 Zannis, V.I., McPherson, J., Goldberger, G., Karathanasis, S.K. and Breslow, J.L. (1984) J. Biol. Chem. 259, 5495 - 5499. 116 Pitas, E., Innerarity, T.L., Arnold, K.S. and Mahley, R.W. (1981) Proc. Natl. Acad. Sci. USA 76, 2311-2315. 117 Bersot, T.P., Mahley, R.W., Brown, M.S. and Goldstein, J.L. (1976) J. Biol. Chem. 251, 2395 - 2398. 118 Innerarity, T.L. and Mahley, R.W. (1978) Biochemistry 17, 1440- 1447. 119 Hui, D.Y., Innerarity, T.L. and Mahley, R.W. (1981) J . Biol. Chem. 256, 5646-5655. 120 Carrela, M. and Cooper, A.D. (1979) Proc. Natl. Acad. Sci. USA 76, 338-342. 121 Sherrill, B.C., Innerarity, T.L. and Mahley, R.W. (1980) J. Biol. Chem. 255, 1804- 1807. 122 Rall, S.C., Weisgraber, K.H. and Mahley, R.W. (1981) J . Biol. Chem. 257, 4171 -4178. 123 Mahley, R.W. and Angelin, B. (1984) Adv. Int. Med. 29, 385-411. 124 Rall, S.C., Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1985) Arteriosclerosis 5, 504a.
125 Blue, M.L., Williams, D.L., Zucker, S., Khan, S.A. and Blum, C.B. (1983) Proc. Natl. Acad. Sci. USA 80, 283-287. 126 Basu, S.K., Brown, M.S., Ho, Y.K., Havel, R.J. and Goldstein, J.L. (1981) Proc. Natl. Acad. Sci. USA 78, 7545 - 7549. 127 Breslow, J.L., McPherson, J., Nussbaum, A.L., Williams, H.W., Lofquist-Kahl, F., Karathanasis, S.K. and Zannis, V.I. (1982) J. Biol. Chem. 257, 14639- 14641. 128 Breslow, J.L., McPherson, J . , Nussbaum, A.L., Williams, H.W., Lofquist-Kahl, F., Karathanasis, S.K. and Zannis, V.I. (1983) J . Biol. Chem. 258, 11422. 129 Wallis, S.C., Rogne, S., Gill, L., Markham, A., Edge, M., Woods, D., Williamson, R. and Humphries, S. (1983) EMBO J. 2, 2369-2373. 130 McLean, J.W., Elshourbagy, N.A. Chang, D.J., Mahley, R.W. and Taylor, J.M. (1984) J. Biol. Chem. 259, 6498 - 6504. 131 Das, H.K., McPherson, J . , Bruns, G.A.P., Karathanasis, S.K. and Breslow, J.L. (1985) J. Biol. Chem. 260, 6240-6247. 132 Paik, Y.K., Chang, D.J., Reardon, C.A., Davies, G.E., Mahley, R.W. and Taylor, J.M. (1985) Proc. Natl. Acad. Sci. USA 82, 3445 - 3449. 133 Zannis, V.I., Breslow, J .L., Utermann, G., Mahley, R. W., Weisgraber, K.H., Havel, R.J., Goldstein, J.L., Brown, M.S., Schonfeld, G., Hazzard, W.R. and Blum, C.B. (1982) J. Lipid Res. 23, 91 1 - 914. 134 Utermann, G., Kinderrnann, I., Kaffarnik, H. and Steinmetz, A. (1984) Hum. Genet. 65, 232 - 236. 135 Assrnann, G., Schmitz, G., Menzel, H.J. and Schulte, H. (1984) Clin. Chem. 30/5, 641 -643. 136 Wardell, M.R., Suckling, P.A. and Janus, E.D. (1982) J. Lipid Res. 23, 1174-1182. 137 Cumming, A.M. and Robertson, F.W. (1984) Clin. Genet. 25, 310-313. 138 Ehnholm, C., Lukka, M., Kuusi, T., Nikkila, E. and Utermann, G. (1986) J . Lipid Res. 29, 229 - 235. 139 Zannis, V.I. and Breslow, J.L. (1980) J . Biol. Chem. 255, 1759-1762. 140 Breslow, J.L., Zannis, V.I., SanGiacomo, T.R., Third, J.L.H.C., Tracy, T. and Glueck, C.J. (1982) J . Lipid Res. 23, 1224- 1235. 141 Schneider, W.J., Kovanen, P.T., Brown, M.S., Goldstein, J.L., Utermann, G., Weber, W., Havel, R.J., Kotite, L., Kane, J.P., lnnerarity, T.L. and Mahley, R.W. (1981) J. CIin. Invest. 68, 1075 - 1085. 142 Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1982) J . Biol. Chem. 257, 2518-2521. 143 Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1982) Proc. Natl. Acad. Sci. USA 79, 4696-4700. 144 Yamamura, T., Yamamoto, A., Hiramori, K. and Nambu, S. (1984)Atherosclerosis 50, 159- 172. 145 Yamamura, T., Yamamoto, A., Sumiyoshi, T., Hiramori, K., Nishioeda, Y. and Nambu, S. (1984) J. Clin. Invest. 74, 1229- 1237. 146 Innerarity, T.L., Weisgraber, K.H., Arnold, K.S., Rall, S.C., Jr. and Mahley, R.W. (1984) J. Biol. Chem. 259, 7261 -7267. 147 Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., Bersot, T.P., Mahley, R.W. and Blum, C.B. (1983) J . Clin. Invest. 72, 1288- 1297. 148 Wardell, M.R. and Janus, E.D. (1985) in: Proceedings of the 7th International Symposium on Atherosclerosis (P.J. Nestel, Ed.), International Atherosclerosis Society. 149 Weisgraber, K.H., Rall, S.C., Jr., Innerarity, T.L., Mahley, R.W., Kuusi, W. and Ehnholm, C. (1984) J. Clin. Invest. 73, 1024- 1033. 150 Ghiselli, G., Schaefer, E.J., Gascon, P . and Brewer, H.B., Jr. (1981) Science 214, 1239- 1241. 151 Zannis, V.I., Ordovas, J.M., Cladaras, C., Cole, F.S., Forbes, G. and Schaefer, E.J. (1985) J. Biol. Chem. 260, 12891 - 12894. 152 Moore, M.N., Kao, F.T., Tsao, Y.K. and Chan, L. (1984) Biochem. Biophys. Res. Commun. 123, 1-7.
153 Lackner, K.J., Law, S.W., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 122, 877-883. 154 Knott, T.J., Eddy, R.L., Robertson, M.E., Priestley, L.M., Scott, J. and Shows, T.G. (1984) Biochem. Biophys. Res. Commun. 125, 299- 306. 155 Scott, J., Knott, T.J., Priestley, L.M., Robertson, M.E., Mann, D.V., Kostner, G., Miller, G.J. and Miller, N.E. (1985) Lancet i, 771-773. 156 Huang, L.S., Miller, D.A., Bruns, G.A.P. and Breslow, J.L. (1986) Proc. Natl. Acad. Sci. USA 83, 644-648. 157 Law, S.W., Lee, N., Monge, J.C. and Brewer, H.B., Jr. (1985) Biochem. Biophys. Res. Commun. 131, 1003 - 1012. 157a Chan, L. Van Tuinen, P., Ledbetter, D.H., Daiger, S.P., Gotto, A.M., Jr. and Chen, S.H. (1985) Biochem. Biophys. Res. Commun. 133, 248 -255. 158 Bruns, G.A.P., Karathanasis, S.K. and Breslow, J.L. (1984) Arteriosclerosis 4, 97- 102. 159 Law, S.W., Gray G., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 118, 934-943. 160 Cheung, P., Kao, F.T., Law, M.L., Jones, C., Puck, T.T. and Chan, L. (1984) Proc. Natl. Acad. Sci. USA 81, 508-511. 161 Lusis, A.J., Taylor, B.A., Wangenstein, R.W. and LeBoeuf, R.L. (1983) J. Biol. Chem. 258, 5071 -5078. 162 Antonucci, T.K., VonDeimling, O.H., Rosenblum, B.B., Skow, L.C. and Meisler, M.H. (1984) Genetics 107, 463 - 475. 163 Meisler, M.H., Wanner, L., Kao, F.T. and Jones, C. (1981) Cytogenet. Cell Genet. 31, 124- 128. 163a Harper, M.E. and Chan, L. (1986) Methods Enzymol. 128, 863-876. 164 Schamaun, O., Olaisen, B., Mevag, B., Gedde-Dahl, T., Jr., Ehnholm, C. and Teisberg, P. (1984) Hum. Genet. 68, 181 - 184. 165 Olaisen, B., Teisberg, P. and Gedde-Dahl, T. (1982) Hum. Genet. 62, 233-236. 166 Whitehead, A.S., Bruns, G.A.P., Markham, A.P., Colten, H.R. and Woods, D.E. (1983) Science 221, 69-71. 167 Lin-Lee, Y.C., Kao, F.T., Cheung, P. and Chan, L. (1985) Biochemistry 24, 3751 -3756. 168 Scott, J., Knott, T.J., Shaw, D.J. and Brook, J.D. (1985) Hum. Genet. 71, 144-146. 169 Jeanpierre, M., Weil, D., Hors-Cayla, M.C., Williamson, R., Junien, C. and Humphries, S.E. (1984) Som. Cell Mol. Genet. 10, 645 - 649. 170 Fojo, S.S., Law, S.W., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 122, 687-693. 171 Humphries, S.E., Berg, K., Gill, L., Cumming, A.M., Robertson, F.W., Stalenhoef, A.F.H., Williamson, R. and Borresen, A.L. (1984) Clin. Genet. 26, 389-396. 172 Myklebost, O., Rogne. S., Olaisen, B., Gedde-Dahl, T., Jr. and Prydz, H. (1984) Hum. Genet. 67, 309-312. 173 Francke, U., Brown, M.S. and Goldstein, J.L. (1984) Proc. Natl. Acad. Sci. USA 81,2826-2830. 174 Gedde-Dahl, T., Jr ., Olaisen, B., Teisberg, P., Wilhelmy, M.C., Mevag, B. and Helland, R. (1984) Hum. Genet. 67, 178- 182. 175 Brook, J.D., Shaw, D.J., Meredith, L., Bruns, G.A.P. and Harper, P.S. (1984) Hum. Genet. 68, 282 - 285. 175a Luo, C.-C., Li, W.-H., Moore, M.N. and Chan, L. (1986) J. Mol. Biol. 187, 325-340. 176 Kessling, A.M., Horsthemke, B. and Humphries, S.E. (1985) Clin. Genet. 28, 296-306. 177 Rees, A., Stocks, J., Sharpe, C.R., Vella, M.A., Shoulders, C.C., Katz, J., Jowett, N.I., Baralle, F.E. and Galton, D.J. (1985) J. Clin. Invest. 76, 1090-1095. 178 Anderson, R.A. and Wallace, R.B. (1985) Clin. Res. 33, 328a. 179 Ferns, G.A.A., Stocks, J., Ritchie, C. and D.J. Galton (1985) Lancet ii, 300-303. 180 Wallis, S.C., Donald, J.A., Forrest, L.A., Williamson, R. and Humphries, S.E. (1984) Hum. Genet. 68, 286 - 289.
3 97 181 Lusis, A.J. and LeBoeuf, R.C. (1986) in: Methods in Enzymology (J.P. Segrest and J.J. Albers, Eds.) pp. 877 - 894, Academic Press, London. 182 Priestley, L., Knott, T., Wallis, S., Powell, L., Pease, R. and Scott, J. (1985) Nucl. Acids Res. 13, 6790. 183 Huang, L.S. and Breslow J.L. (1987) J. Biol. Chem., submitted. 184 Priestley, L., Knott, T., Wallis, S., Powell, L., Pease, R., Brunt, H. and Scott, J. (1985) Nucl. Acids Res. 13, 6793. 185 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1983) Nature 301, 718-720. 186 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1983) Nature 305, 823-825. 187 Karathanasis, S.K., Frazer, C., Zannis, V.I. and Breslow, J.L. (1984) Arteriosclerosis 4, 562a.
‘91987 Elsevier Science Publishers B.V. (Biomedical Division)
359 CHAPTER 12
Lipoprotein genetics and molecular biology JAN L. BRESLOW The Rockefeller University, New York, NY, USA
1. Perspectives and summary Apolipoproteins are important structural constituents of lipoprotein particles and have been shown to participate in lipoprotein synthesis, secretion, processing and catabolism. These subjects have been recently reviewed [ l , 21. In the last 15 years, advances in protein chemistry techniques have allowed the identification, isolation, and characterization of at least eight apolipoproteins (Table 1). Protein sequencing techniques have been used to derive the primary amino acid sequence of the plasma form of six of these polypeptides. In the last few years, cDNA and genomic clones have been derived for each of the apolipoproteins. The DNA sequences combined with cell-free synthesis and tissue and organ culture studies have revealed the presence of apolipoprotein precursors containing NH2-terminal extensions, including prepeptides and in some cases propeptides (Table 2). In addition, the apolipoprotein genes have been mapped in the human genome. Finally, human mutations in the apolipoprotein genes have been identified at both the amino acid and the DNA level. Some of these have profound effects on lipoprotein metabolism and are associated with premature atherosclerosis. This chapter will review current knowledge of human apolipoprotein gene structure and genetic variation. For additional background and references, several other recent reviews should be consulted
[3-51.
2. ApoA-1 ApoA-I is the major protein constituent of high density lipoproteins (HDL). HDL particles are about 50% protein and 50% lipid, and apoA-I is 70% of HDL protein. HDL levels are inversely correlated with susceptibility to coronary artery disease and recently the same association has been demonstrated for apoA-I [6, 71. ApoA-I is abundant in plasma with a concentration of 1 .O to 1.2 mg/ml. ApoA-I is thought to participate by two mechanisms in the reverse transport of cholesterol from tissues to the liver for excretion. ApoA-I can promote cholesterol efflux from tissues.
TABLE 1 Apolipoproteins and their association with human diseases
APOprotein
Plasma concentration (mg/ml)
Isoelectric point (PI)
A-I
1.0- 1.2
A-11
Function
Association with clinical disorders
5.85 - 5.40a 28K
Activates LCAT
Tangier disease; apoA-I - apoCIII deficiency; atherosclerosis
0.3 - 0.5
5.0
8.5K
-
-
A-IV
0.16
5.5
46K
-
-
B-100
0.7 - I .O
-
550K
Receptor-mediated catabolism of LDL
Abetalipoproteinemia; normotriglyceridemic abetalipoproteinemia (B-100 deficiency); atherosclerosis
-
275K
Chylomicron production
7.5
6.5K
Activates (moder-
B-48
M.W.
-
CI
0.04 - 0.06
CI1
0.03 - 0.05
4.9
9K
Activates lipoprotein lipase
Familial Type I hyperlipoproteinemia
CIII
0.12 - 0.14
4.7 - 5.0b
9K
lnhibits catabolism of triglyceride-rich lipoproteins
ApoA-I -apoCIlI deficiency
E
0.025 - 0.050
6.0- 5.7'
34.2K
Receptor-mediated catabolism of apoE-containing lipoproteins
Familial Type 111 hyperlipoproteinemia
-
ately) LCAT
a The isoelectric points of apoA-I isoproteins are: apoA-I, = 5.85; apoA-I, = 5.74; apoA-1, = 5.65; apoA-IS = 5.52; apoA-16 = 5.40. The major plasma isoprotein is apoA-I,. The isoelectric points of individual apo CIII isoproteins are: apoCIII-0 = 5.0; apoCII1-1 = 4.85; apoC1II-2 = 4.65. The isoelectric points of individual apoE3 isoproteins are: apoE3 = 6.02; apo E3,-, = 5.89; a p 0 E 3 , - ~ = 5.78; apoE3,_, = 5.68. The isoelectric points of the common apoE variants are apoE2 = 5.89; apoE4 = 6.18.
361 TABLE 2 Human apolipoprotein amino acids residues ~
ApoA-I ApoA-I1 APOA-1V ApoB-48 ApoB-100 ApoCI ApoCII ApoCI I I ApoE n.p.
-
not present; un.
Prepeptide
Propeptide
Mature protein
18 18 20
6 5 n.p. un . un. n.p. n.p. n.p. n.p.
243 77
un . un.
26 22 20 18 -
316 un. un.
57 79 79 299
unknown.
ApoA-I also displays cofactor activity for the lecithin cholesterol acyltransferase (LCAT) enzyme, which is responsible for almost all plasma cholesterol esterification. This reaction is thought to play a role in transforming nascent HDL to mature HDL particles. In mammals, apoA-I synthesis is approximately equally divided between liver and small intestine, whereas in avians other major sites of synthesis have been identified (references for background material on each of the apolipoproteins are in reviews listed as references 1 and 2, except where specifically provided).
(a) ApoA-I cDNA ApoA-I cDNA clones have been obtained and their DNA sequences derived [8 - 141. From this information, apoA-I mRNA is thought to be 893 bp in length and includes a 5' untranslated region of 35 bp, a region coding for 267 amino acids of 801 bp, a termination codon, TGA, and a 3' untranslated region of 54 bp followed by a poly A tail. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoA-I mRNA in cell-free synthesis experiments, indicates translation initiation at the methionine 24 amino acids upstream of the NH2-terminus of the mature protein. The 18 NH2-terminal amino acid polypeptide, which can be cotranslationally cleaved by microsomal membranes, represents the apoA-I prepeptide. The 6 amino acid polypeptide adjacent to the NH2-terminus of the mature apoA-I is the propeptide and has the rather unusual sequence Arg-His-Phe-Trp-GlnGln [15 - 171. The propeptide is not cleaved intracellularly, but rather is present in secreted apoA-I [16-201. Thus, it is necessary to postulate the existence of a previously unsuspected protease activity in lymph and/or plasma required to cleave this hexapeptide and generate mature apoA-I. This converting protease presumably
plays a role in apoA-I processing and may be an important determinant of apoA-I and, thereby, HDL metabolism. Protease activity, which is inhibited by EDTA, has been demonstrated in human serum and on the surface of human endothelial cells and hepatoma cells [21, 221. The reported cDNA sequences specify an amino acid sequence for mature apoA-I of 243 amino acids, which is very similar to that derived previously by protein sequencing methods [23]. The only difference is at residue 34 where the protein sequence specified Gln and the cDNA sequence indicates Glu. It had been previously noted that the apoA-I amino acid sequence from residues 99 to 230 was composed of six tandem 22 amino acid repeats, and five of the six repeats begin with proline [24, 251. Examination of the DNA sequence in this region confirms this and shows a tandemly repeated DNA structure 66 bp in length [12]. This finding suggests that this portion of the apoA-I gene arose by intragenic duplications. When the six 66 bp repeats are aligned and a consensus nucleotide at each position of the repeat derived, the consensus sequence is 64 to 80% homologous with each of the repeats [12]. Translation of the consensus sequence reveals an interesting underlying protein structure for this region of apoA-I. As noted from the protein structure, proline, an alpha-helix breaker, occurs every 22 amino acids. The intervening amino acids, when placed in an Edmundsen wheel diagram [4, 261, specify an alpha-helix with a nonpolar and a polar face. This is the general character of the amphipathic alpha-helical configuration which is a common feature of the apolipoproteins [27]. It is thought that the nonpolar face interacts with the hydrophobic lipid core of the lipoprotein particle, whereas the polar face interacts with the aqueous plasma environment. In addition, the positively charged residues tend to cluster between the nonpolar and polar faces. The latter has been shown to be important in stabilizing the lipid protein association [28]. A recent, more sophisticated, computer analysis derived to specifically look for DNA repeats in apoA-I reveals that the basic structure is a 33 bp (1 1 codon) repeat and a model based on gene duplication and unequal crossing-over events has been derived [29].
(6) ApoA-I gene The apoA-I gene has been isolated and its DNA sequence derived [12- 14, 301. From the transcription initiation to the polyadenylation site, the gene is 1863 bp in length. A comparison of the sequence of the apoA-I gene with the cDNA reveals three introns (IVS). IVS-1 is 197 bp long and occurs in the 5 ’ untranslated region between bases 20 and 21 upstream of the codon for Met that initiates translation. IVS-2 is 186 bp long and interrupts the codon specifying amino acid - 10, which is in the apoA-I prepeptide. IVS-3 is 588 bp long and interrupts the codon specifying amino acid 43 of the mature protein. The intron locations indicate that apoA-I exons may code for functionally distinct regions of apoA-I. For instance, exon 2 contains most of the apoA-I prepeptide, exon 3 contains the propeptide and the NH2-terminal sequences, whereas exon 4 contains codons for the 200 amino acids
363
which comprise the COOH-terminal portion of the molecule. The latter includes the 6 6 bp tandem DNA repeats. The apoA-I gene transcription initiation site has been designated based on the length of several apoA-I cDNA clones [14]. Upstream of the proposed apoA-I transcription initiation site is a 7 bp long AT-rich region which may be the apoA-I promoter, 'TATA box' [12- 14, 301. Another feature of the apoA-I gene is the presence of an Alu repetitive element approximately 1000 bp 3 ' to the gene [31].
(c) ApoA-I genetic variation ApoA-I is the principal structural protein in HDL. Because of the importance of HDL levels in predicting atherosclerosis susceptibility, extensive population screening for apoA-I structural variants, principally by isoelectric focusing, has been undertaken. Thus far, these studies have resulted in the discovery of at least 1 1 variants (Table 3). These variants have been shown in people who appear to be heterozygotes for one normal apoA-I structural allele and another allele that specifies a gene product that is either one charge unit more acidic or one or two charge units more basic than wild type. The acidic alleles have been designated A'Milano [32, 331, A-lMarburg [34~351, A-lMunster 2A and A-lMUnster 2B 1371. These mutations result from the following single amino acid substitutions: Arg17, Cys, LysIo7 0, Lysl,, 0, and AlaIs8 Glu, respectively. The basic alleles have been designated A-IGjessen 1341 A-lMunster 3.4, A-lMunster 3B1 A-lMunster 3C 1361, A-IMunster3D, A-IMunster4, and A-INorway[37]. These mutations result from the substitutions Prol4, Arg, Asp,,, Asn, Pro4 Arg, Pro3 His, Asp,,, Gly, GluI9, Lys, and Glu,,, Lys, respectively. A-IMiIano and A-IMarburg,
-
-
-
-
wl,
9
-
-
-
-
-
-
-
TABLE 3 ApoA-I genetic variants
Name
Charge difference
Defect
-1 -1 -1
Argl,3
___
A-lMilano A-lMarburg A-1Munster2A A-1Munster2B
A-IGiessen A-1Munsrer3A A'1Munsrer3B A-1Munster3C
A-1Munster3 D A-1Munsrer4 A-INorway
-1
+I
+I +I +I +I t 2 +2
- CYS
-0 -0 Ala,,, - Glu Prold3- Arg Asp,o3 - Asn Pro4 - Arg - His LYSIO, LYSIO,
Pro3
-GlY GluIYs -LYS
ASP213 GluI3,
- LYS
364 but not the other structural variants, have been associated with reduced HDL levels [32, 351. Normal apoA-I activates lecithin cholesterol acyltransferase and A-IGiessen,A-IMarburgand A-IMunster2A have been reported to be defective in this regard [38]. These variants have been discovered because they produce gene products with different net charge from wild type. Presumably, other apoA-I gene mutations exist that affect the protein coding region of the gene, but do not change the charge of apoA-I and are as yet undetected.
3. APOA-11 ApoA-I1 is the second most abundant protein in HDL, comprising approximately 20% of its protein. Plasma apoA-I1 concentrations are 0.3 to 0.5 mg/ml. In vitro, apoA-I1 has been shown to displace apoA-I from HDL particles, as well as both activate hepatic lipase and inhibit LCAT. However, the physiological role of apoA-I1 has not been determined, and primary qualitative or quantitative abnormalities of human apoA-I1 have not been reported. ApoA-I1 is made in the liver and intestine. (a) ApoA-II cDNA
ApoA-I1 cDNA clones have been isolated and sequenced [14, 39- 411. From this information, it has been deduced that apoA-I1 mRNA is 473 bp in length and includes a 5 ’ untranslated region of 58 bp, a region coding for 100 amino acids of 300 bp, a termination codon TGA, and a 3’ untranslated region of 112 bp followed by a poly A tail. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoA-I1 mRNA in cell-free synthesis experiments [42] indicates translation initiation at the methionine 23 amino acids upstream of the NH2terminus of the mature protein. Cotranslational cleavage of the primary translation product by microsomal membranes removes the 18 NH2-terminal amino acids which presumably represent the apoA-I1 prepeptide. The remaining 5 amino acid polypeptide adjacent to the NH2-terminus of mature apoA-I1 is a propeptide with the sequence Ala-Leu-Val-Arg-Arg. The occurrence of two basic amino acids in the propeptide adjacent to the NH2-terminus of the mature protein is rather typical of propeptides and quite different from the apoA-I propeptide [15 - 171. Therefore, apoA-I1 should be cleaved intracellularly, in contrast to apoA-I. However, studies with the human hepatoma cell line, HepG2 [43], indicate that only a fraction of proapoA-I1 is cleaved intracellularly with the rest cleaved extracellularly by cathepsin B. The significance of these extracellular processing events has yet’to be elucidated. The cDNA sequence specifies an amino acid sequence for mature apoA-I1 of 77
365 amino acids, very similar to the previously reported amino acid sequence determined by protein sequencing methods [44]. The difference was at residue 35, where the DNA sequence predicts Glu, and the protein-derived sequence specified Gln. In human plasma, apoA-I1 exists as a dimer, the monomers are connected by a disulfide bridge between the only cysteine in the mature protein which resides at amino acid residue 6 [44]. (b) ApoA-11 gene
The apoA-I1 gene has been isolated and sequenced [45 - 471. From the transcription initiation to the polyadenylation site, the gene is either 1330 or 1343 bp according t o two different reports. A comparison of the apoA-I1 gene and cDNA sequences reveals three introns. IVS-1 is 169 or 182 bp in length and occurs in the 5 ‘ untranslated region between bases 24 and 25 upstream of the codon for Met that initiates translation. IVS-2 is 293 bp long and interrupts the codon specifying amino acid - 6, which is the apoA-I1 prepeptide. IVS-3 is 395 bp long and interrupts the codon specifying amino acid 39 of the mature protein. The intron locations are remarkably similar to those for the apoA-I gene and suggest, as specified for apoAI , that the apoA-11 exons may code for functionally distinct regions. The apoA-I1 transcription initiation site has been designated by using HepG2 cell mRNA for primer extension and Sl nuclease protection experiments. Upstream of this site in the appropriate location there is a 6 bp long AT-rich region, which is presumably the ‘TATA box’ portion of the apoA-I1 gene promoter. Upstream of the ‘TATA box’ is a sequence resembling another regulatory element in some eukaryotic gene promoters, the ‘CAT box’. Finally, an Alu repetitive element has been identified between approximately 300 and 650 bp 3 ’ to the apoA-I1 gene.
4. APOA-IV ApoA-IV was originally discovered as a major constituent of rat HDL. In humans, it is a relatively minor HDL protein with most apoA-IV in the nonlipoprotein plasma fraction. Human plasma apoA-IV concentrations are approximately 0.16 mg/ml. In rats, apoA-IV synthesis occurs both in intestine and liver, and in this animal fat feeding doubles intestinal apoA-IV synthesis [48]. On this basis, it is hypothesized that apoA-IV plays a role in the synthesis and secretion of intestinal triglyceride-rich lipoproteins, but direct proof of this or any other functional role for apoA-IV is lacking.
(a) ApoA-IV cDNA Human apoA-IV cDNA sequence has been recently reported [49, 501. From this in-
formation, human apoA-IV mRNA is thought t o be 1469 bp in length, and includes a 5 ’ untranslated region of 113 bp, a region coding for 396 amino acids of 1188 bp, a termination codon, TGA, and a 3 ’ untranslated region of 165 bp followed by a poly A tail. In other experiments, utilizing intestinal mRNA, the NH2-terminal sequence of the apoA-IV primary translation product was analyzed by microsequencing techniques and compared to the NH2-terminal sequence of the mature protein [51]. In this manner, it was determined that human apoA-IV contains a prepeptide sequence 20 amino acids in length. Unlike apoA-I and apoA-11, apoA-IV does not contain a propeptide. These data combined with the cDNA sequence indicates that the mature apoA-IV protein is 376 amino acids in length. The deduced apoA-IV protein sequence was analyzed and multiple repetitions of 22 amino acid segments with amphipathic alpha-helical character were identified [49]. This is analogous to the previous findings for human apoA-I.
(b) ApoA-IV gene The human apoA-IV gene was recently isolated and characterized [49, 501. It is approximately 2600 bp in length and includes two introns [50]. IVS-1 is 359 bp long and interrupts the codon specifying amino acid - 4 in the apoA-IV prepeptide region. IVS-2 is 777 bp long and interrupts the codon for amino acid 39 of the mature protein. IVS-1 and IVS-2 of the apoA-IV gene are in strikingly similar locations to two of the introns in most of the other apolipoprotein genes, A-I, A-11, CI, CII, CIII and E. However, all of these genes also contain another intron, which interrupts the sequence coding for the 5 ’ untranslated region, which is evidently missing from the apoA-IV gene [50].
(c) Apo-A-IV genetic variation Genetic variation in human apoA-IV has been demonstrated [34, 361. In most people, isoelectric focusing of plasma apoA-IV results in a single major isoprotein of pH 5.50, other people have this isoprotein plus another, which is one charge unit more basic, and a few people have just the more basic isoprotein. In one large German study, the frequencies of these patterns in a normal population were 85.6070, 13.8% and O.6%, respectively [36]. Genetic studies were consistent with a single genetic locus two allele model [52]. From these data, the major allele frequency, specifying the more acidic gene product, was 92.5% and the minor allele frequency, specifying the more basic gene product, was 7.5%. Neither heterozygosity nor homozygosity for the minor allele have been associated with plasma lipoprotein abnormalities or atherosclerosis susceptibility.
367
5. ApoB ApoB is the major protein constituent of low density lipoproteins (LDL), but is also found in chylomicrons and VLDL. LDL particles are approximately 25% protein and 75% lipid and virtually all of the protein is apoB. LDL levels are directly correlated with coronary artery disease susceptibility and recently the same association has been demonstrated for apoB levels [6, 71. ApoB is abundant in plasma with a concentration of 0.7 to 1.0 mg/ml. ApoB synthesis appears to be required for the secretion into plasma of intestinal and hepatic triglyceride-rich lipoproteins. ApoB is also recognized by specific high affinity receptors that mediate clearance of LDL particles from plasma [53]. Human apoB is a glycoprotein which occurs in two forms, designated B-100 and B-48 [54]. B-100 is thought to be a single polypeptide of molecular weight approximately 550 000 produced primarily in the liver, whereas B-48 is approximately half that molecular weight and is produced primarily in the small intestine. Proteolytic degradation products of B-100 called B-74 and B-26 have been described, which are produced by a kallikrein-like activity present in plasma. Whether this happens in vivo remains to be determined [54a]. Classical protein chemistry techniques have provided very little information about the primary structure of apoB. The main problems were that the protein becomes quite insoluble after delipidation and standard methods of proteolytic digestion do not result in a high enough yield of unique peptides for structural studies. It has been suggested that the insolubility problem is due to an abnormal sensitivity of the protein to oxidation after delipidation, and recently, with proper precautions, soluble preparations of delipidated apoB have been obtained [55]. In addition, the use of bacterial proteases has improved the yield of unique apoB peptides [56]. A recent report provided the partial amino acid sequence of two apoB peptides [56]. The paucity of data on apoB structure, using standard protein chemistry techniques, has led to immunochemical studies to gain relevant information about this protein. In this regard, several groups have developed monoclonal antibodies to human apoB that recognize distinct epitopes, and some interesting information has been derived. In one study, the immunoreactivity of apoB in VLDL changed significantly after in vitro lipolysis, suggesting that apoB conformation might change at different stages of lipoprotein metabolism [57]. In another study, epitopes of apoB-100 have been mapped in a linear nonrepetitive array. For a subset of these epitopes, the monoclonal antibodies disrupt apoB binding to the LDL receptor, but d o not bind to B-48. For another subset of epitopes, the monoclonal antibodies bind B-48, but do not disrupt receptor binding [58]. Thus, B-48 and B-100 are antigenically related and it has been suggested that B-48 represents approximately one half of the B-100 protein and that this part of apoB is not involved in receptor binding [54, 581. Recently, a monoclonal antibody has been used to show parallel expression of an
368 epitope in both B-100 and B-48. This is very strong evidence that both B-100 and B-48 are products of the same gene [%a]. (a) ApoB cDNA
ApoB cDNA clones from human liver libraries have recently been isolated by several groups [59 - 64,64al. Two strategies were used. Oligonucleotides were synthesized based on limited apoB protein sequence data and used as probes to screen cDNA libraries. Alternatively, polyclonal antibodies to apoB were used to screen expression libraries and identify clones producing fusion proteins containing apoB epitopes. It appears that two types of cDNA clones have been identified. The first type corresponds to internal parts of the coding sequence and contains the regions specifying the previously reported apoB peptides. Deeb et al. reported a cDNA sequence coding for 197 amino acids of apoB, including the sequence of the previously reported peptide R3-1 [59]. Mehrabian et al. reported another cDNA sequence coding for 104 amino acids of apoB, which included the sequence of the peptide R2-5 [61]. These two sequences do not overlap. The second type of apoB cDNA clone corresponds to the COOH terminal amino acid part of the coding sequence and the 3 ' untranslated region. Knott et al. reported a cDNA sequence coding for the COOH terminal 1955 amino acids of apoB as well as the termination codon TAA and 266 bp of the 3 ' untranslated region [62]. The latter is imcomplete since it lacks the polyadenylation signal and tail. The sequence reported by Deeb et al. is not contained within this cDNA, but the sequence of Mehrabian corresponds to the 5 ' end of this sequence. Wei et al. reported a cDNA clone, coding for the COOH terminal 836 amino acids of apoB, derived from an expression library. This clone produces a protein that reacts with monoclonals that detect B-74 but not B-26 and establishes that B-74 corresponds to the COOH terminal end of apoB [64a]. Knott et al. identified within apoB an amino acid sequence that they propose might be the apoB/E receptor binding region [62]. This region has the sequence ThrThr-Arg-Leu-Thr-Arg-Lys-Arg-Gly-Leu-Lys (residues 276 - 286), which they note is structurally similar to the amino acid sequence of the apoE receptor binding domain (apoE residues 140- 150) [65 - 671. Both these apoB and apoE regions have similarly spaced, positively charged residues and would be ideal for interaction with the postulated ligand binding site of the apoB/E receptor, which is enriched in negatively charged amino acid residues [68]. Further proof that the proposed region of apoB is indeed the receptor binding domain will be required. Thus far, all the apoB cDNA clones described detect on Northern blot analysis of liver mRNA a species of 18 to 22 kb in length. mRNA of this length could contain enough coding information for B-100, which has an apparent molecular weight of 550 000 on SDS PAGE and may contain > 5000 amino acids. These same clones
369 have been used for Northern blot analysis with RNA from adult intestine, which makes B-48 but not B-100. Surprisingly, these studies also show an mRNA species of approximately 20 kb in length [61, 641. The explanation for this, as well as the structural relationship between B-100 and B-48 at the protein, mRNA and gene level, should be forthcoming in the near future.
(b) ApoB genetic variation Genetically determined variation in both the quality and quantity of apoB has been demonstrated. Antisera from multiply transfused patients have been used to define two series of allelic variants called Lp and Ag. Both systems appear to have marginal effects on plasma cholesterol levels and risk of atherosclerosis [69]. Recently, individual variation in apoB reactivity with an anti-apoB monoclonal antibody has been demonstrated [70, 711. Three phenotypes of strong, weak, and intermediate binding have been identified and family studies suggest a genetic basis for this consistent with a single genetic locus with two alleles specifying strong and weak binding forms of apoB. Thus, the intermediate binding pattern is the result of heterozygosity for the strong and weak binding alleles, whereas the other two patterns represent homozygosity for their respective alleles. In a study of the phenotype frequency in unrelated individuals the major allele was the one specifying weak binding, and its frequency was found to be 61% [71]. This apoB antigenic variation presumably results from an alteration in the amino acid sequence of the apoB polypeptide. This variation in apoB does not seem to affect LDL lipid composition or density [71], but whether it is associated with altered plasma lipoprotein levels has yet to be determined. The monoclonal antibody used to detect the apoB variation has been shown not to interfere with receptor binding [71]. Inherited disorders of lipoprotein metabolism associated with diminished plasma levels of apoB have been described. The most striking of these is abetalipoproteinemia [2]. In this disorder, individuals suffer from fat malabsorption and lack apoB-containing lipoproteins in their plasma, including chylomicrons, VLDL and LDL. Parents of affected individuals have normal lipoprotein and apoB levels, however, siblings with this disorder have been described and inheritance is assumed to be autosomal recessive. Immunologically detectable apoB is absent from both the plasma and tissues of these individuals, and the disorder is thought to be a genetic defect in apoB synthesis. A phenotypically similar disorder has been described, termed homozygous hypobetalipoproteinemia, in which parents of affected individuals have half-normal levels of LDL cholesterol and apoB [2]. This condition may also involve a genetic defect in apoB synthesis, but by a different mechanism. Finally, individuals have been described with normal fat absorption and the ability to produce chylomicrons, but low to absent levels of LDL cholesterol [72 - 741. Apparently, these individuals can produce the intestinal form of apoB, B-48, but not the hepatic form, B-100. The existence of this disorder suggests separate genetic con-
370 trol of B-48 and B-100 synthesis. However, at this time, it cannot be determined whether these two gene products are the result of distinct genetic loci or represent differential splicing of the transcript produced from a single genetic locus. Increased apoB levels are associated with atherosclerosis susceptibility [6, 71. In familial hypercholesterolemia (FH), an autosomal dominant disorder characterized by premature atherosclerosis, apoB and cholesterol in LDL are elevated and the genetic lesion is a defect in the apoB/E receptor [53]. A substantial fraction of nonFH individuals with coronary disease have been shown to have increased apoB, but normal cholesterol in LDL. This phenotype has been called hyperapobetalipoproteinemia [75]. A subset of these individuals may have the autosomal dominant disorder associated with premature atherosclerosis, familial combined hyperlipidemia (FCHL) [76]. Plasma from these individuals consistently contains elevated plasma apoB levels but only occasionally is the LDL cholesterol elevated [77]. Genetic abnormalities associated with apoB may be one possible explanation for the hyperapobetalipoproteinemia phenotype, but this has not been proven.
6. ApoCI ApoCI is a constituent of VLDL and HDL. VLDL particles are approximately 10% protein and 90% lipid and apoCI is 10% of VLDL protein. As previously noted, HDL particles are approximately 50% protein and 50% lipid and apoCI is 2% of HDL protein. Human plasma apoCI concentrations are in the range of 0.04 to 0.06 mg/ml. In vitro apoCI has been shown to activate LCAT, but not as efficiently as apoA-I. The physiological role(s) of apoCI has not been defined. Synthesis mainly occurs in liver and to a minor degree in intestine, but has not been evaluated in other organs. Primary qualitative or quantitative abnormalities of human apoCI have not been reported. (a) ApoCI cDNA and gene
ApoCI cDNA clones have been isolated [77], From this information, apoCI mRNA is thought to be approximately 419 bp in length, including a 5' untranslated region of 56 bp, a region coding for 83 amino acids of 249 bp, a termination coding, TGA, and a 3' untranslated region of 111 bp followed by a poly A tail. The DNA-derived amino acid sequence contains the 57 residues of mature apoCI and agrees entirely with the results derived by protein sequencing techniques [79, 801. The DNA sequence also specifies a 26 amino acid NH2-terminal extension [77, 781. This amino acid sequence is compatible with the entire 26 amino acids being the apoCI prepeptide. Recently, apoCI genomic clones have been isolated and partially characterized [81, 821. It appears that there are two copies of the apoCI gene [81]. One of these
37 1
has been sequenced [Sl] and found to contain four exons and three introns and is 4374 bp in length. The exons are 33, 78, 136 and 169 bp in length, and the introns are 138, 1190, and 2630 bp in length. The introns are in strikingly similar locations to several of the other apolipoprotein genes with 1VS-1 interrupting the 5‘ untranslated region, 1VS-2 interrupting the region coding for the prepeptide, and IVS-3 interrupting the region coding for the mature protein. The second apoCI gene has yet to be sequenced. It is not known whether one or both genes are expressed or whether one is a pseudogene. In this regard, it is of interest that Northern blotting of adult human liver RNA with an apoCI cDNA probe indicates two species of apoCI mRNA that differ in length by approximately 20 bp [77]. In addition, primer extension on a template of human liver mRNA indicates two different transcription initiation points, also approximately 20 bp different from each other [77]. It is tempting to speculate that both apoCl genes are transcriptionally active and produce two different size and perhaps different sequence mRNA’s. This remains to be proven.
7. ApoCII ApoCII is a constituent of VLDL and HDL and comprises 10% of VLDL protein and approximately 1% of HDL protein. Human plasma apoCII concentrations are in the range of 0.03 to 0.05 mg/ml. Purified apoCII has cofactor activity for the enzyme lipoprotein lipase, which catalyzes the hydrolysis of triglycerides in chylomicrons and VLDL. The physiological importance of apoCII in activating lipoprotein lipase has been established by the finding of patients with inherited apoCII deficiency, who are severely hypertriglyceridemic and have functional lipoprotein lipase deficiency (see [83]). In such cases, both the hypertriglyceridemia and the lipase deficiency can be relieved by an exogenous source of apoCII [84]. Studies using proteolytic fragments and synthetic peptides of apoCII [85 - 871 have shown that lipoprotein lipase interacts with the COOH terminal amino acids 56 to 79. This interaction is enhanced by residues 44 to 55, which possess phospholipid binding activity. Deletion of the last three residues, 77 t o 79, prevents the protein from activating lipoprotein lipase. Synthesis of apoCII is mainly in liver and, to a minor degree, in intestine. (a) ApoCII cDNA
ApoCII cDNA clones have been isolated and characterized [14, 88 - 901. From this information, apoCII mRNA appears to be 488 t o 494 bp in length, according to different reports. This includes a 5 ’ untranslated region of 38 bp, a coding region for 101 amino acids of 303 bp, a termination codon, TAA, and a 3 ’ untranslated region of from 144 to 150 bp followed by a poly A tail. The heterogeneity in the 3 ’ untranslated region is due t o a minor variation in the polyadenylation site used between different apoCII cDNA clones.
372 The DNA-derived amino acid sequence contains the 79 residues of mature apoCII. This is in agreement with a recent result derived by protein sequencing techniques [91] but differs somewhat from the previously derived amino acid sequence [92]. The DNA sequence also specifies a 22 amino acid NH2-terminal extension compatible with the existence of an apoCII prepeptide. There is a striking homology between the NH2-terminal regions of both apoA-I and apoCII at the amino acid level. ApoA-I amino acids - 2, - 1, + 1, + 2 are identical to apoCII amino acids 5 , 6, 7 and 8 and, in each case, specify Gln-Gln-AspGlu. This region spans the site of cleavage of the protease that converts proapoA-I to mature apoA-I. This suggests that apoCII should be a substrate for the apoA-I converting protease. However, in plasma, mature apoCII is found to be the full 79 amino acids in length and is fully active as a lipoprotein lipase activator in this form. Thus, apoCII is not a physiological substrate for the apoA-I converting protease, for reasons which remain to be elucidated. Preliminary analysis of the apoA-I and apoCII cDNA sequences reveals extensive homology at the DNA level extending between apoA-I bp 102 - 285 (amino acids - 2 to 59) and apoCII bp 79 - 261 (amino acids 5 to 65). Alignment of the DNA sequences in these regions by metric analysis indicates 56% overall matching with runs of matches 11, 8 and 10 bp in length (B. Erickson and J. L. Breslow, unpublished observations). These observations suggest a close phylogenetic relationship between apoA-I and apoCII.
(b) ApoCII gene structure The apoCII gene has been isolated and sequenced “3, 931. The gene is 3320 bp in length. It contains four exons and three introns. The exons are 25, 68, 160 and 241 bp, whereas the introns are 2356, 166 and 304 bp in length. IVS-1 occurs in the 5’ untranslated region between bases 13 and 14 upstream of the codon for Met that initiates translation. IVS-2 and IVS-3 interrupt the codons specifying amino acids - 4 and 50, respectively. There are introns in similar locations in the other apolipoprotein genes, where this has been studied. The two apoCII introns interrupting the protein coding region divide the DNA sequence coding for the signal peptide from that coding for the mature protein, and the DNA sequence coding for the NH, from that coding for the COOH terminal portions of the mature protein, respectively. As previously noted, the COOH terminal amino acids appear to be the functional domain involved in binding lipoprotein lipase. The apoCII gene was found to have four Alu-type repetitive sequences in IVS-1 and another in the 3’ flanking region of the gene. In addition, IVS-3 was composed largely of six copies of a 36 to 40 bp tandem repeat. The significance, if any, of both types of repeats in the apoCII gene is not known. Finally, a 6 bp AT-rich region is present in the appropriate location upstream of the transcription initiation site, which may be part of the apoClI promoter, the so-called ‘TATA box’.
373
(c) ApoCII genetic variation Several patients have been described who lack apoCII in their plasma [83]. They are functionally lipoprotein lipase deficient and severely hypertriglyceridemic. ApoCII deficiency appears to be a recessive genetic disorder with obligate heterozygotes having half-normal apoCII levels, but normal triglyceride concentrations. Southern blot analysis with a probe made from an apoCII cDNA clone reveals that at least a subset of these patients possess the apoCII gene and that it is grossly intact [94]. Two independent families of probands with this disorder have been studied using the TaqI polymorphism t o assess disease linkage with the apoCII gene locus. In each family, cosegregation was observed, which is consistent with the defect in this condition being within or near the apoCII gene [94]. Recently, several probands with apoCII deficiency have been found to have a small amount of a mutant form of the protein in their plasma [95, 961. Presumably the defect in the apoCII gene in these individuals results in the production of a n unstable nonfunctional protein. In most probands with apoCII deficiency, it is possible to fully activate lipoprotein lipase in their post-heparin plasma by adding exogenous pre-heparin plasma and/or purified apoCII [83]. A single family has been described in which several affected individuals are apoCII deficient and where exogenous apoCII fails to activate their lipoprotein lipase [97]. Perhaps in this family the apoCII gene defect is linked t o a defect in the lipoprotein lipase gene or another gene contributing to lipase activation. The gene for lipoprotein lipase is not yet cloned, and it is not known whether it localizes to the same region of the genome as the apoCII gene. A protein polymorphism of apoCII has also been described [98, 991. I n most individuals, plasma a p e 1 1 is a single spot on two-dimensional gels and a single band on one-dimensional size or charge separation gels. However, three hypertriglyceridemic individuals were found who had a normal apoCII component and a n apoCII isoprotein 1 charge unit more acidic than normal. It has been determined that this is due to a substitution of glutamine for lysine at residue 5 5 . The apoCII DNA sequence at this residue is AAA which specifies lysine. It is possible to explain the occurrence of glutaniine at this residue by a single base substitution in the first base of this codon, substituting C for A. Although this mutant form of apoCII was isolated from hypertriglyceridemic patients, it appears to activate lipoprotein lipase normally. Thus, the relationship between the amino acid subsitution and the hypertriglyceridemia probably is not cause and effect.
8. ApoCIII ApoCIII is a constituent of VLDL and HDL and comprises about 50% of VLDL protein and 2% of HDL protein. Human plasma apoCIII concentrations are in the range of 0.12 to 0.14 mg/ml. ApoCIII is a glycoprotein containing 1 mol each of
374 galactose, galactosamine and either 0, 1 or 2 mol of sialic acid. The three resultant isoproteins recognizable by isoelectric focusing are designated CIII-0, CIII- 1 and CIII-2 and comprise 14%, 59% and 27% of plasma apoCIII, respectively. In vitro apoCIII has been shown to inhibit the activities of both lipoprotein lipase and hepatic lipase. ApoCIII has also been shown to decrease the uptake of lymph chylomicrons by the perfused rat liver. These in vitro studies suggest that apoCIII might delay catabolism of triglyceride-rich particles. Recently, hypertriglyceridemic subjects were shown t o have circulation lipoprotein and non-lipoprotein inhibitors of lipoprotein lipase [ 1001. The lipoprotein-associated inhibition correlated best with apoCIII concentration. In the same study, apoCIII was shown to be a noncompetitive inhibitor of the activity of partially purified lipoprotein lipase. In addition, patients with combined apoA-I, apoCIII deficiency were shown to have low plasma triglyceride levels [ 1011, and in vivo studies showed that they rapidly convert VLDL to LDL [102]. In vitro lipolysis of their VLDL was inhibited by added apoCIII [ 1021. Thus, it appears that primary abnormalities in the quantity or quality of apoCIII may affect plasma triglyceride levels and the physiological role of apoCIII may be in the regulation of the catabolism of triglyceride-rich lipoproteins. Functional domains of apoCIII have been demonstrated. The NH2-terminal 40 amino acids do not bind phospholipid, whereas the COOH-terminal 39 amino acids do [103]. Synthesis of apoCIII is mainly in liver and to a lesser degree in intestine. (a) ApoCIII cDNA
ApoCIII cDNA clones have been isolated and sequenced [14, 104- 1061. Based on this information, the apoCIlI genomic sequence (see below), and knowledge of the consensus sequence for transcription initiation, the apoCIII 5 ’ untranslated region is 49 bp. This is followed by a coding region for 99 amino acids of 297 bp, a termination codon, TGA, and a 3 ’ untranslated region of from 183 to 189 bp followed by a poly A tail. The heterogeneity in the 3 ’ untranslated region is due to a minor variation in the polyadenylation site used between apoCIII cDNA clones. Thus, apoCIII mRNA is 532 to 538 bp in length. The DNA-derived apoCIll amino acid sequence differs from the previously reported protein-derived apoCllI amino acid sequence [lo71 at residues 32, 33, 37, and 39. At this location, the DNA sequence predicts Glu, Ser, Gln, Ala, respectively, whereas the previously reported protein-derived, sequence specified Ser, Gln, Ala, Gln, respectively. Several cDNA clones from different cDNA libraries [14, 105, 1061 all have shown the same DNA-derived amino acid sequence. In addition, the DNA sequence coding for residues 32 and 33, GAGTCC, includes a recognition site for the restriction endonuclease HinfI (GANTC), whereas the DNA sequence required to code for the corresponding residues in the protein-derived sequence could not possibly contain a HinfI recognition site [ 1051. Southern blotting analysis of genomic DNA from four normal and two hypertriglyceridemic individuals identifies
315 a HinfI site in this region, also compatible with the DNA-derived amino acid sequence [105]. Thus, the protein sequence, unless it was derived from a person homozygous for a rare apoCIII allele, is probably in error and should be revised. The DNA-derived amino acid sequence indicates a 20 amino acid NH2-terminal extension for the primary translation product of apoCIII [14, 105, 1061. The sequence is compatible with previously reported prepeptide sequences. Cell-free synthesis experiments using mRNA from rat liver and intestine indicates that rat apoCIII is made with a 20 amino acid NHz-terminal extension [108]. This can be co-translationally cleaved by signal peptidase t o yield a product with the same NH2-terminus as the mature protein. Therefore, apoCIII is made as a preprotein and does not contain a propeptide sequence.
(6) ApoCIII gene The apoCIII gene has been isolated and sequenced [31, 1051. The gene is approximately 3133 bp in length and contains four exons and three introns. The exons are 36, 68, 124 and 308 bp in length. IVS-1 is 625 bp long and occurs in the 5 ’ untranslated region between bases 13 and 14 upstream of the codon for Met that initiates translation. IVS-2 is 135 bp long and interrupts the codon specifying amino acid -2, which is in the apoCIII prepeptide. IVS-3 is 1837 bp long and interrupts the codon specifying amino acid 40 of the mature protein. The introns are in similar locations to those in most of the other apolipoprotein genes characterized. The intron locations indicate that apoCIIl exons may code for functionally distinct domains of the protein. For instance, intron 2 separates the prepeptide from the mature protein, and intron 3 seems to separate the phospholipid binding domain from the non-binding domain. Another feature of the apoCIII gene is the presence of an Alu repetitive element in the third intron and another approximately 1500 bp 3 ’ to the gene. Finally, an 8 bp AT-rich region is present in the appropriate location upstream of the proposed transcription initiation site. This may be the apoCIII promoter, ‘TATA box’. Upstream of the ‘TATA box’ is a sequence resembling another regulatory element in some eukaryotic gene promoters, the ‘CAT box’. (c) ApoCIII genetic variation
Variation in the apoCIII gene has been documented. Four apoCIII cDNA clones and a genomic clone have been sequenced [14, 31, 105, 1061. Three sites of variation were identified (Table 4). One site of variation is in the coding region affecting the third base of a codon specifying amino acid residue 14, but in each case, the codon specifies glycine. The other two sites of variation are 31 bp apart in the 3’ untranslated region. The first of these sites affects the first base of the recognition sequence for the restriction enzymes SstI and Sac1 [104, 1091. The sequence is
376 TABLE 4 Human apoCIII genetic variation Clonea
Base pairb 151
388
419
1 (cDNA) 2 (cDNA) 3 (cDNA) 4 (cDNA) 5 (genomic a
cDNA clones were sequenced in three different laboratories. Refers to bp location in apoCIII mRNA.
GAGCTC and those individuals with a G in position 388 would have a restriction enzyme cutting site revealed by Southern blot analysis (see section on RFLP’s). Whether the bp variations identified in the 3 ’ untranslated region are themselves clinically significant remains to be determined.
9. ApoE ApoE in normal plasma is equally divided between VLDL and HDL. It comprises about 10 to 20% of VLDL protein and 1 to 2% of HDL protein. ApoE occurs in a metabolically distinct subfraction of HDL particles where it is a larger fraction of the protein. Human plasma apoE concentrations are in the range of 0.025 to 0.050 mg/ml. Two-dimensional gel electrophoresis of human plasma apoE has shown it to consist of several isoproteins which differ in size and/or charge [ 110, 1111. This is the result of both common genetic variation of apoE in the population (discussed below) and post-translational modification of apoE with carbohydrate chains containing sialic acid [110- 1141. ApoE is synthesized and secreted as sialo apoE and subsequently desialated in plasma [19, 20, 1151, but the physiological significance of this process is unknown. ApoE can be recognized by high affinity receptors and can mediate the binding, internalization, and catabolism of lipoprotein particles. ApoE can serve as a ligand for the LDL (apoB/E) receptor present on hepatic as well as on extrahepatic tissues [ 116 - 1181. Hepatic tissues also possess a high affinity receptor that recognizes particles that contain apoE, but not apoB [ 119 - 1211. This receptor is genetically distinct from the LDL receptor and has been called the chylomicron remnant or apoE receptor. Mature apoE is a 299 amino acid polypeptide [122] and the receptor binding region has been localized to the middle
377 portion of the polypeptide chain between residues 140 and 150, with residue 158 being important for the conformation of the binding domain [65 - 671. Structural mutations in apoE affect receptor recognition and are believed t o underlie Type 111 hyperlipoproteinemia (HLP), a condition associated with increased plasma levels of cholesterol and triglycerides, xanthomas, and premature atherosclerosis (for review, see [123]). Glycosylation of apoE occurs via an 0-glycosidic linkage to a single site a t Thr,9,. The sugar chain contains GlcNAc and GalNAc and one or more sialic acid residues [ 1241. ApoE synthesis occurs in liver and to a minor extent in intestine. However, in contrast to the other apolipoproteins, synthesis has been documented in a wide variety of other tissues including brain, kidney, adrenal gland and reticuloendothelial cells [125, 1261.
(a) ApoE cDNA ApoE cDNA clones have been isolated and sequenced [115, 127- 1301. From the transcription initiation point to the polyadenylation site, apoE mRNA is 1163 bp in length and includes a 5 ' untranslated region of 67 bp, a region coding for 317 amino acids of 951 bp, a termination codon, TGA, and a 3' untranslated region of 142 bp. The cDNA sequence and NH2-terminal microsequencing of the primary translation product of apoE mRNA in cell-free synthesis experiments indicates translation initiation at the methionine 18 amino acid upstream of the mature protein [115]. The NH2-terminal 18 amino acid polypeptide can be co-translationally cleaved by microsomal membranes and represents the apoE signal peptide. There is no propeptide. By analogy with both apoA-I and apoA-IV, human apoE contains eight tandem repetitions of exactly 22 amino acids from residues 62 t o 237 [131]. Only one of these repeats actually begins with proline. However, the sequence of charges of the amino acids in each repeat is strikingly similar. For example, two consecutive acidic amino acid residues occur in the same position in six of the eight repeats. When the DNA segments coding for these repeats are aligned and a consensus nucleotide at each position of the repeat derived, the consensus sequence is 51 to 75% homologous with each of the apoE repeats and 72% homologous to a similarly derived consensus sequence for the six human apoA-I DNA repeats that code for apoA-I residues 99 to 230. Thus, extreme similarity exists with respect to the 66 bp repeats in apoE, apoA-I and apoA-IV, suggestive of a common ancestral origin of this portion of these three apolipoprotein genes. As with the others, the consensus amino acid sequence for apoE, when placed in an Edmundson wheel diagram, specifies an amphipathic alpha-helix [4].
378
(6) ApoE gene structure The apoE gene has been isolated and sequenced by two different groups [131, 1321. In both reported sequences, the exons are 44, 66, 193 and 860 bp in length. There were minor differences between the intron sequences reported. The reported intron lengths were 760,1092 and 582 bp [132], and 757, 1093 and 580 bp [131], respectively. The apoE gene is, therefore, between 3593 and 3597 bp in length, and contains four exons and three introns. IVS-1 occurs in the 5’ untranslated region between bases 23 and 24 upstream of the codon for Met that initiates translation. IVS-2 interrupts the codon specifying amino acid - 4 which is in the apoE prepeptide. IVS-3 interrupts the codon specifying amino acid 61 of the mature protein. The intron locations are strikingly similar to those identified for most of the other apolipoprotein genes and, as previously suggested, may indicate that each exon codes for a functionally distinct region of apoE. The apoE transcription initiation site has been assigned to the A 44 bp upstream of the GT that begins the first intron based on S1 nuclease protection and primer extension experiments with human liver mRNA [ 131, 1321. However, both types of experiments indicated that the start site was heterogeneous and could be within a few bases on either side of the assigned cap site. In addition, the sequence TATAATT occurs beginning 33 bp upstream of the proposed transcription initiation site and is the putative apoE promoter. Furthermore, upstream of the TATA sequence there are four CCCGCC sequences, which have been shown t o play a major role in enhancing SV40 gene transcription. Finally, four Alu repetitive elements have been identified in association with the apoE gene. Two of these are located in the second intron, another is approximately 400 to 700 bp 5 ’ to the gene, and the fourth is approximately 130 to 400 bp 3 ’ to the gene.
(c) ApoE mutations One-dimensional isoelectric focusing of human plasma apoE reveals several bands whose relative concentrations vary between different individuals [112, 1141. Utilizing two-dimensional gel electrophoresis, it was possible to determine that some of these bands were due to sialo apoE isoproteins and others due to variations in the isoelectric point of the major asialo apoE isoprotein(s) [llO, 11I]. Studies of large numbers of individuals revealed six common apoE phenotypes in the population ((1 10, 1111, Fig. 1). Family studies showed that these phenotypes were the result of a single apoE gene locus with three common alleles [llO, 1111. The alleles have been designated €4, €3 and €2 and their gene products from basic to acidic are E4, E3 and E2, respectively. There are three homozygous phenotypes, E4/4, E3/3 and E2/2, and three heterozygous phenotypes, E4/3, E3/2 and E4/2 ([133], Fig. 1). In most of the large studies of apoE phenotype prevalence the range of allele frequencies were E4 14 to 15’70, €3 74 to 78% and €2 8 to 12% ([134- 1371, Table 5). These have
379 Apo E4
E4/4
E
alleles
63 -
E2
-
0 0
E 313 wl
aJ Q
21 c 0 C
E 212
8
E4/3
aJ Q
0
0
0
Fig. I . Schematic presentation of the three-allele model of apoE inheritance and nomenclature of the apoE alleles and phenotypes. The closed circles represent the major asialo apoE isoproteins.
TABLE 5 ApoE phenotype prevalence in population studies; derived apoE allele frequenciesa
Phenotype
Germany
Germany
USA
New Zealand
E4/4 E3/3 E2/2 E4/3 E4/2 E3/2
2.8 59.8 1 .o 22.9 1.5 12.0
2.2 62.2 0.9 19.9 2.9 11.1
1.1 61.3 2.4 20.8 4.2 10.1
51.4 1.4 25.0 1.2 20.0
58.3 0.5 24.8 2.8 12.8
6.3 54.0 0.3 31.9 0.5 6.1
No. of subjects
1031
1557
168
426
400
615
15 77 8
14 78 8
14 71
14 14 12
15 71 8
23 73 4
Allele €4 €3 €2
9
1 .o
Scotland 1 .o
Finland
The apoE phenotype frequencies are given as percentage of that phenotype occurring in the population under study. Allele frequencies were calculated and given as percentage of total (see [134- 1381).
a
been done in diverse geographical areas but primarily in Caucasians. Large studies assessing the frequencies of the apoE alleles in other racial groups have not been reported. A recent report from Finland indicates that this population may have a higher €4 and a lower €2 allele frequency than other Caucasian populations ([138], Table 5). This common apoE polymorphism has been found to play a role in Type I11 hyperlipoproteinemia (HLP) [ 110 - 112, 1391. This disorder is characterized by elevated cholesterol and triglyceride levels, as a result of delayed chylomicron remnant clearance, xanthomas and premature coronary as well as peripheral vascular disease [123]. Over 90% of individuals with Type I11 HLP have the E2/2 phenotype [140], whereas this occurs in only approximately 107'0 (Table 5) of normal individuals [134- 1371. In addition, when E2 is isolated and studied in vitro, it does not bind as well as E3 or E4 to high affinity lipoprotein receptors [141- 1431. It has been suggested that chylomicron remnants with E2 on their surface are recognized poorly by receptors and cleared slowly with the result being an accumulation of these particles in plasma. Chylomicron remnants are potent stimulators of macrophage cholesteryl ester accumulation in vitro and high plasma concentrations of these particles may be involved in the atherogenic process in vivo [123]. These data all suggest that homozygosity for the €2 allele may be the underlying cause of Type I11 HLP . However, the disease frequency is such that only 1 to 2% of people with the E2/2 phenotype actually express the disease. It is known that other hormonal and environmental factors are necessary for disease expression. However, the current belief is that Type I11 HLP is the result of two gene defects. One of these is in the apoE structural gene and the other in another gene that also influences chylomicron remnant synthesis and/or catabolism. The second gene product has yet to be defined. In addition t o the striking involvement of the E2/2 phenotype in Type 111 HLP, it appears that the apoE gene locus may be one of the factors influencing lipid levels in the general population [ 134 - 1371. The €2 allele appears to exert a stepwise gene dosage effect on lowering LDL levels as well as increasing VLDL cholesterol and triglyceride levels. The €4 allele also appears to exert a stepwise gene dosage effect on raising LDL levels. The increased frequency of the €4 allele and the decreased frequency of the €2 allele in the Finnish population may be an underlying genetic factor influencing the exceptionally high LDL cholesterol levels found in this population [ 1381. Amino acid sequence analysis established that the two common variants of apoE, E4 and E2, differ from E3 by single amino acid substitutions [ 1221. E4 differs from E3, at residue 112, because of an arginine for cysteine substitution, and E2 differs at residue 158, because of a cysteine for arginine substitution. Isoelectric focusing and amino acid and DNA sequencing have identified other rare apoE alleles. In all, 11 alleles are known and these are listed in Table 6 . In all but two of these alleles, E7 and E5 [144, 1451, the amino acid substitution underlying the variation has been
identified. Alleles E3** [146], E2 [122], E2* [143], E2** [147], E2*** [148] and E l [ 1491 all involve amino acid substitutions replacing positively charged with neutral amino acids in the region of the apoE receptor binding domain. Where information is available, these gene products have been shown to be defective in receptor binding and/or isolated from individuals with the Type 111 HLP phenotype. This emphasizes the importance of the positively charged amino acid residues in the receptor binding domain. Allele E3* was determined from the DNA sequence [130], but the substitution of proline for alanine at residue 152 might have a significant effect on the conformation of the receptor binding region. Functional studies of this gene product have not yet been reported. The alteration responsible for the E4 allele is not in the receptor binding region, and this gene product is fully functional in receptor binding studies [141- 1431. The E7 and E5 mutations have been found in Japanese, but not Caucasians. Although they are relatively rare alleles, it is said that they appear more frequently in individuals with hyperlipidemia and/or atherosclerosis [144, 1451. In one recent study, four of 58 lipid clinic patients and three of 69 coronary care unit patients were heterozygotes for one of these two alleles, whereas they were not present in 100 normal controls. The molecular basis for these mutations is not known, however, on SDS PAGE, E7 is of normal size, whereas E5 is smaller by 1500-2000 daltons [145]. TABLE 6 Human apoE protein polymorphism
Name"
Charge difference"
Defectb
El E5 E4 E3 E3* E3** E2 E2* E2** E2*** El
f 4 f 2 +I 0 0
? ?
0 -1 -1 -1 -1
-2
Ala,, - Thr, Ala,,, - Pro CYS,,,- Arg, - CYS - CYS Arg,,, - CYS Lys,,, - Gln Arg,,, - Ser C~Y,,,- Asp, Arg,,, - CYS CYSl,,
" Nomenclature for the apoE allele gene products recognized by the isoelectric focusing position of their
major asialo apoE isoprotein as specified by Zannis et al. [133]. The most common allele gene product E3 has an isoelectric point of p H 6.02. Alleles specifying gene products E4, E5, and E7 are 1, 2, and 4 charge units, respectively, more basic and E2 and El are 1 and 2 charge units, respectively, more acidic than wild type. *, **, *** indicate rare apoE variants recently discovered with the same isoelectric focusing pattern as E3 and E2. Amino acid sequence difference with reference t o the most common allele, E3 [122].
3 82
In addition to the apoE mutations just described, a rare form of Type I11 HLP has been reported associated with apoE deficiency [150]. A single family was described in which affected siblings had no immunodetectable apoE in their plasma. Southern blot analysis of genomic DNA from these patients showed that their apoE gene was present and grossly intact [ 15 11. However, monocyte-macrophage cultures from one of these patients was studied and showed a 50-fold decrease in apoE mRNA levels compared to similar cultures from normal people. In addition, two different apoE mRNA sizes were seen with one larger than normal. The apoE mutation in this case appears either to affect the transcription of the apoE gene or the processing of its primary transcript [ 1511.
10. Chromosomal location of apolipoprotein genes (a) Chromosome I : apoA-II Several different groups have used molecular probes and panels of somatic cell hybrids to map the apoA-I1 gene to chromosome 1 ([152- 1541, Fig. 2). Hybrids containing only a portion of chromosome 1, due to translocations, have provided subchromosomal localization. A hybrid containing chromosome lp2 1 to 1qter was positive for the human apoA-I1 gene [153]. Although the data are as yet unpublished, another group claims, in the discussion of one of their papers, that the apoA-I1 gene localizes to the region of lq21 to lq23 [155]. In the mouse, apoA-I1 is in a linkage group with peptidase C and renin found on mouse chromosome 1. In
E
CI
CI I
} A-IZ
I
2
Fig. 2. Chromosomal location of the apolipoprotein genes.
II
19
383 humans, such markers are found on the long arm of chromosome 1, whereas markers for the short arm of human chromosome 1 are found on mouse chromosomes 3 and 4 [153]. These data taken together indicate that the human apoA-I1 gene is on the long arm of chromosome 1, probably in the q21 and q23 region.
(b) Chromosome 2: apoB In recent studies, apoB cDNA probes have been used to map this gene to chromosome 2 ([62, 156, 157, 157a], Fig. 2). Both, in situ and somatic cell hybridization techniques were employed. Hybrids containing translocations involving chromosome 2 were used to provide information about subchromosomal localization. In one study, a hybrid containing 2pter to 2p23 was positive for apoB, whereas a hybrid containing 2p23 to 2qter was negative [157]. In another study, two hybrids containing only 2pter to cen were positive for apoB, whereas a hybrid with a complex rearrangement of the short arm of chromosome 2, due to both terminal and interstitial deletions was negative. The latter hybrid retained the MDH-1 and N-MYC regions [156]. These data taken together indicate that the apoB gene is distal to the MDH-1 and N-MYC loci, which have been mapped to 2p23. The somatic cell hybrid studies are compatible with the in situ hybridization results, which place apoB at the top of the short arm of chromosome 2, between 2pter to 2p24 [62, 1561.
(c) Chromosome 11: apoA-I, apoCIII, apoA-IV Somatic cell hybrids and DNA probes have been used to map the gene for human apoA-I [158 - 1601. In all studies, the apoA-I gene appears to be at the single locus and co-segregates with human chromosome 11 (Fig. 2). Some of the hybrids examined contained only a portion of chromosome 11, and apoA-I cosegregated with p l 1 t o qter [158], p l l to q13 [159], and q13 to qter [160] in three different studies. In the mouse, apoA-I has been mapped to chromosome 9 [161] and shown to be 1.3 & 0.7 centimorgans from uroporphyrinogen I synthetase [162]. The latter has been mapped in humans to chromosome region 1 lq23 t o qter [ 1631. If the linkage group present in the mouse is conserved in humans, it suggests that the apoA-I gene is more distal on the short arm of chromosome 11 than the q13 region. The translocation hybrid thought to contain pl 1 to q13, that was positive for apoA-I, may have to be re-evaluated. Altogether, these data suggest that the apoA-I gene resides on the long arm of chromosome 11 in the region of q23. Recently, the apoA-I gene has been mapped by in situ hybridization to the 1lq22-q23 region [163a]. In addition to the apoA-I gene, other apolipoprotein genes have been localized on chromosome 11 (Fig. 2). cDNA clones and somatic cell hybrids have been used to map the apoCIII gene to this chromosome [158]. A translocation hybrid further
localized the gene to the p l l to qter region. In a family study, the gene for apoA-IV was shown to be linked to the gene for apoA-I [164]. A large Norwegian kindred was identified in which apoA-IV and apoA-I alleles could be distinguished because of isoelectric point differences in their protein products. No recombination was seen in 1 1 observed meioses where this could have been detected. Linkage was suggested by a Lod score of 3.01 at a recombination fraction of 0.00. This study suggests that the apoA-IV gene resides on chromosome 1 1 in the region of the apoA-I gene, q23. The exact relationships of the apoA-I, apoCIII and apoA-IV genes have been determined (Fig. 3). ApoCIII cDNA clones were used to identify the apoCIII gene on human genomic DNA cloned in lambda phage which contained the apoA-I gene [104]. Mapping of the apoCIII gene reveals that it is about 2500 bp from the 3 ’ end of the apoA-I gene. Further mapping and DNA sequence analysis revealed that these genes are coded for by opposite DNA strands [ 1041. The 3 ’ end of the apoCIII gene is located closest to the 3 ‘ end of the apoA-I gene, and the 5 ’ end of the apoCIII gene, containing the apoCIII promoter, is furthest away from the apoA-I gene. Thus, these two genes are convergently transcribed. It is not known where their primary transcripts end, or whether there is any functional significance t o this unusual configuration. Recently, apoA-IV genomic sequences have been located beginning 4 kb 3 ’ to the apoA-I/apoCIII gene complex [49]. The apoA-IV gene is in the same orientation as the apoA-I gene but in the opposite orientation to the apoCIII gene. Thus, the apoCIII and apoA-IV genes share a common 5’ upstream region and are divergently transcribed. Thus, all three apolipoprotein genes at the chromosome 1 1 locus lie within a 14 kb DNA segment.
(d) Chromosome 19: apoE, apoCI, apoCII Family studies have shown that the inheritance of the apoE protein polymorphism cosegregated with the protein polymorphism for the third component of complement [165]. Since the latter had been mapped using DNA probes and somatic cell hybrids to human chromosome 19 [166], the apoE gene was also assigned to this chromosome [165]. ApoE cDNA probes and somatic cell hybrids have now been used to confirm this assignment ([131, 167, 1681, Fig. 2). Limited subchromosomal localization has also been achieved with somatic cell hybrids containing chromo-
* c -
ApoA-I
Apo CIII
___)
Ape A -
IZ
1 Kb
H
Fig. 3. The arrangement of the apolipoprotein genes on chromosome 1 1 . The arrows indicate the direction of transcription.
385 some 19 translocations. Two somatic cell hybrids with chromosome 19’s missing the tip of the short arm and long arm, respectively, were both positive for apoE. In addition to the apoE gene, other apolipoprotein genes have also been localized to chromosome 19 (Fig. 2). cDNA clones and somatic cell hybrids have been used to map the apoCI and apoCII genes to this chromosome [78, 88, 168 - 1701. To determine the relative orientation of the apoE, apoCI and apoCII genes, cosmid clones containing apoE genomic fragments were probed with apoCI cDNA. This has revealed the apoCI gene approximately4 kb 3 ’ to the apoE gene and in the same orientation ([81, 821, Fig. 4). Preliminary evidence indicates a second copy of the apoCI gene 5 to 10 kb 3 ’ to the first apoCI gene [81]. The exact relationship of the apoCII gene to the apoE/apoCI gene cluster has not yet been determined. Genomic clones for apoCII do not hybridize strongly with cDNA clones for apoE or apoCI and vice versa. In spite of this, other genetic evidence indicates a close association. For example, linkage disequilibrium has been demonstrated between the €2 apoE allele and alleles revealed by an apoCII gene-associated RFLP after digesting genomic DNA with the enzyme TaqI [171]. In addition, family studies done by two different groups show linkage of the apoE protein polymorphism and the apoCII TaqI polymorphism [ 171, 1721. In each case, the Lod score was greater than 4.0 at a recombination fraction of 0.00. The combined data from both studies revealed no recombinations in 53 observed meioses. Therefore, apoCII is probably no greater than 2 centimorgans from the apoE/apoCI gene complex. Besides apoE, apoCI and apoCII, other genes have been mapped to chromosome 19 [ 173 - 1751. The gene for familial hypercholesterolemia, the LDL receptor gene, appears to be at the tip of chromosome 19, pter to p13, approximately 20 cm distal to the gene for the third component of complement. The apolipoprotein gene locus is about the same distance from the complement gene, but unlinked to the familial hypercholesterolemia gene. Therefore, the apolipoprotein gene locus must be proximal to the complement gene locus. In addition, the apolipoprotein gene locus appears to be very near the locus for the Lutheran blood group and approximately 15 cm from the loci for myotonic dystrophy and neurofibromatosis.
1 Kb H
Fig. 4. The arrangement of the apolipoprotein genes on chromosome 19. The arrows indicate the direction of transcription. A second apoCI gene is 3 ‘ to the apoE/apoCI gene complex. The dashed lines show uncertainties of the scale in this part of the region. At this point, it is not known whether the apoCII gene is 5 ’ or 3’ to this complex nor is the exact distance from the complex known.
386
1 I . Apolipoprotein gene family In 1977, Barker and Dayhoff, using amino acid sequence data available only for apoA-I, apoA-11, apoCI and apoCIII proposed that the apolipoproteins were all derived from a common evolutionary precursor [24]. In modern genetic terminology, the apolipoprotein genes would be a multigene family, in other words, a group of functionally related genes evolved from an ancestral gene. The data thus far characterizing the genes for apoA-I, apoA-11, apoA-IV, apoCI, apoCII, apoCIII and apoE support this hypothesis. Strong evidence is provided by the existence of gene clusters. The apoA-I, apoCIII and apoA-IV genes all reside within a 15 kb region of chromosome 11. The apoE and apoCI genes reside next to each other, and the apoCII gene is close by on chromosome 19. Other supporting evidence lies in the structural similarities between the genes (Fig. 5). All of these genes but one have four exon, three intron structures with the introns in strikingly similar locations. The exception is the apoA-IV gene in which the IVS corresponding to the IVS-1 of the other genes has apparently been deleted. Other similarities between these apolipoprotein genes include extensive DNA homology in the 5 ’ coding region of apoA-I and CII exon 3 regions that contain a common block of 33 codons, and the existence of multiple 66 bp intragenic duplications in the 3 ’ coding region of the apoA-I, apoA-IV and apoE genes [3, 12, 49, 131, 175al. At the moment it is not certain whether the apoB gene is part of this multigene family. The location of the apoB gene t o a region of the genome quite separate from
-
100 bp
E
v v
v
Fig. 5 . The apolipoprotein mRNA structures are drawn to scale. The length of the protein-coding region is indicated by the rectangles. The thin lines indicate the lengths of the 5 ’ and 3 ‘ untranslated regions o n the left and right, respectively, of the protein-coding region. The locations of the introns are indicated by the triangles. To show the similarities between the genes, the beginnings of all the protein-coding regions are aligned.
387 the other apolipoprotein genes suggests that it may not be part of this gene family. Further work on the structure of the apoB gene will be necessary to establish this conclusion.
12. Apolipoprotein gene associated RFLP’s RFLP’s have been reported for the apolipoprotein genes, and some of these have been associated with lipoprotein abnormalities. RFLP’s have been identified in the apoA-I, apoCIII, apoA-IV gene cluster on chromosome 11 using the enzymes SstI (SacI), PstI and XmnI [104, 109, 1761. The polymorphic site for SstI (SacI) is approximately 2.7 kb 3 ’ to the apoA-I gene and has been shown to be due to a C to G transmutation in the 3 ’ untranslated region of the apoCIII gene, which establishes a n SstI (SacI) cutting site [104]. Utilizing a n apoA-I genomic probe, Southern blotting of genomic DNA after SstI digestion reveals bands of 4.5 kb (S1 allele), 3.2 kb (S2 allele), or both, in heterozygotes. The S2 allele is the minor one and has the additional SstI (SacI) site. In normolipidemic individuals, the frequency of the S2 allele varies in different racial groups. The reported frequency varies from 0 to 6 % in Caucasians, 18% in Indian-Asians, 15% in Africans, 19% in Japanese, and 48% in Chinese [176, 1771. Different clinical studies of Caucasian populations have indicated a n increased frequency of the S2 allele in hypertriglyceridemic individuals (25% and 19%) [109, 1771, Type V hyperlipoproteinemia (27%) [176], low HDL levels (16 YO) [ 1781, and survivors of myocardial infarction (13 Yo) [ 1791. Therefore, the evidence suggests an association between a DNA polymorhpism in the 3 ‘ untranslated region of the apoCIII gene and clinical lipoprotein abnormalities in Caucasians. It is uncertain whether the particular base change involved is causative or merely in linkage desequilibrium with another mutation in the apoA-I, apoCIII, apoA-IV gene cluster that is affecting lipoprotein metabolism. In one study, cDNA clones representing S1 and S2 alleles were sequenced and no associated mutations were found that altered the apoCIII primary amino acid sequence [105]. However, two other base differences were found. One affected the third position of the codon that specifies the glycine a t amino acid residue 14 and the other, affected a base 31 bp 3 ‘ to the SstI polymorphic site, also in the apoCIII 3’ untranslated region. Since the proposed physiological role of apoCIII is in regulating the metabolism of triglyceride-rich lipoproteins, it is attractive to speculate that mutations in and around this gene may be involved in causing hypertriglyceridemia. The polymorphic site for PstI is approximately0.3 kb 3 ’ to the apoA-I gene [176]. Utilizing an apoA-I probe, Southern blotting of genomic DNA after PstI digestion reveals bands of 2.2 kb (P1 allele), 3.3 kb (P2 allele), or both in heterozygotes. The P2 allele, which is missing the PstI site, is the minor one with a frequency of 12%. I n one clinical study, no association between the P2 allele and lipoprotein abnormalities was found [176].
388 The polymorphic site for XmnI is approximately 2.5 kb 5 ’ to the apoA-I gene [176]. Utilizing an apoA-I probe, Southern blotting of genomic DNA after XmnI digestion reveals bands of 8.4 kb (X1 allele), 6.3 kb (X2 allele), or both in heterozygotes. The X2 allele, which is missing the XmnI site, is the minor one with a frequency of 12%. In one clinical study, the X2 allele was associated with Type IIb (25%), Type I11 (25%), and Type V (22Vo), but not with Type IIa or Type IV hyperlipidemia. In the same study, the X2 allele showed an increased frequency in individuals in the highest triglyceride quintile (26%). There was a suggestion of an association with the lowest HDL quintile (19’70), but this did not reach significance [176]. The SstI, PstI, and XmnI RFLP’s in association with the apoA-I, apoCIII, apoAIV gene cluster were found to be in linkage equilibrium and could serve to establish a haplotype [1761. The combined polymorphism information content (PIC value) of these three RFLP’s is 0.5 and they, therefore, could be quite useful in family studies to establish linkage of this gene cluster to clinical phenotypes. In spite of the common protein polymorphism of the apoE gene, RFLP’s have not been reported for the apoE, apoCI gene cluster. Two RFLP’s have been reported for the closely linked apoCII gene using the enzymes TaqI and BglI [171, 172, 1801. The polymorphic site for TaqI is approximately 2 kb 3 ‘ to the apoCII gene. Utilizing an apoCII cDNA probe, Southern blotting after TaqI digestion reveals bands of 3.8 kb (T1 allele), 3.5 kb (T2allele), or both in heterozygotes. The T2 allele, which contains the polymorphic TaqI site, is the minor one with a frequency of 44%. The polymorphic site for BglI is approximately 9 kb 5 ’ to the apoCII gene. Utilizing a 400 bp genomic probe from a unique region approximately 1 to 1.4 kb upstream from the gene, Southern blotting after Bgl I digestion reveals bands of 9 kb (B1 allele), 12 kb (B2 allele), or both, in heterozygotes. The B2 allele, which is missing the polymorphic BglI site, is the minor one with a frequency of 47%. Neither the TaqI nor Bgl I polymorphism have been associated with any clinical lipoprotein abnormalities [171, 172, 1801. The PIC values for both of these polymorphisms separately are approximately 0.37. Utilizing the two together to establish a haplotype should give a PIC value of 0.68;however, due to a degree of linkage disequilibrium the PIC value is reduced to 0.51 [180]. These RFLP’s could be quite useful in family studies to establish linkage of the a p e 1 1 gene, and thereby the chromosome 19 apolipoprotein gene cluster, to clinical phenotypes. The TaqI RFLP alone has already been used to establish linkage to the apoCII deficiency phenotype in families with this form of Type I hyperlipoproteinemia [94]. At the apoA-I1 gene locus on chromosome 1, an RFLP has been reported using the enzyme MspI [155]. The polymorphic MspI site is approximately 0.2 kb 3 ’ to the gene in an Alu-repetitive element. Utilizing an apoA-I1 cDNA probe, Southern blotting after MspI digestion reveals bands of 3.0 kb (M1 allele), 3.7 kb (M2 allele), or both in heterozygotes. The M2 allele, which is missing the polymorphic MspI site,
3 89 is the minor one with a frequency of 19%. In one clinical study, homozygotes for the M2 allele had higher plasma apoA-I1 concentrations and an increased ratio of apoA-I1 to apoA-I when compared to controls or heterozygotes for the M2 allele. HDL cholesterol levels did not differ between the groups [155]. The clinical significance of this observation is as yet unknown. In the mouse, the apoA-I1 locus is also on chromosome 1 and is closely associated with genetic loci that affect HDL composition, the response of HDL cholesterol levels to a diet rich in saturated fat and cholesterol, and atherosclerosis susceptibility [161, 1811. More information will need to be gathered about the effect of the apoA-I1 gene locus or other closely linked loci on these phenomena in man. Several RFLP’s have been reported at the apoB gene locus on chromosome 2. cDNA probes corresponding to the 3 ‘ end of apoB mRNA reveal an EcoRI RFLP [156, 157a, 1821. The polymorphic EcoRI site is located in the coding region of the gene. The two alleles are 11 kb (Rl allele) and 13 kb (R2 allele). R2 is the minor allele and has a frequency of approximately 20%. The same probes also detect a polymorphic region just 3 ’ to the gene, which is revealed with many different enzymes. Presumably, this polymorphism is of the insertion-deletion type. There appears to be significant linkage disequilibrium between these two RFLP’s. The R2 allele is associated with the region of increased length 3 ‘ to the gene [183]. A third RFLP seen with the enzyme XbaI is revealed by more 5 ‘ cDNA probes [184]. The two alleles are 8.6 kb (X1 allele) and 5.0 kb (X2 allele). X2 is the minor allele and has a frequency of approximately 40%. Data on the clinical significance, if any, of these RFLP’s should be forthcoming in the near future.
13. Apolipoprotein gene rearrangements A family has been described in which two sisters had very low HDL but normal LDL levels, xanthomas, severe premature atherosclerosis, and absent plasma apoA-I and apoCIII. First degree relatives of these individuals had half-normal plasma levels of HDL, apoA-I, and apoCIII [loll. Southern blotting of genomic DNA from the probands, after digestion with EcoRI, with an apoA-I cDNA probe revealed a single band of 6.5 kb, whereas normals showed a single 13 kb band. First degree relatives, including the mother and father, of the probands showed one normal band and one abnormal band [185]. Therefore, they appeared to be carriers of a mutant allele associated with the apoA-I gene, and the probands appeared to be homozygous for this mutant allele. Southern blotting of probands’ DNA, after digestion with other restriction endonucleases with an apoA-I cDNA probe, consistently revealed differences from wild type [ 1861. This suggested that the genetic lesion was not a single bp substitution, but rather a more major DNA alteration. Southern blotting with other probes derived from the region of the apoA-I gene indicated that the lesion was a DNA insertion into the fourth exon of the apoA-I gene. This insertion inter-
rupts the normal coding region of the apoA-I gene at approximately the codonspecifying residue 80 of the mature protein and may explain the lack of apoA-I in the plasma of these patients. To define the nature of the DNA insertion, a genomic library was made from the DNA of one of the probands, and clones containing the apoA-I gene and the contiguous up and downstream portions of the insertion were isolated. Southern blotting of genomic DNA from normal individuals, after digestion with EcoRI, with a probe that included apoA-I sequences and insertion sequences, revealed the normal 13 kb apoA-I genomic fragment, as expected, plus another unique band. In the probands the same probe revealed only the 6.5 kb band. This suggested that in the probands a unique piece of DNA, normally present in the genome, has been deleted and was inserted into the apoA-I gene. Further experiments with a probe made to just the insert showed that it hybridized to a region 3 ’ to the apoA-I gene in normal individuals, which contains the apoCIII gene [187]. Thus, it appears that a portion of the apoCIII gene was inserted into the apoA-I gene in the probands and that this is the underlying molecular basis of their apoAI/apoCIII deficiency.
A ckn owledgernents This work was supported by grants from the National Institutes of Health (HL33714, HL32435, AG04727). Dr. Jan L. Breslow is an Established Investigator of the American Heart Association. I would like to express my sincere appreciation to Miss Lorraine Duda for her expert assistance in preparing this review.
References 1 2
3 4 5
6.
7. 8.
Zannis, V.I. and Breslow, J.L. (1985) Adv. Hum. Genet. 14, 125-215. Herbert, P.N., Assmann, G., Gotto, A.M., Jr. and Fredrickson, D.S. (1982) in: The Metabolic Basis of Inherited Disease (J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein and M.D. Brown, Eds.) pp. 589-651, McGraw-Hill, New York. Breslow, J . L . (1985) Annu. Rev. Biochem. 54, 699-727. Breslow, J . L . (1986) in: Biochemistry and Biology of Plasma Lipoproteins (A.M. Scanu and A. Spector, Eds.) pp. 85 - 143, Marcel Dekker, New York. Williams, D.L. (1985) Arteriosclerosis 5, 213 - 227. Heiss, G. and Tyroler, H.A. (1982) in: Proceedings of the Workshop on Apolipoprotein Quantification, pp. 7 - 24, Department of Health and Services, National Institutes of Health (NIH Publication No. 83 - 1266) Bethesda. Brunzell, J.D., Sniderman, A.D., Albers, J.J. and Kwiterovich, P.O., Jr. (1984) Arteriosclerosis 4, 79- 83. Breslow, J.L., Ross, D., McPherson, J., Williams, H . , Kurnit, D., Nussbaum, A., Karathanasis, S . K . and Zannis, V.1. (1982) Proc. Natl. Acad. Sci. USA 79, 6861 -6865.
391 9. 10 I1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
38 39
Shoulders, C.C. and Baralle, F.E. (1982) Nucl. Acids Res. 10, 4873 -4882. Cheung, P. and Chan, L. (1983) Nucl. Acids Res. 1 1 , 3703-3715. Law, S.W. and Brewer, H.B., Jr. (1984) Proc. Natl. Acad. Sci. USA 81, 66-70. Karathanasis, S.K., Zannis, V.1. and Breslow, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 6147-6151. Shoulders, C.C., Kornblihtt, A.R., Munro, B.S. and Baralle, F.E. (1983) Nucl. Acids Res. I I , 2827 - 2837. Sharpe, C.R., Sidoli, A., Shelley, C.S., Lucero, M.A., Shoulders, C.C. and Baralle, F.E. (1984) Nucl. Acids Res. 12, 3917 - 3932. Gordon, J.I., Smith, D.P., Andy, R., Alpers, D.H., Schonfeld, G. and Strauss, A.W. (1982) J. Biol. Chem. 257, 971 -978. Zannis, V.I., Karathansis, S.K., Keutmann, H., Goldberger, G. and Breslow, J.L. (1983) Proc. Natl. Acad. Sci. USA 80, 2574 - 2578. Gordon, J.I., Sims, H.F., Lentz, S.R., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1983) J. Biol. Chem. 258, 4037 -4044. Zannis, V.I., Breslow, J.L. and Katz, A.J. (1980) J. Biol. Chem. 255, 8612-8617. Zannis, V.I., Breslow, J.L., SanGiacomo, T.R., Aden, D.P. and Knowles, B.B. (1981) Biochemistry 20, 7089 - 7096. Zannis, V.I., Kurnit, D.M. and Breslow, J.L. (1982) J. Biol. Cheni. 257, 536-544. Edelstein, C., Gordon, J.I., Toscas, K., Sims, H.F., Strauss, A.W. and Scanu, A.M. (1983) J. Biol. Cheni. 258, 11430 - 11433. Kooistra, T., VanHinsbergh, V., Havekes, L. andKenipen, H.J. (1984)FEBS Lett. 170, 109- 113. Brewer, H.B., J r . , Fairwell, T., LaRue, A., Ronan, R., Houser, A. and Brouzert, T.J. (1978) Biochem. Biophys. Res. Commun. 80, 623 - 630. Barker, W.C. and Dayhoff, M.O. (1977) Conip. Biocheni. Physiol. 57B, 309-315. Fitch, W.M. (1977) Genetics 86, 623-644. Schiffer, M. and Edmundson, A.B. (1967) Biophys. J. 7, 121 - 135. Segrest, J.P., Jackson, R.L., Morrisett, J.D. and Gotto, A.M. (1974) FEBS Lett. 38, 247-253. Segrest, J.P., Chung, B.H., Brouillette, C.G., Kanellis, P . and McGahan, R. (1983) J. Biol. Chem. 258, 2290 -2295. Fitch, W.M., Smith, T. and Breslow, J.L. (1985) in: Methods in Enzymology, Chapter 99, Section VI, pp. 773 - 788, Academic Press, New York. Seilhanier, J.J., Protter, A.A., Forssard, P. and Levy-Wilson, B. (1984) DNA 3, 309-317. Protter, A.A., Levy-Wilson, B., Miller, J., Bencen, G., White, T. and Seilhamer, J.J. (1984) DNA 3, 449-456. Franceschini, G., Sirtori, C.R., Capurso, A., Weisgraber, K.H. and Mahley, R.W. (1980) J. Clin. Invest. 66, 892-900. Weisgraber, K.H., Rall, S.C., Bersot, T.P., Mahley, R.W., Franceschini, G. and Sirtori, C.R. (1983) J. Biol. Chem. 258, 2508-2513. Uterniann, G., Feussner, G., Franceschini, G., Haas, J . and Steinmetz, A. (1982) J. Biol. Chem. 257, 501 -507. Utermann, G., Steinmetz, A , , Paetzold, R., Wilk, J., Feussner, G., Kaffarnik, H., MuellerEckhardt, C., Seidel, D., Volgelbert, K.H. and Zimmer, F. (1982) Hum. Genet. 61, 329-337. Menzel, H.J., Kladetzky, L. and Assniann, C. (1982) J . Lipid Res. 23, 915 -922. Weisgraber, K.H., Rall, S.C., Jr., Mahley, R.W., Ogawa, Y., Fielding, C.J., Utermann, G., Haas, J., Steinmetz, A., Menzel, H.J. and Assmann, G. (1985) in: Proceedings of 7th International Syniposiuni on Atherosclerosis (P.J. Nestel, Ed.) pp. 113 - 114, International Atherosclerosis Society. Rall, S.C., Jr., Weisgraber, K.H., Mahley, R.W., Ogawa, Y., Fielding, C.J., Utermann, G., Haas, J., Steinmetz, A., Menzel, H.J. and Assniann, (1984) J . Biol. Chem. 259, 10063 - 10070. Knott, T.J., Priestley, L.M., Urdea, M. and Scott, J. (1984) Biochem. Biophys. Res. Commun. 120, 734 - 740.
Moore, M.N., Kao, F.T., Tsao, Y .K. and Chan, L. (1984) Biochem. Biophys. Res. Commun. 123, 1-7. 41 Lackner, K.J., Law, S.W. and Brewer, H.B., Jr. (1984) FEBS Lett. 175, 159- 164. 42 Gordon, J.I., Budelier, K.A., Sims, H.E., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1983) J . Biol. Chem. 258, 14054- 14059. 43 Gordon, J.I., Sims, H.F., Edelstein, C., Scanu, A.M. and Strauss, A.W. (1985) J. Biol. Chem. 260, 14824- 14831. 44 Brewer, H.B., Jr., Lux, S.E., Ronan, R. and John, K.M. (1972) Proc. Natl. Acad. Sci. USA 69, 1304- 1308. 45 Knott, T.J., Wallis, S.C., Robertson, M.E., Priestley, L.M., Urdea, M., Rall, L.B. and Scott, J. (1985) Nucl. Acids Res. 13, 6387-6398. 46 Lackner, K.J., Law, S.W. and Brewer, H.B., Jr. (1985) Nucl. Acids. Res. 13, 4597-4608. 47 Tsao, Y.K., Wei, C.F., Robberson, D.L., Gotto, A.M., Jr. and Chan, L. (1985) J. Biol. Chem. 260, 15222 - 15231. 48 Gordon, J.I., Smith, D.P., Alpers, D.H. and Strauss, A.W. (1982) Biochemistry 21, 5424- 5430. 49 Karathanasis, S.K. (1985) Proc. Natl. Acad. Sci. USA 82, 6374-6378. 50 Elshourbagy, N.A., Boguski, M.S., Reardon, C.A., Levy-Wilson, B., Jefferson, L.S., Gordon, J.I. and Taylor, J.M. (1985) Arteriosclerosis 5 , 502a. 51 Gordon, J.I., Bisgaier, C.L., Sims, H.F., Sachdev, O.P., Glickman, R.M. and Strauss, A.W. (1984) J. Biol. Chem. 259, 468-474. 52 Menzel, H. J., Kovary, P.M. and Assmann, G. (1982) Hum. Genet. 62, 349 - 352. 53 Goldstein, J.L., Brown, M.S., Anderson, R.G.W., Russell, D.W. and Schneider, W.J. (1985) Annu. Rev. Cell Biol. 1, 1-39. 54 Kane, J.P. (1983) Annu. Rev. Physiol. 45, 637-650. 54a Cardin, A.D., Witt, K.R., Chao, J . Margolius, H.S., Donaldson, V.H. and Jackson, R.L. (1984) J. Biol. Chem. 259, 8522-8528. 55 Cardin, A.D., Witt, K.R., Barnhart, C.L. and Jackson, R.L. (1982) Biochemistry 21,4503-4511. 56 LeBoeuf, R.C., MIller, C., Shively, J.E., Schumaker, V.N., Balla, M.A. and Lusis, A.J. (1984) FEBS Lett. 170, 105 - 108. 57 Schonfeld, G., Patsch, W., Pfleger, B., Witztum, J.L. and Weidman, S.W. (1979) J. Clin. Invest. 64, 1288- 1297. 58 Marcel, Y.L., Hogue, M., Theolis, R. and Milne, R.W. (1982) J. Biol. Chem. 257, 13165- 13168. 58a Young, S.G., Bertics. S.J., Scott, T.M., Dubois, B.W., Curtis, L.K. and Witztum, J.L. (1986) J. Biol. Chem. 261, 2995-2998. 59 Deeb, S.S., Motulsky, A.G. and Albers, J.J. (1985) Proc. Natl. Acad. Sci. USA 82, 4983-4986. 60 Huang, L.S., Bock, S.C., Feinstein, S.I. and Breslow, J.L. (1985) Proc. Natl. Acad. Sci. USA 82, 6825 - 6829. 61 Mehrabian, M., Schumaker, V.N., Fareed, G.C., West, R., Johnson, D.F., Kirchgessner, T., Lin, H.C., Wang, X.,Ma, Y., Mendiaz, E. and Lusis, A.J. (1985) Nucl. Acids Res. 13, 6937-6953. 62 Knott, T.J., Rall, S.C., Jr., Innerarity, T.L., Jacobson, S.F., Urdea, M.S., Levy-Wilson;, B., Powell, L.M., Pease, R.J., Eddy, R., Nakai, H., Beyers, M., Priestley, L.M., Robertson, E., Rall, L.B., Bertsholtz, C, Shows, T.B., Mahley, R.W. and Scott, J. (1985) Science 230, 37-43. 63 Law, S.W., Lackner, K.J., Hospattankar, A.V., Anchors, J.M., Sakaguchi, A.Y., Naylor, S.L. and Brewer, H.B. (1985) Proc. Natl. Acad. Sci. USA 82, 8340-8344. 64 Shoulders, C.C., Myant, N.B., Sidoli, A., Rodriguez, J.C., Cortese, C., Cortese, R. and Baralle, F.E. (1985) Atherosclerosis , in press. 64a Wei, C.-F., Chen, S.H., Yang, C.-Y., Marcel, Y.L., Milne, R.W., Li, W.-H., Sparrow, J.T., Gotto, A.M., Jr. and Chan, L. (1985) Proc. Natl. Acad. Sci. USA 82, 7265-7269. 65 Innerarity, T.L., Friedlander, E.J., Rall, S.C., Jr., Weisgraber, K.H. and Mahley R.W. (1983) J. Biol. Chem. 258, 12341 - 12347. 40
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80
81 82 83
84 85 86 87 88 89 90 91
Weisgraber, K.H., Innerarity, T.L., Harder, K. J., Mahley, R.W., Milne, R.W., Marcel, Y.L. and Sparrow, J.T. (1983) J. Biol. Chem. 258, 12348 - 12354. Innerarity, T.L., Weisgraber, K.H., Arnold, K.S., Rall, S.C., Jr. andMahley, R.W. (1984) J. Biol. Chem. 259, 7261 -7267. Yamamoto, T., Davis, C.G., Brown, M.S., Schneider, W.J., Casey, M.L., Goldstein, J.L. and Russell, D.W. (1984) Cell 39, 27-38. Berg, K. (1983) in: Progress in Medical Genetics (A.G. Steinberg, A.G. Bearn, A.R. Motulsky and B. Childs, Eds.) pp. 35 -90, W.B. Saunders and Co., Philadelphia. Schumaker, V.N., Robinson, M.T., Curtiss, L.K., Butler, R. and Sparkes, R.S. (1984) J. Biol. Chem. 259, 6423 -6430. Young, S.G., Bertics, S.J., Curtiss, L.K., Casal, D.C. andwitztum, J.L. (1986)Proc. NatLAcad. Sci. USA 83, 1101 - 1105. Malloy, M.J., Kane, J.P., Hardman, D.A., Hamilton, R.L. andDolal, K.B. (1981)J. Clin. Invest. 67, 1441 - 1450. Herbert, P.N., Hoyams, J.S., Bernier, D.N., Berman, M.M., Saritelli, A.L., Lynch, K.M., Nichols, A.V. and Forte, T.M. (1985) J. Clin. Invest. 76, 403 -412. Takashima, Y., Kodama, T., Iida, H., Kawamura, M., Aouratini, H., Itakura, H., Akanuma, Y., Takaku, F. and Kawade, M. (1985) Pediatrics 75, 541 -546. Sniderman, A.D., Shapiro, S., Marpole, D., Skinner, B., Teng, B. and Kwiterovich, P.O. (1980) Proc. Natl. Acad. Sci. USA 77, 604-608. Goldstein, J.L., Schrott, H.G., Hazzard, W.R., Bierman, E.L. and Motulsky, A.R. (1973) J. Clin. Invest. 52, 1544- 1568. Knott, T.J., Robertson, M.E., Priestley, L.M., Urdea, M., Wallis, S. and Scott, J. (1984) Nucl. Acids Res. 12, 3909-3915. Tata, F., Henry, I., Markham, A.F., Wallis, S.C., Weil, D., Grzeschik, K.H., Junien, C., Williamson, R. and Humphries, S.E. (1985) Hum. Genet. 69, 345-349. Jackson, R.L., Sparrow, J.T., Baker, H.N., Morrisett, J., Taunton, O.D. and Gotto, A.M., Jr. (1974) J. Biol. Chem. 249, 5308-5313. Shulman, R.S., Herbert, P.N., Wehrly, K. and Fredrickson, D.S. (1975) J. Biol. Chem. 280, 182 - 190. Paik, Y.K., Lauer, S.J., Levy-Wilson, B., Chang, D.J., Elshourbagy, N.A., Mahley, R.W., Reardon, C.A. and Taylor, J.M. (1985) Arteriosclerosis 5 , 503a. Das, H.K. and Breslow, J.L. (1985) Arteriosclerosis 5 , 503a. Nikkila, E.A. (1983) in: The Metabolic Basis of Inherited Disease (J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein and M.D. Brown, Eds.) pp. 622 - 642, McGrawHill, New York. Breckenridge, W.C., Little, J.A., Steiner, G., Chow, A. and Poapst, M. (1978) N. Engl. J. Med. 298, 1265 - 1273. Musliner, T.A., Church, E.C., Herbert, P.N., Kingston, M.J. and Shulman, R.S. (1977) Proc. Natl. Acad. Sci, USA 74, 5358 - 5362. Kinnunen, P.K.J., Jackson, R.L., Smith, L.C., Gotto, A.M. andsparrow, J.T. (1977)Proc. Natl. Acad. Sci. USA 74, 4848 -4851. Vainio, P., Virtanen, J.A., Kinnunen, P.K.J., Gotto, A.M., Jr., Sparrow, J.T., Pattus, F., Bougis, P. and Verger, R. (1983) J . Biol. Chem. 258, 5477-5482. Jackson, C.J., Bruns, G.A.P. and Breslow, J.L. (1984) Proc. Natl. Acad. Sci. USA 81, 2945 - 2949. Myklebost, O., Williamson, B., Markham, A.F., Myklebost, S.R., Rogers, J., Woods, D.E. and Humphries, S.E. (1984) J. Biol. Chem. 259, 4401 -4404. Fojo, S.S., Law, S.W. andBrewer, H.B., Jr. (1984)Proc. Natl. Acad. Sci. USA 81, 6354-6357. Hospattankar, A.V., Fairwell, T., Ronan, R. and Brewer, H.B., Jr. (1984) J. Biol. Chem. 259, 3 18 - 322.
92 Jackson, R.L., Baker, H.N., Gilliam, E.B. and Gotto, A.M. (1977) Proc. Natl. Acad. Sci. USA 74, 1942- 1945. 93 Wei, C.F., Tsai, Y.K., Robberson, D.L., Gotto, A.M., Jr., Brown, K. andChan, L. (1985) J. Biol. Chem. 260, 15211 - 15221. 94 Humphries, S.E., Williams, L., Myklebost, O., Stalenhoef, A.F.H., Demacker, P.N.M., Baggio, G., Crepaldi, G., Galton, D.J. and Williamson, R. (1984) Hum. Genet. 67, 151 - 155. 95 Maguire, G.F., Little, J.A.. Kakis, G. and Breckenridge, W.C. (1984) Can. J. Biochem. Cell Biol. 62, 847-852. 96 Fojo, S.S., Law, S.W., Baggio, G., Sprecher, D.L., Taam, L. and Brewer, H.B., Jr. (1985) Clin. Res. 33, 569a. 97 Stalenhoef, A.F.H., Casparie, A.F., Demacker, P.N.M., Stouten, J.T.J., Lutterman, J.A. and Van’t Laar, A. (1981) Metabolism 30, 919-923. 98 Havel, R.J., Kotite, L. and Kane, J.P. (1979) Biochem. Med. 21, 121 - 128. 99 Menzel, H.J., Kane, J.P., Malloy, M.J. and Havel, R.J. (1986) J. Clin. Invest. 99, 595-601. 100 Wang, C.S., McConathy, W.J., Kloer, H.U. and Alaupovic, P. (1985) J. Clin. Invest. 75, 384 - 390. 101 Norum, R.A., Lakier, J.B., Goldstein, S., Angel, A., Goldberg, R.B., Block, W.D., Noffze, D.K., Dolphin, P.J., Edelglass, J., Bogorad, D.D. and Alaupovic, P. (1982) N. Engl. J. Med. 306, 1513- 1519. 102 Ginsberg, H.L., Le, N.A., Norum, R.A., Gibson, J. and Brown, W.V. (1984) Clin. Res. 32,396a. 103 Sparrow, J.T., Pownall, H.J., Hsu, F.J., Blumenthal, L.E., Culwell, A.R. andGotto, A.M. (1977) Biochemistry 16, 5427 - 5431. 104 Karathanasis, S.K., McPherson, J., Zannis, V.1. and Breslow, J.L. (1983) Nature 304, 371 -373. 105 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1985) J. Lipid Res. 26, 451 -456. 106 Levy-Wilson, B., Appleby, V., Protter, A., Auperin, D. and Seilhamer, J.J. (1984) DNA 3, 359 - 364. 107 Brewer, H.B., Shulman, R., Herbert, P., Ronan, R. and Wehrly, K. (1974) J. Biol. Chem. 249, 4975 - 4984. 108 Blaufuss, M.C., Gordon, J.I., Schonfeld, G., Straws, A.W. and Alpers, D.H. (1984) J. Biol. Chem. 259, 2452 - 2456. 109 Rees, A., Shoulders, C.C., Stocks, J., Galton, D.J. and Baralle, F.E. (1983) Lancet 1 #8322, 444 - 446. 110 Zannis, V.I., Just, P.W. and Breslow, J.L. (1981) Am. J. Hum. Genet. 33, I 1 -24. 111 Zannis, V.I. and Breslow, J.L. (1981) Biochemistry 21, 1033-1041. 112 Utermann, G., Jaeschke, M. and Mangel, J. (1975) FEBS Lett. 56, 352-355. 113 Jain, R.S. and Quarfordt, S.H. (1979) Life Sci. 25, 1315 - 1323. 114 Utermann, G., Langenback, U., Beisiegel, U. and Weber, W. (1980) Am. J. Hum. Genet. 32, 339- 347. 115 Zannis, V.I., McPherson, J., Goldberger, G., Karathanasis, S.K. and Breslow, J.L. (1984) J. Biol. Chem. 259, 5495 - 5499. 116 Pitas, E., Innerarity, T.L., Arnold, K.S. and Mahley, R.W. (1981) Proc. Natl. Acad. Sci. USA 76, 2311-2315. 117 Bersot, T.P., Mahley, R.W., Brown, M.S. and Goldstein, J.L. (1976) J. Biol. Chem. 251, 2395 - 2398. 118 Innerarity, T.L. and Mahley, R.W. (1978) Biochemistry 17, 1440- 1447. 119 Hui, D.Y., Innerarity, T.L. and Mahley, R.W. (1981) J . Biol. Chem. 256, 5646-5655. 120 Carrela, M. and Cooper, A.D. (1979) Proc. Natl. Acad. Sci. USA 76, 338-342. 121 Sherrill, B.C., Innerarity, T.L. and Mahley, R.W. (1980) J. Biol. Chem. 255, 1804- 1807. 122 Rall, S.C., Weisgraber, K.H. and Mahley, R.W. (1981) J . Biol. Chem. 257, 4171 -4178. 123 Mahley, R.W. and Angelin, B. (1984) Adv. Int. Med. 29, 385-411. 124 Rall, S.C., Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1985) Arteriosclerosis 5, 504a.
125 Blue, M.L., Williams, D.L., Zucker, S., Khan, S.A. and Blum, C.B. (1983) Proc. Natl. Acad. Sci. USA 80, 283-287. 126 Basu, S.K., Brown, M.S., Ho, Y.K., Havel, R.J. and Goldstein, J.L. (1981) Proc. Natl. Acad. Sci. USA 78, 7545 - 7549. 127 Breslow, J.L., McPherson, J., Nussbaum, A.L., Williams, H.W., Lofquist-Kahl, F., Karathanasis, S.K. and Zannis, V.I. (1982) J. Biol. Chem. 257, 14639- 14641. 128 Breslow, J.L., McPherson, J . , Nussbaum, A.L., Williams, H.W., Lofquist-Kahl, F., Karathanasis, S.K. and Zannis, V.I. (1983) J . Biol. Chem. 258, 11422. 129 Wallis, S.C., Rogne, S., Gill, L., Markham, A., Edge, M., Woods, D., Williamson, R. and Humphries, S. (1983) EMBO J. 2, 2369-2373. 130 McLean, J.W., Elshourbagy, N.A. Chang, D.J., Mahley, R.W. and Taylor, J.M. (1984) J. Biol. Chem. 259, 6498 - 6504. 131 Das, H.K., McPherson, J . , Bruns, G.A.P., Karathanasis, S.K. and Breslow, J.L. (1985) J. Biol. Chem. 260, 6240-6247. 132 Paik, Y.K., Chang, D.J., Reardon, C.A., Davies, G.E., Mahley, R.W. and Taylor, J.M. (1985) Proc. Natl. Acad. Sci. USA 82, 3445 - 3449. 133 Zannis, V.I., Breslow, J .L., Utermann, G., Mahley, R. W., Weisgraber, K.H., Havel, R.J., Goldstein, J.L., Brown, M.S., Schonfeld, G., Hazzard, W.R. and Blum, C.B. (1982) J. Lipid Res. 23, 91 1 - 914. 134 Utermann, G., Kinderrnann, I., Kaffarnik, H. and Steinmetz, A. (1984) Hum. Genet. 65, 232 - 236. 135 Assrnann, G., Schmitz, G., Menzel, H.J. and Schulte, H. (1984) Clin. Chem. 30/5, 641 -643. 136 Wardell, M.R., Suckling, P.A. and Janus, E.D. (1982) J. Lipid Res. 23, 1174-1182. 137 Cumming, A.M. and Robertson, F.W. (1984) Clin. Genet. 25, 310-313. 138 Ehnholm, C., Lukka, M., Kuusi, T., Nikkila, E. and Utermann, G. (1986) J . Lipid Res. 29, 229 - 235. 139 Zannis, V.I. and Breslow, J.L. (1980) J . Biol. Chem. 255, 1759-1762. 140 Breslow, J.L., Zannis, V.I., SanGiacomo, T.R., Third, J.L.H.C., Tracy, T. and Glueck, C.J. (1982) J . Lipid Res. 23, 1224- 1235. 141 Schneider, W.J., Kovanen, P.T., Brown, M.S., Goldstein, J.L., Utermann, G., Weber, W., Havel, R.J., Kotite, L., Kane, J.P., lnnerarity, T.L. and Mahley, R.W. (1981) J. CIin. Invest. 68, 1075 - 1085. 142 Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1982) J . Biol. Chem. 257, 2518-2521. 143 Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L. and Mahley, R.W. (1982) Proc. Natl. Acad. Sci. USA 79, 4696-4700. 144 Yamamura, T., Yamamoto, A., Hiramori, K. and Nambu, S. (1984)Atherosclerosis 50, 159- 172. 145 Yamamura, T., Yamamoto, A., Sumiyoshi, T., Hiramori, K., Nishioeda, Y. and Nambu, S. (1984) J. Clin. Invest. 74, 1229- 1237. 146 Innerarity, T.L., Weisgraber, K.H., Arnold, K.S., Rall, S.C., Jr. and Mahley, R.W. (1984) J. Biol. Chem. 259, 7261 -7267. 147 Rall, S.C., Jr., Weisgraber, K.H., Innerarity, T.L., Bersot, T.P., Mahley, R.W. and Blum, C.B. (1983) J . Clin. Invest. 72, 1288- 1297. 148 Wardell, M.R. and Janus, E.D. (1985) in: Proceedings of the 7th International Symposium on Atherosclerosis (P.J. Nestel, Ed.), International Atherosclerosis Society. 149 Weisgraber, K.H., Rall, S.C., Jr., Innerarity, T.L., Mahley, R.W., Kuusi, W. and Ehnholm, C. (1984) J. Clin. Invest. 73, 1024- 1033. 150 Ghiselli, G., Schaefer, E.J., Gascon, P . and Brewer, H.B., Jr. (1981) Science 214, 1239- 1241. 151 Zannis, V.I., Ordovas, J.M., Cladaras, C., Cole, F.S., Forbes, G. and Schaefer, E.J. (1985) J. Biol. Chem. 260, 12891 - 12894. 152 Moore, M.N., Kao, F.T., Tsao, Y.K. and Chan, L. (1984) Biochem. Biophys. Res. Commun. 123, 1-7.
153 Lackner, K.J., Law, S.W., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 122, 877-883. 154 Knott, T.J., Eddy, R.L., Robertson, M.E., Priestley, L.M., Scott, J. and Shows, T.G. (1984) Biochem. Biophys. Res. Commun. 125, 299- 306. 155 Scott, J., Knott, T.J., Priestley, L.M., Robertson, M.E., Mann, D.V., Kostner, G., Miller, G.J. and Miller, N.E. (1985) Lancet i, 771-773. 156 Huang, L.S., Miller, D.A., Bruns, G.A.P. and Breslow, J.L. (1986) Proc. Natl. Acad. Sci. USA 83, 644-648. 157 Law, S.W., Lee, N., Monge, J.C. and Brewer, H.B., Jr. (1985) Biochem. Biophys. Res. Commun. 131, 1003 - 1012. 157a Chan, L. Van Tuinen, P., Ledbetter, D.H., Daiger, S.P., Gotto, A.M., Jr. and Chen, S.H. (1985) Biochem. Biophys. Res. Commun. 133, 248 -255. 158 Bruns, G.A.P., Karathanasis, S.K. and Breslow, J.L. (1984) Arteriosclerosis 4, 97- 102. 159 Law, S.W., Gray G., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 118, 934-943. 160 Cheung, P., Kao, F.T., Law, M.L., Jones, C., Puck, T.T. and Chan, L. (1984) Proc. Natl. Acad. Sci. USA 81, 508-511. 161 Lusis, A.J., Taylor, B.A., Wangenstein, R.W. and LeBoeuf, R.L. (1983) J. Biol. Chem. 258, 5071 -5078. 162 Antonucci, T.K., VonDeimling, O.H., Rosenblum, B.B., Skow, L.C. and Meisler, M.H. (1984) Genetics 107, 463 - 475. 163 Meisler, M.H., Wanner, L., Kao, F.T. and Jones, C. (1981) Cytogenet. Cell Genet. 31, 124- 128. 163a Harper, M.E. and Chan, L. (1986) Methods Enzymol. 128, 863-876. 164 Schamaun, O., Olaisen, B., Mevag, B., Gedde-Dahl, T., Jr., Ehnholm, C. and Teisberg, P. (1984) Hum. Genet. 68, 181 - 184. 165 Olaisen, B., Teisberg, P. and Gedde-Dahl, T. (1982) Hum. Genet. 62, 233-236. 166 Whitehead, A.S., Bruns, G.A.P., Markham, A.P., Colten, H.R. and Woods, D.E. (1983) Science 221, 69-71. 167 Lin-Lee, Y.C., Kao, F.T., Cheung, P. and Chan, L. (1985) Biochemistry 24, 3751 -3756. 168 Scott, J., Knott, T.J., Shaw, D.J. and Brook, J.D. (1985) Hum. Genet. 71, 144-146. 169 Jeanpierre, M., Weil, D., Hors-Cayla, M.C., Williamson, R., Junien, C. and Humphries, S.E. (1984) Som. Cell Mol. Genet. 10, 645 - 649. 170 Fojo, S.S., Law, S.W., Brewer, H.B., Jr., Sakaguchi, A.Y. and Naylor, S.L. (1984) Biochem. Biophys. Res. Commun. 122, 687-693. 171 Humphries, S.E., Berg, K., Gill, L., Cumming, A.M., Robertson, F.W., Stalenhoef, A.F.H., Williamson, R. and Borresen, A.L. (1984) Clin. Genet. 26, 389-396. 172 Myklebost, O., Rogne. S., Olaisen, B., Gedde-Dahl, T., Jr. and Prydz, H. (1984) Hum. Genet. 67, 309-312. 173 Francke, U., Brown, M.S. and Goldstein, J.L. (1984) Proc. Natl. Acad. Sci. USA 81,2826-2830. 174 Gedde-Dahl, T., Jr ., Olaisen, B., Teisberg, P., Wilhelmy, M.C., Mevag, B. and Helland, R. (1984) Hum. Genet. 67, 178- 182. 175 Brook, J.D., Shaw, D.J., Meredith, L., Bruns, G.A.P. and Harper, P.S. (1984) Hum. Genet. 68, 282 - 285. 175a Luo, C.-C., Li, W.-H., Moore, M.N. and Chan, L. (1986) J. Mol. Biol. 187, 325-340. 176 Kessling, A.M., Horsthemke, B. and Humphries, S.E. (1985) Clin. Genet. 28, 296-306. 177 Rees, A., Stocks, J., Sharpe, C.R., Vella, M.A., Shoulders, C.C., Katz, J., Jowett, N.I., Baralle, F.E. and Galton, D.J. (1985) J. Clin. Invest. 76, 1090-1095. 178 Anderson, R.A. and Wallace, R.B. (1985) Clin. Res. 33, 328a. 179 Ferns, G.A.A., Stocks, J., Ritchie, C. and D.J. Galton (1985) Lancet ii, 300-303. 180 Wallis, S.C., Donald, J.A., Forrest, L.A., Williamson, R. and Humphries, S.E. (1984) Hum. Genet. 68, 286 - 289.
3 97 181 Lusis, A.J. and LeBoeuf, R.C. (1986) in: Methods in Enzymology (J.P. Segrest and J.J. Albers, Eds.) pp. 877 - 894, Academic Press, London. 182 Priestley, L., Knott, T., Wallis, S., Powell, L., Pease, R. and Scott, J. (1985) Nucl. Acids Res. 13, 6790. 183 Huang, L.S. and Breslow J.L. (1987) J. Biol. Chem., submitted. 184 Priestley, L., Knott, T., Wallis, S., Powell, L., Pease, R., Brunt, H. and Scott, J. (1985) Nucl. Acids Res. 13, 6793. 185 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1983) Nature 301, 718-720. 186 Karathanasis, S.K., Zannis, V.I. and Breslow, J.L. (1983) Nature 305, 823-825. 187 Karathanasis, S.K., Frazer, C., Zannis, V.I. and Breslow, J.L. (1984) Arteriosclerosis 4, 562a.