Comparison of the human apolipoprotein genes

J. Mol. Biol. (1985)

186, 43-51

Comparison

of the Human Apolipoprotein

Genes

Apo AI1 Presents a Unique Functional Intron-Exon Junction C. Simon Shelley, Colin R. Sharpe Francisco E. Baralle and Carol C. Shoulders Sir William Dunn School of Pathology University of Oxford South Parks Road, Oxford OX1 3RE, U.K. (Received 21 June 1985) The structure and function of the apolipoproteins are of interest because of their central role in lipid metabolism. We report the complete primary structure of the human apo AIT and CIII genes. These, like apo AI, contain three introns located at conserved positions. This may reflect their evolutionary and functional relationship. Indeed, computer-aided analysis shows that these apolipoproteins and apo AIV (rat), CI, CII and E contain homologous amino acid sequences. In the case of apo AI, AI1 and CIII, such sequences are encoded by equivalent exons. Finally, the apo AI1 gene presents a unique intron-exon junction sequence 5’ (G-T),,G-G-G-C-A-G 3’ that, although departing considerably from the accepted consensus (Py),,X-Py-A-G, appears to function efficiently both in vivo and in a transient expression system.

1. Introduction

to be monogenic in character (Fredrickson et al., 1978) and, furthermore, there probably exist many regulatory polymorphic alleles that respond inadequately to such environmental factors as a high fat diet although they may produce an ot,herwise functional protein. Consequently, to increase our understanding of the pathogenesis of these disorders, it is essential to analyse the structure and regulation of the genes involved. The HDL apolipoproteins, apo AI, AII, CIII, CII and CI, may be particularly important, since plasma concentrations of HDL appear to be inversely correlated to the incidence of arterial disease and often, therefore, to coronary disease (Miller & Miller, 1975; Miller et al., 1977). In this regard, we previously characterized the apo AI/C111 gene complex and subsequently ident,ified an associated restriction fragment length polymorphism that segregates with hypertriglyceridaemia in the Caucasian population (Rees et al., 1983, 1985; Shoulders & Baralle, 1984). Scott et al. (1985) have, in addition, identified a re&riction fragment length polymorphism linked to the human apo AI1 gene that appears to associate with elevated plasma levels of HDL. As a prerequisite to future studies of t)hese and other polymorphic apo AIjCTIT and AIT alleles, we have now fully characterized genomic clones of apo AI1 and CIII. Concomitantly we have compiled all amino acid and DNA sequences available for seven

Plasma lipoproteins are essential components of the complex system that regulates the tissue distribution and serum levels of cholesterol and triglycerides. They consist of a hydrophobic lipid core surrounded by a surface monolayer of polar interdigitated with unesterified phospholipids cholesterol and proteins known as apolipoproteins. They are divided by hydrated density into five major classes (CM, VLDL, LDL, IDL and HDL), each of which has a characteristic lipid and apolipoprotein composition. bhe apolipoproteins are During transport, involved in solubilizing hydrophobic lipids and, in addition, some can act as specific cofactors for of the enzymes involved in lipoprotein many or as exchange proteins for lipid metabolism, transfer between lipoproteins or as ligands for cellsurface receptors, thereby influencing the delivery site of each lipoprotein particle (for reviews, see Brown et al., 1981; Galton et al., 1982). However, although the apolipoproteins play a major role in regulating lipoprotein metabolism, only a few alleles polymorphic apoprotein mutant or (Ctermann et al., 1977; Norum et al., 1982) have been found to be associated with specific genetic diseases that are characterized by aberrant plasma lipid levels. This is perhaps not surprising, given that 800/, of lipid metabolism disorders appear not 0022-2836/85/210043~09

$03.00/O

43

0

1985 Academic

Yress Inc

(London)

Ltd.

44

C. S. Shelley et al.

in order to evaluate their of the apolipoproteins evolutionary relationship and to identify the sequences that may have functional significance.

2. Materials and Methods (a) Isolation of apo AII and CIII

genomic clones

Approximately 350,000 recombinants from a Charon 4A human gene library (Lawn et al., 1978) were screened with the isolated primer extended insert of Ml3 AI1 and Ml3 CID (Sharpe et al., 1984) using the plaque hybridization procedure of Benton & Davis (1977): 6 (AI1 1 to 6) and 2 (CIII 1 and 2) independent apo AI1 and apo CID recombinant clones, respectively, were isolated. pRHA1 5.7, pSVed ulW, Ml3 AI1 and Ml3 CIII are subclones that have been described (Shoulders et al., 1983; Higgs et al., 1983; Sharpe et al.. 1984). pHRA1 5.7 is a genomic subclone containing apo AI and 3’ CIII sequences, while pSVed alW, Ml3 AI1 and Ml3 CID contain al globin, apo BIT and CIII sequences, respectively. The methods used to construct and isolate other subclones used in this work are as follows. (1) pAIITS 1 4.0: a total SmuI digest of IA112 was ligated into the PvuII site of pAT153/PvuII/8 (Anson et al.. 1984) and a recombinant containing apo AI1 coding sequences isolated. (2) pS CID 1.9: a 1.9 kbt SstI restriction fragment purified from n CIII 1 was ligated into the SstI site of pBA1 (Shoulders & Baralle, 1982). (3) pR CID: a total RSAI digest of XIII 1 was ligated into the PwuII site of pAT153/PvuII/8 and the recombinant containing sequences 5’ from nucleotide 949 of the apo CIII gene (see Fig. 3) isolated. (4) pSVed alW/AII/I2: a total Pnu 4HI/DdeI digest of pAIITS 1 4.0 was cloned into the BstEII site of pSVedcr1 W and the recombinant containing 12 sequences isolated. The orientation of the 12 insert (extending from nucleotide 1446 to 1803 (Fig. 2)) was such that the alglobin and apo AI1 coding sequences were in the same strand of DNA. (b) DNA sequence determination The nucleotide sequences of the apo AI1 and CT11 genes were determined by the chemical degradation procedure (Maxam & Gilbert, 1977) using 3’ end-labelled restriction enzyme fragments isolated from pAIITS 1 4.0, pRHA1 5.7. pS CIII 1.9 and pR CIII. The sequences were assembled using the computer programs of Staden (1980). (c) Transient expression experiments pSVed ~11W/A11/12 (20 pg) was co-precipitated with calcium phosphate (Chu & Sharp, 1981) and added to subconfluent HeLa cells as described (Grosveld et al., 1983). RNA was extracted by the Nonidet P40 lysis method (Maniatis et al., 1982). size fractionated (Lehrach et al., 1977), and blotted onto Gene Screen membranes as recommended by the manufacturers (New England Nuclear). (d) Analysis of the evolutionary relationship between the apoproteins was

For this purpose, the program Relate (Dayhoff. 1976) used. Briefly, it compares all possible segments of a

t Abbreviations pair.

used: kb, lo3 base-pairs; bp, base-

given length from one sequence with all segments of the same length from the second sequence. Tt then accumulates a segment score from the pair scores of the amino acids occupying corresponding positions within the 2 segments. A segment comparison score is calculated as the difference between the number determined for t’he real sequences and the average determined from the many pairs of randomized sequences. divided by t,he standard deviation of t’he numbers from the randomized sequences. From the comparison segment score (expressed in S.D. units), one is t’hen able to calculat,e the probability of a particular score occurring by chance. The lower the score. the higher the probabilit,y.

3. Results The restriction

enzyme maps of independent

;i

AI1 and L CT11 clones isolated from a human gene library are shown in Figure 1. They bot’h span

regions of 21 kb, and present maps consistent, with those determined by total genomic DNA blotting (Southern, 1975). This characterization suggests that both genes, like that of apo AI, are present in a unique position in the haploid genome and provides no evidence for the existence of related pseudo genes. The sequence of the structural regions of the two genes (see Figs 2 and 3) has been compared with that previously obtained for human apo AI (see Fig. 4; Shoulders et al., 1983; Sharpe et al., 1984). This has led to the identification of a number of common and uncommon features present within these three apolipoprotein genes. These are outlined below. (a) The common features of human ape A I. ,-I I I and CIZI genes

Each of the genes contains sequences shared by other eukaryotic genes (see Figs 1 to 3). For example, they are associated with Alu-t,ype repetitive elements, int’rons, and have a. TATA box structure (Goldberg, 1979) and a polyadenylation signal (Proudfoot & Hrownlee, 1976). In addition, the 5’ flanking region of apo AI1 contains in its anti-sense strand two 3’ T-G-T-T-C-T 5’ sequences. which. in the reverse orientat’ion, have been implicated in other genes in steroid hormone/receptor binding (Van der Ahe rt al., 1985). No such hexanucleotides however are present in t,hp 440 bp upstream from the apo AI gene, despite the recent’ finding that apo AI levels are affected b! oestrogen in human tissue culture cells (Tam et al.. 1985). The conserved location of the introns within the three apolipoprotein genes is particularly st)riking (Fig. 4). All contain three introns that interrupt the transcribed sequences in the 5’ untransla,ted region. t’he signal peptide and the coding portion of t#hr mature plasma prot’ein. More specificallv, in each case the second exon encodes t’he e&e hydrophobic portion of the signal peptide. the third the NH, terminal sequences of the mature apolipoprotein, and the fourth the amphipat)hic helical region implicated in lipid binding.

Human Apo AII

and CIII

45

Gene Structures I kb

PATI53/PvulI/B

opo-CIlI

(b)

Figure 1. Restriction enzyme maps of the human apolipoprotein AI1 (a) and CIII (b) genes. (a) (I) EcoRI restrict,ion enzyme map of the region of chromosome 1 (Knott et al., 1984u; Lackner et al., 1984a; Moore et al., 1984) cont,aining t,he apo AI1 gene. (II and III) Restriction enzyme map of AA11 and pAIITS 14.0, respectively. In the latter. only the restriction sites used for sequencing are indicated. (IV) Organization of the apo AI1 gene. Open boxes denote 5’ and 3’ non-coding sequences; filled boxes, coding sequences; filled bars, introns. The Alu repetitive elements marked are in the same orientation as the apo AI1 gene (Baralle et al., 1980). (V) Partially spliced apo AI1 mRNA extending from nucleotide 1182 to 2269 of the apo AI1 gene (Fig. 2). (VI) Mature apo AI1 mRNA. (b) (I) BarnHI restriction enzyme map of the apo AI/C111 complex. pu’ote the orientation of the genes is reversed to that normally given (Karathanasis rt al., 1983). (II and III) Restriction enzyme maps of the complex. Note. however, there may be further PstI and BgZII sites 5’ from those upstream from the apo AI gene. (IV) Organization of t,he apo CIII gene, see (a) for det,ails of the conventions used. Note, however, the orientation of the Alu-type repetitive DEA is in the opposite orientation to those of the AI1 gene. Sequence hyphens are omitted from Figures for clarity. This st,ructural similarity may support the suggestion by Barker & Dayhoff (1977) that, these proteins and apo CI arose from a common precursor. In order to investigate this possibility, we have searched for DNA sequence similarities between apo AI. AII, AIV (rat) and apo CI, CTT, CIII and I3 complementary DNAs, both by visual inspection and by computer analysis using the programs Diagon and Relate (St,aden, 1982; Dayhoff, 1976). This investigation showed that, apart from the usual eukaryotic regulatory

sequences and the sequence 5’ T-C-(‘-(r-(l(C’)A 3’ in t’he 5’ non-coding region of apo AI. AI1 and E, no significant sequence homology is present in the noncoding regions of any apolipoprot,ein mRNAs characterized to date. In addition, this is the case for the apo AI, AI1 and CIII genes (see the legend to Big. 4). In contrast, the amino acid sequences of t,he apolipoproteins show a remarkable homology. To determine the most homologous regions, we used the computer program Relate (Dayhoff, 1976; see

Figure 2. The nu&wtidt~ srquen~~ of the human apolipoprotein AIT gtwr. The black bars at position 1 LOO to 1108. Sarron- open horrs 1148 to I I.‘,3 and ?Mi to “4X? indicate (‘.AT. TATA and AATAAA box-like struc%urrs. rcqwtivrly. regions. Broad shaded hoxrs denot,r translated sequc~~s. h’arroaintiic*atcs sequencvs cwiing for 5’ and 3’ untranslated hrokrn hoxrs at thr 5’ and 3’ sites intiic,ate possil,lr tranwription start (I,avknrr et crl.. 19816: Knott rt trl.. 19846: the 5’ sites. respectively. i\‘ote that thrrr arr X tvrnlini of out’ longwt (3l)?\‘.A is loc*atrd at position 1188) and poly(A) addition arise from the disc~,rpan~~i~~s u ith t,hr I)re\,iously ~)uhlishrtl 5’ nowcwdinp sequence (Knott ef nl.. 19846). Th ry prohal)ly limitation assocaiattd u ith the trchniqur used for their determination. The arrows at 59 to 61 and 6X to 637 indicate thr X to 5’ orirtltation of the Irc,xarru~lrotide srqutwc~ T-CLT-T-(‘-T in the antisensr strand. The polymorphic~ Nspl site. the 1;tc.k of M t1ic.h gives riw 10 t hr rrstric-tioll fragment length polymorphism asswiatrd with high I~rrls of H l)l, (Scott of ~1.. l!WT,). is l)rost~rlt at l)ositic)li 3Ki3

48

C. S. Shelley et al. . Gene length

cDNA length

i;‘l 1334

ff

I325

468

5’ I-El-Il-E2-I2-E3

13

E4 e

3’

w

apa-AII

1 Length (bp)

29-37

169

76

293

133

395

2-3

13

6

22

10

30

OX,Gene length

% cDNA Length 6-8

Length (bp)

16

230-231 17

20

48-49

15

197

63

185-187

157

586

658-661

% Gene length 1

10

3

10

8

31

34-35

% cDNA Length2

% Gene length 1 V. cDNA Length 5

7

20

16-18

2 13

4

4 24

,a0

69-74

59

10 58

Figure 4. Comparison of the organization of the human apo AI, AI1 and CIII genes. Boxes represent exons (E). connecting lines denote introns (I). Open boxes indicate untranslated sequences; speckled boxes, prepeptide sequences; striped boxes, propeptide sequences; and filled boxes, mature plasma protein sequences. The shaded cross motifs above E4 define regions of the mature protein implicated in lipid binding (McLachlan, 1977; Sparrow & Gotto, 1980; Mao et al.. 1977). The distance from the transcription start site to the poly(A) addition site defines the length of a gene. The range of gene and cDNA lengths reflect the findings of different laboratories (Shoulders et al., 1983; Karathanasis et al., 1983; Knott et al., 19846; Sharpe et al., 1984; Seilhamer et aZ., 1984; Moore et al., 1984; Lackner et al.. 19846). Note that sequences located 5’ from the mRNA capping site, 3’ from the polyadenylation signal and within introns show no sequence homology (see Results for further discussion). The only significant homology detected between these 3 genes. apart from intron/exon boundary sequences and AZuI-type repetitive DNA, occurs between nucleotides 1786 and 1811 of the apo AI gene (Shoulders et al., 1983; sequences coding for amino acids 81 to 89 of the mature protein) and the antisense nucleotides 125 to 147 of the apo CIII gene (Fig. 3). These sequences are 78.8% (26133) homologous.

Materials and Methods for a brief explanation of this program). The segment comparison score (expressed in S.D. units) of rat apo AIV compared with apo AI (22*9), CI (5.4), CII (3*5), CIII (4.8) and E (18.303) are particularly significant, the probability that the scoring homologous stretches of amino acids arose by chance is very low (of the order of P < 10vz3 for apo AI and E. P < lo-’ for CI and CIII). Apo CI, CII, CIIT and E all show greater homology with rat apo AIV than apo AT (s.D. values cf. with apo AI. 4.39, 2.58, 2.65 and 13,98, respectively), only apo ATI shows a similar degree of homology both to apo AI (5.14) and rat apo AI\’ (5.07). Finally, in Table 1, we list some of the most, homologous apolipoprotein sequences identified by the program Relate. These data demonstrate t’hat t,he homologous amino acid sequences of AI, ATT and CIII are encoded by equivalent exons. Whether this is the case for other homologous regions of other apolipoproteins remains to be determined. In this respect, it is significant t,o note that amino acids 82 to 259 of rat apo AIV are highly

homologous apo AI. (b) Atypical

to t,hose encoded

features

present genes

by the fourth

in apo Ail

and

exon of

C?III

Simple repetitive type sequences are present in the second intron of apo AI1 and (.‘lII. Tn the cast of apo CIII, the sequence is based on a repeat of the tetranucleotide 5’ C-T-T-T 3’ that flanks an AZZLtype repetitive sequence. This sequence has been found to associate with such repetitive elements (Baralle et aE., 1980). Here, however, it is repeated many more times and therefore may correspond t,o a nuclease S1 hypersensitive region. In contrast,, t.he repetitive sequence (5’ G-T 3’)16 present in the apo AI1 gene represents potential Z-DNA, and recent studies suggest it may possess weak enhancer-like properties (Hamada et aE., 1984). However, in this gene, the (G-T),, sequence is most notable for its proximity to the acceptor site of intron 2 (12) (see Fig. 2. nucleotides 1711 to 1748). This location. t,ogether with the presence of’ sequences corres-

Human Apo AII

and CIII

Gene Structures

Table 1 Comparison of the homologies within the apolipoproteins, to apo AI, AII and CIII

Residues of the mature proteint

AI AI AI AI AI AI AI1 AI1 AI1 AI1 CIII CIII

4-18 193-208 51-63 66243 -24--5 132-151 l-33 - 2-21 27-37 7-28 2/34 23-40

with particular

reference

o/o Homology at DNA level

Exon

46.6 62.5 53.8 46.0 57.0 53.3 45.0 40.5 51.0 46.9 49.5 47.5

3/3 414 414 a/ND 2/ND 4/ND 3/3 3/ND 3/ND 3/ND 3/ND 3/ND

2nd apolipoprotein

1st apolipoprotein

Apolipoprotein

Residues of the mature proteint

Apolipoprotein AI1 AI1 CIII AIVS E§ E CIII AIV AIV CII AIV E

-1-14 41-56 4652 62-239 -18-O 73-92 l-33 -2-21 113-123 18-39 l-33 U-61

This list is not exhaustive. Only the highest scoring homologous peptides (as identified program) with over 40% homology at the DNA level are shown. t Amino acid numbers are inclusive; ND, not determined. $ Boguski et al., 1984. $ McLean et al., 1984.

ponding to 12 and 13 in 10% of isolated apo AI1 cDNA clones (Sharpe et al., 1984; see Fig. lb(v)), suggested that we may have isolated a defective allele. To determine whether this was the case, we tested the ability of the (G-T),,G-G-G-A-G sequence to act as an acceptor site in a transient 8om

49

by the Relate

expression system. For this purpose, an expression construct, pSVedaIW/AII/I2, that contained 12 of the apo AI1 gene, together with its flanking exons, inserted within the human al globin gene, was transfected into HeLa cells (Fig. 5). RNA was extracted from the transfected cells and analysed

HI

I

2

3

4

5

Hind= 20s

ECORI

*

,x” Y.E 18.5 --+

pSVEDQl

W/AII/IL 930

+

3’ DdeI

(0)

Figure 5. (a) Map of the human al-globinlapo

(b)

(cl

AI1 gene construct used in t’ransient expression experiments, Striped and filled boxes represent human al-globin and apo AI1 exons, respectively. The G-T tract is denoted by an open box. A. Deletion of nucleotides 1426 and 2490 of pBR322. E and 0, Simian virus 40 (SV40) enhancer and origin of replication sequences. (b) Autoradiograph of a Northern blot of RNA from HeLa cells transfected with pSVed alW/AII/I2 and hybridized with the apo-AII-specific cDNA probe, M13AII. To estimate the size of this species, hepatoma G2 RNA was size fractionated in the adjacent slot and hybridized to an apo AI cDNA probe (890 nucleotides in length). The RNA species migrated similarly. (c) Northern blot of total and poly(A)+ liver RNA hybridized with M13AII. Lane 1, IO pg total RNA; lanes 2 to 5, 40, 20. 10 and 5 pg of poly(A)+ RNA.

C. S. Shelley et al

50

by the Northern blotting procedure using apo AI1 cDNA and al -globin-specific probes (see Materials and Methods). Both probes hybridized to a 930nucleotide species, the size of which is consistent with that predicted for a transcript arising from the Al-globin/AII hybrid construct lacking 12. Thus, 12 of this apo AI1 allele appears to be spliced efficiently. This is confirmed by Northern blots of the RNA used to construct the cDNA library from the pre-mRNA which the clones containing sequences were isolated. These failed to reveal any significant amounts of the partially spliced premRNA (see Fig. 5) and, therefore, the high proport’ion of clones in the cDNA library (Sharpe et al., 1984) corresponding to partially spliced apo AH mRNA probably results from the construction of the library with size-selected cDNA molecules and the use of a hybridization probe that detected only those recombmants containing sequences 5’ from nucleotide 1864 (see Fig. 2).

4. Discussion The human genes for apolipoprotein ATT and CIII have been isolated and their nucleotide sequence determined. The organization of these two genes is remarkably similar to that of apo AI. All contain the transcribed three introns, which interrupt sequence in the 5’ untranslated region, the signal peptide encoding region and the coding region of the mature plasma protein. The amphipathic helical regions of apo AI, AIT and CIII, associated with lipid binding, have been localized to amino acids 47 to 240 (McLachlan, 1977), 51 to 70 (Mao et al., 1977) and 48 to 79 (Sparrow & Gotto, 1980), respectively. Thus, in all the apolipoprotein genes characterized to date, the third intron separates t,he helical fourth exon, encoding an amphipathic region, from the amino-terminal portion. In addition, the second exon contains the entire hydrophobic region of the signal peptide. Such an arrangement is consistent with the hypothesis that exons represent functional and structural domains (Gilbert, 1978), although the significance of the conserved intron in the 5’ non-coding region is not clear. The structural similarity of these apolipoprotein genes may, in addition, reflect their evolutionary relationship. Indeed, the data presented in this paper show that the seven apolipoprotein genes analysed encode homologous stretches of amino acids, which are highly unlikely to have occurred by chance. Assuming that the role of convergent evolution is minimal, it, is possible that these stretches arose as a result of exon duplication followed by divergence, or alternatively, that’ a progenitor gene duplicated to give rise to a number of progeny in which the exons may have diverged at significantly different rates. Whatever the case, the identification of such homologous peptides should aid the assignment of biological functions to defined regions of the apolipoprotein molecules. Tn turn, the detailed structural analysis of the apo AI.

AI1 and CIII genes now available will facilitate the identification and characterization of polymorphic alt,ered lipid alleles that may associate with metabolism disorders. Finally, in the second intron of the apo AI1 gene a (dG-dl:), 6 tra,ct replaces t)he polypyrimidines that previously were synonymous with intron 5’ (Py),,X-Py-A-G 3’ acceptor sites (Mount,, 1982), and which have been implicated in cleavage at the 5’ donor site and lariat format’ion (Reed St Maniatis. 1985). This (dG-dU),, repeat stretches from the putative lariat, branch-site (position 1705; Fig. 2: and see Keller & Noon. 1984) to within 3 bp of t,he obligatory 4-G dinucleotide at the 3’ end of t,hr intron. Transient’ expression and Northern blot analysis indicate, however, t,hat I2 is excised in liver and in HeLa cell transient expression systems. Hitherto. mutations involving the conserved (i-T and A-G residues of int,ron donor and acceptor sites have been associated with diseased states (Treisman et al., 1983; Atweh rt al.. 1985). Here, however. the characterization of a naturally occurring variant of the broader 5’ (Pp),,X-Pp-A-G 3’ acceptor site consensus sequence indicates that any role the (PY)~~ tract, performs within the splicing may be mimicked by (G-T),,. mechanism Investigations are underway to define the functionally essential nucleotides of these sequences. We gratefully acknowledge the assistauce of D. J. (.:. Rees in computer

analysis,

and M. H. Murphy

and A.

Cameron for assistance in the earlier stages of the projec%. This work was supported by the Medical Researrh Council (grant, no. G8309498CB) and t’he British Heart Foundation (grants 81151 and 83/30).

References Anson; D. 8.. Choo, K. H.. Rres. D. J. C:.. Giannelli, F.. Gould. K.. Huddleston. J. A. & Brownlee. (:. ($. (1984). EMRO J. 3, 1053-1064. Atweh. G. F., Anagnou. h‘. P.. Shearin. J.. Forget. B. G. & Kaufman R. E. (1985). Xucl. dcids Reu. 13. 777

790. Baralle. F. E.. Shoulders, (‘. C., (~oodbourn, S.. Jrlfrrys. A. & Proudfoot,. Iv. J. (1980). Sucl. Acids BPS. 8. 4393-4404.

Barker, W. C. & Dayhoff. M. 0. (1977). (‘amp. Hiochem. Physiol. 57B, 309-315. Benton, W. D. & Davis, R. W’. (1977). Rcirnce. 196, 180 182. Boguski. M. S., Elshourbagy. h‘.. Taylor. J. M. bz Gordon. J. I. (1984). Proc. AVat. Acad. A’ci.. l’.A.d. 81. 5021 5025. Brown. M. S., Kovanen, P. T. & (Goldstein. J. 1,. (1981). Science, 212. 628-635. Chu, G. & Sharp. P. A. (1981). &nr, 13, 197-202. Dayhoff, M. 0. (1976). Editor of Atlas of Protein SPqwncc and Structure, Nat. Biomed. Res. Fdn, Washington. Fredrickson. D. S., Goldstein. ,J. I,. & Brown. M. S. (1978). In Metabolic Basis of Inherited Disease (Stanbury. ,J. B.. Wvngaarden, J. B. $ Fredrickson. D. S.. eds), 4th edit.. pp. 604-655. McGraw-Hill. New York. Galton, I). J., Stocks, J. & Dodson. 1’. M. (1982). C’lin. Riochem.

Rev. 3. 377-405.

Human Ape AII and CIII Gene Structures Gilbert, W. (1978). Nature (London), 271, 501. Goldberg, M. (1979). Ph.D. thesis, Stanford University, Calif. Grosveld, G. C., de Boer, E., Shewmaker, C. K. & Flavell, R. A. (1983). Nature (London), 295, 120-126. Hamada, H.. Seidman, M., Howard, B. H. & Gorman. C. M. (1984). Mol. Cell. Biol. 4. 2622-2630. Higgs, D. R., Goodbourn, S. E. Y., Lamb, J., Clegg, J. B., Weatherall, D. J. & Proudfoot, N. J. (1983). Natwe

(London),

396, 398-400.

Karathanasis, S. K., Zarmis, V. I. & Breslow. J. L. (1983). Proc. Nat. Acad. hi., U.S.A. 80, 6147-6151. Keller, E. B. & Koon, W. A. (1984). Proc. Nat. Acad. Sci., U.S.A. 81. 7417-7420. Knott. T. J., Eddy, R. L., Robertson, M. E.. Priestley, L. M.. Scott, J. & Shows, T. B. (1984a). Biochem. Biophys.

Res. Cornm,un.

125, 299-306.

Knott. I’. J., Priestley. L. M., Urdea, M. & Scott, J. (1984b). Biochem. Biophys. Res. Commun. 120, 734740. Lackner. K. J.. Law, S. W.. Brewer, H. B. Jr, Sakaguchi. A. P. & Naylor, S. L. (1984a). Bioehem. Biophys. Res. Commun.

122, 877-883.

Lackner,

K. J., Law, S. W. & Brewer, H. B. Jr (1984b). FEBS Letters, 175, 159-164. Lawn, R. M.. Fritsch, E. F.. Parker, R. C., Blake. G. & Maniatis, T. (1978). Cell, 15; 1157-I 174. Lehrach, H., Diamond. D., Wezney, J. M. & Boedtker. H. (1977). Biochemistry, 16, 4743-4751. Maniatis. T.. Fritsch. E. F. & Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press. Cold Spring Harbor. Mao. J. T.. Sparrow, J. T., Gilliam, E. B., Gotto, A. M. Jr & Jackson. R. L. (1977). Biochemistry, 16. 4150.. 4156. Maxam, A. M. $ Gilbert,. W. (1977). Proc. Nat. Acad. Sci., tT.S.il. 74, 560-564. McLachlan, -4. I). (1977). Nature (London), 267. 465-466. McLean, J. W., Elshourbagy, N. A.. Chang, D. J.. Mahley. R. W. & Taylor, J. M. (1984). J. Biol. Chem. 259, 6498-6504.

Miller, Miller, T. Moore.

G. J. & Miller, N. E. (1975). Lancet, i, 16-19. N. E., Weinstein, D. B., Carew, T. E., Koschinsky . & Steinberg, D. (1977). J. Clin. Invest. 60, 78-88. M. N.. Kao, F.-T., Tsao, Y.-K. & Chan, L. (1984). Biochem. Biophys. Res. Commun. 123, l-7. Edited

51

Mount, S. M. (1982). Nucl. Acids Res. 10. 459-472. Norum, R. A., Lakier, J. B., Goldstein, S.. Angel, A., Goldberg, R. B., Block, W. D.. Woffze, D. K., Dolphin, P. J., Edelglass, J.. Bogorand, D. D. & Alanpovic, P. (1982). New Eng. J. Med. 396. 15131519. Protter, A. A.. Lexy-Wilson, B.. Miller. tJ.. Bencen. G.. White, T. & Seilhamer, J. I. (1984). DNA. 3, 449456. Proudfoot, h’. J. 8: Brownlee. G. G. (1976). Nature 263, 21 l-214. (London), Reed, R. & Maniatis, T. (1985). Cell. 41. 955105. Rees, A.. Shoulders. C. C.. Stocks. J.. Galton. D. J. & Baralle, F. E. (1983). Lancet, i. 4444446. Rees, A., Stocks, J., Sharpe. C. R.. Vella. M. A., Shoulders, C. C., Katz, J., Baralle. F. E. & Galton, D. J. (1985). J. Clin. Invest. In the press. Scott, J., Knott, T. J., Priestley, L. M.. Robertson, M. E.. Mann. D. V.. Kostner, G. H.. Miller. G. J. & Miller, N. E. (1985). Lancet, i. 771-773. Seilhamer.
269, 604-607.

Van der Ahe, D.. Janich, S.. Scheidereit. (I.. Renkawitz. R,, Schutz, G. 8: Beato, M. (1985). ,\‘atura (London), 313. TOB--709.

by S. Brenner