Apolipoprotein Multigene Family: Tandem Organization of Human

Proc. Nati. Acad. Sci. USA Vol. 82, pp. 6374-6378, October 1985 Biochemlstry Apolipoprotein multigene family: Tandem organization of human apolipoprotein AI, CIII, and AIV genes (apolipoprotein gene evolution/apolipoprotein AIV sequence/apolipoprotein AIV tissue-specific expression/apolipoprotein AIV developmental regulation) SOTIRIOS K. KARATHANASIS Laboratory of Molecular and Cellular Cardiology, Department of Cardiology, Children's Hospital, and Department of Pediatrics, Harvard Medical School, Boston, MA 02115 Communicated by Michael S. Brown, April 29, 1985

AISTRACT The genes for two of the proteins of the in the bacteriophage vector XL47.1 (8) was carried out as plasma lipid transport system, apolipoprotein AI (apoAI) and detailed (9). An adult human liver cDNA library, in the CIII (apoCIll) are closely linked in the human genome. An bacteriophage vector XgtlO, was a generous gift of E. F. =s30-kilobase (kb) DNA segment containing these genes and Fritsch. Screening of these libraries and preparation of their flanking sequences has been cloned and extensively recombinant X clone DNA was carried out as described (10). characterized. Hybridization studies revealed that a DNA Blotting Methods and Radioactive Labeling of DNA. DNA fragment located 12 kb 3' to the apoAI gene contains sequences and RNA blotting analyses were carried out as reported (11,, homologous to a 1.8-kb mRNA transcript in human fetal 12). Restriction endonucleases were from New England intestine and adult liver but not in fetal liver, kidney, heart, Biolabs. Radioactive labeling of DNA was carried out by brain, or muscle. This DNA fragment was used as a probe to nick-translation (13). isolate a clone from an adult human liver cDNA library. The Human DNA and RNA Purification. Human chromosomal nucleotide sequence of this clone is 74.8% homologous to the DNA was prepared from blood (2), and recombinant M13 cDNA sequence of rat apolipoprotein AIV (apoAIV), another replicative form DNA was isolated as described (14). DNA protein of the lipid transport system, and codes for a protein fragments were isolated from low-melt agarose gels (Bio- that is 58.6% identical to rat apoAIV. These results indicate Rad), phenol-extracted, and precipitated with ethanol. The that apoAI, apoClIl, and apoAIV genes are closely linked in the preparation of total RNA, generously provided by V. I. human genome and suggest that all three of them are derived Zannis, was carried out according to a reported protocol (15). from a common ancestral precursor. Nucleotide Sequencing Determinations and DNA Sequence Comparisons. DNA sequencing analysis was carried out Apolipoproteins (apo) are lipid binding proteins involved in either by the base-specific chemical cleavage method (16) or the transport of cholesterol, triglycerides, and phospholipids by subcloning various DNA fragments in M13 vectors (17) in the plasma. Defects in apolipoprotein structure or biosyn- and using the resulting single-stranded DNA templates for thesis may result in disorders of the plasma lipid transport DNA sequence determination by the dideoxy-chain termina- system and the development of coronary artery disease (1). tion method (18). Comparisons of nucleotide sequence ho- It has recently been shown that mutations in the genes coding mologies were carried out using the SEQ computer program for apoAI and apoCIII are associated with premature coro- (19). nary artery disease and hypertriglyceridemia, respectively (2-4). RESULTS Structural homologies in the amino acid sequences of certain apolipoproteins-namely, Al, All, CI, and CIII- Cloning and Characterization of a Human Genomic DNA have led to the speculation that all of these proteins arose Segment Containing the apoAI and apoClII Genes and Their from a common evolutionary precursor (5). Genes of com- Flanking Sequences. Screening of a human genomic DNA mon evolutionary origin are members of multigene families library with an apoAl cDNA probe resulted in isolation of and in many cases are closely linked (6). The human apoAI several clones, one of which (XApoAI-6) was shown to and apoCIII genes are linked (7), and therefore it is possible contain both the apoAI and apoCIII genes (7, 14). Another of that additional related genes may be located in their vicinity. these clones (XApoAI-8) was purified and mapped with This report shows that a third apolipoprotein gene, restriction enzymes (Fig. 1). This clone contains a portion of apoAIV, is located 12 kilobases (kb) 3' to the apoAIgene and the apoAI gene, the entire apoCIII gene, and ==7 kb of the suggests that apoAI, apoCIII, and apoAIV genes are mem- apoCIIIgene 5' upstream sequences (Fig. 1). A 1.2-kb EcoRI bers of a multigene family generated from a common evolu- DNA fragment, located 12 kb 3' to the apoAI gene, was tionary precursor. In addition, it is shown that a cDNA- isolated from the clone XApoAI-8 and was used as a probe for derived human apoAIV protein sequence contains 22-amino- screening 5 x 105 clones of a human lymphocyte DNA acid-long repeats with the potential of forming amphipathic library. Three positive clones were isolated, and one ofthem helices. (AApoAI-16) was mapped with restriction enzymes (Fig. 1). Genomic mapping of the apoAI and apoClIl genes and their flanking sequences was carried out using as probes a 2.2-kb MATERIALS AND METHODS Pst I, a 1.0-kb BamHI/Pst I, or a 1.2-kb EcoRI DNA Construction, Screening of Bacteriophage Human DNA fragment containing apoAI (14) and apoCIlI (20) gene se- Libraries, and Isolation of Recombinant Clone DNA. Con- quences or sequences located 12 kb 3' to the apoAl gene, struction of a human lymphocyte chromosomal DNA library respectively (Fig. 1 B-D). The resulting map (Fig. LA) is compatible with both the maps ofclones XApoAI-6, XApoAI- 8, and XApoAI-16 (Fig. 1) and the genomic maps constructed The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviations: apo, apolipoprotein; kb, kilobase(s); bp, base pair(s). 6374 Downloaded by guest on September 23, 2021 Biochemistry: Karathanasis Proc. NatL Acad ScL USA 82 (1985) 6375

B. R BI3K C. R B H3 K D. R B H3 K

23 1 - 23 1- 23 1- 94- g.4- 94- lot*qo4 6 5- 6 5- 65- 4 3- 4 3- 4.3-

2 3. 2.3- 2 3- 20- 2 0- 2 0- 1 6- 1 6- 1 6- S ApOA I ApoC III 12Kb EcoR Probe Probe Probe

ApoAIV ApoAl ApoCIlIl 1Kb 5 3' 3 5 5 ,3

BR B BRBKR R BH3 H3 K K H3 R K f f H3 B K a I I I,, 1i Mf f I If 'I I I A. I I1 I I - __I_ -mm IIII

XApoAI-6. XApoAI-8. XApoAI- 16.

FIG. 1. (A) Restriction endonuclease map of a DNA segment containing the human apoAI, apoCIII, and apoAIV genes. The direction of transcription and the relative location of the genes are shown by arrows and solid boxes, respectively. Highly repetitive DNA sequences are indicated by shaded boxes. The extent and the relative location of genomic DNA fragments, contained in the specified X clones, are shown by lines under the map. The nucleotide scale of the map is indicated by a solid bar. (B-D) Autoradiograms resulting by genomic blotting analysis of human lymphocyte DNA using as probes (double-headed arrows) apoAI and apoCI gene sequences and sequences located 12 kb 3' to the apoAI gene, respectively. Numbers along the sides of these autoradiograms indicate the migration of DNA fragments generated by digestion of bacteriophage A and plasmid pBR322 DNAs with HindIII and Hinfl, respectively (size markers). Restriction endonucleases are indicated as follows: B, BamHI; H3, HindIII; K, Kpn I; R, EcoRI.

previously using apoAl or apoCIII cDNA probes (3, 7, 14). quences derived from adult human liver apoAIV mRNA and These results indicate that the apoAI and apoCIII genes and that the apoAIV gene is located 12 kb 3' to the apoAI gene. the 1.2-kb EcoRI DNA fragment, located 12 kb 3' to the In addition, these results suggest that expression of the apoAI gene, are all physically linked and tandemly organized human apoAIV gene is tissue specific (fetal intestine) and in a 17-kb HindIII human genomic DNA fragment (Fig. LA). developmentally regulated in human liver. The highly repetitive DNA elements in this region of the Linkage of apoAl, apoCIII, and apoAIV Genes and Deter- human genome were identified by restriction blotting analysis mination of the Direction of apoAIV Gene Transcription. Both of purified XApoAI-6, XApoAI-8, and XApoAI-16 DNAs the human apoAI and apoCIII genes are located in a 10-kb using as a probe total human chromosomal DNA (Fig. 1). The apoAIV Gene Is located 12 kb 3' to the apoAI Gene. A search for the presence of additional genes in the vicinity of 1 2 3 4 5 6 7 _X __ _~ _ _ the apoAI and apoCIII genes was carried out by hybridization of blotted human tissue mRNA with a probe prepared using the 1.2-kb EcoRI fragment located 12 kb 3' to the apoAI gene (Fig. 1). The resulting autoradiogram shows that fetal intestine and adult liver, but not fetal liver, kidney, heart, 28S- brain, or muscle contain a 1.8-kb transcript with sequences homologous to this probe (Fig. 2). This probe was also used 18S- for screening 1.5 x 105 clones of a bacteriophage XgtlO adult NW .0 human liver cDNA library. Ten positive clones were isolated, and one of them (XgtlO-2) was mapped with restriction enzymes and used for DNA sequence determinations (Fig. 3). Computer comparison (19) of the clone XgtlO-2 and rat apoAIV cDNA sequences revealed that nucleotides 288-768 in the XgtlO-2 cDNA (Fig. 3B) are significantly (E < 0.001) homologous (74.8%) to nucleotides 780-1260 in the rat FIG. 2. Total RNA (20 ,ug per lane) was isolated from human fetal apoAIV cDNA (21). Furthermore, alignment of the XgtlO-2 liver (lane 1), intestine (lane 2), kidney (lane 3), heart (lane 4), brain and the rat apoAIV cDNA-derived amino acid sequences (lane 5), muscle (lane 6), or adult liver (lane 7) and hybridized with showed that 58.6% of their residues are identical while 8.6% a probe prepared using a 1.2-kb EcoRI fragment located 12 kb 3' to of them represent conservative (22) substitutions (Fig. 4). the apoAl gene (Fig. 1). The resulting autoradiogram is shown. The These results indicate that the clone XgtlO-2 contains se- migration of 28S and 18S ribosomal RNAs is indicated. Downloaded by guest on September 23, 2021 6376 Biochemistry: Karathanasis Proc. NatL Acad Sci USA 82 (1985)

A 135 155 Sac Sac Rat: L R P Y A T D L Q AQ I N A Q T Q D M K R Q L T P Y I Q R M QT T I Q DN V EN * S 0 a 3 P.11| EceRi SacSacl Osti Opali 01.11nt1~ o 000 0 *no *w.@o n**@ news F~c.--Il Psi ^c,7Ft Human: L E P Y A D Q L R T Q V N T Q A E Q L R R Q L D P L A Q R M E R V L R E N A 0 S 175 195 Rat: L Q S S M \ P F A N E L K E K F N Q N M E G L K G Q L T P R A N E L K A T I D Q 0O em * S 0*@0 @0.0 @00 0. 0 0 a *S Human: L Q A S L R P H A DE L K A K I D Q N V E E L KGR L T P Y A D E F K V K I D Q 500 800 100 200 300 400 600 700 900 215 235 Rat: N L E D L R S R L A P L A E G V Q E K L N H Q M E G L A F Q M K K N A EE L HT B *memo @05 * *@0@@e@Ee@@ ...... Human: T V E E L R R S L A PY A Q D T Q E K L N H Q LE G L T F Q M K K N A EE L K A CTG GAG CCC TAC GCG GAC CAG CTG CGC ACC CAG GTC AAC ACG CAG GCC GAG CAG 54 Leu Giu Pro Tyr Ala Asp Gin Leu Arg Thr Gin Val Asn Thr Gin Ala Gia Gin 255 275 Rat: K V S T N I D Q L Q K N L A P L V E D V Q S K L K G N T E G L Q K S L E D L NK .mN m 0 *OSOESS * .*.**...... m. Human: R I S A S A E E L RQ R L A P L A E D VRG N LK G N TEG LQK S L AE L GG CTG CGG CGC CAG CTG GAC CCC CTA GCA CAG CGC ATG GAG AGA GTG CTG CGG GAG 108 Leu Arg Arg Gin Leu Asp Pro Lea Ala Gin Arg MET Gin Arg Sal Leu Arg Gia 295 315 Rat: Q L D Q Q V E V F R R AS E PL G DK F N M A L V Q Q M E K F R Q Q L G S D S G 005550 O0 00 e 0@ 0.00.0@-0 Se 0 Human: H L D Q Q V E E F R R R V E P Y G E N F N K A L V Q Q M E Q L R Q K L GP H AG AAC GCC GAC AGC CTG CAG GCC TCG CTG AGG CCC CAC GCC GAC GAG CTC AAG GCC 162 335 355 Asn Ala Asp Ser Leu Gin Ala Ser Leu Arg Pro His Ala Asp Giu Len Lys Ala N P L A LP L PE Rat: 05D V E SH O@0SL S F L E K *@U@@L R E K V S 00-SF M 00S T L Q KK0 G 0SP *D Q 0 Human: D V E G H L S F L E K D L RDK V N S F F S T F K E K E SQD K T L S L P E L E AAG ATC GAC CAG AAC GTG GAG GAG CTC AAG GGA CGC CTT ACG CCC TAC GCT GAC 216 375 Lys Ilie Asp Gin Asn Val G14 Gin Len Lys Gly Arg Leu Thr Pro Tyr Ala Asp Rat: Q VQ EQ VQ EQ VQ PK P LE S * 00 EcoRI Human: Q Q Q E GAA TTC AAA GTC AAG ATT GAC CAG ACC GTG GAG GAG CTG CGC CGC AGC CTG GCT 210 Gin Phe Lys Sal Lys Ilie Asp Gin Thr Sal Giu Gin Lou Arg Arg Ser Len Ala FIG. 4. Alignment of the Xgt10-2 and the rat apoAIV cDNA- derived amino acid sequences (designated by the single-letter code). CCC TAT GCT CAG GAC ACG CAG GAG AAG CTC AAC CAC CAG CTT GAG GGC CTG ACC 324 Identical amino acids and amino acids representing conservative Pro Tyr Ala Gin Asp Thr Gin Gin Lys Len Ann His Gin Len Glu Gly Len Thr substitutions are indicated by solid circles and solid squares, respectively. The residues of the rat apoAIV amino acid sequence are TTC CAG ATG AAG AAG, AAC GCC GAG GAG CTC AAG GCC AGG ATC TCG GCC AGT GCC 378 numbered. PhO Gin MET Lys Lys Ann Ala Gin Gin Lea Lys Ala Arg Ilie Ser Ala Ser Ala

GAG GAG CTG CGG CAG AGG CTG GCG CCC TTG GCC GAG GAC GTG CGT GGC AAC CTG 432 Kpn I DNA fragment. A 5-kb and a 15-kb Kpn I DNA Glu Glu Leu Arg Gln Arg Leu Ala Pro Leu Ala Glu Asp Val Arg Gly Asn Leu fragment flank, sequentially, the 3' end of the apoAl gene (5' end of the apoCIII gene) (Fig. 1A). The 220-bp and a 750-bp AAG GGC AAC ACC GAG GGG CTG CAG AAG TCA CTG GCA GAG CTG GGT GGG CAC CTG 486 EcoRI DNA fragment, containing 5' and 3' apoAIV cDNA Lys Gly Asn Thr Glu Gly Leu Gln Lys Ser Leu Ala Glu Leu Gly Gly His Leu sequences respectively, were isolated from the clone Xgt1O-2 (Fig. 3A) and used as probes for genomic blotting analysis of GAC CAG CAG GTG GAG GAG TTC CGA CGC CGG GTG GAG CCC TAC GGG GAA AAC TTC 540 Kpn I-digested human chromosomal DNA. The resulting Asp Gln Gln Val Glu Glu Phe Arg Arg Arg Val Glu Pro Tyr Gly Glu Asn Phe autoradiograms show that the 220-bp probe hybridizes to both a 5-kb and a 15-kb Kpn I fragment, while the 750-bp AAC AAA GCC CTG GTG CAG CAG ATG GAA CAG CTC AGG CAG AAA CTG GGC CCC CAT 594. probe hybridizes only to a 15-kb Kpn I fragment (Fig. 5). Asn Lys Ala Leu Val Gln Gln MET Glu Gln Leu Arg Gln Lys Leu Gly Pro His These results indicate that apoAI and apoAIV genes are transcribed in the same direction: opposite from the direction GCG GGG GAC GTG GAA GGC CAC TTG AGC TTC CTG GAG AAG GAC CTG AGG GAC AAG 648 of apoCIII gene transcription (Fig. 1A). These probes were Ala Gly Asp Val Glu Gly His Leu Ser Phe Leu Glu Lys Asp Leu Arg Asp Lys also used for genomic blotting analysis of human DNA digested with HindIII. The resulting autoradiograms indicate GTC AAC TCC TTC TTC AGC ACC TTC AAG GAG AAA GAG AGC CAG GAC AAG ACT CTC 702 that, similar to the apoAI and apoCIll genes (Fig. LA), the Val Asn Ser Phe Phe Ser Thr Phe Lys Glu Lys Glu Ser Gln Asp Lys Thr Leu apoAIV gene is located on a 17-kb HindIII DNA fragment (Fig. 5). TCC CTC CCT GAG CTG GAG CAA CAG CAG GAA TAG CAG CAG GAG CAG CAG CAG GAG 756 Ser Leu Pro Glu Leu Glu Gln Gin Gln Glu i. Structure of apoAIV cDNA and Amino Acid Sequences. An C GIn Gin Gln Glu Gln Gln Gln Gli amino acid sequence of the human apoAIV was obtained by in-frame translation of the cDNA sequence of the clone CAG GTG CAG ATG CTG GCC CCT TTG GAG AGC TGA GCT GCC CCT GGT GCA CTG GCC 810 XgtlO-2 (Fig. 3B). The same amino acid sequence was also Gln Val Gln MET Leu Ala Pro Leu Glu Ser obtained by using a different cDNA clone (XgtlO-8; data not shown). This sequence, however, is 18 residues shorter at its CCA CCC TCG TGG ACA CCT GCC CTG CCC TGC CAC CTG TCT GTC TGT CTG TCC CAA 864 carboxyl terminus compared to the sequence obtained by yet AGA AGT TCT GGT ATG AAC TTG AGG ACA CAT GTC CAG TGG GAG GTG AGA CCA CCT 918 a different apoAIV cDNA clone (XgtlO-3; Fig. 3B). It remains CTC AAT ATT CAA TAA AGC TGC TGA GAA TCT AGC CTC AAA AAA AAA AAA AAA AAA 972

A H3_~ K B H3 K b FIG. 3. (A) Restriction map and DNA sequencing strategy of clone XgtlO-2. The bottom scale is in nucleotides. The arms of the vector (XgtlO) and the EcoP sites due to the linkers used for the construction ofthe cDNA library are indicated by Wavy lines and are 2391- 23.14- _ boxed, respectively. Arrows under the map represent the direction 9.4-_ 9.4 _ Intact or dashed and extent of nucleotide sequence determinations. 6.5- 6.5- arrows indicate DNA sequences obtained by using the enzymatic (by subcloning in the appropriate M13 vectors) or the chemical methods. (B) The cDNA and the derived amino acid sequences of the clone 4.3- 4.3- XgtlO-2. Numbers indicate nucleotide residues. The polyadenylyla- tion signal is underlined. Termination codons and proline residues are indicated by dots and stars, respectively. The nucleotide substi- FIG. 5. Genomic blotting analysis of human lymphocyte chro- tution (T to C) found in a different cDNA clone (AgtlO-3) is shown by mosomal DNA using as probes (A) a 220-bp or (B) a 750-bp EcoRI an arrow. The resulting amino acid sequence due to this substitution DNA fragment isolated from the clone XgtlO-2. Size markers and is also shown. restriction endonucleases are as in Fig. 1. Downloaded by guest on September 23, 2021 Biochemistry: Karathanasis Proc. NatL Acad Sci USA 82 (1985) 6377 to be established whether these differences represent differ- DISCUSSION ent human apoAIV gene alleles or are due to cloning artifacts. Inspection of the XgtlO-2 cDNA-derived apoAIV amino An -30-kb DNA segment containing the human apoAI and acid sequence revealed the presence of proline residues at apoCIII genes has been cloned and characterized. Hybrid- regular positions of 22-amino-acid-long intervals (Fig. 3B). ization studies showed that human fetal intestine and adult Alignment of this sequence in blocks starting, whenever liver but not fetal liver, kidney, heart, brain, or muscle possible, with proline resulted in eight (I, II, III, IV, V, VI, contain a 1.8-kb transcript with sequences homologous to a VIII, IX) 22- and one (VII) 18-amino-acid-long segments (Fig. 1.2-kb EcoRI DNA fragment located 12 kb 3' to the apoAI 6A). The nucleotide sequences coding for these amino acid gene. This 1.2-kb EcoRI fragment was used as a probe to segments were similarly aligned (Fig. 6C) and a nucleotide isolate a clone from an adult human liver cDNA library. The consensus sequence was generated (Fig. 6D). Comparison of cDNA sequence of this clone is significantly homologous to this consensus sequence with each of the sequences coding the rat apoAIV cDNA sequence and codes for a protein that for these amino acid segments indicates homologies ranging is 58.6% identical to rat apoAIV. These results indicated that from 50.0% to 72.72% (Fig. 6C). Computer analysis showed this cDNA clone is derived from human liver apoAIV mRNA that, for the nucleotide sequences coding for the amino acid and that the human apoAIV gene is located 12 kb 3' to the segments I, III, IV, V, VII, and VIII, these homologies are apoAI gene. Restriction mapping of human chromosomal highly significant with E values of <0.001 while the E values DNA indicated that the apoAI, apoCIII, and apoAIV genes for the sequences corresponding to segments II, IX, and VI are linked and tandemly organized in a 17-kb HindIII DNA are >1 and 0.0035, respectively. These results indicate that fragment and that the apoAI and apoAIV genes are tran- the 66-bp repeats in the apoAIV gene code for 22-amino acid scribed in the same direction: opposite from the direction of repeats and suggest that the apoAIV gene may have evolved apoCIII gene transcription. by intragenic amplification of an ancestral 66-bp sequence. The human apoAI gene contains 66-bp repeats possibly Comparison of the apoAIV nucleotide consensus sequence evolved by intragenic amplification of an ancestral 66-bp (Fig. 6D) with the similarly derived apoAI (14) or apoE (23) sequence (14). Such amplification of a short ancestral DNA consensus sequences showed 50% and 56% homologies with segment has also been implicated in the evolution of a- E values of 0.95 and <0.001, respectively (Fig. 6 E and F). It fetoprotein and albumin genes, and it has been suggested that is therefore likely that similar intragenic amplification of the the nucleotide sequence of this ancestral DNA segment same or a similar ancestral 66-bp sequence was involved in resembled the consensus sequence of the internal repeats in the generation of the apoAI, apoAIV, and apoE genes. these genes (25). The human apoAIV gene contains 66-bp Alignment of the 22-amino acid repeats in human apoAIV repeats with a consensus sequence 50% homologous to the facilitated the generation of an amino acid consensus se- consensus sequence of the repeats in the apoAI gene. It may quence (Fig. 6A). Plotting the residues in this sequence as a therefore be suggested that the apoAI and apoAIV genes function of their hydropathic index (24) resulted in a pattern arose by intergenic duplication of an apolipoprotein AI/AIV that alternates regularly between hydrophobicity and gene precursor that evolved by intragenic amplification of an hydrophilicity (Fig. 6B). Similar patterns were also obtained ancestral 66-bp sequence. Thus, the extensive homology of with each of the 22-amino acid repeats in human apoAIV the human and rat apoAIV cDNA sequences may indicate (data not shown). These results indicate that the 22-amino- that these intra- and intergenic duplication events occurred acid-long repeats in human apoAIV have the potential of before mammalian radiation and that the apoAI and apoAIV forming amphipathic helices (24). genes may also be closely linked in the rat genome.

A. FMIND ACID REPEATS C. WCLEOTIDE REPEATS HloW0Y I (003) - PYWRTQVNTQALW _ 007) - - 68.18I II (OQ5) - Vl RENA Q-II 073) - I--68.1851.51 III (047) - PHADELKMKID -VELKGRLT III 139) --511- 71.21 IV (069) - "ADEFKVKIDQTVEEI.tSLA IV (205) - _ '-- 12.7272.72 V (091) - "ANDTQEUIH TFQ V (271) - i - 59.09 VI 113 - KNUELKMIS U VI (337) i 660.60 - VII 135 PAEDVRGKTNL E.Q VII -40- 6262.96. VIII 153 - KSLAELGGHLDQQVEEFMVE I0I457 - i- 56.06 IX (175) - PYGENFNI.VQQNEQML G IX (523-,- 5050.00 COIISENSUS PTADEULJKLDQQAEELLX SEQUENCE E V

FIG. 6. (A) Alignment of the clone XgtlO-2 cDNA-derived amino acid sequence (residue numbers are in parentheses) in segments (I-IX) starting whenever possible with proline and an amino acid consensus sequence generated based on the most frequent residue or type of residue per column. Amino acids are identified by the single-letter code. (B) The hydropathic index (solid circles) of each residue in the amino acid consensus sequence ofA was plotted as a function of its residue number. For residues 8 and 22, no hydropathic index was assigned; for residue 14, the average hydropathic index (open circle) ofthose for valine and alanine was used. (C and D) Alignment ofthe clone XgtlO-2 cDNA sequence (residue numbers in parentheses) in segments (I-IX) corresponding to the amino acid segments in A and a derived nucleotide consensus sequence, respectively. (E and F) Consensus sequences derived from the internal repeats in apoAI and apoE genes, respectively. The percent homology of the nucleotide sequence in D to the sequences in C, E, and F is indicated. Downloaded by guest on September 23, 2021 6378 Biochemistry: Karathanasis Proc. Nad Acad Sd USA 82 (1985) Common evolutionary origin of the apoAI and apoClIl 1. Herbert, P. N., Assmann, G. & Gotto, A. M., Jr. (1982) in The genes is suggested by their close physical linkage and by the Metabolic Basis of Inherited Disease, eds. Stanbury, J. B., presence ofintrons interrupting their coding regions at similar Wyngaarden, J. B., Fredrickson, D. S., Goldstein, J. L. & positions (7, 20, 26). The linkage of the human apoCIII and Brown, M. S. (McGraw-Hill, New York). 2. Karathanasis, S. K., Norum, R. A., Zannis, V. I. & Breslow, apoAIV genes and the possible evolutionary relationship of J. L. (1983) Nature (London) 301, 718-720. the apoAI and apoAIV genes may therefore indicate that the 3. Karathanasis, S. K., Zannis, V. I. & Breslow, J. L. (1983) apoCIII gene derived from the same ancestral sequence Nature (London) 305, 823-825. (CIII/AI/AIV ancestor) that gave rise to the AI/AIV gene 4. Rees, A., Shoulders, C. C., Stocks, J., Carlton, D. J. & precursor. Thus, absence of 66-bp repeats in the human Baralle, F. E. (1983) Lancet i, 444-446. apoClIl gene (20, 26) suggests either that the apoCIII gene 5. Barker, W. C. & Dayhoff, M. 0. (1977) Compr. Biochem. diverged from the CIII/AI/AIV ancestor before divergence Physiol. B 57, 309-315. and intragenic sequence amplification of the AI/AIV gene 6. Brown, D. D. (1981) Science 211, 667-674. precursor or that the apoCIIl gene lost its nucleotide se- 7. Karathanasis, S. K., McPherson, J., Zannis, V. I. & Breslow, J. L. (1983) Nature (London) 304, 371-373. quence periodicity during its subsequent evolution. Clearly, 8. Loeney, W. A. & Brammar, W. J. (1980) Gene 10, 249. structural characterization of the apoCIII gene in distantly 9. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular related species may facilitate the unraveling ofthe evolution- Cloning: A Laboratory Manual (Cold Spring Harbor Labora- ary history of the apoAI/apoCIII/apoAIV multigene family. tory, Cold Spring Harbor, NY). Common evolutionary origin of the apoAI and apoE genes 10. Benton, W. E. & Davis, R. W. (1977) Science 196, 180-182. has been suggested by the extensive homology of consensus 11. Southern, E. M. (1975) J. Mol. Biol. 98, 503-517. sequences generated from their nucleotide repeats and by the 12. Thomas, P. (1980) Proc. Natl. Acad. Sci. USA 77, 5201-5205. presence ofintrons interrupting their coding regions at similar 13. Rigby, P. W. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977) (23). The extensive homology (56%) of the consen- J. Mol. Biol. 113, 237-251. positions 14. Karathanasis, S. K., Zannis, V. I. & Breslow, J. L. (1983) sus sequences of the repeats in the human apoE and apoAIV Proc. Natl. Acad. Sci. USA 80, 6147-6151. genes suggests that the apoE gene derived from a sequence 15. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. Y. & evolutionarily related to the AI/AIV gene precursor. How- Rutter, W. J. (1979) Biochemistry 18, 5294-5299. ever, the apoAl and apoAIV genes are located on human 16. Maxam, A. M. & Gilbert, W. (1980) Methods Enzymol. 65, chromosome 11 (27, 28), while the apoE gene is located on 499-560. chromosome 19 (23). The apoCI and apoCII genes are also 17. Messing, J., Crea, K. & Seeburg, H. P. (1981) Nucleic Acids located on human chromosome 19 and the apoCII gene is Res. 9, 309-319. genetically linked with the apoE gene (29-32). It is therefore 18. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. and Acad. Sci. USA 74, 5463-5467. possible that, similar to the apoAl, apoCIII, apoAIV 19. Brutlag, D. L., Clayton, S., Friedland, P. & Kedes, L. H. genes, the apoE gene evolved together with additional genes (1982) Nucleic Acids Res. 10, 279-294. in its vicinity and that another apolipoprotein multigene 20. Karathanasis, S. K., Zannis, V. I. & Breslow, J. L. (1985) J. family may be present on the human chromosome 19. Lipid Res. 26, 451-456. A part of the amino acid sequence of the human apoAIV 21. Boguski, M. S., Elshourbagy, N., Taylor, J. M. & Gordon, protein was obtained by in-frame translation of the cloned J. I. (1984) Proc. Natl. Acad. Sci. USA 81, 5021-5025. apoAIV cDNA sequence. It is noteworthy that this analysis 22. Dickerson, R. E. & Geis, I. (1969) The Structure andAction of suggested that there may be two different apoAIV gene Proteins (Benjamin, Menlo Park, CA). alleles coding for two amino acid sequences, one of them 23. Das, H. K., McPherson, J., Bruns, G. A. P., Karathanasis, S. K. & Breslow, J. L. (1985) J. Biol. Chem. 260, 6240-6247. being shorter at its carboxyl terminus by 18 residues. How- 24. Kyte, S. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132. ever, this difference may be due to cloning artifacts and thus 25. Alexander, F., Young, P. R. & Tilghman, S. M. (1984) J. Mol. more work is needed to clarify this issue. Analysis of the Biol. 173, 159-176. hydropathic properties of the apoAIV amino acid sequence 26. Protter, A. A., Levy-Wilson, B., Miller, J., Bencen, G., indicated that this protein contains 22-amino acid repeats White, T. & Seilhamer, J. J. (1984) DNA 3, 449-456. with the potential offorming amphipathic helices. It has been 27. Law, S. W., Gray, G., Brewer, H. B., Sakaguchi, A.-Y. & suggested that such helices may be responsible for lipid Naylor, S. L. (1984) Biochem. Biophys. Res. Commun. 118, binding and for activation of lecithin:cholesterol acetyltrans- 934-942. The cDNA-derived human 28. Bruns, G. A. P., Karathanasis, S. K. & Breslow, J. L. (1984) ferase by apoAIV (33, 34). Arteriosclerosis 4, 97-102. apoAIV amino acid sequence may facilitate further under- 29. Jackson, G. L., Bruns, G. A. P. & Breslow, J. L. (1984) Proc. standing of these structure-function relationships for this Natl. Acad. Sci. USA 81, 2945-2949. protein. 30. Myklebost, O., Rogne, S., Olaisen, B., Gedde-Dahl, T., Jr., & It should, finally, be pointed out that the increased lym- Prydz, M. (1984) Hum. Genet. 67, 309-312. phatic secretion ofapoAIV by fat feeding ofhumans (35), the 31. Humphries, S. E., Berg, K., Gill, L., Cumming, A. M., significant induction of liver apoAIV mRNA synthesis by Robertson, F. W., Stalenhoef, A. F. H., Williamson, R. & chemically induced porphyria in rats (36), and the synthesis Borresen, A. L. (1984) Clin. Genet. 26, 389-3%. of apoAIV mRNA by adult but not fetal human liver suggest 32. Tata, F., Henry, I., Markham, A. F., Wallis, S. G., Weil, D., factors are involved in of Grzeschik, K. H., Junien, C., Williamson, R. & Humphries, that multiple regulation apoAIV S. E. (1985) Hum. Genet. 69, 345-349. gene expression. 33. Morrisett, J. D., Jackson, R. L. & Gotto, A. M. (1977) Biochim. Biophys. Acta 472, 93-133. I thank Tracy Fitzpatrick for her excellent assistance and Marie 34. Steinmetz, A. & Utermann, G. (1985) J. Biol. Chem. 240, Hager for typing this manuscript. This work was supported by grants 2258-2264. from National Institutes of Health (HL 32032) and from The 35. Green, P. H. R., Glickman, R. M., Riley, J. W. & Quinet, E. American Heart Association-Massachusetts Affiliate, Inc. (13- (1980) J. Clin. Invest. 65, 911-919. 517-845). S.K.K. is an Established Investigator of The American 36. Buchberg, A. M. & Kinniburgh, A. L. (1985) Nucleic Acids Heart Association and a Scholar of the Syntex Incorporation. Res. 13, 1953-1963. Downloaded by guest on September 23, 2021