Nucleotide sequence and the encoded amino acids of human serum

Transcription

1 Proc. Nati cad. Sci. US Vol. 79, pp , January 1982 Biochemistry Nucleotide sequence and the encoded amino acids of human serum albumin mrn (cdn clones/codon usage/prepropeptide/triple-domain structure) CHLLES DUGCZYK, SMON W. LW*, ND OLV E. DENNSON Department of Cell Biology, Baylor College of Medicine, Texas Medical Cen, Houston, Texas Communicated by James F. Bonner, Septemnber 3, 1981 BSTRCT The complete nucleotide sequence of human serum albumin mrn has been demined from recombinant cdn clones and from a primer-extended cdn synthesis on the mrn template. The sequence is composed of 2078 nucleotides, starting upstream from a potential ribosome binding site in the 5' untranslated region. t contains all the translated codons and extends into the poly() at the 3' minus. Part of the translated sequence codes for a hydrophobic prepeptide, Met--Trp-Val- Thr-Phe-le-Ser-Leu-Leu-Phe-Leu-Phe-Ser-Ser-la-Tyr-Ser, followed by a basic propeptide, rg-gly-val-phe-rg-rg. These signal peptides are absent from mature normal serum albumin and, so far, have not been identified in their nascent state in humans. remaining 1755 nucleotides of the translated mrn sequence code for 585 amino acids, which are in agreement, with few exceptions, with the published amino acid sequence for human serum albumin. The mrn sequence verifies and refines the repeating homology in the triple-domain structure of the serum albumin molecule. The gene for serum albumin, which codes for the major plasma protein, is ofparticular inest to study because it is regulated in development. n mammals, serum albumin is synthesized by the adult liver, and its synthesis increases from low levels early in development to a high plateau in adulthood. The embryonic liver and yolk sac, on the other hand, produce predominantly a-fetoprotein, but the synthesis decreases drastically af birth (1-3). This inverse relationship between the expression of the two genes makes it an attractive problem in developmental biology, particularly because the two genes are related. We have recently demined the complete sequence of mouse a-fetoprotein mrn (4). The deduced a-fetoprotein structure revealed extensive homology to mammalian serum albumin, indicating that the two proteins are encoded in the same gene family. Similar conclusions have been reached by others from studies on rat (5) and mouse (6) a-fetoprotein genes. n the present effort we extend our studies to the human genome and report the cloning and sequence demination of DN complementary to the human serum albumin mrn. We deduce the unknown amino acid sequence of the signal peptide and verify a repeating homology in the triple-domain structure of the human serum albumin molecule. METHODS mrn. Human liver mrn was obtained by the procedure of Chirgwin et al (7) from fetal livers removed at weeks of gestation. mmunoprecipitation of albumin-containing polysomes was performed according to Taylor and Tse (8). n vitro translation of mrn was carried out in a cell-free reticulocyte The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C solely to indicate this fact. 71 B C am- -~ FG. 1. utoradiogram of proteins, labeled with [assimethionine in an in vitro reticulocyte translation system, and separated electrophoretically in a sodium dodecyl sulfate/acrylamide gel (9). Lane, no mrn added to the translation system. Lane B, translation products of total human liver mrn. Lane C, translation products of human liver mrn obtained from immunoprecipitated albumin-producing polysomes. Lane D, part of the translation sample that is separated in lane B was immunoprecipitated with antibody prior to its electrophoretic separation on the gel. system, following the instruction of the supplier (New England Nuclear). The translation products were separated electrophoretically according to Laemmli (9). Cloning and Sequence Demination of DN. Doublestranded cdn has been cloned in the Pst site of the plasmid pbr322, as described in detail previously (4, 10). DN sequence was demined according to the procedure of Maxam and Gilbert (11). RESULTS Enriched mrn. The products of in vitro translation of human liver mrn are shown in Fig. 1. On the basis of this translation, mrn that was enriched for serum albumin sequences was estimated to be over 50% pure (Fig. 1, lane C). This enriched mrn was used to obtain a cdn probe to screen the recombinant clones for serum albumin sequences. * Present address: National Heart, Lung and Blood nstitute, National nstitutes of Health, Bethesda, MD D

2 72 Biochemistry: Dugaiczyk et al. Recombinant Plasmids ph36 and ph206. restriction endonuclease map of the largest positive clone, ph36, is shown in Fig. 2, together with a restriction map of the primerextended plasmid clone ph206. The lat was obtained in a second transformation experiment af initiating the cdn synthesis from an innal primer. This primer was a 91-basepair-long DN fragment, Msp (152)-Taq (182/3), isolated from ph36. The two plasmids, ph36 and ph206, share 0.15 kilobase of homologous DN. Together, they encode the entire sequence for human serum albumin, starting with the CTT codon for Leu at position - 10 of the prepeptide and extending into the 3' untranslated region of poly(). Sequence of the lbumin cdn. The entire nucleotide sequence of the serum albumin mrn, as demined from the cloned DN in ph36 and ph206 and from the primer-extended cdn at the 5' minus of the message, is shown in Fig. 3. The inferred amino acid sequence is also indicated. The mrn length is 2078 nucleotides, of which 38 represent the 5' untranslated region, 54 identify a prepeptide of 18 amino acids, 18 identify a propeptide of 6 amino acids, 1755 code for the known 585 amino acids of serum albumin, 189 make up the 3' untranslated region, and 24 are the poly() sequence. Nucleotides 5 to 15 (-34 to -24) in the 5' untranslated region (Fig. 3) are complementary to a 3'-minal region of eukaryotic 18S RN (13) and thus could represent a ribosome binding site: (5')... T-T-C C-T-T-C-T-G-T... albumin mrn (3')... G--G-G---G-G-C-G-U-C-C-m62-m S RN The translated portion of the mrn sequence codes for the signal peptide and the main body of the albumin polypeptide chain. Because prepeptides are removed from nascent secretory proteins (such as albumin) in the endoplasmic reticulum, and the conversion ofproalbumin to albumin takes place in the Golgi vesicles, the presence and the sequence of the signal peptide for normal human serum albumin have not been reported previously. t the 3' end of the message, the putative polyadenylylation signal sequence --T---, is located 16 nucleotides upstream from the beginning of the poly() sequence. nother characistic sequence located near the polyadenylylation site has been identified by Benoist et al. (14); the consensus sequence from several mrns was T-T-T-T-C--C-T-G-C. Proc. Nad cad. Sci. US 79 (1982) similar sequence, T-T-T-T-C-T-C-T-G-T, is located 19 nucleotides upstream from the --T--- hexanucleotide in the human albumin mrn molecule (Fig. 3). Primary Structure of Human Serum lbumin. mino acid sequences for human serum albumin have been published by Behrens et al. (15) and by Meloun et al (16). (See Fig. 4.) There are a few discrepancies between the two reports, often involving sequences surrounding cysteine residues. Perhaps the most serious in ms of the structure of the albumin molecule is the sequence around the 17th and 18th cysteines, because the presence (15) or absence (16) of other amino acids between the two cysteines would affect the polypeptide loops generated by the disulfide bonds ofthese two cysteines. Our results in this critical region are shown by the sequencing gel in Fig. 5, which permits an identification of residues 261 to 290. The 17th and 18th cysteines are residues 278 and 279, and there are clearly no other amino acids between them. Consequently, such a sequence arrangement establishes the near-perfect homology in the triple-domain structure of the albumin molecule (Fig. 4). Our sequence results are in bet agmeement with those of Meloun et al (16), although several discrepancies remain. mino acid positions 94 (Gln), 95 (), 97 (Gly), 170 (Gln), 464 (His), 465 (), and 501 () are specified (16) as, 94 (), 95 (Gln), 97 (), 170 (), 464 (), 465 (His), and 501 (Gln). Residues (Fig. 3), la-sp-pro-his---tyr, are specified (16) as His-sp-Pro-Tyr---la. There are several.possible explanations for these differences, but we do not think the differences are due to erroneous DN sequence deminations. DSCUSSON Demining the complete nucleotide sequence of the cdn has permitted us to identify the pre- and the propeptides of human serum albumin. ctually, the amino acid sequence of the propeptide. The aled protein consequently ceases to be a substrate for the specific'protease that removes propeptides from secretory proteins. t is inesting to note, however, that failure to remove such a propeptide does not prevent the protein from being secreted; at least this is 'true about albumin are presently reporting for the apparently normal protein, rg- Gly-Val-Phe-rg-rg-albumin (Fig. 3). Thus, carriers of albumin Christchurch must be carriers of a CG to C mutation, which changes the codon for rg to Gln in the last position of HpaU (3658) J 1 Taq 1 5' Psol bombo (3611) 16/7 31 Hpa 1 (3548) Pat 182/3 (3611) Taq, H1%~~~~~Hn / /8 Hinf Mbo Mbo Mu 1-1 [3.Pst s0 (3611) Hp&( 2 (3646) / / Taq Mbo Taq Mlo HknU Mbo 1L11 H Kiobases l - 493/4 Taq Hha sat /7 HM H*n41 531/2 472/3 479/0 ph36 O FG. 2. Restriction endonuclease cleavage map of the overlapping human albumin cdn inserts in the recombinant plasmids ph36 and ph26.' The map shows the inserted albumin DN (opent bar) and its orientation in-the pbr322 plasmid DN (black bar). The mrn sequence is presented by the upper DN strands, with the 5' and 3' mini as indicated. Numbers on restriction sites within the human-derived'dn correspond to amino- acid positions in the albumin molecule. Numbers such as 182/3 indicate that the restriction sequence overlaps two amino acids. The Taq (270) site is actually not cleaved in our system'by Taq, due to methylation of this sequence to T-C-G-m6. The inserted sequence in ph206 starts with the'ctt codon for Leu at position -10 of the prepeptide. Numbers on restriction sites within the pbr322 sequence are taken from available sequence data from Sutcliffe (12). One of the two Pst sites in each recombinant plasmid, indicated by *, has not been reconstituted due to loss of a minal in the Pst'-linearized plasmid DN. 666 m T Hhdlll * Pot (3611) 34 Hhfl (3668)

3 Biochemistry: Dugaiczyk et al. Proc. Nati. cad. Sci. US 79 (1982) p r e -10 Met lys trp val thr phe ile ser leu leu phe leu phe ser ()GCTTTTCTCTTCTGTCCCCCCCGCCTTTGGCC TG G TGG GT CC TTT TT TCC CTT CTT m CTC m GC (80) -1-6 p r o ser ala tyr ser arg gly val phe arg arg asp ala his lys ser glu val ala his arg phe lys asp leu gly glu glu asn phe lys TCG GCT TT TCC GG GGT GTG m CGT CG GT GC CC G GT GG GTT GCT CT CGGM GT TTG GG G G T TTC (170) ala leu val leu ile ala phe ala gin tyr leu gin gin cys pro phe glu asp his val lys leu val am glu val thr glu phe ala GCC TTG GTG TTG TT GCC TTT GCT CG TT CTT CG CG TGT CC m G GT CT GT TT GTG T G GT CT G TTT GC (260) lys thr cys val ala asp glu ser ala glu asn cys asp lys ser leu his thr leu phe gly asp lys leu cys thr val ala thr leu C TGT GTT GCT GT GG TC GCT G T TGT GC TC CTT CT CC CTTm GG GC TT TGC C GTT GC CT CTT (350) arg glu thr tyr gly glu met ala asp cys cys ala lys gin glu pro gly arg asn glu cys phe leu gin his lys asp asp asn pro CGT G CC TT GGT G TG GCT GC TGC TGT GC CM G CCT GGG G MT GM TGC TTC TTG CM CC M GT GC C CC (440) asn leu pro arg leu val arg pro glu val asp val met cys thr ala phe his asp asn glu glu thr phe leu lys lys tyr leu tyr MC CTC CCC CG TTG GTG G CC GG GTT GT GTG TG TGC CT GCTm CT GC MT GM GG C TTT TTG M TC TT TT (530) glu ile ala arg arg his pro tyr phe tyr ala pro glu leu leu phe pbe ala lys arg tyr lys ala ala phe thr glu cys cys gin G TT GCC G G CT CCT TC TTT TT GCC CCG G CTC CTT TTCm GCT M GG TT M GCT GCT m C G TGT TGC CM (620) ala ala asp lys ala ala cys leu leu pro lysleu asp glu lie arg asp glu gly lys ala ser ser ala lys gin arg leu lys cys GCT GCT GT M GCT GCC TGC CTG TTG CC MG CTC GT GM CTT CGG GT GM GGG G GCT TCG TCT GCC M CG G CTC G TGT (710) ala ser leu gin lys pbe gly glu arg ala phe lys ala trp ala val ala arg leu ser gin arg phe pro lys ala glu phe ala glu GCC GT CTC C M TTT GG GM G GCT TTC M GC TGG GC GT GCT CGC CTG GC CG G TTT CCC GCT GG TTT GC G (800) val ser lys leu val thr asp leu thr lys val his thr glu cys cys his gly asp leu leu glu cys ala asp asp arg ala asp leu GTT TCC G TT GTG C GT CTT CC M GTC CC CG G TGC TGC CT GG GT CTG CTT GM TGT GCT GT GC GG GCG GC CTT (890) ala lys tyr se cys glu asn gin asp ser ile ser ser lys leu lys glu cys cys glu lys pro leu leu glu lys ser his cys ile GCC G TT TC TGT G MT C GT TCG TC TCC GT CTG MG GM TGC TGT G M CCT CTG TTG GM TCC CC TGC TT (980) ala glu val glu asn asp glu met pro ala asp leu pro ser leu ala ala asp phe val glu ser lys asp val cys lys asn tyr ala GCC G GTG GM MT GT GG TG CCT GCT GC TTG CCT TC TT GCT GCT GT TTT GTT G GT G GT GTT TGC M MC TT GCT(1070) glu ala lys asp val phe leu gly met phe leu tyr glu tyr ala arg arg his pro asp tyr ser val val leu leu leu arg leu ala GG GC G GT GTC TTC TTG GGC TG m TTG TT G TT GC G GG CT CCT GT TC TCT GTC GTG CTG CTG CTG G CTT GCC(1160) lys thr tyr glu thr thr leu glu lys eys cys ala ala ala asp pro his glu cys tyr ala lys val phe asp glu phe lys pro leu MG C TT GM CC CT CT GG MG TGC TGT GCC GCT GC GT CCT CT G TGC TT GCC M GTG TTC GT GM U CCT CTT(1250) val glu glu pro gin asn leu ile lys gin asn cys glu leu phe glu gin leu gly glu tyr lys phe gin asn ala leu leu val arg GTG G GG CCT CG T TT TC M C MT TGT GG CTT m GG CG CTT GG GG TC M TTC CG MT GCG CTG TT GTT CGT(1340) tyr thr lys lys val pro gin val ser thr pro thr leu val glu val ser arg asn leu gly lys val gly ser lys eys cys lys his TC CC MG M GT CCC C GTG TC CT CC CT CTT GT GG GTC TC G MC CT GG M GTG GGC GC M TGT TGT M CT(1430) pro glu ala lys arg met pro eys ala glu asp tyr leu ser val val leu asn gin leu cys val leu his glu lys thr pro val ser CCT GM GC G TG CCC TGT GC GM GC TT CT TCC GTG GTC CTG C CG TT TGT GTG TTG CT GG CG CC GT GT(1520) asp arg val thr lys cys cys thr glu ser leu val asn arg arg pro cys phe ser ala leu glu val asp glu thr tyr val pro lys GC G GTC CC TGC TGC C GM TCC TTG GTG MC GG CG CC TGC TTT TC GCT CTG GM GTC GT GM C TC GTT CCC (1610) glu pbe asn ala glu thr phe thr phe his ala asp ile cys thr leu ser glu lys glu arg gin ile lys lys gin thr ala leu val GG mu MT GCT G C TTC CC TTC CT GC GT T TGC C CTT TCT GG MG GG G C TC MG M CM CT GC CTT GTT(1700) glu leu val lys his lys pro lys ala thr lys glu gin leu lys ala val met asp asp phe ala ala phe val glu lys cys cys lys GG CTC GTG M CC MG CCC MG GC C M GG CM CTG M GCT GTT TG GT GT TTC GCT GCT TTT GT GG MG TGC TGC MG(1790) ala asp asp lys glu thr cys phe ala glu glu gly lys lys leu val ala ala ser gln ala ala leu gly leu GCT GC GT MG GG CC TGC m GCC GG GG GGT M M CTT GTT GCT GC GT C GCT GCC TT GGC TT TM CTCCUMMG(1883) CTCTCGCCTCCTGGTGGGTGGTCGCTTTTCTCTGTTTTTCTTTTTCGTTGGTGTGCCCCCCTGTCTCTTT CTTT (2oo2) TCTTTTGCCTCTTTTCTCTGTGCTTCMTMTMTGGGMTCTM M (2078) FG. 3. Nucleotide sequence of human serum albumin mrn as demined from cloned cdn. The sequence 5'-upstream of the Leu at -10, which is not contained in ph206, was demined as follows. short DN fragment, containing the Taq site (rg at -1 in the propeptide), was obtained from the 5' end of the albumin DN in ph206 (Fig. 2). The fragment was minated by the lu sequence (Ser at -5 in the prepeptide) and the Mnl sequence (-6), and it was 32P-labeled at the 5' minus of the lat. This fragment was annealed with human liver mrn, and the resulting RNDN template/primer was used in the reverse transcriptase reaction to synthesize a cdn copy extended into the 5' minus of the mrn template. This cdn was separated from the deoxyribonucleoside triphosphates on a Sephadex G-50 column and used directly for sequence demination. The first residue was demined from genomic DN. The amino acid sequence shown is deduced from the nucleotide sequence., Termination. the propeptide. The aled protein consequently ceases to be Christchurch, which has reached the bloodstream of its carriers a substrate for the specific protease that removes propeptides (18). from secretory proteins. t is inesting to note, however, that The nucleotide sequence of the albumin mrn has also perfailure to remove such a propeptide does not prevent the pro- mitted us to refine the repeated homology in the triple-domain tein from being secreted; at least this is true about albumin structure of the human albumin molecule. The fact that cys-

4 74 Biochemistry: Dugaiczyk et alp ji~~~~~~~h Proc. Natl. cad. Sci. US 79 (1982) - _ mma Domain )~~~~~~~~~~~~~~~~1 Doai 1 ~~~~~~~~~~~ k Domain11E { FG. 4. ioai _iae eune n iufd odn atr fhmnsrmabmn rpedmi tutr fitra oooyi n varw.tesqec ersnsordeetdt:telvu s codn obon(7.nt h erdretsmer ftedslie bridges throughout the molecule.

5 Biochemistry: Dugaiczyk et al 290 lle His Ser Leu Leu Pro X r cys O k 2 Leu k Ser Ser 'le ;Ser sp Gn sn lle Tyr la 261 FG. 5. utoradiogram of a DN sequencing gel showing the region coding for the 17th and 18th cysteines of human serum albumin; the two cysteines are residues 278 and 279 and there are no other amino acids between them. teines 278 and 279 are not separated by other amino acids, as was originally thought (15), brings the structure of domain closer to the structures of the remaining two domains (Fig. 4). s nucleotide sequences of genes or mrns become available, they prompt renewed speculations about codon usage, because some insight might be gained into rates of mutation or other evolutionary forces, from any nonrandom patn of codons used in a given gene, or a given species. For example, analyzing published nucleotide sequence data of human a- and,b3globin genes, Modiano et al (19) noticed that when several degenerate codons specify one amino acid, the codon that gives rise to a mination codon in a one-step mutation is never used. These "dispensable premination codons" are avoided, it was argued (19), in order to reduce the risk to mutate to mination. Such an argument implicitly postulates that evolution is anticipatory. This argument, however, receives no support from the distribution of codons utilized on the serum albumin gene. Table 1 summarizes the prevalence ofcodons utilized for each of the amino acids in human serum albumin, including the presequence and prosequence. There is no evidence for discrimination against the seven dispensable premination codons: UU, UUG, UC, UCG, CG, G, and GG. We thank Dr. Paul C. MacDonald and Dr. Evan R. Simpson from the University of Texas Southwest Medical Cen, Dallas, for kindly providing human fetal liver samples. We also thank Dr. Brian J. McCarthy and Dr. rthur D. Riggs for critical reading of the manu- c Proc. Natd cad. Sci. US 79 (1982) 75 Table 1. Codon usage in human serum albumin mrn U C G Phe 25 Ser 3 Tyr U u Phe 10 Ser 7 Tyr 6 20 C Leu 10 Ser 6 Ter 1 Ter 0 Leu 13 Ser 3 Ter 0 Trp 2 G Leu 19 Pro 10 His 11 Leu 7 Pro 6 His 5 Leu 3 Pro 7 Gln 11 Leu: 12 Pro 1 Gln 9 ile fle ile Met 4 Thr 7 sn 11 4 Thr 9 sn 6 1 Thr Thr 2 20 Val 12 la 31 G Val 7 la 14 Val 8 la 16 Val 16 la 2 sp 25 sp rg 3 rg 1 rg 3 rg 2 Ser 6 Ser 3 rg 13 rg 5 U C G U C G Gly 3 U Gly 3 C Gly 6 Gly 2 G The numbers indicate the number of times the individual codons are used in the coding region of the mrn. Ter, mination. script. The artwork is by David Scarff. The work was supported by the Baylor Cen for Population Research and Reproductive Biology. 1. belev, G.. (1971) dv. Cancer Res. 14, Gitlin, D., Perricelli,. & Gitlin, G. M. (1972) Cancer Res. 32, Sell, S. & Gord, D. R. (1973) mmunochemistry 10, Law, S. W. & Dugaiczyk,. (1981) Nature (London) 291, Jagodzinski, L. L., Sargent, T. D., Yang, M., Glackin, C. & Bonner, J. (1981) Proc. NatL cad. Sci. US 78, Gorin, M. B., Cooper, D. L., Eiferman, F., van de Rijn, P. & Tilghman, S. M. (1981)J. Biol Chem. 256, Chirgwin, J. M., Przybyla,. E., MacDonald, R. J. & Rut, W. J. (1979) Biochemistry 18, Taylor, J. M. & Tse, T. P. H. (1976) J. Biol Chem. 251, Laemmli, U. K. (1970) Nature (London) 227, Law, S., Tamaoki, T., Kreuzaler, F. & Dugaiczyk,. (1980) Gene 10, Maxam,. & Gilbert, W. (1980) Methods Enzymot 65, Sutcliffe, J. G. (1978) Nucleic cids Res. 5, zad,.. & Deacon, N. J. (1980) Nucleic cids Res. 8, Benoist, C., O'Hare, K., Breathnach, R. & Chambon, P. (1980) Nucleic cids Res. 8, Behrens, P. O., Spiekerman,. M. & Brown, J. R. (1975) Fed. Proc. Fed. m. Soc. Exp. Biol 34, 591 (abstr.). 16. Meloun, B., Moravek, L. & Kostka, V. (1975) FEBS Lett. 58, Brown, J. R. (1976) Fed. Proc. Fed. m. Soc. Exp. Biol 35, Brennan, S. 0. & Carrell, R. W. (1978) Nature (London) 274, Modiano, G., Battistuzzi, G. & Motulsky,. G. (1981) Proc. Natl cad. Sci. US 78,