Complex pattern of alternative splicing generates unusual diversity in the leader sequence of the chicken link protein mRNA

Nucleic Acids Resea ch, Vol. 19, No. 18 4983 -4990 Complex pa e n o al e na i e splicing gene a es unusual di e si y in he leade sequence o he chicken link p o ein mRNA Fe enc Deak1, End e Ba a1 2, Sil ija Mes ic1+, Ma kus Biesold1 and lbolya Kiss'* Ins i u e o Biochemis y, Biological Resea ch Cen e o he Hunga ian Academy o Sciences, H-6701 Szeged, PO Box 521 and 2Ag icul u al Bio echnology Cen e , H-2101 Godillo, Hunga y Recei ed June 5, 1991; Re ised and Accep ed Augus 27, 1991 ABSTRACT We epo he e he isola ion o he 5' end and he p omo e egion o he gene o chicken ca ilage link p o ein, and demons a e ex ensi e he e ogenei y o he leade sequence a ising om di e en ial u iliza ion o mul iple splice si es wi hin he 5'-mos exon. The 500-base pai s (bp) exon 1 consis s o solely un ansla ed sequence and is ollowed by an in on >33 kilobase pai s (kb). Toge he , he i e exons p edic a gene size longe han 100 kb. Mul iple ansc ip ion ini ia ion si es we e mapped 34, 46, 56, 66 and 76 bp downs eam o a TATA-like mo i . Sequence analysis e ealed ha in addi ion o he non- spliced a ian , mul iple mRNA species we e gene a ed by al e na i e splicing esul ing in he exclusion o 92, 166, 170, 174 and 263 nucleo ides (n ), espec i ely, om exon 1. Polyme ase chain eac ion con i med he exis ence o a ious splice o ms, and showed cell ype- and de elopmen al s age-speci ic exp ession o one g oup o hem. Seconda y s uc u e p edic ions indica ed ha he leade s o he splice o ms could o m s able hai pin s uc u es wi h di e en ee ene gies o o ma ion (up o AG = - 1 10 kcal/mol), sugges ing ansla ional con ol. The splice a ian de ec ed in he la ges amoun had he leas s able p edic ed hai pin (AG = - 31.7 kcal/mol). INTRODUCTION Link p o ein (LP), an abundan glycop o ein o he ca ilagenous ex acellula ma ix, s abilizes he p o eoglycan agg ega es ia simul aneous binding o hyalu onic acid and chond oi in sul a e p o eoglycan (1-5). Two o h ee o ms o LP ound in se e al species (6-8), seem o di e only in p o eoly ic clea age o a sho pep ide and in he deg ee o glycosyla ion (7, 9). The p ecu so o chicken LP consis s o 355 amino acids including a sec e o y signal pep ide (10). The ma u e p o ein is di ided in o h ee s uc u al and unc ional domains s abilized GenBank accession no. M35040 by disul ide b idges (11, 12). The immunoglobulin-like domain in e ac s wi h p o eoglycan, while he wo andemly epea ed domains bind o hyalu ona e (13, 14). Link p o eins may ha e ye o he unc ions as hey a e also ound in issues o he han ca ilage (15-18). Fu he mo e, an al e na i ely spliced exon encoding an ex a segmen be ween he signal pep ide and he i s domain o a chond osa coma LP has been epo ed (19). In chicken ca ilage, mRNA size classes o 5.8 -6.0 kb and 3.0 kb, di e ing in hei 3' un ansla ed egion (3'-UTR), we e de ec ed (10). The LP mRNA species a e encoded by a single gene pe haploid genome (12). The ou p o ein coding exons, which e ol ed by duplica ion and exon shu ling, a e sca e ed o e a genomic egion longe han 70 kb in chicken (12). The 5' end o he gene, howe e , has no been isola ed p e iously. In he p esen s udy, we cha ac e ize he i s exon and he 5'- lanking egion o he chicken LP gene. A compa ison o he nucleo ide sequences o genomic and cDNA clones, and polyme ase chain eac ion (PCR) p oduc s e ealed ha al e na i e splicing wi hin he 5'-UTR ga e ise o he o ma ion o mul iple LP mRNA species in chond ocy es. The po en ial ole o he he e ogeneous 5'-UTR in he egula ion o gene exp ession is discussed. MATERIALS AND METHODS P ime ex ension, cDNA and genomic cloning All posi ions a e gi en om he i s nucleo ide o he ansla ion s a codon o he LP p ecu so (10, 12). Oligonucleo ides (Table I) we e syn hesized by phospho amidi e chemis y (20) and labeled wi h T4 polynucleo ide kinase (21). RNA was p epa ed om s e na o 14-day-old chicken emb yos, and om s e na and a icula ca ilage o 6-week-old (ju enile) chicks as desc ibed (22). Fo analy ical p ime ex ension, 0.1 pmoles o end-labeled p ime was annealed wi h 4 yg o poly(A)+ RNA immedia ely a e dena u a ion wi h 10 mM me hyl-me cu ic hyd oxide, and elonga ed wi h 40 uni s o M- MLV e e se ansc ip ase (BRL) in 50 il o 50 mM T is-HCl * To whom co espondence should be add essed + P esen add ess: Pli a Resea ch Ins i u e, 41000 Zag eb, I.L. Riba a 89, Yugosla ia .. 1991 Ox o d Uni e si y P ess 4984 Nucleic Acids Resea ch, Vol. 19, No. 18 (pH=8.3), 75 mM KCI, 3 mM MgCl2, 10 mM DTT, supplemen ed wi h 0.5 mM dNTP, 40 Ag/ml ac inomycin D and 20 uni s o placen al RNase inhibi o . Fo he cons uc ion o he p ime ex ension lib a y, 0.25 pmoles o P3 p ime and 5 yg poly(A)+ RNA we e used. The cDNA-RNA hyb ids we e ailed using dGTP and e minal ans e ase and annealed wi h 1.15 pmoles o oligo(dC)- ailed BamHI linke in he p esence o 1.5 uni s o RNase H. The second cDNA s and was syn hesized wi h 10 uni s o DNA polyme ase I. The ends we e polished wi h T4 DNA polyme ase Table I. Oligonucleo ide p ime sequences Name Sequence Posi ion PA PB1 PB2 PC Fo wa d p ime s 5'-CAAGTGGTCAGAAGTATCACGT-3' -605 o -584 5'-TAGTTCGGGACTGGTGTGCG-3' -488 o -469 5'-TTAGTTCGGGACTGGTGTGCGGTGCAGAGC-3' -489 o -460 5'-ACTTGGGAGCTCCACACAA-3' -46 o -28 Re e se p ime s PD 5'-GTCTCTTTGTGGCTCTGGGTGGCAGAGGAC-3' - 159 o - 188 PE 5'-GGTGGCAGAGGACCCTAAAAAAGCTCTGCAC-3' - 176 o - 188 and -452 o -468 PF 5'-CTGCCCCAGCCTCCGTGCCCCTCAGCTTCT-3' -404 o -433 PG 5'-GGTGGCAGAGGACTCACACACAGAGTGTCC-3' - 176 o - 188 and -359 o -375 P2 5'-CTTGTCATCTTCACAGTCAC-3' 8 o - 12 P3 5'-ACCACAAGTAGGCGGGGTCC-3' 128 o 109 and liga ed o phospho yla ed SalI linke . A e BamHI and Sall diges ion, he cDNA was size ac iona ed by Sepha ose CL-4B column ch oma og aphy and inse ed in o pUC 12 ec o . Recombinan plasmids we e in oduced in o Esche ichia coli DH5 cells (23). Colonies we e sc eened by hyb idiza ion o end-labeled P2 and P3 as desc ibed (10). A chicken genomic lib a y (12) was sc eened using he 277-bp EcoR l-SacI agmen ( om posi ion -312 o -35) o pLP7G12. Re e se ansc ip ion coupled wi h PCR (RT-PCR) The p ocedu e was simila o published me hods (24, 25) and was ep oducible in epea ed expe imen s. cDNA was syn hesized by ex ension o p ime P3 using 2 jig poly(A)+ RNA as empla e. Following alkaline hyd olysis o RNA and e hanol p ecipi a ion, he cDNA was ampli ied, as indica ed, using 2.5 uni s o Taq polyme ase (NEBL) and 100 pmoles o p ime s in 67 mM T is-HCl (pH=8.8), 16.6 mM (NH4)2SO4, 10 mM 2-me cap oe hanol, 6.7 mM MgCl2, 0.2 mM o each dNTP, 10% ( / ) DMSO in a inal olume o 100 1d. The las i e cycles we e ca ied ou by adding 10 i Ci o adioac i e e e se p ime o analysis on sequencing gels. In o he expe imen s p ime PD was annealed o 2 jg poly(A)+ o 20 , g o al RNA a 64°C and e e se ansc ibed in a inal olume o 50 yd. A e dena u a ion o 5 min a 100°C, 4 $1 o his eac ion mix u e was added o 10 mM T is-HCl (pH=7.9), 2 mM MgCl2, 10% T i on X-100, 1 y M o each p ime , 100 y g/ml bo ine se um albumin, and 200 ,uM o each CTGAGGGCTG TGTGGTGCCA GTTCGGTAAG CAAGGCACAC TCTGTA GTGTGATG AGAGAAGCTG GACCCAGTTC CTGGAE iI:CTCPg j -692 G!AGGTGAGG TAAAACAAAA GCACAAGGGG ATCCTGTCAG GTCCAGCTGT TTCCACCCCT GCACACCCCT GCCTTTCATG CATCACTTTC TCCCTGTCTG ==========e===========_ - --------------------------------------------------- -592 AAACCACCAC GTTCCCACCC TTGGAGCTAA TGGCAGGGTG GGGGCAAGTA AGGGGGGGGG GGGGGGGGGC TGCGAGCGAG CAGAGGCAAG TGGTCAGAAG I I I 1 -492 TATCACGTTT CTCAGGGTGT CCCACACC AT ATCCCTTCAT GAGGAACAGT TCTTTTTTTT TTTCTTTTTT TTTCTTTTTT TTTCTTTTTT OG1 5S3 -392 TTTTAGTTCG GGACTGGTGT GCGGTGCAGA GCTTTTTTAG GTGCAGCTCT GGGAGGGCAG AAGCTGAGGG GCACGGAGGC TGGGGCAGGG ACTTGGCGGA -------------------- - -----------------------~~~~~ 8C7 -292 ACAGGACAGC GTGCTGGGAC ACTCTGTGTG TGAGTGAGTG AGACAGACCG CGTGTGTGGG TTTATTTTTT CCTCTCATTC CTGGGGTCTG AATGGAAAGC 7G12 -192 ACGTTAAGCT GTAATTAAGA CATCTTGGAG GCGTACTTAA CTTATGACGT GTTGCTTTTC TTTTCCTCTC GGTGAGAAAG TGGTTCTCTT TTCCCTCAAT I1B4 -92 **** ***** **** **** ********** *** ** *** * *** ****** *C*T***A***** ******** * * * * ** **** CAGGTCCTCT GCCACCCAGA GCCACAAAGA GACGCTCGGG ACATAGGCAC ACACACGCGC ACACCAACTT AAGGTGAGCA CCATAAACTT CCTCTGCACT -27 * ******** ************* *** ** ***** TCGCTGTCCT GTTTCAAACT GTTGGAGAGT TCTGAGCGC.. TCTCGACTTG GGAGCTCCAC ACAAGg aaa gaaa aa ca aac gggg ..In on 1 -26 +1 ******* ****** ****** **** ccaaaaaa ga gagcc c g a g ga gccc acagTGAAGA AGATTCTTGT GACTGTGAAG A3CACAAGTC TACTCTTTCT ... Exon 2 Me Th Se L euLeuPheLe all clones( Fig. 1. Nucleo ide sequence o he 5' end o he chicken LP gene. Posi ions a e gi en om he ansla ion s a si e wi hin exon 2, no including in on 1 (nucleo ides in lowe case le e s). Nucleo ide sequences ep esen ed by cDNA clones a e indica ed by solid lines. All sequences we e de e mined om bo h di ec ions. TATA and CCAAT mo i s a e boxed. Homopu ine and homopy imidine ac s longe han 30 n a e deno ed by double and single b oken lines, espec i ely. A ows poin o he majo ansc ip ion s a si es. Nucleo ides iden ical wi h he human sequence (31) a e ma ked wi h as e isks abo e he sequence. Physical map o cDNA (A) and genomic (B) clones spanning he 5' end o he LP gene. Solid ba s ep esen he ec o s. Hyb idiza ion o p ime s P2 (-) and P3 (=) is indica ed below he map o each cDNA. A mo e de ailed es ic ion map o he 1.7-kb SalI-EcoRI genomic agmen is shown in an expanded o ma . Open box indica es he posi ion o exon 1. Ho izon al a ows deno e he sequencing s a egy. Abb e ia ions o es ic ion enzymes: A, A aI; B, BamHI; C, Sacl; E, EcoRI; R, EcoRII; S, SalI; V, A al. b -1 5S3 5S3 8C7 7G12 7G12 8G1, 5S3, 8C7 ilB4 all bu IlB4 10/- Nucleic Acids Resea ch, Vol. 19, No. 18 4985 dNTP in a inal olume o 50 Al. Ampli ica ion was pe o med wi h 1.25 uni s o Taq polyme ase using a Coy TempCycle . DNA analysis and seconda y s uc u e p edic ion The nucleo ide sequence was de e mined by he me hod o Sange e al. (26) using M13 subclones cons uc ed as desc ibed (10). T ansc ip ion s a poin s we e mapped wi h T4 DNA polyme ase on cloned genomic DNA as sugges ed by Hu and Da idson (27). F agmen s sepa a ed by polyac ylamide gel elec opho esis we e ans e ed o ni ocellulose il e s a e dena u a ion by boiling o 10 min in 1 M ammonium-ace a e, 20 mM NaOH. DNA agmen s we e labeled by andom p ime ex ension (28). RNA seconda y s uc u e was p edic ed wi h he FOLD p og am o Zuke (29) included in he Sequence Analysis So wa e Package o he Gene ics Compu e G oup, Uni e si y o Wisconsin (30). RESULTS Al e na i e splicing in he 5'-UTR o he chicken LP gene Genomic clones o chicken LP isola ed p e iously ca ied he en i e p o ein coding egion and 3'-UTR, bu no he i s exon o he gene (12). In o de o isola e he 5' end o he gene, we Table H. Pu a i e splice si es a he 5' end o he LP gene dono si ea posb accep o si eC posb cDNA clone agTTC -485 gc agGTG -451 TAGg gcag -451 cc caa cagGTC - 188 pLP 8G1 TGTg gag -362 cc caa cagGTC -188 pLP D48 TGAg gag -358 cc caa cagGTC -188 pLP 5S3 TGAg gaga -354 cc caa cagGTC - 188 pLP 8C7 TCGg gaga -220 AAGg gagc -118 gccc acagTGA in onI pLP 1 lB4 AAGg aaag -26 gccc acagTGA in onI all pLP clones excep 1 IB4 A AGg ag yyyyyynyagG consensus d amo i s con aining AGg o g ag sequence. : a o g. bposi ion o he splice si e in he sequence as in Fig. 1. Cmo i s di e ing by no mo e han wo nucleo ides om he sequence yyyyyynyagG and only a posi ions no unde lined. y: c o . dShapi o and Senapa hy (32). cons uc ed a cDNA lib a y using poly(A)+ RNA pu i ied om s e na o chicken emb yos as empla e, and a speci ic p ime P3 (Table 1), complemen a y o he LP mRNA om posi ion 109 o 128 ela i e o he ansla ion s a si e. The lib a y was sc eened wi h a mix u e o end-labeled P2 and P3 oligonucleo ides (Table I), speci ic o exons 2 and 3, espec i ely. Fi e o he posi i e cDNA clones wi h he longes inse s we e sequenced. The 277-bp EcoRII-SacI agmen o he mos 5'-end clone pLP7G12 hyb idized o RNA species o he same mobili y as desc ibed o LP mRNA (10) (da a no shown). The same agmen was used as a hyb idiza ion p obe o sc een a chicken genomic lib a y. One posi i e clone, XgLP100 was isola ed and cha ac e ized. Res ic ion mapping and Sou he n hyb idiza ion e ealed ha he clone ca ied exon 1 alongwi h 800 bp o ups eam sequence and 19.7 kb o he pu a i e i s in on (da a no shown). E en hough he p e iously desc ibed genomic clone XgLP12.1 (12) ex ended 13.6 kb ups eam o he 5'-mos ansla ed exon, named exon 2 in his pape , he es ic ion maps o he wo clones did no o e lap, no did he wo clones c oss hyb idize. The e o e, we concluded ha he i s in on is la ge han 33 kb. Since he p o ein coding exons including he 3'-UTR co e a genomic egion o 70 kb o longe (12), hence his indica es a p ima y ansc ip o e 100 kb in leng h. The nucleo ide sequence o he i s exon and lanking egions was de e mined (Fig. 1). Alignmen o he genomic and cDNA sequences e ealed comple e iden i y along exon 2; howe e , only one o he new cDNA clones, pLP7G12 o e lapped con inuously wi h he co esponding genomic sequence. Sho e o longe in e nal segmen s bo de ed by consensus dono and accep o splice si es (Table H) we e absen om he 5'-UTR o he o he cDNA clones. This obse a ion s ongly indica es ha he empla es o he cDNA molecules we e gene a ed by al e na i e splicing e en s wi hin exon 1, de ining a leas i e di e en splice a ian s o LP mRNA. A homology sea ch e ealed ha in he leade egion, wo ou o ou po en ial accep o si es we e u ilized, while ou o he se en pu a i e dono si es, six ha e con ibu ed o he gene a ion o obse ed di e si y (Table II). Two sho open eading ames we e ound in he 5'-UTR om -300 o -276 and om -248 o -122. Bo h o hese ups eam ini ia ion codons a e emo ed by excision o he al e na i e in ons om he 8C7-, 5S3- and 8GI- ype messages. Polymo phism was ***** *** Ex ension o PE °0 ° 0 o*00 00o00 0000 0*0 Ex ension o PF V VV V S1 nuc ease p o ec ion I III 1I 1l I T4 po yme ase s ops -562 -552 -542 -532 -522 -512 -502 -492 -482 -472 | AAAAAT GATCCCTTCA TGAGGAACAG TTCTTTTTTT TTTTCTTTTT TTTTCTTTTT TTTTCTTTTT TTTTTAGTTC GGGACTGGTG TGCGGTGCAG JT AAJ TTA CTAGGGAAGT ACTCCTTGTC AAGAAAAAAA AAAAGAAAAA AAAAGAAAAA AAAAGAAAAA AAAAATCAAG CCCTGACCAC ACGCCACGTC Fig. 2. Composi e ep esen a ion o he ansc ip ion ini ia ion si es de e mined by a ious echniques. The TATA-like sequence is boxed. Thickness o he ho izon al a ows below he sequence e lec s he ela i e equency o ini ia ion. Nuclease S I mapping o he 5' end o he chicken LP gene. Uni o mly labeled, single s anded p obes, complemen a y o he mRNA om posi ions indica ed by solid black ba s in he schema ic diag am, we e made by ex ension o he sequencing p ime on M13 subclones. The p obes we e hyb idized o 60 g o o al RNA and ea ed as desc ibed (43). The leng h o he p o ec ed agmen s, ep esen ed by hin lines, a e gi en in nucleo ides. A, B and C e e o espec i e p obes. Lanes G and T, dideoxy sequencing ladde ; lane 1, undiges ed p obe; lanes 2 and 3, p obe/RNA hyb id ea ed wi h nuclease S1 a 30°C o 37°C, espec i ely; lane 4, p obe diges ed wi hou RNA added. 4986 Nucleic Acids Resea ch, Vol. 19, No. 18 de ec ed be ween -370 and -366, whe e some o he cDNA clones lacked one o he epea ed CT dinucleo ides. Sequence elemen s o po en ial unc ional signi icance we e iden i ied in he 5'- lanking egion o he gene. Two TATA-like (a -806 and -562) and wo CCAAT mo i s (a -845 and -795) we e ound (Fig. 1). The sequence om -387 o -370 ma ches he consensus clea age si e o chicken opoisome ase (33). Pu ine and py imidine ich segmen s al e na e h oughou he sequence. The nucleo ide sequence om posi ion - 193 o -27 has 69 % iden i y wi h he co esponding human sequence (31). Mapping o ansc ip ion ini ia ion si es o he LP gene Th ee di e en echniques we e used o de ine he ansc ip ion s a si e and he esul s a e summa ized in Fig. 2. S1 nuclease mapping using p obes o h ee di e en leng hs (da a no shown) indica ed mul iple ini ia ions in a egion om posi ion -529 o -485. A s ong p o ec ion was obse ed a he 3' end o a poly(T) ac a ound -486 and wo addi ional equen ini ia ion si es we e mapped a - 517 -i-1 and - 508 1 by each p obe (Fig. 2). ACOG A ... . M13 -- -- - ~ mm DNA a _ _ _ _ Saoi m - 1 1- __ -4.I a 'I' k i I LLi (AP 'A, .7 AT'A Pe o ming SI p o ec ion assays wi h o he p obes, we could no , howe e , de ec any ansc ip o igina ing om he TATA mo i a - 806 (da a no shown). These esul s indica ed ha he TATA- like mo i a -562 may se e as a p omo e o he LP gene. SI nuclease sensi i i y o he poly(U) s e ch a he 5' end o he LP mRNA may ha e con ibu ed o he S I pa e n obse ed. The e o e, we also mapped he 5' end using T4 DNA polyme ase (Fig. 3). A speci ic p ime , PA, was annealed o a single s anded an isense DNA spanning he 5' end o he LP gene, and he p ime was ex ended wi h T4 DNA polyme ase in he p esence o hyb idized RNA. Mul iple s a si es we e de e mined wi hin he poly(T) s e ch om -529 o -487. The majo si es we e loca ed a -497, -507 -+1 and -517 + 1, and wo mino si es a -529 and a -487. The ex ension p oduc ep esen ing he las si e was ain bu isible in he o iginal au o adiog am. In p ime ex ension expe imen s (Fig. 4) he p ime PE , complemen a y o he splice o m ep esen ed by pLP8G1, esul ed in wo majo clus e s o ex ension p oduc s 60 + I and 69 ± 2 n in leng h, hus de ining ansc ip ion ini ia ion a -498 ± 1 and -507 ±2, espec i ely. In con a y, when PF , a p ime speci ic o all bu 8G 1- ype splice a ian s was used, he ex ension p oduc s indica ed mul iple s a si es wi hin a wide egion, be ween -529 and -487 (Fig. 2). The mos abundan ansc ip s (1 13 n ) ini ia ed a ound -517. The di e ence be ween he p ime ex ension esul s in epea ed expe imen s sugges s ha he LP splice o ms may di e in he ela i e equency o u iliza ion o he mul iple s a si es. Howe e , due o he complex na u e and high s abili y o he leade (see below) I. .1_. . a -P :' I0 Fig. 3. Mapping o ansc ip ion s a si es by T4 DNA polyme ase. The diag am o he le depic s he expe imen al s a egy. An M13 DNA ca ying he RNA- complemen a y s and o exon 1 and he p omo e egion was used as empla e. Solid ba wi h as e isk ep esen s he 32P end-labeled PA p ime , b oken a ow indica es he p ime ex ension p oduc s. Lanes A, C, G and T, dideoxy sequencing ladde s o he same empla e DNA om unlabeled PA p ime in he p esence o [35S]dATP. Lane 1, 4 ug poly(A)+ RNA was hyb idized o he empla e be o e p ime ex ension. Lane 2, p ime ex ension in he absence o hyb idized mRNA. The numbe s a he igh ma gin o he au o adiog am deno e he nucleo ide posi ions. Tha pa o he au o adiog am which ca ies he ex ension p oduc a posi ion -487 is shown a e p olonged exposu e, in an inle o he igh . Analysis o PCR p oduc s on a dena u ing polyac ylamide gel. D, PC/P3 p ime pai was used in RT-PCR unde condi ions desc ibed o Fig. 6. A and B. Lane 1, p oduc s o RT-PCR. Lane 3, PCR pe o med om I ng inse o pLP8Gl as empla e. M, pUC12 HaeIIl diges used as size ma ke . I a Fig. 4. P ime ex ension analysis o he 5' end o he LP gene. 4 pg poly(A)+ RNA was hyb idized a 60°C o 5 32P end-labeled oligonucleo ides, comple- men a y o exon 1 a posi ions shown in he schema ic diag am. Exon 1 is boxed. cDNA was syn hesized a 42'C o 1 h wi h M-MLV e e se ansc ip ase. The ex ension p oduc s om p ime s PE (lane 1) and PF (lane 2) we e analysed on sequencing gels. Lanes A, C, G and T, dideoxy sequencing ladde s. Numbe s a he ma gins o he au o adiog ams deno e he leng h in nucleo ides. , .. .A Nucleic Acids Resea ch, Vol. 19, No. 18 4987 which impeded he design and ex ension o p ime s speci ic o he dis inc splice o ms, we could no es his hypo hesis. To sum up, he esul s ob ained by h ee di e en s a egies a e gene ally consis en wi h he conclusion ha he LP gene is ansc ibed om mul iple s a poin s loca ed wi hin he poly(T) s ech 34 + 1, 46 + 1, 56 + 1, 66 ± 1 and 76 ± 1 bp downs eam o he TATA-like mo i (Fig. 2). Nucleo ide sequences a he ini ia ion si es ma ch he YYCAYYYY consensus sequence (34) desc ibed o o he euka yo ic genes excep o he lack o A a posi ion 4. The ini ia ion si es a -507, -517 and -497 a e used mo e equen ly han he o he s. The majo disc epancies in SI mapping ( wo mino si es and inc eased equency o clea age a ound -486) can be explained by he SI sensi i i y o he 5' end o he hyb id. Indeed, mo e accu a e da a we e ob ained wi h T4 DNA polyme ase as compa ed o he o he wo echniques, due o he ac ha his analysis was nei he in luenced by he base composi ion no he seconda y s uc u e o he RNA. PCR analysis o he mul iple LP mRNA species The expec ed agmen sizes o a ious combina ions o o wa d and e e se p ime s in RT-PCR analysis a e depic ed in Fig. 5. Using wo di e en p ime pai s (Fig. 6A and B), splice a ian ep esen ed by cDNA clone 8G1 was ound o be he mos abundan among he LP mRNA species in emb yonic s e nal ca ilage, ollowed by he 8C7/5S3- ype a ian s. Bands co esponding in size o 7G12 and 1 1B4 splice o ms we e no de ec able in PCR e en upon p olonged o e exposu e o he - ,-- >33 kb - >39 kb -4-11.6 kb-U- 72 kb -{ s as s- PB1 A ~PB1 7G12 4 - 11B4 I 404 8C7 = 330 5S3 326 D48 322 8G1 _233 Hypl 230.238 Hyp2 141 Hyp3 _1n0168 Hyp4 71 c P2 PD 7012 331 SC7 e 165 583 ca=161 D48 , 157 801 68 au o adiog ams; howe e , se e al sho e p oduc s we e ound, co esponding in size o he hypo he ical messages Hyp2 -4 (Fig. 6A, lane 2). In o de o es i addi ional exons hidden in he long i s and second in ons ha e con ibu ed o he di e si y obse ed, ano he RT-PCR was ca ied ou wi h he PC/P3 pai o p ime s. A single p oduc o 174 bp was moni o ed, iden ical in size o ha syn hesized on he cDNA clone pLP8G1 (no shown). Longe PCR p oduc s we e no obse ed e en a e long exposu e. The e o e, we concluded ha no al e na i e exons loca ed be ween exon 1 and exon 3 we e de ec able in chicken ca ilage. The PCR p oduc s and he cDNA clones we e u he co ela ed by designing he PB2/PD pai o p ime s, each 30 n in leng h, hus dec easing he complexi y and minimizing he nonspeci ic backg ound in RT-PCR. Fo y cycles p oduced DNA agmen s isible in e hidium-b omide s ained gels i espec i e o whe he poly(A)+ o o al RNA was used as empla e (Fig. 7A). All he bands co esponding o he a ious cDNA clones we e obse ed, excep o he 7G12- ype PCR agmen . In ac , hyb idiza ion o a speci ic p obe clea ly indica ed he p esence o he 7G12- ype PCR p oduc (Fig. 7B, lane 2) appea ing as agmen s o e e -inc easing leng h. The explana ion o his phenomenon is he p esence o di ec epea s, which led o he ou -o - egis e hyb idiza ion o compe ing PCR p oduc s a subsequen cycles o ampli ica ion. In addi ion o se e al sho e epea s, a andem epea o 23 n con aining ou misma ches was ound a -338 and -243. Simila a e ac s we e de ec ed by o he s when DNA empla e con aining andem epea s was used in PCR (35). Su p isingly, apa om he expec ed PCR p oduc s de ec ed in dena u ing gels (Fig. 6C), wo addi ional bands (ma ked X A M32 1 P? gmen sizes s In A . 120 bp 8C7 a _ 5S3 X- 8G1 Hypl I / -'VI H yP3 X Hyp2 0 _ B M- - 1 M C 4 5 6 - -587 u:_458 -434 - -298 - -267 - -251 8C7 5S3 A W D48 M - 587 * 458 -434 - -298 w- 267 ` -251 am -1 74 --174 _ - 102 100 bp X -80 Hyp4 8G1- _ - -102 PC P3 all '-' 174 Fig. 5. Diag am o al e na i e splicing in he LP leade egion. The s uc u e o he gene is shown a he op o he igu e. Exons (E l -E5) a e numbe ed om he 5' end o he gene (no e he di e ence om e . 12). Open boxes ep esen he 5'- and 3'-UTRs and black boxes he p o ein coding exons. A ows indica e he posi ions o he p ime s used in PCR. The PCR agmen s assumed o be o med wi h a ious p ime pai s, based on he s uc u e o he cDNA clones, a e depic ed. The egions ha bo ed by he cDNA clones a e shown as do ed ba s o compa ison. P ime pai s PB1/P2 (A), PB1/P3 (B), PB2/PD (C) and PC/P3 (D) we e chosen. F agmen sizes expec ed o he a ious splice o ms a e shown in base pai s a he igh side o he diag ams. Fig. 6. Analysis o he PCR p oduc s on dena u ing polyac ylamide gels. A, B and C ep esen p ime combina ions as in Fig. 5. A and B, cDNA sy hesized om poly(A)+ RNA o chick emb yo s e na was subjec ed o 30 cycles o ampli ica ion a 94°C o 1.5 min, 48°C o 2.5 min and 75°C o 3.5 min in he p esence o 10% DMSO. Lane 1, P oduc s o RT-PCR. Lane 2, as lane 1, bu 10- old amoun was loaded. Lane 3, PCR pe o med om 1 ng inse o pLP8G1 as empla e. M, pUC12 HaeIll diges used as size ma ke . C, To al RNA om emb yonic s e nal (lane 4), ju enile s e nal (lane 5) and ju enile a icula (lane 6) chond ocy es was used in RT-PCR employing he PB2/PD p ime pai . 30 cycles o ampli ica ion we e pe o med a 94°C o 1 min, 65 °C o 2 min and 75°C o 3 min. Numbe s o he igh indica e he agmen sizes in nucleo ides. D -4 / -i 4988 Nucleic Acids Resea ch, Vol. 19, No. 18 N4 ;5 & S 3 I-) 4 B Fig. 7. Analysis o he PCR p oduc s on nondena u ing gels. 40 cycles o ampli ica ion we e pe o med in RT-PCR unde condi ions as in Fig. 6C. A, UV- luo escen pic u e o he e hidium-b omide s ained gels. To al RNA ob ained om emb yonic s e nal (lane 2), ju enile s e nal (lane 3) and ju enile a icula (lane 4) chond ocy es was used. Con ol eac ion was pe o med employing a cloned genomic DNA agmen ca ying exon 1 (lane 1). B, Sou he n hyb idiza ion. RT-PCR p oduc s ob ained using RNA isola ed om ju ene s e al chond ocy es we e sepa a ed on a 5% polyac ylamide gel (lane 1), blo ed and hyb idized o a con inuously labeled 7G12-speci ic p obe (lane 2) ex ending om posi ion -347 o -188, and o he 5' end-labeled oligonucleo ide PG (Table I), speci ic o he 5S3 o m (lane 3). M, posi ions o he pUC 12 HaeLI agmen s as in A. and Y in Fig. 7) we e isualized in non-dena u ing gels. The appa en mobili ies o hese bands dec eased by inc easing he empe a u e du ing elec opho esis, and hese bands hyb idized o he 5S3-speci ic oligonucleo ide (Fig. 7B, lane 3). To con i m he iden i y o he PCR agmen s, all he bands we e excised om he non-dena u ing gels, cloned and sequenced. Sequencing o he 160 bp agmen con i med he p esence o he 5S3 and 8C7 ypes, and in addi ion, e ealed he exis ence o a new splice a ian (D48), which u ilized he dono si e loca ed a posi ion -362. Sequence analysis also p o ed ha band X ep esen ed he he e oduplexes o 5S3/D48 and 8C7/5S3 PCR p oduc s, di e ing in leng h by 4 n . On he o he hand, due o he polymo phism o he emb yonic popula ion, he 2 n di e ence esul ed in he he e oduplex band Y. This band was no obse ed when RNA om s e nal o a icula ca ilage o a homozygo ic animal was used in RT-PCR (Fig. 7A, lanes 3 and 4). The o ma ion o he e oduplexes can be explained by compe i ion o he accumula ing PCR p oduc s wi h he oligonucleo ide p ime s o annealing o empla e DNA. To sum up, in spi e o he di icul ies due o he complexi y o he 5'-UTR, PCR con i med he di e si y o he LP leade sequence and e ealed di e ences in he ela i e amoun s o he a ious splice o ms. The 68-bp agmen , ep esen ing he 8G I splice a ian , was he mos abundan p oduc s a e a ious ampli ica ion egimens. The 8C7/5S3/D48 splice o ms, howe e , showed cell ype-speci ic exp ession, as being de ec able in signi ican ly lowe amoun s in ju enile a icula compa ed o emb yonic and ju enile s e nal ca ilage, in epea ed expe imen s (Fig. 6C and 7A). Finally, he 7G12- ype PCR p oduc was ound in e y small amoun s when ei he RNA sou ce was used. 5S3 7G12 Fig. 8. P edic ed seconda y s uc u es o he LP leade s. The p og am FOLD (29) was used o p edic he mos s able s em-loop s uc u e in he 5'-UTR o LP mRNAs as shown om posi ion -529 o +35 o he 7G12- and 8Gl- ype a ian s, and om -529 o + 128 o he 5S3- ype. The o e all ee ene gy o o mna ion (AG) alues a e -150.3, -71.8 and -140.4 kcal/mol o splice o ms 7G12, 8G1 and 5S3, espec i ely. A, B and A' ep esen di e en domains wi h p edic ed ee ene gies o -110.2 kcal/mol, -38.8 kcal/mol and -31.7 kcal/mol, espec i ely. As e isks ma k he cap o he longes ansc ip , solid ba s and hin lines show he posi ions o he ini ia ing and ups eam AUGs, espec i ely. Po en ial seconda y s uc u e o he LP leade s Fig. 8 shows compu e models o he mos s able seconda y s uc u es p edic ed o he 5' end o h ee LP splice o ms. The 7G12- ype a ian appea s o possess he mos ex ensi e seconda y s uc u e, which can be di ided in o wo egions: 1/ domain A om posi ion -477 o -126 con ains mos o he leade including bo h ups eam open eading ames and is p edic ed o o m a e y s able s em-loop s uc u e wi h AG=-10.2 kcal/mol; and 2/ domain B om -120 o +30 includes he ansla ion s a codon o he LP p ecu so and has a p edic ed AG= -38.8 kcal/mol. The compu e gene a ed hai pin s uc u e o he 8G1 splice o m also in ol es domain B bu domain A is signi ican ly unca ed wi h a p edic ed AG= -31.7 kcal/mol. In con a y, he seconda y s uc u es p edic ed o he 5S3, 8C7 and D48 a ian s a e en i ely di e en om ha o 7G12- ype. I sugges s ha dele ion o he leade sequence om posi ion -358 o - 189 may in luence he olding pa e n a ound he ansla ion s a si e. DISCUSSION The p esen s udies demons a e ha he ansc ip ion uni o chicken LP has a complex s uc u e. E en hough he gene comp ises o only 5 exons, he size o he ansc ip ion uni is la ge han 100 kb, due o he p esence o long in ons and unusually long UTRs (529 bp o 5 '-UTR and 4.8 kb o 3'-UTR). Long leade s a e, howe e , known o mRNAs o impo an egula o y p o eins in ol ed in g ow h and de elopmen (36-40). The LP gene is ansc ibed om mul iple ini ia ion si es loca ed be ween 34 and 76 bp downs eam o a TATA-like mo i . Al hough his mo i di e s om he canonical sequence o RNA- polyme ase H-dependen p omo e s (41), p elimina y expe imen s indica e ha a 140 bp agmen ca ying his mo i is indeed able o wo k as a p omo e in ansien exp ession assays. The occu ence o mul iple s a si es and he high GC con en o he AG Nucleic Acids Resea ch, Vol. 19, No. 18 4989 LP p omo e esembles o he ea u es o housekeeping genes (42), bu di e s om hose in he lack o he classical Spl binding si es. GC- ich sequences we e also ound in he 5'- lanking egion o o he ca ilage-speci ic genes (43, 44). One o he mos s iking ea u es o he LP gene is he complex s uc u e o he 5'-UTR (Fig. 5) a ising om al e na i e splicing wi hin exon 1, ia di e en ial u iliza ion o six dono and wo accep o splice si es. The gene a ion o six di e en splice o ms con i med by sequencing o cDNA clones and PCR agmen s, howe e , p o ides only a minimum es ima e. Addi ional splice a ian s may exis as sugges ed by he appea ance o PCR agmen s Hypl-Hyp4 (Fig. 5A and 6A). The eason why hese o ms a e no ep esen ed by he cDNA clones, p obably is ha he cDNA clones, ha bo ing sho inse s, ha e no been selec ed o u he analysis. The occu ence o mul iple s a si es, and he u iliza ion o dis inc polyadenyla ion si es (10) may also inc ease he he e ogenei y. Fu he mo e, i is possible ha al e na i e splicing in he ansla ed egion, al hough no de ec ed in chicken ca ilage, may con ibu e o he di e si y o LP mRNA in o he species o issues, as obse ed ecen ly in a chond osa coma (19). Complex s uc u e o he 5' leade has been desc ibed in a ew cases (45 -48). Al e na i e splicing wi hin he 5'-UTR is o en combined wi h al e na i e p omo e usage (47, 48). In case o he LP gene, whe e he mul iple LP mRNA species a ise om he same p omo e , he gene exp ession may u he be con olled by egula ed splicing by p o iding speci ic splicing ac o s. I is possible ha some o he LP splice o ms (7G 12, 1 1B4) may in ac ep esen p ocessing in e media es, which a e kep non unc ional by empo a ily e aining he in on. Examples o egula ed splicing ha e been e iewed ecen ly (36). The ac ha he e ogenei y is con ined o he highly conse ed 5'-UTR o he LP mRNA which shows cell ype- and s age- speci ic exp ession, opens he possibili y ha he a ious splice o ms a e subjec ed o some kind o ansla ional con ol. The di e ences in he p edic ed seconda y s uc u e o he a ious splice o ms seem o suppo his assump ion. The splice o m which had he leas s able p edic ed seconda y s uc u e (8G1), was ela i ely mo e abundan in chond ocy es o di e en o igin compa ed o he o m wi h po en ially mo e s able leade s uc u e (7G 12; AG= -110 kcal/mol). The hypo hesis o ex ensi e seconda y s uc u e o ma ion in he LP leade is subs an ia ed conside ing he di icul ies in p ime ex ension and RT-PCR analysis o he mRNA, and also in sequence analysis. The a ious LP mRNA species may ha e di e en s abili ies. Al e na i ely, hey may be ansla ed a di e en a es, since hai pin s uc u es wi h p edic ed ee ene gies g ea e han -50 kcal/mol, loca ed ups eam o he ini ia o AUG, ha e been p oposed o inhibi he ansla ion o he mRNA by in e e ing wi h he mig a ion o he ibosomal ini ia ion complex (49, 50). I has been con i med ecen ly o he mRNAs o o ni hine deca boxylase and se e al p o o-oncogenes ha long leade s, wi h ex ensi e seconda y s uc u e and ups eam AUG codons, impai ansla ion (36, 39, 40, 51, 52). Binding o egula o y p o eins, as demons a ed o he leade s o e i in mRNA and polio i us RNA (53 -56), may also be in ol ed in ansla ional con ol o al e na i ely spliced LP mRNAs. Sho open eading ames loca ed ups eam o he ac ual ansla ion s a si e may also in luence he ansla ion e iciency o an mRNA (50, 52). Two o he LP splice a ian s (7G12 and IIB4) ca y ups eam open eading ames in e ained in ons. Since all h ee AUG codons, including he ini ia o codon o LP, occu in un a o able con ex o ini ia ing ansla ion (50, 57), he ups eam AUG codons may also ha e a egula o y ole. Da a p esen ed he e p o ide a basis o u he analysis o he ole o he 5'- lanking egions and a ious leade sequences in he ansc ip ional and ansla ional egula ion o he exp ession o LP in di e en cell- ypes and du ing de elopmen . ACKNOWLEDGEMENTS We hank F.Solymosi o aluable discussion, L.Szilak o he syn hesis o he oligonucleo ides, I.Feke e and A.Simon o excellen echnical assis ance, and A.Bo ka o he a wo k. This wo k was suppo ed by G an OKKFT T 1986/296 om he Na ional Founda ion o Technical De elopmen o Hunga y and G an OTKA 168 and 896 ( o I.K.) and a special und om he Biological Resea ch Cen e ( o F.D.). REFERENCES 1. Hascall, V.C. and Sajde a, S.W. (1969) J. Biol. Chem. 244, 2384-2396. 2. Heinega d, D. and Hascall, V.C. (1974) J. Biol. Chem. 249, 4250-4256. 3. Oegema, T.R.,J ., B own, M. and Dziewia kowski, D. D. (1977) J. Biol. Chem. 252, 6470-6477. 4. Ha dingham, T.E. (1979) Biochem. J. 177, 237-247. 5. Tho n on, D.J., Sheehan, J.K. and Nieduszynski, I.A. (1987). Biochem. J. 248, 943-951. 6. De Luca S., Heinega d, D. and Hascall, V. C. (1977) J. Biol. Chem. 252, 6600-6608. 7. Bake , J.R. and Ca e son, B. (1979) J. Biol. Chem. 254, 2387-2393. 8. Roughley P.J., Poole, A.R. and Mo , J.S. (1982) J. Biol. Chem. 257, 11908-11914. 9. Le Gledic, S., Pd in, J-P., Bonne , F. and Jolles, P. (1983) J. Biol. Chem. 258, 14759-14761. 10. Deak, F., Kiss, I., Spa ks, K.J., A g a es, W.S., Hampikian, G. and Goe inck, P.F. (1986) P oc. Na l. Acad. Sci. USA 83, 3766-3770. 11. Neame, P.J., Ch is ne , J.E. and Bake , J.R. (1986) J. Biol. Chem. 261, 3519-3535. 12. Kiss, I., Deak, F., Mes ic, S., Delius, H., So6s, J., D6kany, K., A g a es, W.S., Spa ks, K.J. and Goe inck, P.F. (1987) P oc. Na l. Acad. Sci. USA 84, 6399-6403. 13. Goe inck, P.F., S i pe, N.S., Tsonis, P.A. and Ca lone, D. (1987) J. Cell Biol. 105, 2403-2408. 14. Pe in, J.-P., Bonne , F., Thu ieau, C. and Jolles, P. (1987) J. Biol. Chem. 262, 13269-13272. 15. Ga dell, S., Bake , J., Ca e son, B., Heinega d, D. and Roden, L. (1980) Biochem. Biophys. Res. Commun. 95, 1823-1831. 16. Poole, A.R., Pidoux, I., Reine , A., C6s e , L. and Hassell, J. R. (1982) J. Cell Biol. 93, 910-920. 17. Fi e, R.S., Ca e son, B. and Mye s, S.L. (1985) J. Cell Biol. 100, 1050-1055. 18. S i pe, N.S., Dicke son, K.T. and Goe inck, P. F. (1990) De . Biol. 137, 419-424. 19. Rhodes, C., Doege, K., Sasaki, M. and Yamada, Y. (1988) J. Biol. Chem. 263, 6063-6067. 20. Sinha, N.D., Bie na , J., McManus, J. and Kos e , H. (1984) Nucleic Acids Res. 12, 4539-4557. 21. Maxam, A. and Gilbe , W. (1980) Me hods. Enzymol. 65, 499-560. 22. A g a es, W.S., Deak, F., Spa ks, K.J., Kiss, I. and Goe inck, P.F. (1987) P oc. Na l. Acad. Sci. USA 84, 464-468. 23. Hanahan, D. (1983) J. Mol. Biol. 166, 557-580. 24. Saiki, R.K., Gel and, D.H., S o el, S., Scha , S.J., Higuchi, R., Ho n, G.T., Mullis, K.B. and E lich, H.A. (1988) Science 239, 487-491. 25. Lipson, K.E. and Base ga, R. (1989) P oc. Na l. Acad. Sci. USA 86, 9774-9777. 26. Sange , F., Nicklen, S. and Coulson, A.R. (1977) P oc. Na l. Acad. Sci. USA 74, 5463-5467. 27. Hu, M. C-T. and Da idson, N. (1986) Gene 42, 21-29. 28. Feinbe g, A.P. and Vogels ein, B. (1983) Anal. Biochem. 132, 6-13. 29. Zuke , M. and S iegle , P. (1981) Nucleic Acids Res. 9, 133-148. 30. De e eux, J., Haebe li, P. and Smi hies, 0. (1984) Nucleic Acids Res. 12, 387-395. 4990 Nucleic Acids Resea ch, Vol. 19, No. 18 31. Dudhia, J. and Ha dingham, T.E. (1990) Nucleic Acids Res. 18, 1292. 32. Shapi o, M.B. and Senapa hy, P. (1987) Nucleic Acids Res. 15, 7155 -7174. 33. Spi zne , J.R. and Mulle , M.T. (1988) Nucleic Acids Res. 16, 5533 -5556. 34. Buche , P. and T i ono , E.N. (1986) Nucleic Acids Res. 14, 10009- 10026. 35. Je eys, A.J., Wilson, V., Neumann, R. and Key e, J. (1988) Nucleic Acids Res. 16, 10953-10971. 36. Kozak, M. (1988) J. Cell Biol. 107, 1-7. 37. Kim, S-J., Glick, A., Spo n, M.B. and Robe s, A.B. (1989) J. Biol. Chem. 264, 402-408. 38. Wen, L., Huang, J-K. and Blackshea , P.J. (1989) J. Biol. Chem. 264, 9016-9021. 39. P ops , F., Rosenbe g, M.P., Iye , A., Kaul, K. and Vande Woude, G.F. (1987) Mol. Cell. Biol. 7, 1629-1637. 40. Rao, C.D., Pech, M., Robbins, K.C. and Aa onson, S.A. (1988) Mol. Cell. Biol. 8, 284-292. 41. Co den, J., Wasylyk, B., Buchwalde , A., Sassone-Co si, P., Kedinge , C. and Chambon, P. (1980) Science 209, 1406-1414. 42. Dynan, W.S. (1986) TIG 2, 196-197. 43. Kiss, I., Deak, F., Holloway, R.G.,J ., Delius, H., Mebus , K.A., F imbe ge , E., A g a es, W.S., Tsonis, P.A., Win e bo om, N. and Goe inck, P.F. (1989) J. Biol. Chem. 264, 8126-8134. 44. Kohno, K., Sulli an, M. and Yamada, Y. (1985) J. Biol. Chem. 260, 4441 -4447. 45. Reynolds, G.A., Golds ein, J.L. and B own, M.S. (1985) J. Biol. Chemll. 260, 10369-10377. 46. Pe al a, E.G., Winslow, J.W., Pe e son, G.L., Smi h, D.H., Ashkenazi, A., Ramachand an, J., Schime lik, M.I. and Capon, D.J. (1987) Science 236, 600-605. 47. Luo, X., Pa k, K., Lopez-Casillas, F. and Kim, K. (1989) P oc. Na l. Acad. Sci. USA 86, 4042-4046. 48. Nakajima, H., Yamasaki, T., Noguchi, T., Tanaka, T., Kono, N. and Ta ui, S. (1990) Biochem. Biophys. Res. Commun. 166, 637-641. 49. Kozak, M. (1989) Mol. Cell. Biol. 9, 5134-5142. 50. Kozak, M. (1989) J. Cell. Biol. 108, 229-241. 51. G ens, A. and Sche le , I.E. (1990) J. Biol. Chem. 265, 11810-11816. 52. Manzella, J.M. and Blackshea , P.J. (1990) J. Biol. Chem. 265, 11817-11822. 53. Leibold, E.A. and Mun o, H.N. (1988) P oc. Na l. Acad. Sci. USA 85, 2171 -2175. 54. Rouaul , T.A., Hen ze, M.W., Caughman, S.W., Ha o d, J.B. and Klausne , R. D. (1988) Science 241, 1207 - 1210. 55. del Angel, R.M., Papa assiliou, A.G., Fe nandez-Tomas, C., Sil e s ein, S.J. and Racaniello, V.R. (1989) P oc. Na l. Acad. Sci. USA 86, 8299-8303. 56. Andino, R., Rieckho , G.E. and Bal imo e, D. (1990) Cell 63, 369-380. 57. Kozak, M. (1987) Nucleic Acids Res. 15, 8125-8148.