Proc. Natd. Acad. Sci. USAVol. 86, pp. 5242-5246, July 1989Biochemistry

Vasotocin and isotocin precursors from the white sucker,Catostomus commersoni: Cloning and sequence analysisof the cDNAs

(neuropeptide/hypohlamic hormone precursor/teleos fish neurophysin/glycopeptide/copeptin)

JORG HEIERHORST*, STEVEN D. MORLEY*, JAIME FIGUEROA*, CHRISTIANE KRENTLER*, KARL LEDERISt,AND DIETMAR RICHTER*t*Institut fur Zellbiochemie und klinische Neurobiologie, Universitits-Krankenhaus Eppendorf, Martinistrasse 52, 2000 Hamburg 20, Federal Republic ofGermany; and tDepartment of Pharmacology and Therapeutics, Health Science Centre, The University of Calgary, 3330 Hospital Drive N.W., Calgary,Alberta, Canada T2N 4N1

Communicated by Berta Scharrer, March 27, 1989

ABSTRACT The nucleotide sequences of cloned cDNAsencoding the precursors for vasotocin and isotocin have beenelucidated by analyzing a Agtll library constructed frompoly(A)' RNA from the hypothalamic region of the teleost fishCatostomus commersom'. Screening of the library was carriedout with synthetic oligonucleotide probes deduced from theamino acid sequences of the nonapeptides vasotocin and iso-tocin. The cDNA nucleotide sequences predict isotocin andvasotocin prohormone precursors each consisting of a signalpeptide, a hormone moiety, and a neurophysin-like molecule.However, in comparison to their mammalian counterparts,both fish neurophysins are extended at their C termini by an=30 amino acid sequence with a leucine-rich core segment.These extensions show striking similarities with the glycopep-tide moiety (the so-called copeptin) present in mammalianvasopressin precursors, except that they lack the consensussequence for N-glycosylation. These data suggest that mam-malian copeptin is derived from the C terminus of an ancestralneurophysin.

The vasopressin/oxytocin hormone superfamily is widelydistributed throughout the animal kingdom. It comprisesmore than a dozen variants, all containing a 9-amino acidpeptide backbone with several highly conserved residuesincluding the two cysteines at positions 1 and 6, which forma disulfide bridge: Cys-Xaa-Xaa-Xaa-Asn-Cys-Pro-Xaa-Gly-NH2 (1). Their endocrine activities in mammals-namely, control of water retention (vasopressin) and smoothmuscle contraction (oxytocin)-are well documented,whereas their functions in invertebrates such as the coelen-terate Hydra (2) or the fish-hunting cone snail Conus (3)remain to be elucidated.Sequence analysis of cDNAs encoding the vasopressin or

oxytocin precursors have been carried out for several mam-malian species (4). They predict that the vasopressin precur-sor consists of the hormone, a cysteine-rich protein calledneurophysin, and a glycopeptide. While the neurophysinprobably acts as carrier protein in the transport of thehormone from the hypothalamus to the neurohypophysis, theglycopeptide, or "copeptin," is thought to function as aprolactin-releasing factor (5). The oxytocin precursor has asimilar structure except that the copeptin moiety is absentand the neurophysin sequence is slightly shorter than itsvasopressin-associated counterpart. The central regions ofthe two neurophysins are extremely well conserved betweendifferent mammals, with variations occurring predominantlyat the N and C termini.

In teleost fish, vasotocin and isotocin are, respectively, thecounterparts of mammalian vasopressin and oxytocin (1, 6).However, neither their precursor structures nor their neuro-physin sequences have been analyzed to date. Here we reportthe identification of cDNA clones encoding the teleost fishvasotocin and isotocin precursors, the structural organiza-tion of the predicted proteins, and their relationship to theirmammalian counterparts. §

MATERIALS AND METHODS

Materials. Tetramethylammonium chloride was purchasedfrom Merck-Schuchardt (Hohenbrunn, F.R.G.). Endonu-clease-free Escherichia coli DNA polymerase I (Klenowenzyme) and restriction enzymes were obtained from Boeh-ringer Mannheim; reverse transcriptase, RNase H, phage T4DNA ligase, and T4 DNA polymerase, from P. H. Stehelin(Basel); E. coli DNA ligase and T4 polynucleotide kinase,from BRL (Neu Isenburg, F.R.G.); and (dT)12_18, oligo(dT)-cellulose, and EcoRI adaptors (9-mer, CTCGTGCCG; 13-mer, AATTCGGCACGAG), from Pharmacia. Thermusaquaticus (Taq) DNA polymerase was from Cetus. EcoRI-digested, dephosphorylated Agtll arms were from PromegaBiotec. Gigapack Gold in vitro packaging extracts werebought from Stratagene. 2'-Deoxyadenosine 5'-[a-35S]tri-phosphate (500 Ci/mmol; 1 Ci = 37 GBq) was from Du Pont,and adenosine 5'-[-32P]triphosphate (6000 Ci/mmol) camefrom Amersham-Buchler (Braunschweig, F.R.G.). Oligonu-cleotides were synthesized in an automated DNA synthesizerusing phosphoramidites.

Construction of a Agtll Library. Total RNA was isolatedfrom hypothalamic fragments, containing the nucleus preop-ticus regions, of 225 suckers (Catostomus commersoni) (7),and poly(A)+ RNA was purified by oligo(dT)-cellulose chro-matography (8). cDNA was prepared from 5 ,ug of poly(A)+RNA according to the method of Gubler and Hoffmann (9)and inserted via EcoRI adaptors into the EcoRI site of Agtll(10). The resulting library consisted of 107 independentrecombinants and was screened without resort to amplifica-tion.

Identification of cDNAs Encoding the Sucker Vasotocin andIsotocin Precursors. For each screen 2.5 x 105 recombinantswere analyzed by in situ hybridization on nitrocellulosereplica filters (=25,000 phage per 150-mm plate) with fullydegenerate synthetic 20-mer oligonucleotide pools corre-sponding to the first 7 amino acids of the respective hormone

Abbreviation: PCR, polymerase chain reaction.tTo whom reprint requests should be addressed.§The sequences reported in this paper have been deposited in theGenBank data base (accession nos. M25144 and M25145).

5242

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Biochemistry:Heierhorstetal.Proc. Nati. Acad. Sci. USA 86 (1989) 5243

sequences. For isotocin two pools were generated: pool I,5'-GG-RCA-RTT-NGA-DAT-RTA-RCA-3'; pool II: 5'-GG-RCA-RTT-RCT-DAT-RTA-RCA-3'. For vasotocin asingle pool was synthesized: 5'-GG-RCA-RTT-YTG-DAT-RTA-RCA-3'. (R = A or G; N = A, C, G, or T; D = A, G,or T; Y = C or T.) The oligonucleotides were 5'-end-labeledwith T4 polynucleotide kinase to a specific activity of >108cpm4&g and filters were screened using conditions underwhich only exact matches with the target sequence should bedetected (11). In brief, after nonstringent hybridization, fil-ters were washed in a solution containing tetramethylammo-nium ions, the binding of which eliminates the preferentialmelting of A-T versus G-C base pairs, allowing the stringencyof washing to be controlled solely as a function of probelength and independently of base composition. The hybrid-ization and washing conditions used reproduced those de-scribed previously (11), except that 50 pmol of end-labeledoligonucleotide pool was used per hybridization; the finalstringent wash steps were carried out at 580C, a temperature2-40C below the dissociation temperature calculated for20-mer oligonucleotides in 3 M tetramethylammonium chlo-ride.

Nucleotide Sequence Analysis. Inserts from positive cloneswere subcloned into Bluescript plasmids (Stratagene) or intoM13 phage vectors (12) and their identities were confirmed bysequence analysis on both strands (13). The precursor struc-tures of sucker vasotocin and isotocin were compared withthose for their mammalian counterparts, vasopressin andoxytocin, by alignment of nucleotide and predicted proteinsequence, using the ALIGN program of the DNASTAR (Mad-ison, WI) sequence analysis software package, which isbased on the algorithm of Wilbur and Lipman (14).

isotocinvasotocin

Polymerase Chain Reaction (PCR) and Genomic DNA Se-quencing. One microgram of sucker liver genomic DNA (15)was subjected to 30 cycles of PCR amplification (16) usingoligonucleotide primers deduced from the vasotocin- or iso-tocin-encoding cDNA sequences. The reaction mixture (1001ud) contained 2.5 units of Taq polymerase, 100 pmol each ofan appropriate "forward" and "reverse" oligonucleotideprimer (Fig. 1, arrows). Annealing was at 550C for 120 sec,extension was at 720C for 75 sec (except for the last cycle,where 5 mmn was used), and denaturation was at 920C for 75sec. Agarose gel (2%) electrophoresis and ethidium bromidestaining revealed in each case a single amplified DNA bandcorresponding to the expected size of 160 base pairs (bp) forvasotocin and 195 bp for isotocin.For direct sequencing of amplified genomic DNA, single-

stranded DNA template was generated (17) and the PCRprotocol was modified by using 0.5 pmol and 50 pmol ofappropriate forward and reverse primers and 35 amplificationcycles. After an ethanol/ammonium acetate precipitationstep, the amplified DNA was sequenced by using the forwardoligonucleotide as primer (13). Sequencing of the oppositestrand was performed by reversing the oligonucleotide con-centrations in the PCR and using the reverse oligonucleotideas pnimer.Northern Blot Analysis. Poly(A)' RNA (10 ,&g) was dena-

tured with glyoxal (18), electrophoresed in a 1.2% agarosegel, and transferred to Hybond-N membranes (Amersham)according to the supplier's instructions. Hybridization wascarried out with labeled (19) cDNA probes (106 cpm/mi)specific for either vasotocin (nucleotides 440-535; Xba I-EcoRI fragment) or isotocin (nucleotides 420-692; Pvu II-EcoRI fragment) mRNA.

5' CTGAATCGTTTCACTACCTGTCAATTTCAGCCTTTGCAAAATC

i s ot oc in N Pm s g s m f s v f 5 11 y 1 1 s v c s a cy i s n cp i g g k r ai

43 AGCGAGAGTTATTTCCGTTTTCGCGCGTACTTAACCACGCCTGTGAAGGC2 GGTTTATTTTGCAGGACTTTTCCCCCCTTCTCAACAACGCGGGAGAGGTT

msd s f1 p tc ii1 cli a i s s a cy i qn cp r gg k r s qv as ot o ci n N P

q d sp srq cm s cg pg dr gr cf gps i cc g eg 1 g cl143 TCGATACTTGCGGAGCTCGCGGGTGGCGTTTGTCATTTCGGTAGTTGCGC10? AGCCAGACACC- --AGCAGAGAGTGTGTGTCGTGTGGTCCTGGGAATGCAGGCCGTTGTTACGGTCCCAGTATCTGCTGTGGAGCGGCTCTGGGTTGTCT

p dt s re cv s cgp g nag r cy g psi c cga al1gcI

34

33

67

65

1 g s p e t q r c 1 e e d f 1 p s p c e a g g k v c g y e g r c a 100243 GTGCCCGAATACAGCGAGGATTTCCCCGGGGCGTGAATTCGTTAGCGTTC199 CTGTCCAAAGTACGAGAGGACCTCGGCCGGGCGAGAACTTGGTAGAGTTC

v g s p e t . s c m e e n h 1 p s p c e t g g r p c g d e g r c a 98

a p g v c c d s e g c s v q9s vd g d g d-at- a v s q p a s s q 134343 GCCAGGCGTTATAAGCGA 7TACGCT-- ATGGCGGCCAACGCGCACGCGA C299 GCCCGAGCTTGTACCGTATTCGTTGACCGATT__GAACTCGTCGACGTTAAATCAGAGCCAG

a p g v c c d s v s c v m d s e c 1 e d v r s d q s e d p s r 1 k t 132

d Ill k 11I hI s np a h py r h q

399 CTTTCGGGAACCTCGGTTTTAATGGGTTGAGCGCGGCTCTAACTTATCTGGTTGCCACCGCv sg ei 1 In 1 nlIa s rgr r df

154

152

499 GTGTtTCCAGTGTGTAAATATGTAATAAATGGTCAAATGTGAAAAAAAAAAAAAA

643 TTTACTGTGCAGCTATTAAAGTCAGAGAGAACGTAAAAAAAAAAAAAAA 3'

FIG. 1. Comparison of nucleotide sequences and deduced amino acid sequences of the sucker isotocin and vasotocin precursor cDNAs.Polyadenylylation signals are underlined. Nucleotide sequences used for synthetic PCR primers (Fig. 2) are marked by arrows. Amino acidsare represented by lowercase one-letter symbols. Hormone amino acid sequence [residues 21-29 (isotocin) or 20-28 (vasotocin)] is followedby a Gly-Lys-Arg sequence and then the neurophysin (NP) sequence, which extends to the C terminus. Broken lines show amino acid sequenceswhere a G -.+ A point mutation would generate a potential glycosylation site (additional sites are located further upstream, but these are notmarked because they are not the right distance from the leucine-rich core segment). Numbers on the left indicate nucleotide positions; thoseon the right are for amino acid sequences. Three dashes in the vasotocin cDNA sequence indicate a gap introduced to optimize the alignment.

Biochemistry: Heierhorst et al.

5244 Biochemistry: Heierhorst et al.

RESULTSCloning and Nucleotide Sequence Analysis. A cDNA library

representing mRNA from the sucker brain hypothalamicregion was constructed in Agtll and was screened with two20-mer oligonucleotide pools deduced from the first 7 aminoacids of the isotocin and vasotocin sequences, respectively.In each case, analysis of 2.5 x 105 recombinants gave rise to4 positively hybridizing clones. Subsequent sequence anal-ysis confirmed the validity of our oligonucleotide screeningstrategy by revealing two families ofclones, one encoding theisotocin precursor and the other the vasotocin precursor.The largest cDNA clone encoding the isotocin precursor

consists of 691 bp, while that for the vasotocin precursor is553 bp long (Fig. 1). The isotocin-encoding insert exhibits asingle open reading frame of 462 nucleotide residues predict-ing a protein of 154 amino acid residues with Mr 16,055. Thevasotocin counterpart encodes 152 amino acid residues giv-ing rise to a calculated Mr of 16,011.The isotocin cDNA includes a putative 5' untranslated

region with the first possible initiation codon at position 43.However, the 5' region ofthe vasotocin cDNA is incomplete,with the first ATG triplet encountered at position 5. Neitherof the potential initiation codons are preceded by the con-sensus sequences for initiation found in most eukaryotes (20),but they are succeeded by putative signal sequences (seebelow). The 3' untranslated region of isotocin cDNA issignificantly longer than that for vasotocin, suggesting asomewhat larger isotocin mRNA. This was confirmed byNorthern blot analysis of sucker poly(A)+ RNA, whichrevealed RNA bands of 775 bases for vasotocin mRNA and920 bases for isotocin mRNA (Fig. 2). Provided that theisotocin cDNA represents the full-length mRNA, it shouldcontain a poly(A) tail of -220 residues. Consensus sequencesfor polyadenylylation are present at positions 573-578 and657-662 in isotocin cDNA and at positions 522-527 in vaso-tocin cDNA (Fig. 1).

Structural Organization of the Isotocin and Vasotocin Pre-cursors. As shown in Fig. 1, the two precursors possesssimilar structures, each consisting ofa putative signal peptidefollowed directly by the hormone moiety and then by acysteine-rich protein with many characteristics of the neu-rophysin family.The predicted signal sequences of the two precursors

contain a high proportion of hydrophobic amino acids, afeature characteristic for signal peptides. Each nonapeptidehormone is connected to the rest of the precursor by theresidues Gly-Lys-Arg, known to serve as potential signals forhormone amidation (Gly) and precursor processing (Lys-Arg), respectively (21).Comparison of the sucker vasotocin- and isotocin-

associated neurophysins reveals a highly conserved centralregion at amino acid residues 10-76 (isotocin neurophysin) or9-75 (vasotocin neurophysin), which also shows consider-able homology to their mammalian (1), avian (22), andamphibian (23) counterparts. Unexpectedly, the cDNA se-

IV _

_1-2-3-4

FIG. 2. Northern blots of sucker poly(A)' RNA. Blots werehybridized with either a vasotocin (V)- or an isotocin (I)-specificprobe. Hae 111-digested, 32P-labeled OX174 DNA provided molec-ular size markers: 1, 1353 bp; 2, 1078 bp; 3, 872 bp; 4, 603 bp. Thedifferences in the intensity of the signals for the two mRNAs are dueto different specific activities of labeled probes used.

quences for both precursors predict larger neurophysin-likeproteins that are extended at their C termini by a stretch of=30 amino acid residues including a leucine-rich core seg-

A 1 2 3 4 5 6 7

_,195 bp' 160 bp -

B 3' 3'T L e u T\C C

A A,,G A r g G \

A A

TGA

CCC

A

AAG

.T

\C

5

-1'S e r G

A

P1 r o CC

A s p AG

AG I u A

G

S e r C/I

5

FIG. 3. (A) Ethidium bromide staining (Left) and Southern blotanalysis (Right) of PCR-amplified sucker liver genomic DNA encod-ing the vasotocin precursor. Lane 1; Hae 111-digested 4X174 DNAmarkers; the following lengths in base pairs are indicated (top tobottom) by horizontal lines: 310, 281/271, 234, 194, 118, and 72; lanes2 and 5, cDNA encoding the vasotocin precursor; lanes 3 and 6,cDNA encoding the isotocin precursor; lanes 4 and 7, sucker livergenomic DNA. The "forward" and "reverse" primers used foramplification are indicated by arrows in Fig. 1. Amplification of theisotocin cDNA (lane 3) gave rise to the expected band of 195 bp. Theamplified DNA was electrophoresed in a 2% agarose gel and South-ern blotted (15). The blot was hybridized to the vasotocin-specificprobe (see Fig. 2). No cross-hybridization was observed with iso-tocin cDNA (lane 6). (B) Comparison of cDNA nucleotide residues374-394 (Fig. 1) and amplified genomic DNA sequence encoding thesucker vasotocin precursor. (Right) Genomic DNA was amplifiedand sequenced as outlined in Materials and Methods. (Left) Corre-sponding cDNA sequence. Lanes (from left to right) show productsof T, C, G, and A sequencing reactions.

Proc. Natl. Acad. Sci. USA 86 (1989)

Proc. Natl. Acad. Sci. USA 86 (1989) 5245

rat AVPbovine AVPhuman AVPtoad vasotocinsucker vasotocinsucker isotocin

AREQ$:NA~~~~AtQlOCP.A::RElLLRLVQLAGTQESVDSAKPRVY

NORE-S: NA- LL : At LL R L V QtLAC A P E P A E PAQ P G V YA sRKA....:^S| t $: " A ~TQt -0 6 A- :0 Lei L R L V; Q: L A 6 A P E fP F f P A Q P D A YVTP E Q : " M X:IQ M :0 A S A. :S : D -t t t Rt L M V M- A N 07RQ Q Q S K :H Q F ::YS Q$E PS R [K T V : E LRLINLA S R G R R D F::D G: : A. T A V SQP A S 5 Q0 tt .I I At: : S: N::: P H P Y R L H Q

FIG. 4. Comparison of glycopeptide (copeptin) sequences of various species with the C-terminal region of sucker vasotocin and isotocinprecursors. Aligned amino acid residues that are homologous in the sucker and the glycopeptide sequences are shown in bold type. Colonsindicate gaps introduced to optimize homology. The sequences are from refs. 4 and 23. AVP, arginine vasopressin.

ment. This part of the neurophysin sequence shows a re-markable resemblance to that of the glycopeptide (copeptin)of the vasopressin precursor family except that the teleostfish sequence lacks a potential glycosylation site.

Analysis of the C Terminus of Sucker Neurophysin. Inspec-tion of the amino acid and nucleotide sequences of thevasotocin and isotocin precursors reveals sites within theC-terminal regions that, by a single mutation, could be alteredto a glycosylation site (Fig. 1, broken lines). In the case oftheisotocin precursor, the respective codons (nucleotides 409-417) read Asp-Ala-Thr, and those of the vasotocin precursor(nucleotides 380-388) read Asp-Pro-Ser. In either case, asingle G -- A mutation would change the aspartic codon(GAC or GAT) into an asparagine-encoding triplet (AAC orAAT), thus establishing the N-glycosylation consensus se-quence Asn-Xaa-Thr/Ser.To show that the sequence Asp-Pro-Ser determined from

the vasotocin cDNA did not derive from a cDNA synthesisartifact but was also present in the genomic DNA encodingthe respective precursor, sucker liverDNA was amplified bythe PCR technique (16). Two synthetic 24-mer oligonucleo-tide primers were synthesized from the previously deter-mined vasotocin cDNA sequence up- and downstream of thesite encoding Asp-Pro-Ser (Fig. 1, arrows). The amplifiedDNA, consisting of 160 bp, was shown to have a sequenceidentical to that obtained for the respective cDNA region(Fig. 3). This experiment independently confirmed at thegenomic level that the sequence reads Asp-Pro-Ser. In anal-ogy the sequence Asp-Ala-Thr of the isotocin precursor wasfound to be encoded in the amplified genomic DNA fragment(195 bp) ofthe isotocin gene (data not shown). Hence both thevasotocin and isotocin precursor truly lack an N-linkedglycosylation site.

DISCUSSIONThe data show that vasotocin and isotocin are expressed ascomposite precursors, together with their respective neuro-physins, in the hypothalamic neurosecretory system of theteleost fish C. commersoni. Although the primary structureof the two identified neurophysins indicates an organizationsimilar to their counterparts from mammals (1), birds (22),and amphibians (23), they exhibit features not previouslyfound in other neurophysin molecules. Both sucker neuro-physins are extended by -30 amino acids at their C termini,and these extensions show striking similarities with theglycopeptide ofthe vasopressin precursor family, except thatthe sucker sequences lack a glycosylation site (Fig. 4). In ratsthis glycopeptide, also known as copeptin, has been shown tobe involved in prolactin release (5), a function that has notbeen determined in the sucker endocrine system.The cDNA sequences encoding the two sucker precursors

predict that a single nucleotide mutation could change anaspartic residue into an asparagine residue to generate apotential glycosylation site (Asn-Xaa-Thr/Ser) within theC-terminal region of the neurophysin molecules (Fig. 1). Thepotential glycosylation site and the leucine-rich core se-quence would be separated by roughly the same number ofamino acid residues as they are in the mammalian (4) and toad(23) glycopeptides. Also, upstream of the potential glycosyl-

ation site in the sucker vasotocin precursor, there is anarginine residue that may have become a processing signal inthe corresponding precursors of higher eukaryotes.

Further, our data refine the hypothesis that the two hor-mone precursors are derived from an ancestral molecule,presumably by duplication ofthe vasotocin gene encoding thenonapeptide and a larger neurophysin protein. In the courseof evolution the C-terminal part of the neurophysin eventu-ally became a separate glycopeptide with distinct biologicalfunction(s) (vasopressin precursor) or was simply discardedbecause of lack of need (oxytocin precursor). Nevertheless,it should be kept in mind that teleost fishes have evolved fora long time (>400 million years; ref. 24) in a phylogeneticbranch quite remote from the line ofdescent of the tetrapods.

We thank Mrs. Helen Wilson of Nanton (Alberta) for help in fishcollection; Henk Zwiers, Evelyn and David Ko, Ilona Berbeckar,and Estrella Gonzalez for assistance with the collection of fishhypothalami and preparation of total RNA; Werner Rust and MarionDaumigen-Kullmann (Hamburg) for technical assistance; and Dr.Willi Kullmann (Hamburg) for synthesizing the oligonucleotides.This work was supported by the Deutsche Forschungsgemeinschaft(D.R.) and the Medical Research Council of Canada (K.L.).

1. Acher, R. & Chauvet, J. (1988) Biochimie 70, 1197-1207.2. Grimmelikhuijzen, C. J. P., Dierickx, K. & Boer, G. J. (1982)

Neuroscience 7, 3191-3199.3. Cruz, L. J., de Santos, V., Zafaralla, G. C., Ramilo, C. A.,

Zeikus, R., Gray, W. R. & Olivera, B. M. (1987) J. Biol. Chem.262, 15821-15824.

4. Rehbein, M., Hillers, M., Mohr, E., Ivell, R., Morley, S.,Schmale, H. & Richter, D. (1986) Biol. Chem. Hoppe-Seyler367, 695-704.

5. Nagy, G., Mulchahey, J. J., Smyth, D. G. & Neill, J. D. (1988)Biochem. Biophys. Res. Commun. 151, 524-529.

6. Gill, V. E., Burford, G. D. & Lederis, K. (1977) Gen. Comp.Endocrinol. 32, 505-511.

7. Okawara, Y., Morley, S. D., Burzio, L. O., Zwiers, H., Led-eris, K. & Richter, D. (1988) Proc. Natl. Acad. Sci. USA 85,8439-8443.

8. Aviv, H. & Leder, P. (1972) Proc. Natl. Acad. Sci. USA 69,1408-1412.

9. Gubler, U. & Hoffman, B. J. (1983) Gene 25, 263-269.10. Huynh, T. V., Young, R. A. & Davis, R. W. (1985) in DNA

Cloning, ed. Glover, D. M. (IRL, Oxford, U.K.), Vol. 1, pp.49-78.

11. Wood, W. I., Gitschier, J., Lasky, L. A. & Lawn, R. M. (1985)Proc. Natl. Acad. Sci. USA 82, 1585-1588.

12. Yanish-Perron, C., Vieira, J. & Messing, J. (1985) Gene 33,103-119.

13. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.Acad. Sci. USA 74, 5463-5467.

14. Wilbur, D. J. & Lipman, W. J. (1983) Proc. Natl. Acad. Sci.USA 80, 726-730.

15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) MolecularCloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

16. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi,R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science239, 487-491.

17. Gyllensten, U. B. & Erlich, H. A. (1988) Proc. Natl. Acad. Sci.USA 85, 7652-7656.

18. Thomas, P. S. (1983) Methods Enzymol. 100, 255-266.

Biochemistry: Heierhorst et A

5246 Biochemistry: Heierhorst et al.

19. Feinberg, A. P. & Vogelstein, B. (1984) Anal. Biochem. 137,266-267.

20. Kozak, M. (1984) Nucleic Acids Res. 12, 857-872.21. Richter, D. (1986) in Molecular Cloning ofHormone Genes, ed.

Habener, J. F. (Humana, Clifton, NJ), pp. 173-206.

Proc. Nadl. Acad. Sci. USA 86 (1989)

22. Lazure, C., Saayman, H. S., Naud6, R. J., Oelofsen, W. &Chrdtien, M. (1987) let. J. Peptide Protein Res. 30, 634-645.

23. Nojiri, H., Ishida, I., Miyashita, E., Sato, M., Urano, A. &Deguchi, T. (1987) Proc. Natl. Acad. Sci. USA 84, 3043-3046.

24. Acher, R. (1980) Proc. R. Soc. London Ser. B 210, 21-43.