Edward Kislauskis et al- The Rat Gene Encoding Neurotensin and Neuromedin N

8/3/2019 Edward Kislauskis et al- The Rat Gene Encoding Neurotensin and Neuromedin N

1/6

THE OURNALF BIOLOGICALHEMISTRY0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No . 10, Issue of April 5 , pp. 4963-4968, 1988Printed in U .S.A.

The Rat Gene Encoding Neurotensinand Neuromedin NSTRUCTURE,TISSUE-SPECIFICEXPRESSION, AND EVOLUTION OF EXON SEQUENCES*

(Received for publication, November 3, 1987)

Edward Kislauskis, Bryant Bullock, Sandra McNeil, and PaulR. DobnerSFrom the Dewrtment of Molecular Geneticsand Microbiology, Uniuersity of Massachusetts Medical Center,Worcester, Massachusetts 01655

Recombinant DNA clones encoding the neurotensinlneuromedin N precursor protein have been isolatedfrom both bovine hypothalamus cDNA and ra t genomiclibraries using a heterologous canine cDNA probe. NU-cleotide sequence analysis of these clones and compar-ison with the previously determined canine sequencehas revealed that 76% of the amino acid residues ar econserved in a ll three species. The protein precursorsequences predicted from bovine hypothalamus andcanine intestine cDNA clones vary at only 9 of 170amino acid residues suggesting that within a speciesidentical precursors re synthesized in both the cen tralnervous system and intestine. The rat gene spans ap-proximately 10.2 kilobases (kb) and is divided intofour exonsby three introns. The neurotensin andeu-romedin N coding domains ar e tandemly positioned onexon 4. RNA blot analysis has revealed that the ratgene is transcribed to yield two distinct mRNAs, 1.0and 1.5kb in ize, in all gastrointestinal and all neuraltissues examined except he cerebellum. There is astriking variation in the relative levels of these twomRNAs between brain and ntestine. The smaller1.0-kb mRNA greatly predominates in intestine whileothmRNA species a re nearly equally abundant in hypo-thalamus, brain stem, and cortex. Sequence compari-sons and RNA blot analysis indicate that these twomRNAs result from the d ifferential utilizationf twoconsensus poly(A) addition signals and differ in theextent of their 3 untranslated regions. The relativecombined levels of the mRNAs in various brain andintestine regions correspond roughly with the relativelevels of immunologically detectable neurotensin ex-cept in the cerebral cortex wheremRNA levels are 6times higher than anticipated.

Neurotensin (NT) is a member of a family of structurallyrelated peptides which cause contraction of smooth muscleand, when injected into theperiphery of anesthetized rats, anacu.tehypotensive response. The carboxyl-terminal portion ofNT is the major determinant of biological activity (1)and isthe most conserved region of the different family members.

*T hi s work was supported in part by Grant HL33307 from theNational Institutes of Health (toP. R. D.). The costs of publicationof this article were defrayed in par t by the payment of page charges.Thi s article must therefore be hereby marked uduertisement inaccordance with 18U.S.C. Section 1734 solely o indicate this fact.The nucleotide sequence(s) eported in this aper hasbeen submitted503185.to the GenBankTM/EMBL Data Bank with accession umber(s)$ Supported by National Inst itutes of Health Biomedical ResearchSupport Gran t S07RR5712. To whom correspondence should be sent.The abbreviations used are: NT, neurotensin; NGF, nerve growthfactor; kb, kilobase; N, neuromedin N.

The frog skin peptide, xenopsin (2), shares four of five car-boxyl-terminal amino acid residues with NT. Similarly, the 4carboxyl-terminal amino acids of NT and a 6-amino acidpeptide isolated from porcine spinal cord, neuromedin N, areidentical (3 ) .Recently, we have determined the primary struc-ture of a 170 amino acid precursor protein which encodes bothNT and neuromedin N by sequencing cDNA clones derivedfrom the canine intestine (4).NT was first isolated from bovine hypothalamus (5) and islikely to serve as a neurotransmitter or neuromodulator inthe central nervous system (6). The development of a radio-immunoassay for NT led to he discovery of NT in thegastrointestinal tract (7) where it maybe nvolved in theregulation of fat metabolism (8). Discrete NT-containingendocrine cells are found dispersed throughout the intestinalmucosa, most prominently in the distal small intestine (9).NT has also been detected in the adrenal medulla of a varietyof mammals (10). In cats, a specific subpopulation of nor-adrenaline-containing chromaffin cells also contain NT (11).Interestingly, the at pheochromocytoma PC12 ell line,which displays some of the phenotypic characteristics of ad-renal chromaffin cells, can be induced to produce high levelsof NT immunoreactive material by combinations of hormones(12, 13).NT content and production are increased up to 600-fold by the combined action of nerve growth factor (NGF),dexamethasone, and activators of adenyl cyclase. The induc-tion is highly cooperative with NGF exerting primarily apermissive effect. This is exciting in view of the importanceof NGF as a urvival and differentiation factor affecting cellswhich migrate from the neural crest during development (14).PC12 cells are thought to be representative of pleuripotentneural crest cells and differentiate along a sympathetic neu-ronal pathway in response to NGF (15).The isolation of cDNA clones encoding NT has madepossible molecular approaches to understanding the biosyn-thesis of the peptide and the regulation of the gene whichencodes it. To examine these questions in tractable experi-mental models, we have isolated and characterized the ratgene encoding the NT/neuromedin N (NT/N) precursor andexamined its tissue-specific expression. We also report that itis likely that an identical NT/N precursor is synthesized inthe gut and central nervous system.

E X P E R I M E N T A LP R O C E D U R E SMaterials-Avian myeloblastosis virus reverse transcriptase wasfrom Life Sciences (St. Petersburg, FL); restriction and other enzymeswere from Boehringer Mannheim and New England Biolabs; 32P-labeled nucleotides from Amersham Corp.; and deoxy and dideoxy-nucleotides from Pharmacia LKB Biotechnology Inc.Construction of Recombinant Libraries-DNA was isolated fromrat testes, partially digested with Sau3A, and size-fractionated onsucrose gradients. Digestion products 15-20 kb in size were ligated

with BamHI digested XEMBL4 phage arms, and packaged in vitro as4963


2/6

4964 Rat Gene Encoding Neurotensin and Neuromedin Ndescribed (16). A cDNA library was constructed from poly(A)+ RNAisolated from bovine hypothalamus in Xgtll as described (4).Screening Procedures-Escherichia coli (K802 or Y1088) were in-fected with recombinant phage a nd plated onto 15-cm bacterial platesat a density of 20,000-30,000 plaques/plate. Th e EcoRI insert of afull-length canine neurotensin precursor cDNA (4) was 32P-labeledby nick translation and used to screen recombinants as described(16).Hybridization was performed in 5 X NaCl/sodium citrate (1 XNaCl/sodium citrate = 0.15 M NaC1, 0.015 M sodium citrate, pH 7) ,10 X Denhardt's solution (1 X Denhardt's solution = bovine serumalbumin, polyvinylpyrrolidone, and Ficoll all at 0.2 mg/ml), 50 mMsodium phosphate buffer (pH 7), 0.5% sodium dodecyl sulfate at 60 "Cand the filters were washed in several changes of 2 X NaCl/sodiumcitrate, 0.1% sodium dodecyl sulfate at room temperature followed bya wash in the same solution a t 50 'C. Positively hybridizing bacteri-ophage were plaque purified and XDNA was prepared as described(16).DNA Sequence Analysis-Restriction fragments of rat genomicclones encompassing exons were identified by Southern blotting singa 3ZP-labeledcanine neurotensin precursor cDNA probe and sub-cloned into either pGEM4 (Promega Biotec) or pUC12 (17). Variousdefined fragments were subcloned into M13mplO or m pl l (17) andsequenced by the dideoxy method (18).RNA Analysis-RNA was extracted from various tissues by theguanidine thiocyanate procedure (19) and poly(A)+RNA was selectedby passage over oligo(dT)-cellulose (20). RNA blot analysis, andnuclease protection experiments were performed as described (21,22). For primer extension, 32P-labeled DNA primer (100,000 cpm)was hybridized to RNA in 0.1 M NaCl, 20 mM Tris-HC1 (pH 7.9), 0.1mM EDTA for 10 h at 60 "C.An equal volume of a solution containing80m M Tris-HC1 (pH 7), 10 mM MgC12, 400pM each of dGTP, dCTP,dATP, TTP, and10 units of reverse transcriptase were added on ice.The reactions were incubated at 37 "C for 30 min and the productswere analyzed as described (22). Uniformly labeled single-strandedprobes were synthesized using an EcoRIlHinfI fragment of rNT19subcloned in M13mplO. The probes used for primer extension andnuclease protection were generated by cleavage with Sst I andEcoRI,respectively, followed by isolation of the labeled strand on 6% acryl-amide, 7 M urea gels.

RESULTSScreening of Libraries-Recently, we have demonstratedusing a canine NT/N cDNA probe that different sized NT/NmRNAs are expressed in bovine hypothalamus and canine

intestine (4). A 1.5-kb poly(A)+ RNA is detected in bovinehypothalamus while a 1.0-kb poly(A)+ RNA is detected in thecanine ntestine. These results opened the possibility thatthese mRNAs encoded different precursor proteins specific toeither neural or gastrointestinal tissues. They also suggestedthat the canine cDNA probe could be used to isolate corre-sponding sequences from other mammalian species. Southernblot analysis of dog, rat, and human genomic DNA revealedthat an nique gene was dentified by the canine NT/N probein each case (data not shown). To examine the structure ofthe larger 1.5-kb RNA detected in bovine hypothalamus andto characterize the rat NT/N ene, we screened a Xgtll cDNAlibrary derived from bovine hypothalamic poly(A)+ RNA anda rat genomic library constructed using XEMBL4 with a 32P-labeled canine NT/N cDNA probe.Screening of approximately 240,000 plaques from the bo-vine hypothalamus cDNA library resulted in thedentificationof 14 hybridization-positive clones. Based on preliminaryrestriction analysis, two of these (bhNT3 andbhNT12) wereselected for detailed analysis. Approximately 360,000 inde-pendent rat genomic recombinants were screened resulting inthe isolation of three hybridization-positive clones. TWOfthese were nearly identical (rNT18 and rNT19) while thethird (rNT23) overlapped throughout half of the cloned se-quences. Southern blot analysis using the canine cDNA probeindicated that rNT18 and rNT19 contained the entire ratNT/N precursor gene and rNT19 was chosen for detailedanalysis.

Equivalent Protein Precursors in Hypothalamus and Zntes-tine-The composite structure of the bovine hypothalamuscDNA clones is depicted in Fig. l.4 along with the structuresof the ratgene and previously described canine cDNA clones.The complete cDNA sequences and comparable exon se-quences from the rat gene are presented in Fig. 2. The NT/Nprotein precursor predicted from the bovine hypothalamuscDNA sequence is nearly identical to that determined previ-ously from canine intestine cDNAs. The general features ofthe precursor are he same with the neuromedin Nandneurotensin coding domains located in tandem near the car-boxyl terminus of a 170-amino acid protein, bounded andseparated by Lys-Arg basic amino acid pairs. There is alsoconsiderable similarity in the untranslated regions; however,the 3' untranslated region of the bovine sequenceextends 402bases beyond the comparable region of the canine sequence.In addition to a consensus poly(A) signal common to the wocDNA sequences, the bovine sequence contains a consensussignal near the end of the extended region.Structure of the Neurotensin/Neuromedin N Gene-Re-

Ra t NeurotenrinGene500 bp

E n n n n

U50 bp

FIG.1. Structure of rat neurotensin/neuromedin N geneand sequencing strategy. A, the structure of rNT19 isolated froma rat genomic recombinant library. In the ratene, black boxesdenoteEcoRI (E) and HindIII ( H ) restriction sites are indicated. Theexons and thin lines denote either introns or

flanking sequences.structure of the gene is projected onto schematic representations ofcow and dog NT/N cDNA clones. Boxes represent coding sequencesand known functional domains are indicated by diamonds, signalsequence; uertical lines, neuromedin N-like; diagonal lines, neurome-din N black, neurotensin; open, regions of unknown function. B ,regionsof rNT19 corresponding to exons were identified by Southernblot analysis and subcloned into plasmid vectors. Exon sequenceswere determined by the dideoxy method using the strategy depicted.Exon sequences are indicated by either black boxes (coding) or openboxes (untranslated); other sequences are indicated by a thin line.


3/6

Rat Gene Encoding Neurotensin and Neuro medin 4965

(OW I l W 1110 1120 1 1 J O* u I I I I U W ~ I I ~ I U c I * ~ ~ G ~ G c ~ c ~ c ~ C A I ~ ~ ~ ~ c ~ c ~ ~ ~ G . ~ - - - . . - . - . . - . - . - . - . . - C ~ O ~ c ~ ~ ~ G G - - c ~ ~ c ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ O ~ ~ ~ ~ ~ , , ~ , ~ " , , ~ ~ ~ ~A w. '1~""""""" L G T C , i i G C C l i i C L l i l C i l l 1 u ( ( I l AA 16 T IT T c 6"I l l 0 l rOO lZ l0 l i 2 D 1210 1140 1150 I 260 1))" ,110

11 % i l i 0 ,160 I l l 0 '180

LUL~11111111 '1"- I I I I I IT~ ' IG IT I6 r l l r ,~..,~,~ ' , ~ ~ U ~ ~ G ~ ~ ~ , , , ~ , ~ , ~ ~ ,. . . . ,~~~ . . . .~~~~~,~ ,O ~O c , , G , , ~ ~ , - c ~ , c ~ ,LC""" K T AG - 7.". T I LTGG G il""" GUG 11146 n l c T CA 1 6 1 Tr T(I I W lliU I 2 0 ,130 , 3 4 0 135" 1160 ,,io 118" i1wl i r r T 1 1 6 L " " " " " " - ~ W ~ ~ ~ C ~ ~ ~ ~ ~ ~ , C , ~ , ~ ~ ~ , ~ ~ ~ ~ ~ , , ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ , ~ - ~ ~ ~ ~ ~ ~ ~ ~ , c , ~ ~ ~ ~ * ~ , ~ ~ , ~ ~ , " ~ , ~ , " ~ ~ ~ ~ , , , ~ , , ~ ~ , , ~.* 6 T L I T L I T M 7W 7 i i l b l i l i G ~ " . " " ~ 71 T 7 71 M.. t

FIG. . Sequence of the rat neurotensin/neuromedin N pre-cursor gene and corresponding sequences of bovine hypothal-amus and canine ntestine cDNAs. Rat gene sequences weredetermined by the dideoxy method using the strategy depicted in Fig.1B. Bovine hypothalamus cDNA sequences were compiled from twoindependent recombinants (bh NT3 and hNT12) by strategies simi-lar to those used previously to determine the canine intestine cDNAsequence (4). The canine sequence has been extended using datacompiled from two recombinants not previously reported (ciNT11Eand ciNT5B). The numbering of the rat sequence begins at the firstnucleotide of the Sau3A cloning site and oes not include interveningsequences which have not been completely determined. The bovine(midd le) and canine ( lower) cDNA sequences starting at positions20 0 and 207, respectively, are presented below the rat sequence. Thepositions of the three intervening sequences are indicated by arrowsand the onserved consensus splice donor and acceptor sequences areshown. The cap site (arrowhead) , TATA homology (bold underl ine ) ,and a sequence resembling a cAMP response element (bold linesabove and below) are indicated. The first and last nucleotides of thecDNA sequences are indicated by an asterisk and a p l w sign, respec-tively. The positions of two direct 21-base pair repeatsof two poly(A)addition signals are indicated y thin underlines. The predicted aminoacid sequence of the rat NT/N precursor protein is shown above therat DNA sequence. Differences between the rat andither cow or dogcDNA sequences are indicated with gaps necessary to obtain maxi-mum alignment indicated by dashes. Sequences were aligned usingDNAstar (Madison, WI) software. The lengths of the composite cowand dog cDNA sequences are 1166 and 759 nucleotides, respectively.strict ion fragme nts f rNT 19 harboring exons ere ident ifiedby Southern blot a nalysis and ubcloned for sequence analy-sis.Nucleotidesequences were determ ined by th e dideoxymethod using the s t rategy shown in Fig. 1B. Th e gene spansapproximately 10.2 kb of DN A and comp arison with heterol-ogous can ine andbovine cDN As has evealed the posi t ionsofthree in terven ing sequenceshich divide he mR NA equenceinto four xons. The sequenc es t the junct io ns etween exonsand in t rons co nform to c onsensus p li ce donor and accep torsequences (2 3). Exon 1 encode s the putat ive s ignal peptide;however, the rem ainin g divisions of the gene do not clearlydemarcate funct ional domains . Th e neuro tens in and neuro-

medin N coding domains are both ocated in tandem on exon4. Evolution of Exon Sequences-The predicted p recurso r pro-tein sequences from dog, cow, and rat are ompared in Fig. 3.Th e can ine an d bovine sequences are closely conserved dis-playing 95% a mino acid iden t i t ies . Th e rat sequence is mored ivergent asudged by comparison with eith er thedog or cowsequence with 77 and 78% am ino acid identities, respectively.However, most of the su bstitu tions between species are c on-servative (as defined by the Dayhoff PAM250 matrix (24)).At t he nucleotide level, the sequence con servation in thecoding region is roughly the same as that at the amino acidlevel. In addi t ion, there is substant ial homology in the un-translated regions. T he 5 ' and 3 ' untranslated regions of thecow and dog are closely relatedwith 92 and93% of thecom parable positions conserved, respectively. T he ra t geneuntranslated sequences ar e more divergent and substant ialgaps must be introdu ced to obtain o pt imal al ignm ent . How-ever, sequences t r ikingly s imilar to both theroximal portionof the 3 ' untranslated egion shared between the dog and cowcDN As a nd the ex tended por t ion con ta ined n ly in the cowcDN A are contiguous on exon 4 of the rat gene. Two consen-sus poly(A) addition signa ls are prese nt in both the rat genean d bovine cDNA sequences.Expression of the Rat NeurotensinlNeuromedin N Gene-To define the s tart point of t ranscript ion, primer extensionan d nuclease protection exper iments were performed usingRNA isolated from both gastrointest inal and neuronal t issues(Fig. 4). Hy bridization of a single-stranded probe extendingfrom exon 1 hrough to the ' end of the sequences containedin rNT19 wi th po ly(A)+ R NA from ei ther t i s sue source andsubseq uent digestion with mun g bean nuclease resulted in agroup of closely spaced protected fragm ents Fig. 44); he twomost prom inent bands map to posi t ions 152 an d 154 of therat NT /N gene. Hybridizat ion of a s ingle-stranded primer(see Fig. 4C) with RN A from ei ther t issue source and su bse-quent extension with reverse t ranscriptase resul ted n heformation of two prom inent extension products hich ma p toposi t ions 154 and 157 (Fig. 4B ). Thus, the cap s i te is ei therat or near posit ion 154. This a ssignme nt is su pported y theoccurrence of a "TATA" homology (p ositions 125-132) 29base pairs upstr eam f the assigned cap site. Also noteworthy,is a sequence beginning at posi t ion 99 which bea rs a s t r ikingresemblance to the cA M P response eleme nt previously de-fined or the at som atosta t in (25) andh u m anpreproen-kepha lin (26) genes. In addition, two 21-base pair imperfectdirect repeats are encountered upstream of the TAT A ho-mology beginn ing at posi t ions 27 an d 59.T o examine the t issue pecifici ty of ra t NT /N gene expres-sion, poly(A)+ RNA was prepared from various t issues andexam ined by RN A b lot analy sis (Fig. 5). Using a 32P-labeledsingle-stranded probe corre spon ding to theoding region an dpart of the 3 ' untranslated region of exon 4 (rNTB61, Fig.5C ), two prominen t bands of approximately 1.0 and 1.5 k bwere observed in a ll neu ral tissue s e xcept theerebellum (Fig.5A). In gastrointest inal issues, he 1.0-kb RN A s clearlyidentified bu t the 1.5-kb RN A is only fa intly visible (Fig. 5A,most evident in ane 7) suggesting tha t the rat io f these twomR NA s may vary in a tissue-specific ma nne r (see "Discus-sion").Comparison of the rat gene sequence with the bovine andcanine cDNA sequences suggestedh a t t h ewo mRN As mightarise as a resu lt of the utiliza tion of both of the poly(A)addition signals encounte red in exon 4 of the rat gene. T otes t this possibility, a 32P-labeled single-stran ded probe (3'UT -1, Fig. 5C) corresponding to the dis tal port ion of th e 3 '


4/6

4966 Rat Gene Encoding Neurotensin and Neuromedin NFIG. 3. Comparison of the pre-dictedeurotensinlneuromedin N lT

precursor sequence% from rat*'Ow*and dog. The complete precursor se-quences are depicted using the singleletter amino acid code. Amino acidswhich are identical in all three speciesare boxed.1 4 0 1 5 0 16 0

KIIVlKll~~YlLKROLYlYK?RR?YlLKKIIVIIUISYILKlOLYIYKSRRSYlLKK I I V l I l l I l ? T I LK RQ LY I NK ? l R P T I LK

1 2 3 4 5 6 G A T C

C50 bp

TATATATAEco R Ir-.r imer

1 S1 probe1 pro tec ted

FIG. 4. Determination of transcription initiation site.A, nu-clease protection. A "'P-labeled single-stranded probe was hybridizedto 10pg of rat hypothalamus (lanes and 4 ) or jejunum-ileum (lanes5 and 6 ) poly(A)+ RNA, or total yeast RNA (lane2) followed bydigestion with either 50 (lanes3 and 5 ) or 100 (lanes2, 4, and 6)units of mung bean nuclease. The products of the reaction wereanalyzed on a 6%polyacrylamide, 7 M urea gel. Two arrows denotethe major protection products. Undigested probe, lane 1; dideoxysequencing reactions using the template used to synthesize the probe(GATC). B , primer extension. "P-Labeled single-stranded primerwas hybridized to 10 pg of jejunum-ileum (lane ) ,cerebrum (lane ) ,or testis (lane5) poly(A)+ RNA, or total yeast RNA (lane2) andextended with reverse transcriptase. Reaction products were analyzedas described above. Unreacted primer, lane 1; dideoxy sequencingreactions (GATC). The extension products can be directly positionedon the gene sequence and are indicated by an arrow. C, he positionsof the probe used in A and he primer used in B are depictedschematically. 5' Flanking (thick line), 5' untranslated (open bar) ,coding region (diagonal lined bar), and vector ( thin l ine) sequencesare indicated. The extent of the probe, primer, protected fragment,and the extent to which the primer was extended (2%-zag line) arealso indicated. To make the template, an EcoRIIHinfI fragment ofrNT19 was subcloned into M13mpll.

untra nslate d region (as defined by t he bovine cDN A sequence)was used to probe the sam e RN A blot (Fig. 5B). his probeidentifies only one mR NA pecies corresp ond ing to the upp erband identifiedby th e coding egionprobe. This band spresent in al l gastrointest inal and neuronal t issues with theexception of th e cerebellum. As wit h the revious probe, hereis no detectable NT/N mR NA expressed in test is , l iver, orkidney. These results strongly support the contention thatt he tw o N T /N mRN A s a r e t he r e su l tf the utilizationof twodifferent poly(A) addition signals.DISCUSSION

W ehave previously characterized cDNA clones derivedfrom the canine intest inalmucosa which encode neurotensinand neuromedin N (4). y screen ing rat genomic and bovinehypothalamic cDNA librar ies w ith the caninerobe, we haveisolated the rat NT/N gene and bovine hypothalamic NT/N

1 2 3 4 5 6 7 8 91011A4- 1.5 kb4- 1.0

B cI 4- 1.5

C B R B c

10 0 bpH tAATAAA AATAAA1 rNTB61

1 3 'UT-1FIG. 5. RNA blo t ana lys is. Poly(A)+ RNAwas isolated fromvarious adult rat tissues, size-fractionated on a 1% ormaldehyde-agarose gel, transferred to a Zetabind filter (Cuno, Meriden, CT), andhybridized with either of two R2P-labeled ingle-stranded probes(panel C).Hybridizing bands were visualized byutoradiography withan intensifying screen for 70 h. 5 pg of poly(A)+ RNA from brainstem, lane ; cerebellum,lane 2; cerebrum, lane 3; hypothalamus, lane4; t o t a l brain, lane 5;duodenum-jejunum,lane 6; ejunum-ileum, lane

7; arge intestine, lane 8; estis, lane 9; kidney, lane IO ; and liver, lane11 . The blot was hybridized with a probe (rNTB61) spanningcodingregion and proximal 3' untranslated sequences (pane l A ) , subse-quently stripped, and re-hybridized with a probe (3 ' UT-1) spanningonly the distal 3' untranslated region (panel E ) . The probes aredepicted schematically in panel C. ntron ( thin l ine) ,oding (diagonallined bar), 3' untranslated (open bar) , and 3' flanking (thin ine)sequences are indicated. The positions of two consensus poly(A)addition signals and the extentf the antisense probes, rNTB61 and3' UT-1, are also indicated.cDNA clones. Nucleotide sequence analysis hasrevealed th atth e general features of th e predicted NT/N precu rsor are thesame in all three species. T he cow an d dog sequences areclosely conserved differing at only 9 of 170 amino acid resi-dues. The ra t sequence is missing one of the two tand emmethionine codons which begin the cow and dog sequencesand is more divergent. The near identi ty of the precursorpredicted from either canine intestin e or bovine hypothala-mus cDNA clones indica tes tha t both NT andeuromedin Nare processed from an equivalent precursor in the gut and thenervous system.T h e rat N T/ N gene coding egion is divided into four exonsby thre e ntro ns as revealed by comparison with he twoheterologous cDNA sequences. Although exon 1encodes onlythe puta tive signal peptide (amino acids 1-23) with the as-par tate codon spli t by intron 1 probably comprising part ofth e cleavage signal (27), he positions of the remaining twointron s do not divide the gene into obvious function al do-mains. The remaining exons ncode amino acids 24-44 exon2) , 45-119 (exon 3 ) , a n d 120-169 (exon 4). T h e N T a n d


5/6

Rat Gene Encoding Neuroten sin and Neurom edin N 4967neuromedin N coding domains are ocated in tandem near hecarboxyl terminus of the predicted precursor on exon 4. Thisfeature of the gene, in addition to the act that we have thusfar isolated only cDNAs which encode both N T and neuro-medin N, indicates that distinct mRNAs encoding only NTor neuromedin N cannot be generated by alternative splicingmechanisms.Comparison of the complete NT/N precursor protein se-quence from dog, cow, and rathas revealed a tight evolution-ary conservation. Of 169 comparable positions, 76% are iden-tical in all three species and most of the substitutions areconservative. Based on our data and previously reported se-quences for bovine neurotensin (28) and porcine neuromedinN (3), he sequences of these two peptides appears to beinvariant among mammals. The high degree of conservationof the entire precursor could reflect structural requirementsfor appropriate precursor processing to yield the biologicallyactive peptides. Alternatively, the precursor might be proc-essed to yield biologically active peptides other than neuro-tensin and neuromedin N which are subject to high selectiveconstraints. The precursor sequence is much more highlyconserved between dog and cow (95%) than etween either ofthese species and rat (77 and 78%, respectively). This couldbe due to the higher rate of nucleotide substitution which hasbeen postulated for rodents (29, 30).In addition to the close conservation of the coding regionbetween the three species, there is also substantial similarityin the untranslatedegions (see Fig. 2) . Comparison of the 3'untranslated regions of the canine and bovine sequences hasrevealed that, inaddition to a common proximal region span-ning approximately 200 nucleotides which is >90% identical,the bovine sequence extends approximately 400 nucleotidesfurther where a second consensus poly(A) addition signal isencountered. This extended 3' untranslated region accountsnicely for the difference in the sizes of canine intestine andbovine hypothalamus NT/N mRNAs we have observed pre-viously (4). The composite bovine cDNA sequence is 1166nucleotides in length and,llowing for a 150-200-base poly(A)tail, t he mRNA size of 1316-1366 is similar to our previousestimate of 1.5 kb based on RNA blot analysis. Comparisonof the bovine cDNA with the rat gene sequence reveals se-quences strikingly similar to both the proximal part of the 3'untranslated region (as defined by dog and cow cDNAs) andthe distal part (contained only in the bovine cDNA). Thesesequences are contiguous on exon 4 of the rat gene.

We have examined the tissue specificity of rat NT/N geneexpression by RNA blot analysis using probes derived fromexon sequences of the rat gene. Using a probe containing thecoding and proximal part of the 3' untranslated region ofexon 4, two poly(A)+ RNAs were detected in all gastrointes-tinal and neuronal tissues examined with the exception ofcerebellum (Fig. 5A). Quantitation of the two bands by scan-ning densitometry (data not hown) revealed that the atio ofthe level of the 1.5-kb to the1.0-kb species ranges from 1.1 inthe cerebral cortex to 0.8 in hypothalamus and brain stem.Thisratio sdrasticallyaltered ngastrointestinal issueswhere the 1.0-kb mRNA is at least 10 times more abundantthan the 1.5-kb mRNA. The sizes of these two mRNAs (1.0and 1.5 kb) correspond roughly to the sizes of the canineintestine (1.0 kb; previously reported smaller RNA specieshave not been observed in subsequent experiments)' andbovine hypothalamus (1.5 kb) NT/N mRNAs (4). Reprobingthe same blot with a probe corresponding to only the distalportion of the 3' untranslated region (Fig. 5B) resulted in theidentification of only the 1.5-kb mRNA species, strongly

* D. L. Barber and P. R. Dobner, unpublished observations.

indicating that thedifference between the two mRNAs is theextent of their 3' untranslated regions. Thus, in the rat,bothof the mRNA types defined by heterologous cDNAs are pro-duced apparently by the utilization of different poly(A) addi-tion signals. Furthermore, the proximal site is highly favoredin intestinal issues. Whether ornot the inclusionor exclusionof the distal 3' untranslated region has any functional con-sequence remains to be determined.

The relative levels of NT/N mRNA (quantitated by scan-ning several different exposures of the blot shown in Fig. 5Aand summing the values obtained for both the 1.0- and 1.5-kb bands, data not shown) are roughly equivalent to herelative levels of NT in these tissues as determined by radio-immunoassay (7). For instance, the level of NT/N mRNA inthe hypothalamus is 5.7 times greater than in brain stemwhich is similar to he 4.7-fold difference in NT contentbetween these two tissues. However, although the NT ontentof the hypothalamus is 30 times that of the cortex, the NT/N mRNA evels in cortex are only &fold lower than nhypothalamus. This striking difference opens the possibilitythat the NT/N protein precursor is differentially processedin cortex and hypothalamus resulting in the disparity betweenthe levels of immunologically detected NT and the relativeNT/N mRNA levels.The best example of such tissue-specificprocessing is the differential processing of the proopiomelan-ocorticotropin precursor in the anteriorand ntermediatelobes of the pituitary (31).We have mapped the rat NT/N gene cap site to position154 by both primer extension and nuclease protection exper-iments using RNA isolated from both intestine andbrain. Asis the case for many eukaryotic genes transcribed by RNApolymerase I1 (32), a TATA homology is encountered starting29 base pairs upstreamof the cap site. In addition, a sequencematching the previously defined cAMP response element (25,26) at seven of eight positions is located between positions103 and 110 of the rat gene and is immediately preceded byanother half-repeat of this palindromic sequence (99-102).This cAMP response element could be involved in he controlof NT levels in the rat PC12 cell line by combinations ofnerve growth factor, dexamethasone, and activators of ade-nylate cyclase (13). The isolation of the rat NT/N gene willenable investigations of the molecular mechanisms underlyingthis complex synergistic control.

Acknowledgments-We thank Dr. Susan E. Leeman orhelpfuldiscussions and critical evaluation of the manuscript. We also thankDrs. Andrea J. Pereira and Louis De Gennaro for critical commentson the manuscript.REFERENCES

1. Carraway, R., and Leeman, S. E. (1975) in Peptides: Chemistry,Structure, and Biology (Walter, R., and Meienhofer,J. , eds) pp.679-685, Science Publishers, Ann Arbor,MI2.Araki, K., Tachibana, S., Uchiyama, M., Nakajima, T., andYasuhara, T. 1973) Chem.Pharm. Bull. (T o k yo )21, 801-28043. Minamino, N.,Kangawa, K., Matsuo, H. 1984) Biochem. Bio-phys. Res. Commun. 122, 42-5494. Dobner, P. R., arber, D. L., Villa-Komaroff, L., and McKiernan,C. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 3516-35205. Carraway, R., and Leeman, S. E. (1973) J. Biol. Chem. 248,

6. Leeman, S. E., Aronin, N., and Ferris, C. (1982) Recent Prog.7.Carraway, R., and Leeman, S. E. (1976) J. Biol. Chem. 261,8. Peterson, B. J. , Christiansen, J. , Rokaeus, A,, and Rosell, S.9. Sundler,F., Hakanson, R., Hammer,R.A., Alumets,J ., Carraway,

6854-6861Horm. Res. 38.93-1327045-7052(1984) Scand. J. Gastroenterol. 19. 8-51


6/6

4968 Rat Gene Encoding Neuroten sin and Neurom edinR., Leeman, S. E., and Zimmerman, E.A. (1977) Cell Tissue 21. Dobner, P. R., Kawasaki, E. S., Yu, L.-Y., and Bancroft, F. C.Res. 178,313-321 (1981) Proc. Natl. Acad. Sci. U . S.A . 78,2230-223410. Goedert, M., Reynolds, G. P., and Emson, P. C. (1983)Neurosci. 22. Dobner, P. R., Kislauskis, E., Wentworth, B. M., and Villa-

11. Terenghi, G., Polak, J. M., Varndell, I. M., Lee, Y. c., Wharton, 23. Padgett, R. A., Grabowski,P. J., Konarska, M. M., Seiler, S., andJ., and Bloom, S. R. (1983)Endocrinology 11 2 , 226-233 Sharp, P. A. (1986)Annu. Reu. Biochem. 55, 1119-115012. Tischler, A. S., Lee, Y. C., Slayton, V. W., and Bloom, S . R. 24- DaYhoff, M. (1978)AtfnsO f Protein sequence and Structure, VOl-(1982)Regul. Pep t. 3, 415-421 5, Suppl. 3, National Biomedical Research Foundation, Silver13. Tischler,A. S.,Lee, Y.C., Costopoulos, D., Slayton, V. W., Jason, Spring, MDW. J., and Bloom, S. R. (1986) J.Neurosci. 6, 719-1725 25. Montminy, M. R., Sevarino, K. A., Wagner, J. A. , Mandel, G.,14. Levi-Montalcini, R. (1987)Science 237, 1154-1162 and Goodman,R. H. 1986)Proc. Natl. Acad. Sci. U. S. A . 83,6682-6686

man, H. M. (1986) Nature 323, 353-356

Let t . 35, 155-160 Komaroff, L. (1987) Nucleic Acidses . 16 , 199-218

15. Greene, L. A.9 and Tischler, A. S. (1976)ProC. Nat L SCi. u. 26. Comb,M., i ~ ~ b ~ ~ ~ ,. c., seasholtz, A., ~ ~ ~ b ~ r t ,., and ~ ~ ~ dS. A . 73,2424-242816. Maniatis, T., ritsch, E. F., and Sambrook, J. (1982)Molecular 27. von ~ ~ i j ~ ~ ,. (1983) J, i o ~ h e ~ .33 ,17-21

CloniW "A Laboratory SpringHarbor 28 . Carraway, R., and Leeman, S. E. (1975) J. Biol. Chem. 250,Cold Spring Harbor, NY17. Messing, J. (1983) Methods Enzymol. 10.20-7818. Sanger, F., Nicklen, s.,and Coulson, A. R. (1977) P roc. N d . 30. WU,c.-I.,nd Li, w.-H. (1985) Proc. Natl. A C ~ .ei. u. S . A .19. Chirgwin, J. M., Przybyla, A. E. , MacDonald, R. J., and Rutter, 31. Douglass, J., Civelli, O ., and Herbert, E. 1984) Annu. Reu.20. Aviv, H., and Leder, P. (1972)Proc. Natl. Acad. Sci.U. S. A. 69, 32. Breathnach, R., and Chambon, P. (1981) Annu. Reu. Biochem.

1907-191129. Britt en, R. J. (1986) Science 231,1393-1398Acad.ci . U. S. A . 74,5463-5467 82,1741-1745W. J. (1979)Biochemistry 18, 5294-5299 Biochem. 53, 665-7151408-1412 50,349-383

Edward Kislauskis et al- The Rat Gene Encoding Neurotensin and Neuromedin N

Documents