Molecular cloning and carboxyl-propeptide analysis of human type III procollagen

Volume 12 Number 24 1984 Nucleic Ac ids Research

MoleculaT (Honing and carboxyl-propeptide analysis of human type III procollagen

Helen R.Loidl1, Jane M.Brinker1, Mary May2, Taina Pihlajaniemi3, Scott Morrow2, JoelRosenbloom2 and Jeanne C.Myers1-"

•Connective Tissue Research Institute and 'School of Dental Medicine, University of Pennsylvania,Philadelphia, PA 19104, JDept. Biochemistry, Rutgers Medical School, Piscataway, NJ 08854, and4Dept. Medicine and Human Genetics, University of Pennsylvania, Philadelphia, PA 19104, USA

Received 10 September 1984; Accepted 14 November 1984

ABSTRACT

Two human cDNA l i b r a r i e s prepared from normal f i b r o b l a s t (GM3348) andrhabdomyosarcoma (CCL136) mRNAs were screened under cross h y b r i d i z a t i o ncond i t i ons w i th a genomic fragment coding f o r exons 2 and 3 o f the aviantype I I I procol lagen C00H-propept1de (Yamada, Y . , Mudryj , M., S u l l i v a n , M.and deCrombrugghe, B. (1983) J . B1o l . Chem. 258, 2758-2761). One cDNAclone con ta in ing a 1.12 kb I n s e r t was I s o l a t e d from the CCL136 l i b r a r y andl a t e r used to I d e n t i f y a GM3348 der ived clone w i t h a 2.4 kb I n s e r t .Comparison o f the human and avian type I I I C-term1nal propept ides revealedmuch more divergence 1n the f i r s t 54 amino a d d s f o l l ow ing the terminalcys te ine o f the t r i p l e h e l i c a l region than 1s present 1n the a l ( i ) anda 2 ( I ) proco l lagen chains of these spec ies. Ana lys is of poly (A+) RNA fromnormal human f i b r o b l a s t and tumor c e l l l i n e s showed tha t they d i f f e r e dg r e a t l y 1n the r e l a t i v e amounts of a l ( l ) , a 2 ( l ) , and a l ( l l l ) mRNAs.Furthermore, as we p rev ious ly repor ted f o r the a l ( i ) and a 2(1)t ranscr ip ts , mul t ip le mRNAs also hybridize to the cloned a l ( I I I ) DNA.

INTRODUCTION

Five major and at least f i ve minor types of collagen comprise the most

abundant protein family found 1n vertebrates (reviews 1,2; and ref. 3-7).

In general, collagen 1s synthesized in the form of procollagen molecules

composed of three polypeptide chains consist ing of a central t r i p l e hel ical

domain with the repeating Gly-X-Y sequence, and shorter carboxy and amino

propeptides. The d i f fe rent chains are the products of probably more than

f i f teen genes which have recently begun to be Isolated from several species

(reviews 8,9; and ref . 10-17). Determination of the DNA sequences of the

genes and thei r encoded amino add sequences are fundamental to under-

standing the function of the various types of collagen and the factors

cont ro l l ing the i r synthesis. Previously, human cDNA clones coding for the

al and a2 chains of the type I procollagen heterotrimer (pro al >2 (pro a2)

were Ident i f ied (18,19) and used to Isolate overlapping genomic fragments

(20,21). Analysis of these two genes, which are located on d i f fe ren t

chromosomes (22), showed that the coding region 1s frequently Interrupted

© IRL Press Limited, Oxford, England. 9383

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

by about 50 sometimes very la rge Int rons (20,21) . Although the mRNA

t ransc r i p t s of these genes are normally found at a two to one r a t i o

r e f l e c t i n g the type I collagen subunit composit ion, the leve ls can vary 1n

disease states (23) .

To study the coordinate and d i f f e r e n t i a l expression of several

procol l agens, we constructed two cDNA l i b r a r i e s contain ing sequences

corresponding to d i f f e r e n t types of procollagen 1n add i t ion to type I .

Here we report the I s o l a t i o n and charac te r i za t ion of cDNA clones coding for

about ha l f of the human a l ( m ) procollagen cha in . From the DNA sequence

of these c lones, part of the t r i p l e he l i ca l and 3' untranslated region and

the en t i re C-term1nal propeptide was der ived. While the te lopept lde and

adjacent 51 par t of the C00H-propept1de of avian (24,25) and human type I I I

are more divergent than seen 1n the ccl(I) and a2 ( I ) chains of these

species (26-28) , the major par t of the C00H-propept1de 1s conserved.

Comparison of t h i s domain 1n three human procollagen chains also shows

s i g n i f i c a n t homology In the regions where the cysteine residues and

g lycos lya t ion attachment s i te are loca ted . Together with e luc ida t ing the

s t i l l undetermined ro le o f type I I I , these probes w i l l be valuable 1n

def in ing mutations and regulatory abnormal i t ies 1n genetic and acquired

diseases of connective t i s sue .

MATERIALS AND METHODS

Enzymes and Isotopes

Res t r i c t i on endonucleases were purchased from New England Bio labs, DNA

polymerase I and proteinase K from Boehringer Mannheim, T4 DNA Ugase and

rabb i t r e t i cu locy te lysate from Bethesda Research Laborator ies. A l l

enzymes were used according to manufacturers suggestions. Formamide was

obtained from Fluka, o l i go dT ce l l u l ose and primers for DNA sequencing from

Col laborat ive Research, and n i t r o c e l l u l o s e f i l t e r s from Schleicher S

Schuei l . Labeled Isotopes were purchased from Amersham. Cell l i n e 013348

was obtained from the Human Genetic Mutant Cell Repository, Camden, New

Jersey, and CCL136 and HT1080 from the American Type Culture C o l l e c t i o n ,

Rockv i l le , Maryland.

RNA P u r i f i c a t i o n and cONA Cloning

Poly (A+) RNA was Iso lated by o l i go dT ce l lu lose chromatography and

size f rac t iona ted on a sucrose-SDS gradient (19). F i r s t strand DNA

synthesis was performed 1n a 150 ul react ion contain ing 5ug of mRNA, 0.7ug

o l i go dT 12-18 primer, 1.30mH dGTP, .53CTM TTP, .32 mM dCTP, .27mM dATP and

55 uni ts /ml of AMV reverse t ranscr ip tase p u r i f i e d by DEAE and CM sepharose

9384

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

chromatography (29). Second strand DNA was synthesized with the Klenow

fragment of DNA polymerase I after heating the reaction mix at 100° for 2

minutes (30). Following Incubation for 1 hr. at 38*, the reaction was

passed through a G50 sephadex column and the peak fractions were ethanol

precipitated. The double-stranded DNA (lOOng) was treated with 3500

units/ml of Sj nuclease (Miles) at 25* for 30 minutes, and then size

fractionated on a 5-20% neutral sucrose-SDS gradient. The pool containing

the largest DNA was dC-ta1led for 6 minutes at 11° with terminal

transferase. 3ng of DNA was annealed with 5ng of pBR322 which had been

cleaved with PstI and dG-ta1led (New England Nuclear). The hybrid

molecules were used to transform Escher1ch1a col l strain MC1061 (31).

Screening of the Bacterial Colonies

The clones were replica plated onto nitrocellulose f i l t e r s , grown for

four hours and then the plasmid DNA was amplified by overnight Incubation

of the f i l t e r s at 37* on Luria plates containing 12.5ug/ml tetracycline and

150 ug/ml chloremphenicol. The f i l t e r s were processed (30), baked at 76*

for 2 hours, and hybridized under low stringency (35% formaraide, 6xSSPE,

34* for 40 hours) or high stringency (50% forniamide, 5xSSPE, 40* for 20

hours). Final washing conditions were lxSSC, 55° or .2xSSC, 65°

respectively. The DNA probe used for Isolating the 1n1tal human type I I I

clone was lOng/ml of the Insert from a subcloned genomic fragment

containing exons 2 and 3 coding for part of the C00H-propept1de of avian

type I I I procollagen (24,25) provided by Y. Yamada. The human type I I I

cDNA clone, E6, was Isolated using 2ng/ml of the 5' PstI restr ict ion

fragment of RJ5 which was Isolated by electroel ution from an agarose gel.

Inserts and plasmids were nick translated with 32p nucleotides to a

specific act iv i ty of 4-8xlO8 cptrt/ug.

DNA Sequencing

Plasmid DNA was restricted with PstI , or PstI and EcoRI and the

fragments were Isolated by electroel ution after electrophoresis 1n 1%

agarose gels. The DNA was extracted with phenol /chloroform and then cloned

Into the phage M13mp8,9 system (32). Restriction fragments derived from

the cDNA Inserts, now contained 1n the recombinant phage, were sequenced by

the Sanger dideoxy method using a universal primer of 17 nucleotides.

Northern Blot Hybridization

Electrophoresis of poly (A+) RNA 1n .7% agarose, 2M formaldehyde gels,

transfer to nitrocellulose and hybridization conditions were essentially as

described (19,33).

9385

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

. . ( IH)Uon jTI, COOH-PROPEPTIDE .3 '

5' i_

Figure 1 . Restr ic t ion Map of Human Type I I I cOHA Clones.The regions of the pro al ( I I I ) chain encoded by the cDNA clones are

shown on the top of the f igure. RJ5 terminates 5' at residue 892 of thet r i p l e hel ical region (35) and by comparison E6 extends to about 592. Onthe 3' side, RJ5 1s missing the last 30 amino adds of the COOH propeptidewhich have been derived from E6. The human type I I I cDNA clones, RJ5 andE6, were Independently analyzed for the presence of endonucleaseres t r i c t ion s i tes . Al l sites localized 1n RJ5 by res t r i c t i on mapping werealso found at the same position 1n E6: EcoRI, Ps t I , Aval/Xhol, Aval, andPvuII. Other enzymes that were determined negative with both clones wereBamHI, Xbal, H i n d l l l , EcoRV, Smal, C la l , Sa i l , and B g l l l . The 51 1100 bpand 3' 930 bp PstI fragments of E6 were separately hybridized to Southernblots of RJ5 cleaved with the di f ferent res t r i c t i on enzymes in order tocompletely ver i fy the Ident i f icat ion and regions of overlap. Both clonesare Inserted Into the PstI site of pBR322 1n the u or ienta t ion. The bottompart of the f igure shows the strategy for the DNA sequencing of RJ5 and E6.Both strands of the four fragments of RJ5, 51 Ps t I , middle Ps t I , and 3'Pstl-EcoRI and EcoRI-PstI were completely sequenced. The las t 90nucleotides coding for the C00H-propept1de and 78 nucleotides of the 3'untranslated region were determined from the 3' EcoRI-PstI fragment of E6.

RESULTS

Isolat ion of cONA Clones

Two sources of mRNA were investigated for the cloning of human type I I I

procollagen sequences. One was a normal f ibrob last cel l l i ne , GM3348, and

the other a human rhabdomyosarcoma cel l l i n e , CCL136, that was reported to

synthesize type I I I , and to a lesser extent types IV and V (34). Poly (A+)

RNA was size fract ionated and the 28-35S RNA analyzed by ce l l free

translat ion for the presence of type I I I procollagen. The gradient

f ract ion giving the highest y ie ld of a protein that was sensitive to

bacterial collagenase and which was s l i gh t l y larger 1n size than an a l ( I )

standard by polyacrylamide gel electrophoresis was used as template for AMV

reverse transcMptase. Double stranded cDNA was prepared and Inserted Into

the PstI s i te of the plasmid PBR322. Transformation of the recombinant

molecules was carried out using E. co l l s t ra in MC1061 (31) which gave an

eff ic iency of 0.5- lx lO8 colonies per ug of supercoiled pBR322. The I n i t i a l

screening of approximately 8000 colonies under cross hybridization

9386

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

I '<L

al(m) al(I) a2(I)A B C

Figure 2. Hybridization of Human Procoilagen cDNA Clones to PolyRNAs.

Four tenths microgram of poly (A+) RNA Isolated from normal fibroblasts(GM3348) lane 1, a rhabdomyosarcoma (CCL136) lane 2,and a fibrosarcoma(HT1080) lane 3 was electrophoresed 1n a .1% agarose 2M formaldehyde gel at30V for 20 hours (33). The RNA was transferred to a nitrocellulose f i l t e rat 4° for 16 hours, baked at 76* for two hours and hybridized to 5ng/ml of32P labelled DNA probe (specific act iv i ty 4-8xlO8 cpm/ug). Autoradiographyof the type I I I f i l t e r was about f ive times longer than for a l ( I ) ando2( I ) .

Panel A Hybridization with the a l ( I I I ) cDNA clone, RJ5. The sameresults were obtained with the clone E6.

Panel B Hybridization with the ccl(i) cDNA clone, a 12, (42)containing a 2.4 kb insert extending about 150 nucleotides 5'and 450 nucleotides 31 to Hf404 (28).

Panel C Hybridization with the a2( i) cDNA clone, Hf32, containing a2.2 kb Insert (19).

conditions with avian type I I I DNA sequences (24,25) yielded eight positive

signals from the GM3348 and one from the CCL136 l ibrar ies. Since two of

these GM3348 clones were found to be coding for a l ( I ) and a2( I ) , the one

CCL136 positive clone was next characterized. This cDNA clone, RJ5,

contained a 1.12 kb insert that showed a different restr ict ion pattern

(F1g. 1) than that characteristic of the human a l ( I ) (18,28) or a2(I)

(19,27) cDNA clones. When RJ5 was hybridized to Northern blots of S43348

and CCL136 poly (A+) RNAs, two mRNAs of about equal Intensity and different

from a l ( I ) and a2(I) were observed in each lane (Fig. 2).

In a subsequent screening, the 5' 216 bp PstI fragment of RJ5 was

9387

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

( U i u ACT i t c CCT H T CCC CCT K I CCT K l CCC CCT CCT K l ( 1 1 TCC C U CCT CCT CCI CCT CCT C U CCC C M CCI CCT CAC AAACLY CLI SER CL1 PRO U t O.Y PRO ALA CLY AlA PIO CLT PRO ALA 6LT SCR ARC CLT ALA PRO CLY PRO CLR CLT PRO ARC CLY ASP LTS

«J »51U I CAA ACA U T U A CCT CCA CCT CCT U C ATC AA> U A CAT CCA WA TTC CCT H T AAT CCA CCT CCC CCA CtT TCT C O CCC CCT CCTCLT CLU THA GLT CLI ARG CLY ALA ALA CLT ILE LTS CLT HIS ARC CLT PHI PRO GLT AS! PRO CLT ALA PRO CLT SCR PRO CLT PRO ALA

PRO

til MlCCT CAC CAG CGT CCA ATC CCC ACT CCA CCA CCT CCA GCC CCC ACA CCA CCT CTT CCA CCC ACT CCA CCT CCT CCC AAA CAT CCA ACC AGTCtT U I C L I CL< ALA H E U.1 SCR PRO CLT PRO ALA t L I PRO ARC CLT PKO »AL CLT PRO SCR 111 PRO H O CLT LTS ASP U T THR SCR

ALA THR

M; ionCCA CAT CCA CCT CCC ATT CCA CCA CCA GCC CCT CCA CCT AAC AGA CCT CAA ACA C W TCT CAC CCC TCC CCA CCC CAC CCA CCC CAA CCACLT HIS PRO CLT PRO ILE CLT PRO PflO CLY PKO AJIG CLT AS! ARC CLT CLU ARC CLT SCR CLU CLT SER PRO CLY HIS PRO CLT CLH PRO

SCR

1 0 1 * 1013GCC CCT CCT CCA CCT CCT CCT CCC CCT U T CCT TCCGLT PRO PRO CLT PRO PRO CLY ALA PRO CLT PRO CTS

•

Figure 3. The DMA Sequence Encoding 132 Amino Adds of the Triple HelicalRegion of Human Type I I I Procollagen.

The protein sequence obtained from the DNA sequence of the type I I Iclones, RJ5 and E6, was compared with the published human type I I I proteinsequence derived by Edman degradation (35). Only four differences werefound and they are shown on the bottom l ine . These occur at residues 924,980, 981 and 983.

puri f ied and hybridized under stringent conditions to the same set of

f i l t e r s and to about 4000 new colonies from a later transformation. Eight

positive clones were Identif ied and the largest clone, and the one

extending most 51 , E6, was later characterized by restr ict ion mapping (Fig.

1) and Southern blot hybridization. E6 contains a 2.4 kb Insert that

overlaps RJ5 about 400 bases 3' and 900 bases 5 ' , and therefore entirely

encompases the original Isolate. The Inserts of the other type I I I clones

terminate 3' to the end of E6 as does one other positive from the original

eight GM3348 clones and these were not examined further.

Comparison to Human Amino Add Sequence Determined by Protein Sequencing

The 5' 216 bp PstI fragment of RJ5 contained entirely Gly-X-Y sequences

as did 180 nucleotides of the 5' region of the adjacent 382 bp PstI

fragment. Immediately at the 3' end of the 180 nucleotides were found two

adjacent codons, TGC and TGT, that designate the cysteine residues

responsible for formation of the disulf ide bonds of type I I I collagen.

According to the published amino add sequences of type I I I , the t r i p le

helical region contains 1023 residues terminating at the 3' end with the

f i r s t of the double cysteines (35). Therefore, the most 5' amino add

encoded by RJ5 1s residue 892 (F1g. 3) and by comparison E6 extends to

about 592.

Only a few differences were apparent when we compared the type I I I

sequences derived from the clones with the reported human amino acid

sequence determined by Edman degradation (34): at 924 a threonine versus

9388

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

CC I I I C CTC CS A I CCA Tt CSC 1*1 tCH LAA C C AA AICI C*T CST t i l CCA U C K I CCC »TI CCI O l MT U i H I W U l d T d «ST TT7 CCC C t l TAJ-TAT CU UI Ui CU ATC CATCIS (Lr CLT I U CLT ALA ALA ALA III I I I 6LI I U t i l UT CII ITJ » U t i l CIT H I ALA MO [11 ,111 U l . t t f CLI PI© HEI AW• U.1 TU U l LEB UT A A A B.T • » TU UT TTI B.I • £ • AMI I I I UB

U K C CTC C A C r i s A * T C U C « T M A C A A A GITC AAA AIC AAC ACC tAT | « ATI ATC ACT ICA CTC A M TCT CCI AAI U A t U ATA |AA A K CTC ATI A l l CCT CAT U T UT t i l AJUMI US III U l THI UP U l HE «T TKI U l LEI I U HI >U U l 111 U l !U U l HI IEI HE HI Pt© UP UT HI All IUAM CLI Ltl U l IEI «T HE AH IU ll£ l£i

il n

U C LCC CCT A*A AAC TCC MA tAC CT« AAA TTC T « CAT CCI I U CTt AJU U T CtA tAA TAC I t t S l T t « CCI AAC CAA u i I K AAAA H Fttt ALA A l l AW C I I A l t ISP L [ | ITJ PHt CTS HIS m t L I LEI LTI U l U l U f l IT1 H M U U T W A U U l ttt CTJ I U

* • HE +

* A I » I A T ( C C J C j £ c C C A A116 *AT CCI AIC AAI 6IA U C U T AAT *T* CAA ACI U t 1AA ACA I K AIA ACT CCC AAT CCI TT1 JUT %T1 CCA C M AAA O U TCfi T ULEI AIP « .A lit LTJ I.I HI CTJ U l V I U l TW1 U.T U l H I CTI HE S I U * U l Ml U l AM »U « • UC LTI IIJ TIP TlJ" • ' TTI * * Ltl » u TW AM

111 IMT A C C A A A C t A I A I A 6 CC I A C CC

A U CAT ICT ACT CCT U C AAC AAA CAC GIT TCC TTT C U U C IXC ATC CAT GCT CCI I I I CAC I I I I U I U MC M l CCT fcU CTI CCT

THI CLI U l U.T LTS HI UP Alt

Kl INC I C A T C C A C IC CC «C

(UA CAT U C CTT U T CTC CAC CTC CCA TTC CIT C M CTT CTC TCC * « CCA CCT I U CAC AAC ATC ACA TAT CAC I « AAA AAT A K ATTt i e U P m LEI ASP m u i IEI ALA P«E LEO U C LEI LEI U I M I U I ALA H I U I m i n m TTI HII e n LTI ASI H I ILE

SEI CLI ILE •

in noC A A C C T A A A C I O T I A C I A C

OCA TAC ATI SA1 CAI « C AIT «** AAT «TA AAG AAG GCC CTG AAG CTG ATC ttt Tt* U T I U U I O U TIC AAt K I tAA I U AAT A U

U k U l T U T i l U P HE

111 1 «A T C T G T * C T T I C T I C A T A A A

AAA TTC ACI TAC ACA t ! T CTC U C CAT GCT TIC ACS AAA CAC ACT CCC CAA TCC ACC AAA ACA CTC ITT CAA TAT CCA ACA CCC A U tCTU S ME T O TTI m U L LEI U . I U P CLT CTS THI ITS I I S T i l Ctl U l TIP U l LTS TO T U P>E UQ TTI A1G THI A1G LTS U l

TTI Hit ALA * U T T U

t*[ tnA » T * * T A T A T C t ( C C

I » L A H L E I P I O I U T U I V ILE ALA P H T T I A S * ILE U T U T P t O U P t i l U S P * U T * U W T U U . I P H T U C T I P I C I E !PUT T U ILE *

I ( A T CC K T C T I T I C ACC CC K AAC t A T C CIC CAA I t TtC TCC I I C I t i C C I IIAA ACC A U C T t TAT C T I AAA TCC C U CAA A U AAA TTT AAC TCC ATA I * f t I T CCI CTT 6 1 1 CIA ATC T T I ICA ACC

Figure 4. Comparison of the Human and Avian a l ( I I I ) Telo- and C00H-P ropeptide Regi ons~ ~~

The middle two lines are the nucieotides and corresponding amino addsencoded by the human type I I I clones. The nucleotide differences found 1nthe avian are shown on the top Hne, and where these designate a differentamino add 1s shown on the bottom l ine. Amino add 1 1s the most 3' of thedouble cysteines (the 5' cysteine 1s residue 1023 of the t r i p le helicalregion shown 1n figure 2). The symbols Indicate the deletion A 1n the

avian sequence of three amino adds, Insertion v of one amino a d d , thecysteine residues*, and the cleavage site + between the telo- andC00H-propept1de, and the glycosylation attachment site which is underlined.The last 78 nucleotides are the beginning of the 3' untranslated region.

proline, at 980 a threonine versus alanine, at 981 a serine vs. threonine,

and at 983 a h1st1d1ne vs. serine. The remaining 128 amino adds were

Identical.

Analysis of the Telo- and Carboxy- Propeptides of a l ( I I I ) , a2(I ) and a l ( I )

Alignment of the human and chick type I I I (24,25) sequences beyond the

collagenous region revealed considerable dissimi lar i ty 1n the f i r s t 54

amino adds (Fig. 4) . In the human, three amino adds (Gly I le Gly)

corresponding to residues 11-13 were Inserted and one amino add (Glu)

between the two tryosines at positions 23 and 24 was deleted. Within the

telopeptide (cleavage Is presumed to occur between residues 25 and 26), 12

9389

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

a l ( l l l ) PIO ASP « l l SO AtC ITS Ai« PI0 ALA AM * U CIS UE AIT LCI ITS PHI CIS HIS PIO CLU LEU LIS SCI BLI SLUa l l I) - CLU IK« - - - - U ( LEU SE« - - - IIP S£l - - T i la l l l l - CLU - Tut THI KIT - - SCI ASP TIP -

* *

tl IMa l l I I I ) T l * TIP »AL ASP P«0 AS! U.1 S l l CIS LIS LCU U P ALA 1LC LTS »Al PDC CIS AS! RET U.U Till 6LI U.% TNI CTSa l ( l ) - - III TKI MT CIO - - - - T i l - ASP PHI P«0all l l - lie »si

* • *

B i« 1 M

AlA H I Oi l A i l I H T i l T i l HIS CIS LTS ASK S t l ILC ALA T i l HIT ASPa l l I I I ) K T TCC CM T J O n t l TAT CAC TCC AAA AAT ACC ATT ICA TAC «T0 SATa i l I) --c --T c —c c --ca i m --c -c --c s --c -— 8-e --e c

«AL•

Figure 5. Conserved Regions 1n the COOH Propeptide of Human o t l ( I I I ) ,a2(I) and a_KI).

Figure 5A shows residues 55-106 of the COOH propeptide of type I I Iprocoiiagen (top l ine) and the corresponding region of a2(I) (27) anda l ( i ) (28). Dashes Indicate no difference from the type I I I amino addsequence. Asterisks designate the cysteines found 1n c t l ( I I I ) .

Figure 5B shows residues 168-184 of the C00H-propept1de Including thegiycosylation attachment site (underlined). The two top lines show thetype I I I amino add and nucleotide sequence. The two bottom lines are thenucleotide sequences of the a2( I ) (27) and a l ( I ) (28) cDNA clones.Dashes Indicate no change from the type I I I DMA sequence. Only one change1n the amino acid sequence 1s noted: Val at position 180 1n the ocl(I).

differences were observed between chick and human and one at the

propeptidase cleavage site Involved a charge change, Arg to Gly. In the

region containing residues 26 to 54, 13 additional changes occurred

Including three charge differences, the most notable being a Lys 1n the

human at residue 32 and a Glu 1n the chick. Therefore, a total of 25

changes 1n 54 amino acids (not counting the deletion and Insertion) have

been inst i tu ted. In the corresponding region of the chick (26) and human

a l ( I ) (28) there are only two changes, and eight are present 1n the a2(I)

not Including the four amino acid insertion in the human (26,27). The fact

that the human and avian type I I I genes have diverged more than a l ( I ) or

a2(l) suggests that the lat ter two have evolved more recently or that

changes in this region of ctl ( I I I ) are tolerated.

However, when residues 55-106 were compared, very strong homology

between the type I I I chick and human protein sequence was Immediately

evident (Fig. 4) . Only three amino adds have diverged and none of these

involved a difference in charge. Since 1t seemed Important that the

protein sequence in this area be conserved, we Investigated the comparable

region in the human a l ( I ) (28) and a2(I) (27). As can be seen 1n F1g.

5A, this segment of the three different procollagen COOH-propeptides also

9390

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

shows str iking homology with each other. Interestingly, and perhaps most

signif icant 1s the presence in this domain of the f i r s t f ive of the eight

(seven in a2) cysteine residues 1n a K H I ) , a2( I ) , and a l ( I ) . These

amino adds are Involved 1n both Inter- and 1ntracha1n disulf ide bonds and

thus the assembly of the procollagen molecules (9,36-38). Because of these

functions there 1s apparently a selective pressure for maintaining a

specific conformation.

Previously, other investigators have compared several di f ferent

procollagen chains and found a conserved nucleotide and amino add sequence

surrounding positions 171-173 (Asn He Thr) that designate the

glycosylation attachment site (25). In the DNA sequence coding for amino

acids 170 to 179, there are th i r ty Identical nucleotldes 1n the avian

ocl(III) (25), and human a2(I) (27) and a l ( I ) (28). This homology 1s

probably why we observed hybridization of the chick type I I I genomic

subclone with both human a l ( l ) and cc2(I) mRNAs even at stringent

hybridization and washing conditions (not shown). However, since

hybridization of the avian a l ( H I ) DNA with the human a l ( I I I ) mRNA in the

CCL136 cell l ine was very fa int and inconsistent, we were not surprised to

find that changes in these th i r ty nucleotides of the human type I I I had

occurred. Although four out of th i r ty nucleotides have diverged, the amino

add sequence 1s identical (Fig. 3). This region contains the 6th cysteine

residue. These results Indicate that there does not appear to be

conservation of the DNA sequence as long as the protein is relat ively

unchanged (Fig. 5B).

In the final third of the COOH-propeptide, inclusive of the last two

cysteines, homology between human and avian c t l ( I I I ) (Fig. 4) and human

cd(I) and a2(I) 1s also apparent (25,27,28). Especially notable 1s the

maintenance of the charged residues (Arg, Lys, Glu, Asp). The last f i f teen

amino adds of human and avian a l ( I I I ) are Identical and more l ike the

human a l ( I ) sequence (28) than the o2(I) (27).

Northern Blot Hybridization

Transcription of the human a l ( I ) and a2(I) procollagen genes results

1n multiple mRNAs 1n cultured fibroblasts (18-21). The difference 1n size

(200-1000 nucleotides) is due to the use of different polyadenylation

attachment sequences 1n the 3' untranslated region. To determine whether

this feature also applied to the ocl(II I) gene, we hybridized the cloned

a l ( I I I ) DNA to poly (A+) RNA from normal f ibroblast and rhabdomyosarcoma

cell l ines. As shown 1n Fig. 2A, two mRNAs of about equal intensity are

present. These are similar 1n size to the a2(I) mRNAs which range from

9391

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

5500-6200 nucleotides (Fig. 2C, and ref. 19,20). Whether these dif ferent

procollagen RNA transcripts are related to tissue specif ici ty s t m remains

to be determined. I t 1s unlikely that their occurence 1s fortuitous since

we have recently Identif ied a cDNA clone coding for another human

procollagen chain which also hybridizes to multiple mRNAs. (tyers, J.C. £ t

a i , manuscript in preparation).

Hybridization of a l ( I I I ) , a l ( I ) and a2(I) human procollagen cDNA

clones to poly (A+) RNAs purified from normal and tumor cell lines also

showed a signif icant difference 1n their representation (Fig. 2). Although

a l l three procollagen genes are expressed 1n normal human fibroblasts, we

have never detected a2(I) mRNA 1n the CCL136 tumor cell l ine (Fig. 2C) and

only trace levels of a l ( I ) can be seen after long exposure (Fig. 2B). The

amount of type I I I 1s comparable 1n both the normal and CCL136 cell l ines

(F1g. 2A), and 1n the lat ter vastly exceeds the small amount of a l ( l ) .

None of the three genes seem to be expressed 1n a human fibrosarcoma cell

l i ne , HT1080, which synthesizes type IV procollagen (Fig. 2 and ref. 39).

DISCUSSION

Analysis of the type I I I sequences provides a great deal of information

about regions of the protein that have been maintained between species and

the dif ferent procollagen chains. Comparison of the C-term1nal propeptides

of human and avian a l ( I I I ) shows strong homology after the variable 5'

region, but 1n the three human procollagen chains two divergent and two

conserved regions can be seen. The telopeptide and 5' part of the

C00H-propept1de are divergent but the adjacent 3' region where the f i r s t

f ive cysteines (four 1n a2(D) are located has been essentially

maintained. The second variable region 1s followed by the conserved end of

the C00H-propept1de which Includes the other three cysteines and the

glycosylation attachment s i te. The c r i t i ca l role of this domain 1n

selection and association of the procollagen chains 1s probably responsible

for the conservation of these regions (9,32). Support for this concept

comes from our studies of an osteogenesis Imperfecta patient whose DNA

contains a four nucleotide frameshift deletion 1n the pro a2(I) collagen

gene coding for the end of the carboxy-propeptide (38). Change of the last

33 aroino adds Including the position of the final cysteine prevents

Incorporation of pro a2(I) chains Into the type I procollagen molecule and

results exclusively 1n ( 01)3 homotrimer formation (38,40).

Whereas type I 1s the major constituent of bone and tendon (1,2,9),

type I I I seems to predominate 1n tissues rich in smooth muscle cel ls (41).

9392

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

The role of type I I I collagen has not been clearly defined but in another

Inherited disease of connective tissue, Ehlers Danlos type IV, rupture of

viscera, uterus and large blood vessels 1s attributed to reduced amounts

and/or structural ly abnormal type I I I collagen (9,42). Several type I

procollagen mutations resulting 1n skeletal abnormalities have recently

been defined by the use of a l ( I ) and a 2( I ) DNA clones (37,38,43).

Similar studies with these type I I I cDNA probes, and in the future genomic

clones, w i l l certainly elucidate the function and regulation of type I I I

procollagen.

ACKNOWLEDGEMENTS

We are very g r a t e f u l to Dr. Y. Yamada f o r p rov id ing the avian type I I I

genomic subclones. The encouragement and he lp fu l suggest ions of Drs. N.

Ke fa l i des and E.J . Macarak are deeply app rec ia ted . We also thank M. Mason

fo r e x c e l l e n t t yp ing o f the manuscript and DNA sequences. These s tud ies

were supported by Nat ional I n s t i t u t e s of Health Research Grants AM33348,

AM2O553, DE-02623 and HL-29702.

*To whom correspondence should be addressed:Connective Tissue Research I n s t i t u t e and U n i v e r s i t y of Pennsylvania3624 Market S t r e e t , P h i l a d e l p h i a , PA 19104

REFERENCES1 . B o r n s t e i n , P. and Sage, H. (1980) Ann. Rev. Biochem. 49,957-1003.2. M i l l e r , E .J . and Gay, S. (1982) Methods Enzymol. 84.37J2.3. Burgeson, R.E., Hebda, P.A., Mo r r i s , N.P. « H o l H s T e r , D.W. (1982) J .

B1o l . Chem. 257,7852-7856.4 . Jander , R., 'K l u te rbe rg , J . , and G l a n v i l l e , R.W. (1983) Eur . J .

Biochem. 133,39-46.5. Bentz, H.7~Morr1s, N.P. , Murray, L.W., Saka l , L.Y. H o l l i s t e r , D.W. and

Burgeson, R.E. (1983) Proc. N a t l . Acad. Sc1. USA 80,3168-3172.6. Sage, H., Trueb, B. and Borns te in , P. (1983) J . BTol . Chem.

258,13391-13401.7. UTFson, G . J . , K i e l t y , C M . , Garner, C , Schor, S .L . , and Gran t , M.E.

(1983) Biochem. J . 211,417-426.8. Tate, V . , F i ne r , M . T f o e d t k e r , H. and Doty, P. (1982) Cold Spr ing

Harbor Symp. 47^1039-1049.9. Prockop, D.J. and K1v1r i kko , K . I . (1984) New Eng l . J . Med. 311 ,

376-386.10. Lukens, L .N . , F r i s c h a u f , A .M. , Pawlowski, P . J . , B H e r l e y , G.T. and

Lehrach, H. (1983) Nuc le ic Acids Res. 11,6021-39.11 . Kurk inen, M., Barlow, D.P. , Helfman, D.'HT, W i l l i a m s , J .G. and Hogan,

B.L.M. (1983) Nuc le ic Ac ids . Res. 11,6199-6209.12. Wang, S.Y. , and Gudas, L . J . (1983 )T roc . N a t l . Acad. Sc1. USA

80,5880-5884.13. UTnomuja, Y . , Showal ter , A . , vander Rest, M., Seidah, N.G.. C h r e t i e n ,

M., and Olsen, B.R. (1984) Biochemistry 23,617-624.14. Ninomuja, Y . , and Olsen, B.R. (1984) P r o c T N a t l . Acad. S c i . USA

9393

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research

81,3014-3018.15. Strom, CM. and Uphol t , W.B. (1984) Nucleic Acids Res. 12, 1025-1038.16. Sandei l , L . J . , Pren t ice , H.L., Kravis, 0. and Upholt, W7¥. (1984) J .

B i o l . Chem. 259, 7826-7834.17. Yamada, Y. , TTa"u, G., Mudryj , M., Ob ic i , S., and deCrombrugghe, B.

(1984) Nature 31£, 333-337.18. Chu, M.L., Myers, J . C . , Bernard, M.P., Ding, J . F . , and Ramirez, F.

(1982) Nucleic Adds Res. ^£,5925-5934.19. Myers, J . C . , Chu, M.L., Faro, S.H., Clark, H .J . , Prockop, D .J . , and

Ramirez, F. (1981) Proc. Natl . Acad. Sc i . USA 78,3516-3520.20. f lyers, J . C . , Dickson, L.A., de Wet, W.J. , BernaTd", M.P., Chu, M-L.,

D IL ibe r to , M., Pepe, G., Sangiorgi , F.O., and Ramirez, F. (1983) J .B i o l . Chem. 258,10128-10135.

2 1 . Chu, M.-L. , "dT~Wet, W., Bernard M., Ding, J . F . , Morabi to, M., Myers,J .C . , Wi l l iams, C , and Ramirez, F. (1984) Nature 3J£, 337-340.

22. Huerre, C , Junien, C , Weil, D., Chu, M-L, Morabito, M., VanCong, N.,Myers, J . C . , Foubert, C , Gross, M-S., Prockop, D.J . , Boue, J . , Kaplan,J .C . , dela Chapelle, A. , and Ramirez, F. (1982) Proc. Na t l . Acad. Sc i .USA 79_,6627-6630.

23. de Wet, W.J . , Chu, M-L., and Prockop, D.J. (1983) J . B1ol . Chem.258,14385-14389.

24. Tamada, Y. , Mudryj , M., Su l l i van , M. and de Crombrugghe, B. (1983) J .B i o l . Chem. 258,2758-2761.

25. Yamada, Y. , KuFn, K., and de Crombrugghe, B. (1983) Nucleic Adds Res.11,2733-2744.

26. F u l l e r , F. and Boedtker, H. (1981) Biochemistry. 20_,996-1006.27. Bernard, M.P., Myers, J .C. , Chu, M.L., Ramirez, F., Elkenberry, E.F.,

and Prockop, D.J. (1983) Biochemistry 22,1139-1145.28. Bernard, M.P., Chu, M.L., Myers, J .C . , Ramirez, F., Elkenberry, E., and

Prockop, D.J. (1983) Biochemistry 22,5213-23.29. Myers, J . C . , Ramirez, F., Kadan, D.L., Flood, M. and Spiegelman, S.

(1980) Ana l . Biochem. 101,88-96.30. Mania t is , T . , F r i t s c h , TJ., Sambrook, J . (1982) Molecular Cloning: A

Laboratory Manual, Cold Spring Harbor Laboratory, W.3 1 . Casadaban, M.J. and Cohen, S.N. (1980) J . Mol. B i o l . 138,179-207.32. Messing, J . (1983) Methods Enzymol . 101,20-78.33 Thomas, P.S. (1980) Proc. N a t l . A c a d T S d . USA ]]_, 5201-5205.34. Kr ieg , T . , T impl , R., A l i t a l o , K., Kurkinen, M. and Vaher i , A. (1979)

Febs. L e t t . 104,405-410.35. Seyer, J.M. "and'Kang, A.H. (1981) Biochemistry 20_,2621-2627.36. 01 sen, B. and Dickson, L. 1n Chem. and B i o l . Mineral ized Connect. T i ss .

(ed. Ve1s. A.) E lsev ier /Nor th-Hol land, NY (1981) 143-153.37. Dickson, L.A., P1hl a janiemi, T. , Deak, S.B., Pope, F.M., N i cho l l s , A.,

Prockop, D .J . , and Myers, J.C. (1984) Proc. N a t l . Acad. S d . USA, 81_,4524-4528.

38. P1hlajan1em1, T . , Dickson, L.A., Pope, F.M., Korhonen, V.R., N icho l l s ,A. , Prockop, D .J . , and Myers, J.C. (1984) J . B i o l . Chem., 1n press.

39. P1hlajan1em1, T . , M y l l y l a , R., A l i t a l o , K., Vaher i , A. and K i v i r i k k o ,K . I . (1981) Biochemistry 20,7409-7415.

40. Deak, S.B., N i c h o l l s , A . , T o p e , F.M. and Prockop, D.J. (1983) J . B1ol.Chem. 258, 15192-15197.

4 1 . Hayne,~RT, Vail , M.S. and M U l e r , E.J. (1978) Biochemistry 17, 446-452.42. Pye r i t z , R.E., S t o l l e , C.A., Parfrey, N.A. and Myers, J .C. TT984) Amer.

J . Hed. Genet., In press.43. Chu, M.L., Wi l l i ams, C .J . , Pepe, G., H i rsch, J . L . , Prockop, D.J. and

Ramirez, F. (1983) Nature 304, 78-80.

9394

by guest on August 24, 2016

http://nar.oxfordjournals.org/D

ownloaded from