the garp geneencodesanewmember ofthefamilyof leucine...
TRANSCRIPT
Vol. 5, 213-219, February 1994 Cell Growth & Differentiation 213
The GARP Gene Encodes a New Member of the Family ofLeucine-rich Repeat-containing Proteins’
Vincent Ollendorff, Tetsuro Noguchi,Odile deLapeyriere, and Daniel Birnbaum2
Laboratoire d’Oncologie Mol&ulaire, Unite i 1 9, Institut National de IaSante et de Ia Recherche Medicale, 27 Boulevard Let Roure, 13009Marseille, France
Abstract
We have characterized a new human gene, namedGARP, localized in the 1 1qi 4 chromosomal region.GARP comprises two coding exons, is expressed as twomajor transcripts of 4.4 and 2.8 kilobases, respectively,and encodes a putative transmembrane protein of 662amino acids, the extracellular portion of which is almostentirely made of leucine-rich repeats. The molecularweight of the protein immunoprecipitated fromtransfected cells is 80,000. The CARP protein hasstructural similarities with the human GP lba and GP Vplatelet proteins, and with the Chaoptin, Toll, andConnectin adhesion molecules of Drosophila.
IntroductionAs more proteins are being identified, their sequences showthat many of them are members of larger families whichshare one or more common structural motifs. These motifscan be considered a perfect functional fit obtained long agoin evolution, and they constitute a repertoire from whichmolecules with different functions can be assembled. Forexample, motifs such as immunoglobulin-like loops,cysteine-nich repeats, fibronectin type Ill domains, tyrosinekinase domains, and many others, are often found in cellsurface molecules. When studying the human 1 lql 3-ql 4chromosomal region, which is amplified in breast carcino-mas (see Ref. 1 for a review; 2), we isolated a new genewhich we named GARP(3). The mouse homologue gene islocated on chromosome 7, in a region conserved betweenhumans and mice. We show here that this gene encodes aputative transmembrane protein that is composed mostly ofa particular type of motifs, named LRR.3
LRR have been identified in a wide variety ofproteins fromspecies as distantly related as yeast and humans. The averagelength of a leucine-nich unit is 24 amino acids. It is char-actenized by a periodic distribution of hydrophobic amino
Received 10/8/93; revised 1 1/22/93; accepted 12/2/93.
1 This work was supported by the Institut National de Ia Sante et de Ia Re-
cherche M6dicale (INSERM). V. 0. was supported by a Fellowship from theLigue Nationale contre le Cancer; this work was part of his doctoral thesis.T. N. was the recipient of a “Poste Vert” INSERM.2 To whom requests for reprints should be addressed.3 The abbreviations used are: LRR, leucine-rich repeat(s); cDNA, comple-mentary DNA; poIy(A)�, polyadenylated; kb, kilobase(s); kDa, kilodalton(s);ADNIV, autosomal dominant familial neovascular inflammatory vitreoreti-nopathy; ADFEVR, autosomal dominantfamilial exudative vitreoretinopathy;SDS, sodium dodecyl sulfate.
acids, especially leucine residues, and folds into an amphi-pathic �3 structure (4). The number of units varies from 1 toover 40.
Proteins with LRR appear to be involved in a variety offunctions. They are often membrane bound but can alsoeither be secreted (5) or exhibit a cytoplasmic (6) or nuclear(7) localization. LRR are found in molecules as diverse asproteoglycans, adhesion molecules, enzymes, tyrosine ki-nase receptors, or G-coupled hormone receptors. The LRRdomain seems to be involved in protein-protein interactions.A group of several LRR proteins has been identified in Dro-sophila (8). The Toll, Chaoptin, and Connectin adhesionmolecules and the Slit secreted protein seem to play impor-tant roles in cell differentiation, morphogenetic events, andmigration of cells and axons (5, 9-i 2). In humans, severalmembers of the LRR family are known. The proteoglycansdeconin (1 3) and biglycan (14), the platelet GP Iba receptorfor the von Willebrand factor (see Ref. 1 5 for a review), anda number ofother molecules (1 6-23) all contain LRR. Thesemotifs appear to be necessary for the activity of the protein(24-27).
Results
Cloning of the Human GARP Gene and cDNA. The initialcloning of GARPgenomic sequences has been reported (3).These sequences were derived from clones cos9 and cosi 05(Fig. 1A), isolated from cosmid libraries that were made fromDNAs extracted from normal blood cells and the MDA-MB-1 34 breast carcinoma cell line, respectively. This cell linehas an amplified 1 1qi 3-qi 4 region (28, 29).
Several cDNA clones were then obtained by screening ahuman cDNA placenta library with genomic subclones offragments PR98 and D, identified as carrying exonic se-quences (3) (Fig. 1 A). Overlapping cDNA clones were ob-tamed and characterized (Fig. 1 D). Two types ofcDNA wereobtained (Fig. 1 C). Accordingly, Northern blot hybridizationof human placenta poly(AY� RNA showed that CARP wasexpressed as two major transcripts of 4.4 and 2.8 kb, re-spectively (3). Fig. 1 B shows a tentative map of the CARPlocus. The GARPgene seems to comprise only two codingexons, the first one containing the signal peptide and nineamino acid residues, and the second one containing the restof the coding sequence and the 3’ untranslated region (seebelow).
Nucleotide and Deduced Amino Acid Sequence of theHuman GARP cDNA. The complete nucleotide sequencewas determined for the longest cDNA overlap of4.i kb (longform). The largest open reading frame is able to code for aprotein of 662 residues, starting with a methionine residuelocated at the beginning of a stretch of hydrophobic aminoacids, which putatively constitutes a signal peptide. This do-main is followed by a putative extracellular region whichconstitutes most of the molecule. Twenty LRR are present inthe extracellular region. A second hydrophobic stretch ofamino acids may correspond to a transmembrane domain.Finally,a short putative intracellularregion of 15 residues
R NC E R
R
I
Ic�
I.
�AAMAA
‘� )� NC
H�
�P3
SHPE
214 Human GARPGene
A.
PR9O
0
cos9. coslO5
B.� �
C.
D. ______________
Fig. 1. Structure ofthe human GARPgene. A, the inserts ofcosmid clones cos9 and coslOS are partially represented. Two genomic fragments derived from cos9and cosl 05, PR98 and D, of 2.4 and 4.5 kb, respectively, are shown below. Subclones of these fragments (described in Ref. 27) were derived to screen a cDNA
library. B, tentative physical map of the GARPgene. Boxes, the two exons. The location of the first exon within an EcoRl fragment of 7 kb is not known. Selectedrestriction enzymes are as follows: E, EcoRV; Nc, Ncol; R, EcoRI. C, representation ofthe two forms of GARPcomplete cDNAs. The positions of some restrictionenzymes indicated in B are shown. D, inserts of overlapping cDNA clones.
terminates the primary translational product of the CARPgene. The deduced amino acid sequence of the putativeGARP protein, shown in Fig. 2, thus exhibits several specificfeatures of a transmembrane molecule. It has a calculatedmolecular weight of 72,000. In the long form, a long 3’ un-translated tail of 2052 nucleotides is followed by a poly-adenylation signal. In the short cDNA form, which was alsosequenced, a region of 1545 nucleotides is absent (Fig. 2,arrowheads). The nucleotide sequence of the CARP genewas also obtained from genomic clones and allowed us todetermine the organization of the gene. Most of the extra-cellular region, the transmembrane domain, and the intra-cellular region of the putative GARP protein are encodedwithin one large exon (Fig. 1 B).
The 20 leucine-nich units are grouped in two blocks, sepa-rated by a short proline-rich segment, and constitute about70% ofthe molecule. They are aligned for comparison in Fig.3. They show a typical structure with interspersed hydro-phobic (predominantly leucine) and hydrophilic residues.
Comparison of the GARP Gene Product with Other LRRProteins. Due to the presence of leucine-rich repeats, theGARP protein (hereafter designated as Garpin) has structuralsimilarities to other proteins. Fig. 4A shows comparativealignments of the consensus motifs found in some LRR pro-teins. Some of the LRR proteins are known to share aminoacid similarity extending in either amino-terminal orcarboxy-terminal LRR-flanking regions, or both. Compari-son of these flanking regions is shown in Fig. 4B. Specificresidues found in other LRR molecules in the NH2-terminalLRR-flanking region, as well as in the COOH-terminal LRR-flanking region (especiallyfour cysteines that are highly con-served among these proteins), were also found in Garpin.Therefore, in all portions ofthe molecule, Garpin resemblesthe other LRR proteins. This can be seen in the schematicrepresentation of some of these proteins (Fig. 5).
Biochemical Characterization of the GARP-encodedProduct. The cDNA insert SHPE (see Fig. 1), subcloned intoa Bluescnipt vector to obtain pC-huGARP, was used as atemplate for in vitro translation experiments. pC-huGARP
was transcribed in vitro, either from the T3 (antisense tran-script) or T7 (sense transcript) promoters. RNAs were trans-lated in vitro in a rabbit neticulocyte lysate (see “Materialsand Methods”). The results are shown in Fig. 6A. One bandwith an apparent molecular mass of 72 kDa was detected inthe sense lane. This corresponds to the calculated molecularmass of the unglycosylated CARP gene product.
NIH 3T3 cells were transfected with a CARP-containingplasmid. Clones expressing the CARP mRNA (data notshown) were selected, and one such clone, called NG-i4,was radiolabeled with [35S]methionine. Immunoprecipita-tion of the CARP gene product from NG-14, carried outusing the rabbit antiserum, is shown in Fig. 6B. A band of 80kDa was observed. As a control, in vitro translated productswere immunoprecipitated with the rabbit anti-Garpin im-mune serum. A band with the expected molecular mass of72 kDa was observed.
The difference between the molecular mass of the in vitrotranslated product and the immunoprecipitated proteincould be explained by the presence of N-linked glycosyla-tion. This was directly tested by treating NG-14 CARP-expressing cells with tunicamycin, an inhibitor of N-linkedglycosylation. After such a treatment, the apparent molecu-lan weight of Garpin was reduced to 72,000, consistent withthe predicted molecular weight ofthe amino acid backbone.
[xpression ofthe GARPGene in Human Tissues. Expres-sion of CARP in human tissues was studied by hybridizinga commercial Northern blotfiltencontaining poly(A)� RNAs,from eight different tissues, with a human probe (Fig. 7). Aspreviously observed (3), two bands of 4.4 and 2.8 kb, re-spectively, were detected in placenta. The same level of ex-pression was observed in lung and kidney. Weaker signalswere detected in heart, liver, skeletal muscle, and pancreas.No expression was evidenced in brain.
Discussion
We have characterized a new gene which codes for a pu-tative protein with leucine-nich repeats. The deduced pni-
r-----C.
LLALLTLGLAAQHQDXVPCX
-r�--��- r��ir-r--
!�6 L V 0 N E I
GG�GCCACQGGG�
1495
467
1555
cATGkG�.ccccAGATccTacrG
XRPQILL
�LwL� � �V � FL� � L� aL� .T�
x&V 0 B K V S C Q V L a L L Q V P 3 V L
I
�
p p D ? B T L �
I.-.�CCCTGGGCT�rCTACACGGCAC�rrCGTCACCTGGACCTGAaCACCAkTGkGATCAGCrrC
115
7
175
27
235
47
295
67
35587
415
107
475
127
535
147
595
167
655
187
715
207
775
227
835
247
895
267
955
287
1015
307
C. ? E I. 0 L S S � 48717
GGCCTGGAGGCCTCC?�GGAGGTC .�TGGCA T6 �kC �GC �AC tOG �TC �TG .‘TC T0 AC 1615
0 L S A S L B V �.j. j� � � L .�.. .L. .L. S... .k. Q. 507
1
1675
V 0 L P C F I C L K B L $ L A B N B L S 527
19
CACC��rCocQCCTG0ACACA00CT0�rcTCACTQGA.GG?GCTaGACCT0�GkAACAACAGC 1735
ELPAWTQAVSLBVLDLRNXS 54720
rrCAGCCTCCTGocAOGCAGTGCCATCGOTGGCCTOGAGACCAGCCTCCCGCGcCTCTAC 1795
?SLLPGSAflOGLBTSLRRLY 567
p L 0 F V T A L B H L 0 I. S LL.L.L.L.L2
CTCC.AGCCAGGkGCC�N’CCA0GCCCTGAcCCACCTGGAGCACCTCAGCC�rGGCTCACAAC
LQPGA?OALTHLBHLSLAHW
3
CGGCTGGCGATGGCCACTOCGCTGAGTGCTGGTGGCCT000CCCCCTGCCACGCGTGACC
RLAMA?ALSAGGLGL.�LIVT
TCCC�1’GGACCTGTCTGGGAACAGCCTGTACAG�0CCCTGCTGG�.GC00CTGCTGGGG0AG
CToCAGGGG&kTCCACTCAGCTGCTOCGGCAATGGCTGGCTGGCAGCCcAGC�rGCAcCAG 3.855
LQGBPLSCCGBGWLAAQLBQ 587
cGCCG�’rGGACGTGGACGCCACCCAGGA.CCTGATCTGCCGCTTCAGC�rCcCAGaAGGAG 1915
GRVDVDATQDLICR?SSQBB 607
GTGTCCCTGAGCCACGTC�G?CCCGA0GACTGTGAGAAGGGGGGACTGAAGMCATCAAC 1975
VSLSHVRPBDCBRGGLININ 627
C�rCA��kTCATCCACCrATACr00TCTCTQCCATCCTCCTCAcCAcaCTGGC�GCC 2035
LIIXLTVILVSAILLT?LAA 647
-TGC�CTGCGVCGCCGGC&GA1’1�AACCA&CSCTATAMGCC?A&MAAGCCGGGA0 2095
CCCVRRQXFXQQYKA- 662
�
ACACTCTAG);TCAGTOGGGGAOCCTGAGG?ACAGAGAAGA.GTGAGGkCTGACTCAAGGTC 2155
ACAc�GTG�TCcGGATCCCke�ACTCTGGrCTCCAAATTACkGCCCkGGACkCC!1’frcrC 2215
TG�CGCCTGC�CATCA0G0TGkC TCccGGGCTGC�CI?1’G0GTCCkGCTGTG 2275
GAkGCCAGAkGrrGGGcGa?rrCAo0GkCAGC�AGAA?Ak1�’GACCTGTcAaATCAA 2335
� 2395
COAAATGAGATACA1?CCC0CCCTCAAG�MVTCCCAGTCTCG�28QGAG�.GAGTGC?0CAG 2455
AGCCACGTGGccGocAC0CAe�?GCTTAGG0ccTGAGGTGTCOAAAG0cCAG0GcTCC&G 2515
AoCrCGGCkaGCCCCGCTGcrN�0TGCGG?GkG’ZCCTGCCCC0Ge1’0T0CA0GGZG&Ga 2575
� 2635
� 2695
AGCTGTACGAGTGTGT�CGQ�TCTCOGTATTGTPA0QTGGGCCkC0GcGGC�CCkGC&GG 2755
� 2815
� 2875
CCCTGCCCAGCCCCACTBCTCC8TCOCA000CC&CC?IGGCTQCTG&GCCkOGAGC?MA 2935
� 2995
S L 0 L S 9 N S L V S 0 L C. LL.L..iC_.L_L
�
GCACCCAGCCTGC�TACCCTCTCACTGGCGGAG�CkGTCTGACTCOCCTC�CCCGCCkC
A P 5 L H T � S I� A LN S �
ACC�CCGG�CAGCCTGCGGGAGCGCTACGCATGC0TGCTGATGG�C
TFRDMPALEQLDLH�LL.�..�L6
ATCGAGG�.TGGCOCC?ZCGAG0GCCTGCCCCGCCTGkCCCkTCTCAACCTCTCCkGG�kT
IBDGAFBGLPRLTH�NLSRN
7 *
TCCCTCACCTGCATCTCCGACTTCAGCCTCCAGCAGCTOCGGQTOCrAGe.CCTGAOCTGC
SLTCISD?SLQ0LRV�DLSC
N S I B � F 0 T A S 0 P 0 A � � � j� j� �
C?rGACCTGCGGGAGAACAAACTOC�rCCM�rrCCCCGACCTGGCcGCGCTCCcGAGAC�rC
� 0 L R B N X L L H P P 0 L j � � j � �
ATCTACCTGMCTZ0TCC�AC�ACCTCATCC0GCTCCCCA.CkGGGCC&CCCCkGGkC�GC
L__i f K I. S N N C. X B L P ? �_. i... .2� .a.p �
AkGGGCATCCAeGCACCTTCCGA0GGCT0GTCAGcCCT0CCCC�rCTCAGCCCCCAGCGGG
K 0 I H A P 8 B G B S A L P L S A P S 0
MTGCCAG�G0CcGCCCCCrrrCCCAGC�rCTTGAA?CTGG8?rroAGCTACAATGAGATr
N A S a R P L S Q L L 11 I, 0 I. S Y N B I
4 11GAGCTCATCcCCGACAGerrrC�rrG�GCACCTGACCTccCTOT0C?rcCTaAACcTCAGC.
1075
327
1135
347
1195
367
1355
387
1315
407
3055
� 3115
� 3175
AGTGG&.CAGGAAGTGGCcCC5.CATAO�CG�CAQG�GMOCAWCTGTGCTG?GC 3235
ACA.CCTGCTCTCTCCTCTCTCCC&OGCAGGCAGCTGCAGGCOCTCTCCTCCT?CTCTOCC 3295
�
5�C51CT�rCAICCC�BTTCCTCTATGGAQC�GAOcCTGGWBM’COAG&CTATGGAA? 3415
� 3475
� 3535
G5TGGACcA0AaTCOATGTcCTC5’cTCAGCTATCAcATCACAAGATAkTaC�1’aGC?�CAAA 3595
�
C�rCC?rWTGTOcCTCATCATOCMOG&�C?rB1?1’cCCTC?MCAAAAACAth?AAM 3655
AaccTCA�C.AaATaACccCcAwcCTCATAccA�aG>cATGAaCTaTCToGGAkaAA 3715
T0G&,OGTGCTGGCOACCAACTC8AGACC1�N?B?BGCTGTCflCATC�TCTTACCTG’PGCT 3775
� 3835
� 3895
BLIPDSPLBaL?SLC?�NLSL7 �
AGAAACT0C’rrGCGGACC?rrGA.GGCCcGGCGC�1�rAGGCTCCCT0CCCTaCCT0AT0CTC
R N C I( R T P 5 A B 1 1. 0 8 j� � � .� � �
�kCTZAQCCACAAT0C�C0GAGACACTUGMCTO00CGCCAGAGCCCT000GTCT
L 0 I� S H N A I� B ? L B I. CO � �j. � � #{231}�1:3
CTGCG0&e0CTGCTCCTAcA.GGGCAAT0CCCTGCGOG5.CCTGCCCCCATACACC�frrG�C
L R ? f L L 0 0 11 A I. B 0 6. g � � j� � �
4
AATC’r0GCCA.GCCTGCAGCGQCTCAACCT0CA0GGGAACC0AGTCAG�CCCTaTGGGGGa
It C. A S L 0 A � N L Q 0 N B 3� J P C G 91375
427AATG9GC&GAGCC&GGGCTA.OOCCC?COGIOGTOTOTGCOAGCACCCACOCAGACCCW!CC 3955
� 4015
CAAGGTGTGGGTCCCCAC0CCTGATCTt�’GAAAA.CACTACACACoGGCT0CT0TCAC�N’CC 4075
PDIPGPSOCVAPSGI?SLRS
1435
447
4135
4163
Cell Growth & Differentiation 215
Fig. 2. Nucleotide and deduced amino acid sequence ofthe human GARPcDNA. The GARPnucleotide sequence codes for a putative transmembrane protein.
The coding sequence is characterized, from 5’ to 3’, by the presence of a putative signal peptide (underlined in bold), two series of 1 0 blocks of LRR (underlined
and numbered), separated by a proline-rich sequence, a putative transmembrane domain (underlined in bold) followed by a short, potentially intracellular,carboxy-terminal tail. Asterisks, five potential N-linked glycosylation sites. In the 3’ untranslated region, a polyadenylation signal is indicated (underlined). Thearrow at position 178 indicates the 3’ limit of exon 1 and the 5’ limit of exon 2. Arrowheads delimitate the 3’ untranslated region spliced out in the short formof the CARP transcripts. Numbers relative to nucleotide and amino acid positions are shown at right.
amino acid
54-77
78-101
102-128
129-153
154-177
178-201
202-222
223-247
248-269
270-292
320-343
344-3 67
368-390
391-4 14
415-438
448-470
471-495
496-518
519-54 0
541-565
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
D
D
L�]G
L[�]T
L�S
L Q G
L A E
L R N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Q E�1E I
R L
S L
S L
R
S
A
Y
T
S I
F �
N A
SGL
R L
L
Q
T
L
T
A
P
A
E
R
S
G
L
R
H
P
A
S
L
T
F
A
�
F
G
Q
G
G
R
F
A
G
E
D
Y
L
L
A
M
T
T
G
P
P
A L
H L
P L
S L
A L
R
E
P
H
E
H
H
R
T
Q
V L
S L
S I
K �
L I
E I
M
T
E
L
R
E
D I
C I
A F
H F
E�IL I
E
S
Q
P
�
P
D
D
T
D
T
D
G
F
A
G
S
A
S
S
P
F
F
�
Q
�
P
L
E
Q
P
A
Q
E
G
Q
Q
A
D
H
L
L
A
�
S
L
P
R
E
P
K
T
R L
V
FQL
R L
G I
S I,
T
T
I
H
C
H
W
Y
A
F
C L R T F E A R R L G S L P C L M L
A L E T L E L G A R A L G S L R T
A L R D L P P Y T F A N L A S L Q R
R V
E I
PGL
G L
S
E
E
M
P C
L �
V
V L
G
R
A
Q
G
A
T
V
P
G
G
D
D
A
A
E
F
L
L
P
L
G
P
G
H
G
C
P
T
�
F
S
P
E
I
G C
L
ASL
C L
V
T
E
K
A
E
V
R
R L
S F
S
S
H L
L L
P
P
A
G
W
S
T
A
Q
M G G
A
�
V
E
S L
TSL
E
R
V
R
216 Human GARPGene
CONSENSUS: L D L S - N - L - - L - - - - L - - L - - L - -Fig. 3. Comparison of the leucine-rich repeats of the CARP-encoded protein. The amino acid sequence of the 24 amino acids of LRR comprising the main bodyof the extracellular region are aligned to allow for comparison. Leucine residues are boxed. Residues other than leucine found in 50% or more of the units arealso boxed. Space is a gap introduced to maximize alignment. As in other proteins of the family, the asparagine residue (N) is invariant. A derived consensus
sequence is shown below.
mary structure of the human protein, which we have namedGanpin, shows the presence of several repeats made of 24amino acids with a remarkable peniodicity of leucine resi-dues. In vitro translation experiments yielded a single prod-uct of 72 kDa, in good agreement with the calculated mo-lecular mass ofthe GARPgene product. A protein of similarsize was immunoprecipitated when products from the in
vitro experiments were treated with a rabbit anti-Garpin im-mune serum made against a TrpE-Garpin fusion protein. Aprotein of higher molecular weight was observed in cells
stably transfected with a CARP full-length cDNA. The dif-ference in molecular masses could be attributed to N-linkedglycosylation.
Leucine-rich motifs are found in a number of proteins in-volved in important biological processes such as embryonicdevelopment, cell morphogenesis, cell and axon migration,and blood coagulation. In these molecules, LRR often makeup a large portion of the protein. In several instances, LRRhave been shown to be essential for the specific activity ofthe protein (24-27, 30). Most of the CARP sequence is en-coded with i n one exon . The genom ic structu re of some otherhuman LRR proteins is known. In the case of OMgp and GPIba, two LRR proteins characterized in humans, the codingsequence is not interrupted by introns (20, 31 ). In contrast,LRR ofthe extracellular region ofthe thyroid-stimulating hor-mone receptor are encoded by separate exons (22).
Owing to the presence of leucine-rich motifs, Garpin hasstructural similarity with other LRR proteins. It resembles GPIba and GP V (32) and the Drosophila adhesion moleculesToll, Chaoptin, and Connectin, in which the LRR constitutea large part of the molecule. This suggests that Garpin alsocould be involved in protein-protein interactions at the cell
surface, especially in adhesion processes. The structure andpattern ofexpression ofthe CARPgene are compatible witha possible role as an adhesion molecule of its encoded prod-uct. This hypothesis is currently being investigated.
CARP is located in chromosomal region 1 1 ql 3.S-qi 4 (3),which is amplified in about 1 7% of breast carcinomas (2,33). The CCNDJ/PRAD1 cyclin gene, located in liql3.3,is thought to be the key gene of the ilq amplification inbreast cancer (34). The region containing CARPcould eitherdefine a separate amplicon or merely represent the telomericextension of a large amplification. For the time being, therole of CARPwithin the amplicon is not known. However,CARP does not exhibit structural similarity with the onco-genes, and it is unlikely that it is involved in the cancerousprocess. It probably merely represents a “passenger” gene inthe amplicon. In addition to DNA amplification and rear-rangement in malignant diseases, the 1 1 qi 3-qi 4 region isinvolved in several nontumoral pathologies, some of whichare hereditary diseases. ADNIV, ADFEVR, and Usher syn-drome type I, in particular, appear to map close to or withinthe 1 1qi 4 region (35-37). In that respect, itisinterestingtonote that Chaoptin and COnnectin, to whkh Garpin is struc-turally related,are involved in the development of sensoryorgans and nervous system in Drosophila (1 0). WhetherCARPis implicated in pathology, and is a possible candidatefor ADNIV or ADFEVR, has yet to be determined. Moreover,the human olfactory marker protein OMP is encoded by agene located at 1 1qi 4-q21 which is a candidate for Ushersyndrome type I (38). A cluster ofgenes involved in patholo-gies in relation with the development of sensory organsmight thus exist in the ilqi3.S-q2i region. This possiblyincludes the CARP gene.
Leucine Rich Family
5- N
A.
GARPIN
CONNECTIN
TOLL
CHAOPTIN
CYCLASE
GPIba
PG4 0
LRG
- NN
--N
L
L
L
L
L
L
D
N
D
D
L
L
L
L�
L
-Iii- -
-ILl- -
III1L1
-L�d--- h- -- LTT
KI S --L- -
LIL
I
h
L
V
L
5-
5-
SG
L
- - - AF
P--LF
F
P-- h
P-GLL
- -G -h
PPGLL
N
N
N
N
N
-H- -N
--Iii--- -ILl-
- ILH -Q G[�j_ Q
L
L
L
L
L
L
L
L R-CONSENSUS L-L--N-L--h-
I
NH2 Flanking sequence
C QV
CSD
CS D
C DN
C SG
C DK
Q V P S V L P P D T�T
K V P K D L P P D T�L
S V P K E I S P D T_TIL
YLP-FVPSRMKY
TLPSGLQENIIH
A L P P D L P KID_T_TJI
Cell Growth & Differentiation 217
B.
GARPIN
PG4 0
BIGLYCAN
FM
OMGP
GPIba
VS
VQ
VQ
MY
VD
VN
LGLL
LG L D
LG LK
RNLK
RNLS
RN L T
CONSENSUS V-C--LGL--LP--LPPDTT-RN V
C. COOH Flanking sequence
** * *
GARPIN GNPLSC-CGNGWLAAQLHQGR VDVDA----TQDLICRF--SSQEEVSLSHV HVRPEDCEKGGLKNINTRKB GNPFTCSCDIMWLK-TLQETKSSPD TODLYCLN-ESSI<NMP-LANL QIPNCGLHR TYPSHC-CAFRNLPKKEQNFSFSIF ENFSKQCESTVRKADNETLYSAIFEENELSGWDYDYGFCSpKGPIbcz GNPWLCNCEILYFRRWLQDNAENVYVWKQGVDVKAMTSNVASVQCDNSDKFPVYKYPGKG CPTLGDGPIbB ANPWRCDCRLVPLRAWLGRPER APYRDLRCVAPPALRGRLLPYLAE DELRAACAPGPLCTOLL1 DNPLVCDCTILWFIQLVRGVHKPQYSRQFK LRTDRLVCSQPNVLEGTPVROIEP QTLICPLDTOLL2 GNPWMCDCTAKPLLLFTQDNFERIG DRNEMMCVNAENPTRNVELS TNDICPAEKCONNECTIN DNKFICDCRLQWIFELKNRTRHLQLRD SLEDLHCTLQEPKLSHFVDPVPP TILDVLNI
Fig. 4. Comparison of Garpin with other LRR proteins. A, alignment of consensus sequences of LRR derived from eight different LRR proteins. LRC, leucine-rich
a2 glycoprotein of the serum; h, hydrophobic residues; space is a gap introduced to maximize alignment. B, alignment of the amino-terminal LRR flanking region
of Garpin with the corresponding regions from other LRR proteins. FM, fibromodulin. Conserved residues are boxed. C, alignment of the carboxy-terminal LRRflanking region of Garpin with the corresponding regions from other LRR proteins. LHR, lutropin receptor. The four cysteine residues found in all proteins are
in bold and indicated by asterisks. Dash, a gap introduced to maximize alignment.
Materials and Methods
Cells, Tissues, and Embryos. NIH 3T3 cells were grown inDulbecco’s modified Eagle’s medium supplemented with1 0% fetal calf serum.
cDNA Screening. Two cDNA libraries made from humanplacenta were screened, and several clones were isolated:HG clone from a Clontech (Palo Alto, CA) library and SHPclones from a Stratagene (La Jolla, CA) library (see Fig. 1).Mouse cDNA clones were obtained by screening a mouseembryo library as described (3). The “BigLow” fragment of0.5 kb (described in Ref. 3), corresponding to the transmem-brane domain, the carboxy-terminal region, and a portion ofthe 3’ untranslated region, was used as a probe.
Nucleotide Sequencing. Nucleotide sequences were de-termined with the dideoxynucleotide method, using Blue-script (Stratagene) single-stranded or double-stranded tem-plates and the modified T7 polymerase protocol (Sequenase;U.S. Biochemicals). Both short and long cDNA clones weresequenced, as well as corresponding portions of genomicclones derived from cosmids. Clone SHPE was sequencedon both strands. Sequence analysis was carried out usingPC-gene software (Intelligenetics). The accession number forthe nucleotide sequence of the CARP gene is Z24680.
Antibody Production. A TrpE-Carp construct was madeby inserting a 1 .7-kb EcoRI fragment of the murine Carpcoding region (corresponding, in the human protein, toamino acid residues 467 to 662; see Fig. 2) into the EcoRlsite ofthe pATH3 expression vector. A 58 kDa fusion proteinwas produced in RR1 bacterial cells as described (39). Thebacterially expressed protein was injected into rabbits togenerate polyclonal antibodies against the product of theCarp gene.
In Vitro Translation Experiments. A 2.2-kb BamHI frag-ment containing the full-length coding sequence of humanCARP, derived from the cDNA clone SHPE (see Fig. 1), wassubcloned intoa Bluescnipt KS� vector (Stratagene)to obtainplasmid pC-huGARP and was used for in vitro transcriptionand translation assays. In vitro transcription was carried outwith the T3 (antisense transcript) and T7 (sense transcript)RNA polymerases from the Stratagene RNA transcription kit,according to the manufacturer’s instructions. Aliquots of 0.5to 1 pg of synthesized RNAs were translated in a rabbit re-ticulocyte lysate system (Promega), in the presence of[35S]methionine (Amersham) according to the recom-mended protocol. Products of in vitro translation were re-solved by electrophoresis in a 1 0% SDS-polyacrylamide gel.
TOLL
9.5-.
7.5-.
4.4-0
2.4�
A Bg NIH 3T3 NG-14‘�I 11 1
:� � TunlcamycinAS S .�- + - +
106-
80-
50-
218 Human CARP Gene
We thank our colleagues C. Mawas, F. Birg, F. Coulier, ). Jacquemier, and P.
Gaudray (Nice), for helpful discussions and enthusiastic support.
GARPIN .ftUW!IIWWI![;
GPIba IHIHII
OMGP #{149}HH1H�
CONNECTIN L LHHHHI �1
CHAOF’ltN � � I � � � �1�����I1�� ��I������I���II�
Fig. 5. Schematic representation of Garpin and other LRR proteins. Two
types of LRR proteins are represented, i.e., the transmembrane type ( TM,transmembrane domain) and the Pt-anchored type (P/, glycosyl-
phosphatidylinositol structure). Three mammalian proteins (Garpin, GP be,
and OMgp) and three Drosophila adhesive molecules (Toll, Connectin, and
Chaoptin) are shown. Putative signal sequences and transmembrane domainsare shown in black. Blocks of LRR can be separated by a short segment (hori-
zontal stripes). Each unit is shown in open boxes. LRR flanking regions are
shaded.
Fig. 6. Biochemical characterization of Garpin. A, RNA transcribed in vitrofrom the human CARPcDNA clone pC-huGARP were translated in a rabbitreticulocyte lysate system and analyzed by 10% SDS-polyacrylamide gel
electrophoresis. Translation was from either antisense (AS lane) or sense (Slane) transcripts. B, immunoprecipitation ofGarpin (arrowheads)from CARP-
transfected NIH 3T3 cells (NC-14). Control NIH 3T3 cells and in vitro trans-lated products are shown. Cells were treated (+) or not (-) with tunicamycin.Migration of molecular mass markers (molecular mass standards from Bio-
Rad) is indicated between the two panels.
The gel was then fixed, treated fonfluonognaphy with Amplify(Amensham), dried, and exposed to Fuji on Kodak films.
Cell Transfections. The full-length cDNA clone derivedfrom SHPE (see Fig. 1 ) and encoding the human CARP genewas inserted in both sense and antisense orientations in themammalian expression vector pMexNeo (40) to obtainpMexNeo-GARP and pMexNeo-PRAG, respectively. NIH
IItBPILULI MKPa
Fig. 7. Expression ofthe CARPgene in human tissues. A commercial North-em blotfilter (Clontech), containing 2 pg/lane ofpoly(A)’ RNA extracted fromvarious human tissues (as indicated above each lane), was hybridized with
a human CARPprobe (see “Materials and Methods”). Two major transcripts,of 4.4 and 2.8 kb, respectively, were detected (arrowheads). Ht, heart; B,
brain; P1, placenta; Lu, lung; Li, liver; M, skeletal muscle; K, kidney; Pa,pancreas.
3T3 cells were seeded at 3 x 1 O� cells/i 00-mm dish 1 8 hprior to transfection. Cells were transfected with 10 pg of
either pMexNeo-GARP or pMexNeo-PRAG, using Lipofec-tin reagent (GIBCO-BRL) according to the manufacturer’sindications.
Immunoprecipitation Experiments. Approximately 3 xio� NIH 3T3 cells were labeled for 3 h with 10 pCi[35S]methionine/mI in Dulbecco’s modified Eagle’s mediumminus methionine supplemented with iO% dialyzed fetalcalf serum. The labeled cells were lysed in 200 p1 of dena-tuning buffer (SO m� Tnis-CI, pH 7.5-O.5% SDS-70 mt’�if3-mencaptoethanol), boiled for 5 mm, and diluted in 800 p1
of nadioimmunoprecipitation assay buffer (1 0 m� Tnis-Cl, pH7.5-1% sodium deoxycholate-1% Nonidet P-40-130 mt�iNaCI-O.2S mM phenylmethylsulfonyl fluoride). All sampleswere initially preadsonbed with 5 p1 of preimmune serum at
4#{176}Cfor 1 h and then incubated with 40 p1 of ProteinA-Sepharose CL-4B (Phanmacia) for another 1 h at 4#{176}C.The
supennatants of these incubations were then used for im-munoprecipitation with anti-Ganpin rabbit serum. Sampleswere incubated with 5 p1 of antiserum for 1 h on ice, fol-lowed by incubation with 40 p1 of Protein A-Sepharose for1 h at 4#{176}C,on a rocking system. Immune complexes werewashed once with buffer A (1 0 m�.i Tnis-CI, pH 7.5-1 50 ms�iNaCI-2 m� EDTA-0.2% Nonidet P-40), once with buffer B(same as buffer A but with 500 mt�s NaCI), and once withbuffenC(iO mMTris-Cl, pH 7.5). After boiling in 2X Laemmlibuffer, the samples were nun in a 7.5% polyacrylamide gel.In some experiments, cells were treated with 1 0 pg/mI tu-nicamycin throughout the labeling period.
Northern Blot Hybridizations. A blot filter containing 2pg/lane of poIy(AY� RNAs extracted from eight human tis-sues (Clontech) was hybridized with the 2.2-kb insert frag-ment from pC-huGARP, using the conditions recommendedby the manufacturer.
Acknowledgments
Cell Growth & Differentiation 219
References
1 . Gaudray, P., Szepetowski, P., Escot, C., Birnbaum, D., and Theillet, C.DNA amplification at 1 1 ql 3 in human cancer: from complexity to perplexity.Mutat. Res., 276: 31 7-328, 1992.
2. Szepetowski, P., Ollendorff, V., Grosgeorge, J., Courseaux, A., Birnbaum,D., Theillet, C., and Gaudray, P. DNA amplification at 1 lql 3.5-q14 in humanbreast cancer. Oncogene, 7: 2513-2517, 1992.
3. Ollendorif, V., Szepetowski, P., Mattel, M-G., Gaudray, P., and Birnbaum,D. New gene in the homologous human 11q13-q14 and mouse 7F chro-mosomal regions. Mamm. Genome, 2: 195-200, 1992.
4. Krantz, D., zidovet�ki, R., Kagan, B., and zipursky, L. Amphipathic f3
structure ofa leucine-rich repeat peptide. j. Biol. Chem., 266:16801-16807,1991.
5. Rothberg, J., Jacobs, R., Goodman, C., and Artavanis-Tsakonas, S. Slit: anextracellular protein necessary for development of midline glia and com-missural axon pathways contains both EGF and LRR domains. Genes & Dev.,4: 2169-2187, 1990.
6. Kataoka, T., Broek, D., and Wigler, M. DNA sequence and characteriza-tion ofthe S. cerevisiaegene encoding adenylate cyclase. Cell, 43: 493-505,
1985.
7. Ohkura, H., and Yanagida, M. S. pomb#{233}gene sds22� essential for a mid-mitotic transition encodes a leucine-rich repeat protein that positively modu-
lates protein phosphatase-1 . Cell, 64: 1 49-1 57, 1 991.
8. Hortsch, M., and Goodman, C. Cell and substrate adhesion molecules inDrosophila. Annu. Rev. Cell. Biol., 7: 505-557, 1991.
9. Hashimoto, C., Hudson, K., and Anderson, K. The Toll gene of Drosophila,required for dorso-ventral embryonic polarity, appears to encode a trans-membrane protein. Cell, 52: 269-279, 1988.
10. Reinke, R., Krantz, D., Yen, D., and zipu,�ky, L. Chaoptin, a cell surfaceglycoprotein required for Drosophila photoreceptor cell morphogenesis, con-tains a repeat motiffound in yeast and human. Cell, 52: 291-301, 1988.
1 1 . Nose, A., Mahajan, v., and Goodman, C. Connectin: a homophilic celladhesion molecule expressed on a subset of muscles and the motoneuronsthat innervate them in Drosophila. Cell, 70: 553-567, 1992.
1 2. Ollendorff, V., Noguchi, T., and Birnbaum, D. Des proteines a motifsriches en leucines definissent une cinqui#{232}me famille de moleculesd’adh#{233}rence. Med./Sci., 9: 1 102-1 109, 1993.
1 3. Krusius, T., and Ruoslahti, E. Primary structure of an extracellular matrixproteoglycan core protein deduced from cloned cDNA. Proc. NatI. Acad. Sci.USA, 83: 7683-7687, 1986.
1 4. Fisher, L., Termine, J., and Young, M. Deduced protein sequence of bonesmall proteoglycan I (biglycan) shows homology with proteoglycan II(decorin) and several nonconnective tissue proteins in a variety of species. J.Biol. Chem., 264: 4571-4576, 1989.
1 5. Roth, G. I. Developing relationships: arterial platelet adhesion, glyco-protein Ib, and leucine-rich glycoproteins. Blood, 77: 5-1 9, 1991.
16. Takahashi, N., Takahashi, Y., and Putnam, F. Periodicity of leucine andtandem repetition of a 24-amino acid segment in the primary structure ofleucine-rich a2-glycoprotein ofhuman serum. Proc. NatI. Acad. Sci. USA, 82:1906-1910, 1985.
1 7. Lee, F., Fox, E., zhou, H. M., Strydom, D., and Vallee, B. Primary structureof human placental ribonuclease inhibitor. Biochemistry, 27: 8545-8553,1988.
1 8. Schneider, R., Schneider-Scherzer, E., Thurnher, M., Auer, B., and Sch-
weiger, M. The primary structure of human ribonuclease/angiogenin inhibitor(RAI) discloses a novel highly diversified protein superfamily with a commonrepetitive module. EMBO J., 7: 4151-4156, 1988.
19. McFarland, K., Sprengel, R., Phillips, H., KOhIer, M., Rosemblit, N., Ni-kolics, K., Segaloff, D., and Seeburg, P. Lutropin-choriogonadotropin recep-tor: an unusual member of the G protein-coupled receptor family. Science(Washington DC), 245: 494-499, 1989.
20. Mikol, D., Alexakos, M., Bayley, C., Lemons, R., LeBeau, M., and Stefans-
son, K. Structure and chromosomal localization of the gene for theoligodendrocyte-myelin glycoprotein. J. Cell Biol., 1 1 1: 2673-2679, 1990.
21. Tan, F., Weerasinhe, D., Skidgel, R., Tamei, H., Kaul, R., Roninson, I.,Schilling, I., and Erd#{246}s,E. The deduced protein sequence of the human car-boxypeptidase N high molecular weight subunit reveals the presence ofleucine-rich tandem repeats. J. Biol. Chem., 265: 13-19, 1990.
22. Gross, B., Misrahi, M., Sar, S., and Milgrom, E. Composite structure of the
human thyrotropin receptor gene. Biochem. Biophys. Res. Commun., 177:
679-687, 1991.
23. Cutler, M., Bassin, R., zanoni, L., andTalbot, N. Isolationofrsp-1, a novelcDNA capable of suppressing v-Ras transformation. Mol. Cell. Biol., 12:3750-3756, 1992.
24. Lee, F., and Vallee, B. Modular mutagenesis of human placental ribo-
nuclease inhibitor, a protein with leucine-rich repeats. Proc. NatI. Acad. Sci.USA, 87: 1879-1883, 1990.
25. Susuki, N., Choe, H. R., Nishida, Y., Yamawaki-Kataoka, Y., Ohnishi, S.,Tamaoki, T., and Kataoka, T. Leucine-rich repeats and carboxyl terminus arerequired for interaction of yeast adenylate cyclase with RAS proteins. Proc.NatI. Acad. Sci. USA, 87:8711-8715, 1990.
26. Fresco, L., Harper, D., and Keene, J. Leucine periodicity of U2 smallnuclear ribonucleoprotein particle (snRNP) A’ protein is implicated in snRNPassembly via protein-protein interactions. Mol. Cell. Biol., 11: 1578-1589,1991.
27. Braun, T., Schofield, P., and Sprengel, R. Amino-terminal leucine-rich
repeats in gonadotrophin receptors determine hormone selectivity. EMBO J.,10: 1885-1890, 1991.
28. Lafage, M., Nguyen, C., Szepetowski, P., Pebusque, M-J., Simonetti, J.,Courtois, C., Gaudray, P., deLapeyriere, C., Jordan, B., and Birnbaum, D. The1 1 qi 3 amplicon of a mammary carcinoma cell line. Genes Chrom. Cancer,
2: 171-181, 1990.
29. Lafage, M., Pedeutour, F., Marchetto, S., Simonetti, J., Prosperi, M-T.,Gaudray, P., and Birnbaum, D. Fusion and amplification of two originallynon-syntenic chromosomal regions in a mammary carcinoma cell line. Genes
Chrom. Cancer, 5: 40-49, 1992.
30. Schneider, D., Hudson, K., Lin, T. Y., and Anderson, K. Dominant andrecessive mutations define functional domains of Toll, a transmembrane pro-tein required for dorsal-ventral polarity in the Drosophila embryo. Genes &Dev., 5: 797-807, 1991.
31 . Wenger, R., Kieffer, N., Wicki, A., and Clemetson, K. Structure of thehuman blood platelet membrane glycoprotein Iba. Biochem. Biophys. Res.Commun., 156:389-395, 1988.
32. Hickey, M., Hagen, F., Yagi, M., and Roth, G. Human platelet glyco-protein V: characterization ofthe polypeptide and the related Ib-V-Ix receptorsystem of adhesive, leucine-rich glycoproteins. Proc. NatI. Acad. Sci. USA,90: 8327-8331, 1993.
33. Karlseder, J., zeillinger, R., Schneeberger, C., Czerwenka, K., Speiser, P.,Birnbaum, D., Gaudray, P., and Theillet, C. Patterns of DNA amplification atband qi 3 ofchromosome 1 1 in human breast cancer. Genes Chrom. Cancer,
in press, 1994.
34. Lammie, A., Fantl, v., Smith, R., Shuuring, E., Brookes, S., Michalides,
R., Dickson, C., Arnold, A., and Peters, G. Dl 1 5287, a putative oncogene onchromosome 1 1ql 3, is amplified and expressed in squamous cell and mam-mary carcinomas and linked to BCL-1 . Oncogene, 6: 439-444, 1991.
35. Ii, Y., MUller, B., Fuhrmann, C., Van Nouhuys, E., Laqua, H., Humphries,P., Schwinger, E., and Gal, A. The autosomal dominant familial exudativevitreoretinopathy locus maps on 11q and is closely linked to D11S533. Am.
J. Hum. Genet., 51: 749-754, 1992.
36. Stone, E., Kimura, A., Folk, J., Bennett, S., Nichols, B., Streb, L., and
Scheffield, V. Genetic linkage of autosomal dominant neovascular inflam-matory vitreoretinopathy to chromosome 11q13. Hum. Mol. Genet., 1:685-689, 1992.
37. Kimberling, W., M#{246}Iler, C., Davenport, S., Priluck, I., Beighton, P.,
Greenberg, J., Reardon, W., Weston, M., Kenyon, I., Grunkemeyer, I., PiekeDahI, S., Overbeck, 1., Blackwood, D., Brower, A., Hoover, D., Rowland, P.,and Smith, R. Linkage of Usher syndrome type I (USH1 B) to the long arm ofchromosome 1 1 . Genomics, 14: 988-994, 1992.
38. Evans, K., Fantes, J., Simpson, C., Arveiler, B., Muir, W., Fletcher, ).,VanHeyningen, V., Steel, K., Brown, K., Brown, S., St. Clair, D., and Porteous, D.Human olfactory marker protein maps close to tyrosinase and is a candidategene for Usher syndrome type I. Hum. Mol. Genet., 2: 1 1 5-1 18, 1993.
39. Moran, M., Koch, C., Sadowski, I., and Pawson, T. Mutational analysisof a phosphotransfer motif essential for v-fps tyrosine kinase activity. Onco-gene, 3:665-672, 1988.
40. Stanton, B. Ph.D. thesis. Characterization and analysis of an alternativec-fms transcript expressed in human tumor cell lines. Baltimore: Universityof Maryland, 1989.