Expression and Alternative Splicing of the Deleted in ... · The DCC (deleted in colorectal cancer) gene was identified because it is affected by somatic mutations in colorectal tumors,

[CANCER RESEARCH 54. 4493-4501. August 15. 1994|

Expression and Alternative Splicing of the Deleted in Colorectal Cancer(DCC) Gene in Normal and Malignant Tissues1

Michael A. Reale, Gang Hu, Abrahim I. Zafar, Robert H. Getzenberg,2 Stuart M. Levine, and Eric R. IVanni '

Departments of Medicine Â¡M.A. R.Â¡,Pharmacology [G. H.Â¡,and Pathology Â¡R.H. G., A. l. Z., S. M. L, E. R. F.], Molecular Oncology and Development Program, Boyer Centerfor Molecular Medicine, Yale University School of Medicine, New Haven, Connecticut 06536-0812

ABSTRACT

The DCC (deleted in colorectal cancer) gene was identified because it isaffected by somatic mutations in colorectal tumors, including allelic lossesin greater than 70% of cancers and localized mutations in a subset ofcases. The DCC gene also may be inactivated in other tumor types,including cancers of the pancreas, stomach, breast, prostate, and brain, aswell as some leukemias. We have characterized DCC complementaryDNAs obtained from human fetal brain tissues and IMR32 human iicn-

roblastoma cells. Based on the fetal brain complementary DNA sequence,the predicted transmembrane DCC protein product has 1447 amino acids.The extracellular domain of about 1100 amino acids has four immuno-globulin-like domains and six fibronectin type Ill-like domains; the 325-

amino acid cytoplasmic domain does not show similarity to previouslycharacterized proteins. Comparison of DCC complementary DNAs fromIMR32 cells to those from fetal brain identified two potential alternativesplice sites. Studies of adult mouse tissues revealed that DCC trancriptswere present at very low levels in all tissues studied, and alternativesplicing of DCC transcripts was seen in some tissues. Immunoblotting andimmunoprecipitation studies with DCC-specific antisera identified proteinspecies with molecular weights of approximately 175,000-190,000 in some

rodent tissues and human tumor cell lines. DCC protein expression washighest in brain tissues and neural crest-derived cell lines and markedly

reduced or absent in the majority of cancer cell lines studied. Treatmentof DCC-expressing cells with tunicamycin decreased the apparent molec

ular weight of the immunoreactive proteins, establishing that DCC is aglycoprotein. The studies presented here demonstrate that the DCC geneencodes several related glycoprotein species that are likely to be expressedat very low levels in many normal adult tissues. Furthermore, the absenceof DCC expression in some of the cancer cell lines studied may result fromgenetic inactivation of DCC.

INTRODUCTION

Allelic losses are thought to signal the existence of a tumor suppressor gene in the affected chromosomal region (1, 2). In previousstudies of human colorectal carcinomas, chromosome 18q was frequently affected by allelic loss with 18q loss of heterozygosity detectable in more than 70% of cases (2, 3). A common region of allelicloss on chromosome 18q in colorectal cancers that included18q21 -qter was identified. Subsequent studies identified two tumors

with somatic mutations, one with homozygous loss and the other witha point mutation, affecting sequences at the 18q21 locus known asD18S8 (3, 4). Positional cloning studies then identified a gene thatincluded the DÃŒ8S8locus within one of its introns. This gene wastermed DCC (for deleted in colorectal cancer), and its predicted amino

Received 4/26/94; accepted 6/28/94.The cosls of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1Supported in part by a Patrick and Catherine Weldon Donaghue Postdoctoral Fel

lowship (to M. A. R.), Swebilius Cancer Research Award (to R. H. G.), the Lucille P.Markey Charitable Trust (to E. R. F.). a James S. McDonnell Foundation MolecularMedicine in Cancer Research Award (to E. R. F.). and Yale Cancer Core Grant 5P30-CA16359-19.

2 Present address: Pittsburgh Cancer Institute. University of Pittsburgh School of

Medicine, 10th Floor, Biomedicai Sciences Tower, Pittsburgh, PA 15213-2582.3 To whom requests for reprints should be addressed, at Boyer Center for Molecular

Medicine. Room 354, Yale University School of Medicine, 295 Congress Avenue, NewHaven, CT (16536-0812.

acid sequence specified a protein with similarity to the N-CAM4

family of cell surface proteins (4). At present, localized somaticmutations predicted to inactivate the remaining DCC alÃelehave onlybeen identified in a fraction of colorectal cancers (4, 5). However,recent studies have established that the 29 exons encoding the DCCopen reading frame span nearly 1.4 megabases (5); thus, only a smallsubset of the DCC sequences have been surveyed for mutations.Consistent with its possible function as a tumor suppressor gene, DCCgene expression has been shown to be markedly decreased or absentin the majority of colorectal cancer cell lines and tumors (4, 6). Inaddition, evidence has been presented that inactivation of the DCCgene may also occur in breast, pancreatic, brain, bladder, endometrial,and prostatic cancers, as well as in some leukemias (7-22).

The sequence of human DCC cDNAs isolated from fetal brainpredicts a 1447-amino acid transmembrane protein with four immu-nogiobulin-like and six fibronectin type Ill-like extracellular domains.The 325-amino acid cytoplasmic domain of DCC shares little simi

larity with previously characterized proteins (23). Based on studies ofcells transfected with the full-length DCC cDNA as well as human

tumor cell lines (i.e., the 577MF nonseminomatous germ cell tumorand IMR32 neuroblastoma cell lines) that express DCC transcriptsand protein endogenously, the DCC gene encodes a cell surfaceprotein of about M, 180,000 (23, 24). In addition, data from immu-

nohistochemical and in situ hybridization studies suggest that DCC isexpressed in central and peripheral nervous system axons and in somedifferentiated cell types of the intestine (23, 24).

To date, DCC expression studies have been focused primarily onDCC gene expression. Moreover, the very low abundance of DCCtranscripts in most tissues and cell lines has often necessitated verysensitive assays for expression, such as RT-PCR-based approaches. A

single survey of DCC gene expression in normal mammalian tissueshas been presented. Using an RT-PCR-based approach, DCC tran

scripts were found to be present at low levels in essentially all adultrat tissues with greatest abundance in the brain (25). As noted above,DCC transcripts are markedly reduced or absent in the majority ofprimary colorectal cancers and colorectal cancer cell lines (4, 6;. Amarked reduction in the abundance of DCC transcripts has also beenobserved in human pancreatic (10) and prostatic (15) adenocarcino-

mas, gliomas (19), and leukemias (21, 22). All published studies haveonly examined the abundance of transcripts containing DCC extracellular domain sequences. This may represent a rather limited viewof DCC expression since alternative splicing of transcripts has beenobserved in many transmembrane proteins, including the N-CAMs

(26, 27).In this paper, we describe an analysis of DCC cDNAs from human

fetal brain tissues and the IMR32 human neuroblastoma cell line. Tworegions affected by alternative splicing, one in the extracellular domain and one in the cytoplasmic domain, were identified in cDNAsobtained from IMR32 cells. The relative abundance of alternativelyspliced transcripts was characterized in normal adult tissues from the

4 The abbreviations used are: N-CAM, neural cell adhesion molecule; cDNA, complementary DNA; RT-PCR. reverse transcription-polymerase chain reaction; SDS-PAGE,sodium dodecyl sulfate-polyacrylamide gel electrophoresis; ECL, enhanced chcmilumi-nesence; GST, glutathionine S-transferase.

4493

Research. on December 1, 2020. © 1994 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

DCC TUMOR SUPPRESSOR GENE EXPRESSION

mouse. In addition, we have examined the expression of DCC proteinsin normal adult rodent tissues, as well as human and rodent tumor celllines, and have found that DCC protein expression was greatest inbrain tissues and neural crest-derived cell lines. Our results suggestthat the DCC gene encodes several closely related A/r 175,000-

200,000 glycoproteins that are expressed at very low levels in anumber of normal adult tissues.

MATERIALS AND METHODS

DCC cDNA Cloning and Sequencing. Approximately 1.0 X 10* plaques

from each of two commercially prepared human fetal brain cDNA libraries(Clontech, Palo Alto, CA and Stratagene, La Jolla, CA) were plated and liftedonto Hybond N+ nylon filters (Amersham, Arlington Heights, IL), accordingto the manufacturer's instructions. The filters were hybridized to a '"P-labeled

4.5-kilobase DCC cDNA fragment containing the human DCC open reading

frame [Ref. 23; the DCC cDNA clone was generously provided by Drs. K. R.Cho and B. Vogelstein (Johns Hopkins University School of Medicine)].Hybridization and washing conditions were as described (28) except that thehybridization solution contained 0.5% nonfat dried milk. Clones were isolatedfollowing three rounds of hybridization selection. Their cDNA inserts wereeither recovered and subcloned into the EcoRI site of pBluescript II KS â€”

(Stratagene), or plasmids were excised using helper phage (Stratagene).Unique plasmid subclones were sequenced in their entirety with both externaland internal primers by a modification of the dideoxy chain terminationmethod of Sanger et al. (29) and as described by others (30-33). To charac

terize DCC cDNAs from IMR32 human neuroblastoma cells, total RNA wasisolated, and first-strand cDNA was prepared by priming with random hex-amers (34, 35). The full-length DCC open reading frame and approximately450 base pairs from the 5' untranslated region and 150 base pairs from the 3'

untranslated region were then amplified in overlapping 500-1400-base pair

fragments using PCR (36) and various oligomers derived from the fetal brainDCC cDNA sequence. The PCR products were then either sequenced directlyor following subcloning in Bluescript II KS-. Sequence discrepancies noted

between the IMR32 and human fetal brain DCC cDNA sequences wereconfirmed by independent RT-PCR experiments on IMR32 RNA.

RT-PCR Assay of DCC Gene Expression. Total RNA was prepared from

IMR32 human neuroblastoma cells and adult mouse tissues using the acidguanidinium thiocyanate method (34). First-strand cDNA was prepared fromtotal RNA using random hexamers and the BRL superscript preamplifi-

cation kit (Gibco/BRL Life Technologies, Grand Island, NY). The cDNAwas then amplified by PCR using primer pairs derived from the human DCCcDNA sequence (the location of the primers is indicated relative to theinitiating ATG at base pair 604). Three primers were used to analyze theexpression of extracellular domain sequences (Fig. 2A ): sense primer (for bothextracellular pairs a and b, DCK2834S, 5'-CCCAGACTAACTGCATCAT-CATGAG-3'; antisense for pair a, DCK3088A, 5'-CGAGGTGGGGAAAT-CATCAAGCA-3'; and antisense for pair b, DCK3151A, 5'-CACCTACTG-GTGGGAGCAT-3'.

Similarly, three primers were used to analyze the expression of the cyto-

plasmic domain sequences (Fig. 2B): sense (for both cytoplasmic pair a and b),DCST3090S, 5'-CACAGTGCTGGTAGTGGTCAT-3'; antisense pair a,DCK4498A, S'-AGTTGGTCCTTCACTCACTGACTG-S'; antisense pair b,DCK4504A, 5 ' -TTGGGTTGATGGTCCTTCACTCAC-3 '.

For the experiments shown in Fig. 2, A and B, RT-PCR products were

electrophoresed on 1.2% agarose gels and visualized with UV light followingethidium bromide staining. Their identities were confirmed by Southern transfer and hybridization with a -12P-labeled DCC cDNA probe. For the RT-PCR

experiment shown in Fig. 2C, the PCR products were radiolabeled with32P-dCTP, electrophoresed on a 5% LongRanger gel (AT Biochem, Malvern,

PA), and subjected to autoradiography.DCC Peptides and Recombinant Fusion Proteins and Generation of

Polyclonal Antisera. Three Xenopus DCC (XDCCa) peptide sequences5 that

were highly conserved with human DCC and that were predicted to be

5 W. E. Pierceall, M. A. Reale, A. F. Candia, C. V. E. Wright, K. R. Cho, and E. R.

Fearon. Neural specific expression of a homologue of the deleted in colorectal cancer(DCC) gene in developing Xenopus embryos, submitted for publication.

antigenic were synthesized and purified in the Yale core facility. Followingsynthesis, one extracellular (no. 507, KVEHPDKMANDQGRHGDGGY-WSC) and two cytoplasmic (no. 219, AQQRKKRASHSAGKRKGSQKDL-

RPC and no. 511, GRDSPRQSCQDITPVSHSQSESQC) peptides were eachconjugated to maleimide-activated keyhole limpet hemocyanin (Pierce Chem

ical Co., Rockford, IL). Human fetal brain DCC and XDCCa cDNAs wereused for production of bacterial fusion proteins. DCC cDNA sequences encoding approximately 295 amino acids of the 325-amino acid cytoplasmic

domains of the human and Xenopus DCC products were cloned into the ÃŸamHIsite of the pDS-MCS vector [kindly provided by Dr. C. Dang (Johns Hopkins

University School of Medicine)]. This vector encodes six histidines at theamino terminus of the fusion protein, and the histidines allow for purificationof the fusion protein on a nickel-agarose column (37). Large scale production

of each fusion protein was achieved by expansion of bacterial cells into 1 literof Luria-Bertani broth prior to isopropyl-ÃŸ-D-thiogalactopyranoside induction.The hexahistidine fusion proteins were purified on nickel-agarose columns andthe GST-fusion protein on a glutathione-agarose column. One liter of bacterialcells generally yielded about 1-3 mg of each fusion protein. Eleven rabbit and

10 mouse antisera were prepared according to standard immunization procedures at Hazelton Research Products, Inc. (Denver, PA) and the Yale animalfacility, respectively. Six of the rabbit antisera (nos. 641, 645, 647, 721, 723,and 724) underwent affinity purification on the appropriate antigen-agarose

column. Antisera and their respective antigens were: no. 641, peptide 219; no.645, peptide 507; no. 647, peptide 511; no. 721, hexahistidine human cytoplasmic domain fusion protein; and nos. 723 and 724, hexahistidine Xenopuscytoplasmic domain fusion protein.

Cell Lines. Ten of the human neuroblastoma cell lines [SJNB-1, SJNB3-SNJB8, and SNJB10-SNJB12 were a generous gift from Dr. T. Look (St. JudeChildrens' Research Hospital)]. The keratinocyte and head and neck squamous

cancer cell lines (Hl^neo and FaDuHy, respectively) and glioblastoma lines(Sz, J889H, and H5683) were gifts from Drs. Michael Reiss and JosephPiepmeier (Yale University School of Medicine), respectively. The rat Dun

ning prostate cancer cell lines and a transfected Chinese hamster ovary (CHO)cell line expressing full-length DCC cDNA were gifts from Drs. J. Isaacs and

K. R. Cho (Johns Hopkins University School of Medicine), respectively. Theremainder of the cell lines were obtained from the American Type CultureCollection (Rockville, MD).

DNA Transfections. NIH3T3 and Rail cells expressing full-length DCC

were obtained by transfection with the mammalian expression vector pCMV/DCC-S (23, 38). In brief, the cells were transfected with a mixture of 3 figplasmid DNA and cationic liposomes [Lipofectin reagent (Gibco/BRL-Life

Technologies)]. Approximately 48 h after transfection, selection in G418 at600 fig/ml was begun. Single G418-resistant clones were trypsinized and

established as clonal cell lines. The cell lines were maintained in G418 at400 fig/ml.

Preparation of Tissue and Cell Line Protein Lysates. Cell lines or tissuehomogenates were solubilized in Tris-buffered saline [25 mM Tris(hydroxym-

ethylaminomethane), pH 8] with detergents (1% deoxycholate, 1% NonidetP-40, and 0.1% SDS) and protease inhibitors (50 fig/ml antipain, 5 /J.g/ml

aprotinin, 2 /xg/ml leupeptin, UK)ng/ml phenylmethylsulfonyl fluoride, and 1mM EDTA). Protein concentrations were determined with the BCA proteinassay (Pierce).

Western Blot and Immunoprecipitation Analyses of DCC Expression.Unprocessed cell line/tissue lysates or lysates that had been immunoprecipi-tated underwent SDS-PAGE and Transblot semi-dry transfer (Bio-Rad, Her

cules, CA) to Immobilon (Millipore, Bedford, MA) membranes. The primaryDCC-specific antisera were used at 10-100 ng/ml, and the secondary goat-anti-rabbit or goat-anti-mouse antisera coupled to horseradish peroxidase(Pierce) was used at a 1:20,000 dilution. The antigen-antibody complex was

detected by ECL (Amersham) and subsequent exposure to Hyperfilm (Amersham). For immunoprecipitation, lysates were precleared by incubation withpurified rabbit immunoglobulin and protein A:Sepharose (Immunopure Plus,Pierce, Rockford, IL). The supernatant was then incubated with the purifiedDCC-specific antisera (10 fig/ml). DCC-specific immunoprecipitates were

recovered following incubation with protein A:Sepharose. They were thenwashed four times with Tris-buffered saline/1% Triton, resuspended in Lae-mmli's sample buffer, and then subjected to SDS-PAGE. DCC protein in the

immunoprecipitates was detected by the ECL-immunoblot assay described

4494




above. For the studies of DCC glycosylation, tunicamycin (Sigma) was storedat â€”20Â°Cand dissolved in aqueous solution immediately prior to use.

RESULTS

Sequence Analysis of DCC cDNAs from Human Fetal Brainand the IMR-32 Neuroblastoma Cell Line. Screening of 1.0 X IO6

clones from each of two human fetal brain cDNA libraries with aDCC cDNA probe identified more than 20 unique clones. ThesecDNA clones were assembled into a consensus DCC cDNA sequencespanning more than 6.0 kilobases. Based on the genomic structure ofthe DCC gene (5), all observed sequence differences between thevarious isolated fetal brain cDNA clones were thought to have arisenfrom the generation of chimeric cDNA clones during library construction rather than through alternative splicing of DCC transcripts. Thesequence of the 4341-base pair open reading frame identified by

Hedrick et al. (23) was confirmed. The predicted open reading frameencodes a protein with four immunoglobulin-like domains, six fi-bronectin type Ill-like domains, a hydrophobic transmembrane region,

and a cytoplasmic domain of about 325 amino acids that does notshow similarity to other characterized proteins (Fig. 1). There are sixpotential sites for N-linked glycosylation in the extracellular domainof the molecule (39). No arginine-glycine-aspartate (ROD) tripeptide sequence was noted in the fibronectin type Ill-like extracellular

domains. The DCC cytoplasmic domain contains several potentialphosphorylation sites: there are four tyrosines; two serines are potential targets for protein kinase C (S-X-R/K) (40); and there are tenserines that may be phosphorylated by casein kinase II (S-X-X-D/E;

Ref. 41).Previous studies have demonstrated that IMR32 human neuroblas

toma cells synthesize DCC transcripts and protein (24). To characterize the sequence of DCC transcripts in the IMR32 cell line, anRT-PCR-based approach was used to obtain DCC cDNAs for thefull-length open reading frame and some of the 5' and 3' untranslated

regions. Sequence analysis of the cDNA products obtained from

IMR32 cells through RT-PCR studies revealed three differences com

pared to the DCC cDNA sequence from human fetal brain. Alldifferences were confirmed in products obtained from at least twoindependent RT-PCR experiments. One of the changes was a single

nucleotide substitution at codon 951 in the sixth fibronectin type IIIdomain. This change resulted in a substitution of leucine (codon TTG)for phenylalanine (TIT). Given the conservative nature of this aminoacid substitution and the fact that a highly conserved Xenopus homologue of DCC codes for leucine at this position, this sequence difference is likely to be a polymorphism. The other two sequence differences appeared to reflect alternative splicing of DCC sequences inexons 17 and 26.

Sequence analysis of IMR32 RT-PCR products containing fibronectin type Ill-like domains four and five consistently demon

strated an internal deletion of exactly 60 base pairs or 20 codons, ascompared to the corresponding cDNA sequence obtained from fetalbrain DCC cDNAs. These transcripts presumably arise through theuse of an alternative splice acceptor site located within exon 17(Fig. 1). The remaining sequence difference noted in some RT-PCR

products of the IMR32 DCC cytoplasmic domain was an internaldeletion of 6 base pairs. This difference appeared to result from useof an alternative splice donor site at the 3' end of exon 26 and resulted

in the deletion of two codons and an amino acid substitution at athird (amino acids RSQSV versus RTâ€”V; Fig. 1). This alternative

splice would be predicted to remove a potential casein kinase IIphosphorylation site.

Expression and Alternative Splicing of DCC Transcripts inAdult Mouse Tissues. As described above, previous RT-PCR studies

using extracellular domain primers have suggested that DCC transcripts are present at very reduced levels in many normal adult rattissues with greatest abundance in brain tissue (5, 25). Therefore, wechose to use RT-PCR studies to characterize the pattern of expression

and alternative splicing of DCC transcripts in adult mouse tissues.Two primer sets derived from the sequences encoding the fourth and

ttcatttggtgggtttttaaacccag AT CCC ACT GAC CCA OTT GAT TAT TAT CCT TTGDPTDPVDYYPL

830 Y 840CTT GAT GAT TTC CCC ACC TCG GTC CCA GAT CTC TCC ACCLDDFPTSVPDLST

1295GGA GCA GGA AGA ACT CAG gtaatgcat_G A G R S Q

EXONS

PROTEINIg-Like Domains FN Type Ill-Like

DomainsTM Cytoplasmic

Fig. 1. DCC gene exon map and alternative splice sites. The predicted protein product has four immunoglobulin-like (Ig-like) domains, six fibronectin (FN) type Ill-like domains,a transmembrane (TM) region, and approximately 325 amino acids of cytoplasmic sequence. The 29 exons encoding the 4341-base pair open reading frame of DCC are indicated, withtheir size proportional to the amount of cDNA sequence encoded. Analysis of the 5' and 3' untranslated regions has not been completed; thus, exons 1 and 29 are indicated by filledboxes. Exons 17 and 26 are affected by alternative splicing. The splice acceptor sequence at the 3' region of intron 16 is shown (lower case), as are codons 819 to 842 of the DCC

cDNA sequence (upper case), with the corresponding amino acids shown beneath. The alternative splice acceptor site within exon 17 is indicated by the arrowhead. For exon 26, codons1294 to 1299 of the DCC sequence are indicated with the corresponding amino acids beneath. The alternative splice donor site for exon 26 is indicated by an arrowhead, and intron26 sequences are indicated in lower case (see text).

4495




b a baba_1 2 3 4 5 6__7 8 9 10 11

Mabababababababababab a b M

B1018-

_1 __ 2__ 3__4__5__6 __ 7 8 9 10 11Mabababababababababab a b M

Mababababababab ab

220/201154/134-

Fig. 2. RT-PCR analysis of DCC transcripts in IMR32 cells and normal mouse tissues. First-strand cDNA was prepared from total RNA of various mouse organs/tissues with randomhexamers and reverse transcriptase. PCR was carried out with oligomers derived from the human DCC sequence. The RT-PCR products were then electrophoresed on 1.2% agarose

gels, and the products were visualized by UV light following ethidium bromide staining. A, analysis of DCC extracellular domain transcripts. Oligomers were: Lane a, DCK2834S andDCK3088A; Lane b, DCK2834S and DCST315IA. DCK3151A is derived from sequences 3' to the exon 17 alternative splice site, and DCK3088A is contained within the 60-nucleotide

alternatively spliced region (see Fig. 1 and text). Lane numbers/samples: 1, no template; 2, nonspecific RNA control; 3, IMR32 human neuroblastoma cells; 4, stomach; 5, smallintestine; 6, lung; 7, liver; 5, kidney; 9, colon; 10, brain; 11, bladder. All tissues were obtained from mouse. Size markers (M) in base pairs are indicated. Oligomer pair "a" detectsa 301-base pair product in 1MR32 and all mouse tissues; oligomer pair "b" detects a 341-base pair product in all mouse tissues, as well as a 281-base pair alternatively spliced product

in IMR32 (Lane 3b) and mouse bladder (Lane lib). B, analysis of DCC cytoplasmic domain transcripts. The oligomers used were; Lane a, DCST3930S and DCK4498A; Lane b,DCST3930S and DCK4504A. The 3' end of DCK4498A is contained in the alternatively spliced sequences of exon 26, and this oligomer will not amplify cDNAs lacking these

sequences. In contrast, DCK4504A will amplify cDNAs with or without the six alternatively spliced nucleotides. The analysis was carried out with cDNA prepared from RNA isolatedfrom tissues of two different mice (upper and lower panels). Samples/lane numbers are as indicated above in (A). DCC-specific RT-PCR products of about 600 base pairs were detectedin all mouse tissues, except the colon RNA sample from the first mouse (upper) did not amplify efficiently with either oligomer pair "a" or "b" or with control oligomers (not shown).

C, alternative splicing of cytoplasmic domain transcripts in IMR32. Samples: 1. IMR32; 2, mouse liver; 3, mouse colon; 4, mouse stomach. RT-PCR analysis was carried out asdescribed above for (B), except the products were labeled with [32P]dCTP during the PCR and were then analyzed on a sequencing gel. The autoradiograph exposure time for Lanes

Ja and Ib was approximately 8-fold less than that for the samples in Lanes 2-4.

fifth fibronectin type Ill-like domains of human DCC were chosen.

The primers were derived from cDNA sequences in different DCCexons and would not be predicted to amplify genomic DNA orunspliced transcripts. Based on the human DCC sequence, primer set"a" was predicted to yield an RT-PCR product of 301 base pairs from

DCC transcripts containing the 60-nucleotide alternatively spliced

region but would not amplify DCC transcripts lacking these sequences. Primer set "b" was predicted to generate a product of 341base pairs if the more 5' splice acceptor at exon 17 was used (60-

nucleotide sequence present) or a product of 281 base pairs if the more3' splice acceptor was used (Fig. 1). In studies of IMR32, primer set"a" generated a 301-base pair DCC product, while primer set "b"

generated both 341- and 281-base pair products (Fig. 2A, Lanes 3aand 3b). As predicted from our prior RT-PCR, cloning, and sequenc

ing studies of DCC transcripts in IMR32, transcripts lacking the60-nucleotide region predominated in IMR32 (i.e., the 281-base pair

product; Fig. 2A, Lane 3b). DCC transcripts containing extracellulardomain sequences were detected by RT-PCR analysis in all mouse

tissues studied. Bladder was the only mouse tissue observed to havealternative splicing of the sequences located between fibronectin typeIll-like domains 4 and 5 (Fig. 2A, Lane lib). In the bladder, therelative levels of the 341- and 281-base pair RT-PCR products were

nearly equivalent, suggesting that about 50% of the transcripts werealternatively spliced for exon 17 sequences.

We also analyzed the tissue distribution and abundance of transcripts containing DCC cytoplasmic domain sequences. As describedabove, we noted alternative splicing affecting the 3' region of DCC

exon 26 in IMR32. The alternative splice affects the presence of sixnucleotides in the DCC mRNA (Fig. 1). Two primer pairs weresynthesized to determine if these alternatively spliced transcripts werepresent in adult mouse tissues. The 3'-end of the antisense primer forcytoplasmic domain primer set "a" was derived from the sequencesaffected by the alternative splice. Thus, cytoplasmic primer set "a"

would only amplify DCC transcripts containing the six nucleotides. Incontrast, cytoplasmic primer set "b" was predicted to amplify tran

scripts containing the six nucleotides, as well as those lacking thesequences. As predicted from our prior RT-PCR cloning and sequencing studies of DCC transcripts in IMR32, both primer sets "a" and "b"

generated the expected RT-PCR products from IMR32 (Fig. 2B,Lanes 3a and 3b). Both primer sets generated DCC-specific productsin RT-PCR studies of the mouse tissues (Fig. 25). Most tissues

studied appeared to have reduced amounts of transcripts containingthe extra six nucleotides from exon 27 (primer set "a") compared to

the total abundance of DCC cytoplasmic domain transcripts (detectedby primer set "b").

In order to provide an estimate of the relative abundance of thesetwo alternatively spliced transcripts, we carried out RT-PCR studies inthe presence of [32P]dCTP using two sets of primers. To distinguish

4496




NIH-3T3 CHO

Fig. 3. DCC protein expression in transfecled celllines. Lysales (approximately 100 fig of total protein/lane) from NIH3T3 and CHO cells (A ) and Railcells (B) were studied by SDS-PAGE on 8% gelsand an ECL-immunoblot analysis with the indicatedprimary rabbit antisera: IgG. control rabbit immu-noglobulin; nos. 721, 723, and 724, DCC-specificantisera (see "Materials and Methods"). DCC and

the relative mobility of protein markers in kilodal-

tons are indicated.

200-

Rat1

>DCC200-

97.4-

69-

46-

46-

PCR products differing in size by only six base pairs, we analyzedthe RT-PCR products on a sequencing gel. In IMR32, the relative

abundance of the two alternatively spliced products was aboutequal (Fig. 2C, Lane Ib). Previous studies have suggested that therelative abundance of the two spliced products was nearly equal innormal fetal brain tissues.6 Although in all adult mouse tissues

studied transcripts containing the six nucleotides were present(Fig. 2B), transcripts lacking the six nucleotides were of muchgreater abundance (note only the smaller product is detected byprimer set "b" in Fig. 2C, Lanes 2b, 3b, and 4b).

It should be noted that, despite our attempts to control for thequantity of input RNA in each experiment as well as the efficiency ofcDNA synthesis [using oligomers derived from murine glyceralde-hyde 3-phosphate dehydrogenase gene sequences (data not shown)],the RT-PCR assay used for DCC gene expression is clearly a semi-quantitative one. In addition, we found that the levels of DCC-specificRT-PCR products generated from any given tissue showed some

variation in independent experiments (Fig. 25 and data not shown).Nevertheless, our studies suggested that DCC transcripts were mostabundant in adult mouse brain, stomach, liver, and colon tissues(Fig. 2, A and 5).

Expression of the DCC Protein in Cell Lines and NormalRodent Tissues. A series of rabbit and mouse polyclonal antiserawere generated by immunization with DCC peptide antigens andrecombinant bacterial fusion proteins containing human or XenopusDCC cytoplasmic domain sequences (see "Materials and Methods").

Our initial studies of DCC protein expression were performed oncell lines that had been transfected with a mammalian expressionconstruct encoding human DCC. DCC-expressing cell lines de

rived from transfection of CHO cells and murine NIH3T3 and Ratifibroblast cells were studied. Although the open reading frame ofthe DCC cDNA predicts a protein of approximately M, 155,000,

1K. R. Cho, personal communication.

immunoblot studies revealed that the predominant form of theDCC protein in the transfected lines migrated in a rather broadregion with an apparent molecular weight of about 180,000-

210,000 (Fig. 3). A doublet was often seen with several of theantisera, such as no. 723 (Fig. 3).

A sizable number and variety of cell lines (37 in total), the majorityof them derived from human tumors, were studied for endogenousDCC protein expression by immunoblotting or a combined immuno-

precipitation and immunoblotting assay. Three of 11 human neuroblastoma cell lines studied (IMR-32, SJNB-5, and SJNB-6) demon

strated an approximately M, 180,000 protein that was detected byseveral anti-DCC antisera but not by control rabbit immunoglobulin

(Fig. 4). The doublet pattern seen in the transfected lines was not seenin these lines. DCC protein expression was not detectable by ourassays in any of the other cell lines studied (Table 1).

Several normal mouse and rat tissues, including brain, lung, stomach, liver, pancreas, kidney, small intestine, colon, and bladder, werealso examined for DCC protein expression by immunoblotting or acombined immunoprecipitation and immunoblotting assay. The DCCprotein was most abundant in mouse and rat brain, although it wasalso detectable in mouse bladder (Fig. 5, A-C). Using a number ofdifferent DCC-specific antisera, a Mr 175,000-190,000 doublet was

detected in both mouse and rat brain. Several more rapidly migratingproteins were also detected with several of the DCC-specific antisera

in brain tissue, particularly mouse brain (Fig. 5A and data not shown).These smaller immunoreactive species are likely to represent DCCdegradation products generated during sample preparation/processingor perhaps cross-reactive proteins. Nevertheless, it is possible that

some of the forms arise from alternative splicing of DCC mRNA orfrom in vivo processing of the DCC protein(s).

A semiquantitative estimate of the sensitivity of the immunoblotassay was obtained through immunoblotting studies of extracts thathad been supplemented with varying amounts of a purified recombinant bacterial fusion protein containing the DCC cytoplasmic domain

4497



DÂ« MAÂ«Â«SU'HRISSOR (.1 NI lAI'RISSION

Fig. 4. DCC protein expression in human neuro-blastoma cell lines. DCC expression was studiedusing a combined immunoprecipitation and ECL-immunoblot analysis. Protein lysates from humanneuroblastoma cell lines, IMR32 (A, 600 /ig of totalprotein) and SJNB-5 and SJNB-6 (B. 720 /ig totalprotein from each line), were subjected to immunoprecipitation with control rabbit immunoglobulin(IgG) or DCC-specific antisera (nos. 641, 645,and 721: see "Materials and Methods"). The immu-noprecipitales were then analyzed by SDS-PAGEon 8% gels and ECL-immunoblotting using a mousepolyclonal antisera raised against the His,,-humanDCC fusion protein (A) or rabbit antiserum no.723 (B).

A IMR-32

IgG 641 645 721

200-

SJNB-5 SJNB-6IgG 645 721 IgG 645 721

97.4-

-DCC217-

130-

72-

42.2-

-DCC

(GST-DCC). In our immunoblotting assay with DCC-specific anti-

serum no. 723, we could detect as little as 156 pg of the recombinantDCC fusion protein admixed with 250 p.g of total mouse brain protein(Fig. 6, Lane 5). Although the signals obtained by ECL-immuno

blotting do not often have a wide range over which they are linearlyrelated to the quantity of protein present, the signal of endogenousDCC in 250 /xg of total mouse brain protein is roughly equivalent tothat of about 1 ng of the DCC fusion protein (Fig. 6 and data not

Table 1 Expression of Â¡heDCC protein in cell lines

Cell line type and identity"No. of

samplesDCC

protein

Neuroblastoma IMR32. SJNB-5. SJNB-6 3 +Neuroblastoma SJNB-1,3,4,7,8,10,11,12 8Glioblastoma J889H, H5683, Sz 3Breast HBL-100 1Breast cancer BT-20. MCF7 2Colorectal cancer HT-29, Caco-2, Lovo, Lim 1863 4Pancreatic cancer 5

HS700T, H5766T. MIA PaCa-2, BxPc-3, AsPC-1Melanoma 5

SKMEL3. SKMEL24, SKMEL28, SKMEL31, G3612Head and neck squamous cancer FaDuHy 1Keratinocyte HKcneu 1Proslatic cancer AT-2, AT-3, AT-6 3Phcochromocytoma PC12 1

Total 37 3+34-

" Human cell lines with the exception of prostate cancer and pheochromocytoma cell

lines.h Detectable (+) or undetectable (-) by immunoblot or combined immunoprecipita

tion and immunoblot analysis.

shown). Thus, in mouse brain (the most abundant tissue source forDCC studied in the adult mouse), DCC protein(s) represents less than0.0004% of the total protein.

Posttranslational Modification of DCC. As described above, wenoted a discrepancy between the apparent molecular weight of DCCin the various tissues and cell lines studied and that predicted by thecDNA sequence. This result, as well as the multiple Mr 175,000-

210,000 species detected in different tissues and cell lines, suggestedthat DCC might be affected by a posttranslational modification, suchas glycosylation. In addition, it was noted that the DCC extracellulardomain sequences contain six potential AMinked glycosylation sites.To determine if DCC was modified by /V-linked glycosylation,

NIH3T3 cells that had been transfected with a DCC expressionconstruct and that were expressing DCC were treated with tunicamy-cin. The antibiotic tunicamycin is known to inhibit ^-linked glycosylation by blocking the formation of the dolichol phosphate-linked

oligosaccharide donor (42). Following this treatment, the apparentmolecular weight of the most rapidly migrating form of DCC in theNIH3T3 cells was decreased from approximately M, 200,000 to170,000 (Fig. 7). All concentrations of tunicamycin used failed toinhibit fully the production of at least some of the larger species.Nevertheless, the tunicamycin studies suggest that DCC is modifiedby glycosylation at W-linked sites.

DISCUSSION

The DCC gene has been implicated as a tumor suppressor gene onthe basis of frequent allelic losses in colorectal cancer (2-4), identi-




MOUSE BRAINBRAT

DDAII

200-

tO <0 <Ã» I-. N N. 0. !=

3-DCC217-

130-

72-

30-

MOUSEBLADDER

200-

97.4-

69-

46-

-DCC

Fig. 5. DCC expression in normal rodent tissues. Combined immunoprecipitation and ECL-immunoblot analysis of DCC expression in mouse brain (A), rat brain (B), and mouse

bladder (C). Lysates from the tissues were immunoprecipitated with control antisera (IgG, purified rabbit immunoglobulin; PB #4, preimmune sera from mouse #4) or the indicatedrabbit (nos. 641, 645, 647, 723, 721, and 724) or mouse (IM #4) DCC-specific antisera. A, 320 /xg total protein lysate/immunoprecipitating antisera; B, 325 jig total proteinlysate/immunoprecipitating antisera; C, 650 /xg total protein lysate/immunoprecipitating antisera. Immunoprecipitates were then electrophoresed on 8% gels and analyzed by anECL-immunoblot analysis with a polyclonal mouse antisera raised against the His6-human DCC fusion protein (A) or a rabbit polyclonal antiserum no. 723. Relative mobility of

molecular weight markers (in kilodaltons) and DCC protein are indicated.

fiable somatic mutations in the gene in a subset of cancers, and themarked reduction or absence of its transcripts in the majority ofprimary colorectal cancers and cell lines (4-6, 23). It has also been

implicated in the pathogenesis of a variety of other tumors by similarstudies (7-22). At present, however, little is known about the specific

genetic alterations that may inactivate DCC in colorectal cancers andother tumors or the means by which DCC inactivation may contributeto the cancer cell phenotype. Moreover, the role of the DCC gene inthe regulation of normal cell growth and/or differentiation remainspoorly understood. To address these and other questions about theDCC gene, we have begun in these studies to characterize DCCexpression and alternative splicing in normal and neoplastic tissues.

Based upon comparison of the sequence of DCC cDNAs fromIMR32 human neuroblastoma cells to cDNAs isolated from fetalbrain, we have identified two alternatively spliced regions. Using anRT-PCR-based approach, we have demonstrated that DCC transcripts

are expressed at low levels in many normal adult tissues of the mouse.Alternative splicing of DCC transcripts was also noted in some mousetissues. Although DCC transcripts were detected in all normal mousetissues studied, DCC protein was only detected by our immunoblot-

ting or combined immunoprecipitation and immunoblotting studies inmouse and rat brain and mouse bladder. Of the many human androdent tumor cell lines studied, only three neuroblastoma lines haddetectable DCC protein expression in our assays. The failure to detectDCC protein expression in the many tumor cell lines studied mayreflect the relative insensitivity of our assays. Alternatively, the absence of DCC protein expression in some cell lines may be attribut-

1 2 3i

4i

6i

200 â€”-^ ^

Â»â€¢Â»â€¢â€”Â»â€¢-â€¢Â»-.â€” DCC

97 â€”

69 â€”

46 â€”

-Â«-GST-DCC

Fig. 6. Sensitivity of the ECL-immunoblot analysis for DCC expression. Varying amountsof a recombinant GST-DCC fusion protein were added to lysates containing approximately250 ng of total mouse brain protein. The samples were then electrophoresed and studied byECL-immunoblot analysis with anti-DCC antiserum no. 723. The amounts of GST-DCC

fusion protein added were: Lane 1, 40 ng; Lane 2, 10 ng; Lane 3, 2.5 ng; Lane 4, 625 pg; Lane5, 156 pg; Lane 6, 40 pg. Relative mobility of protein markers in kilodaltons, endogenousDCC (DCC), and the GST-DCC recombinant fusion protein are indicated.

4499




II I

3 4 5

200- DCC

97 -

69 -

scripts also bears some resemblance to alternative splicing of theN-CAM family member LI. LI transcripts are affected by a 12-base

pair deletion in the cytoplasmic domain that results in the loss of apotential casein kinasc II phosphorylation site (46). Phosphorylationof the cytoplasmic domain of N-CAM and LI has been demonstrated,

and in all cases it has been serine phosphorylation (47, 48). Given theabundance of potential phosphorylation sites in the DCC cytoplasmicdomain (e.g., 10 casein kinase II consensus sites) and evidence thatalternative splicing affects one of these sites, it is possible thatphosphorylation may play a role in the biological function of the DCCcytoplasmic domain.

Additional studies will be necessary to gain a more completeunderstanding of the expression and alternative splicing of the DCCgene in normal and neoplastic tissues. Nevertheless, the reagents andobservations described here should prove useful for generating furtherinsights into the role of the DCC gene in normal cell growth anddifferentiation as well as tumorigenesis.

46 -

Fig. 7. DCC is a glycoprotein. Tunicamycin-trealment of a DCC-transfected NIH3T3cell line. Cells were incubated in complete media with varying concentrations of lunica-mycin for 24 h. Lysates were then obtained, electrophoresed, and subjected to ECL-

immunoblot analysis with rabbit antiserum no. 723. Lane 1, no treatment; Lane 2, 1fig/ml; Lane 3, 5 Â¿Â¿g/ml;Lane 4, 10 ng/ml; Lane 5, 20 fig/ml. Mobility of DCC proteinspecies and protein markers in kilodaltons are indicated.

able to mutational inactivation of the DCC gene. Finally, loss of DCCexpression may be selected for during the generation/establishment ofpermanent cell lines from primary tumor tissues or may simply reflecta change in expression that results from the I'Mvitro growth of cells.

Our studies of DCC expression in human neuroblastoma cell lines,however, suggest that the absence of detectable DCC expression in 8of the 11 cell lines studied is not likely to reflect merely a tissueculture artifact.

Several similarities between DCC and previously characterizedmembers of the N-CAM family can be noted (27). DCC has bothmultiple immunoglobulin-like and fibronectin type Ill-like extracel

lular domains. The highest levels of DCC expression appear to be inthe brain and neural crest-derived tissues, and DCC is a glycoprotein.

In addition, alternative splicing of DCC transcripts affects both extracellular and cytoplasmic domain sequences. Alternative splicing ofN-CAM transcripts in sequences encoding fibronectin type Ill-like

domains has been noted previously, and, as we observed for the DCCextracellular domain sequences, it occurs in a tissue-specific fashion(26, 27, 43, 44). In particular, a muscle-specific domain (MSD) exonlies between N-CAM exons 12 and 13 and encodes an additional 37amino acids that lie amidst fibronectin type Ill-like domains. Thisregion is known to undergo significant O-linked glycosylation (45),and it has been speculated to introduce a "hinge" into the molecule

(27, 43). Although the 20-codon region in DCC that is alternatively

spliced is not similar at the amino acid level to the predicted productof the MSD exon, their similarity in location might reflect somefunctional similarity.

Another alternative splice site in the DCC gene affects six nucle-

otides in the sequences encoding the cytoplasmic domain. DCC transcripts lacking these sequences have a loss of two codons and anamino acid substitution at a third codon. with a resultant loss of apotential serine phosphorylation site for casein kinase II. The alternative splicing of the cytoplasmic domain sequences in DCC tran-

ACKNOWLEDGMENTS

We thank Drs. Kathleen R. Cho and Bert Vogelstein for providing DCCplasmids and cDNA sequence information prior to publication and Drs. Kathleen R. Cho, John Issacs, A. Thomas Look, Michael Reiss, and JosephPiepmeier for generously providing cell lines.

REFERENCES

1. Cavenee, W. K., Dryja, T. P., Phillips, R. A., Benedict, W. F., Godbout, T., Gallic.B.. Murphree. A. L., Strong, L. C., and White, R. Expression of recessive alÃelesby chromosomal mechanisms in retinoblastoma. Science (Washington DC), 288:501-503, 1983.

2. Vogelstein. B.. Fearon, E. R.. Kern. S. E., Hamilton, S. R., Preisinger, A. C.,Nakamura. Y., and White, R. AJlelotype of colorectal carcinomas. Science (Washington DC), 244: 207-211, 1989.

3. Vogelstein, B., Fearon, E. R., Hamilton, S. R., Kern, S. E., Preisinger, A. C., Leppert,M.. Nakamura, Y., White. R., Smits. A. M. M., and Bos, J. L. Genetic alterationsduring colorectal lumor development. N. Engl. J. Med., 319: 525-532, 1988.

4. Fearon, E. R.. Cho, K. R., Migro. J. M.. Kern. S. E.. Simons. J. W.. RuppeÂ«,J. M.,Hamilton. S. R.. Preisingcr. A. C.. Thomas. G.. Kinzler. K. W., and Vogelstein. B.Identification of a chromosome 18q gene that is altered in colorectal cancers. Science(Washington DC), 247: 49-56, 1990.

5. Cho. K. R., Oliner, J. D., Simons, J. W., Hcdrick, L., Fearon, E. R.. Preisinger, A. C.,Hedge, P.. Silverman, G. A., and Vogelstcin. B. The DCC gene: structural analysisand mutations in colorectal carcinomas. Genomics, 19: 525-531, 1994.

6. Kikuchi-Yanoshita. R., Konishi, M. Fukunari, H.. Tanaka, K.. and Miyaki. M. Lossof expression of the DCC gene during progression of colorectal carcinomas in familialadenomatous polyposis and non-familial adenomatous polyposis patients. CancerRes., 52: 3801-3803. 1992.

7. Cropp, C. S., Lidereau, R., Campbell, G., Champene, M. H., and Callahan, R. Lossof hcterozygosity on chromosomes 17 and 18 in breast carcinoma: two additionalregions identified. Proc. Nati. Acad. Sci. USA, 87: 7737-7741, 1990.

8. Devilee, P.. van Vieil, M.. Kuipers-Dijkshoorn, N., Pearson, P. L., and Cornelisse,

C. J. Somatic genetic changes on chromosome 18 in breast carcinomas: is the DCCgene involved? Oncogene, 6: 311-315, 1991.

9. Thompson, A. M., Morris. R. G., Wallace, M., Wyllie, A. H., Steel, C. M., and Carter,D. C. AlÃeleloss from 5q21 (APC/MCC) and 18q21 (DCC) and DCC mRNAexpression in breast cancer. Br. J. Cancer, 68: 64-68, 1993.

10. Hohne, M. W., Halatsch, M-E., Kahl, G. F., and Weinel, R. J. Frequent loss ofexpression of the potential tumor suppressor gene DCC in ductal pancreatic adeno-carcinoma. Cancer Res., 52: 2616-2619, 1992.

11. Uchino, S., Tsuda. H., Noguchi, M., Yokota, J.. Terada, M., Saito, T.. Kobayashi, M.,Sugimura, T., and Hirohashi, S. Frequent loss of heterozygosily at the DCC locus ingastric cancer. Cancer Res., 52: 3099-3102, 1992.

12. Huan, Y., Boynton, R. F., Blount, P. L., Silverstein, R. J., Yin, J., Tong, Y.,McDaniel, T. K., Newkirk, C., Resau, J. H., Sridhara, R., Reid, B. J., and Meltzer. S.Loss of heterozygosity involves multiple tumor suppressor genes in human esopha-geal cancers. Cancer Res.. 52: 6525-6530, 1992.

13. Imamura. T., Arima. T., Kalo. H.. Miyamoto, S.. Sasazuki. T., and Wake, N.Chromosomal deletions and K.-ras mutations in human endometrial carcinomas. Int.J. Cancer, 51: 47-52, 1992.

14. Bergerhei, U., Nordenskjold, M., and Collins, V. P. Deletion mapping in human renalcell carcinoma. Cancer Res., 49: 1390-1396, 1989.

15. Gao, X., Honn, K. V., Grignon, D.. Sakr, W., and Chen. Y. 0. Frequent loss ofexpression and loss of heterozygosity of the putative tumor suppressor gene DCC inprostatic carcinomas. Cancer Res., 53: 2723-2727, 1993.

16. Wu, S-0., Storer. B. E., Bookland, E. A.. Klingelnutz, A. J., Gilchrist, K. W.,

Meisner, L. F., Oyasu, R., and Reznikoff, C. A. Nonrandom chromosome losses in

4500




slepwise neoplastia Iransformation in vitro of human urocpithelial cells. Cancer Res.,51: 3323-3326, 1991.

17. Chenevix-Trench, G., Leary, J., Kerr, J., Michel, J., Kefford, R., Hurst, T., Parsons,P. G., Friedlander, M., and Khoo, S. K. Frequent loss of heterozygosity on chromosome 18 in ovarian adenocarcinoma which does not always include the DCC locus.Oncogene, 7: 1059-1065, 1992.

18. Yamaguchi, T., Toguchida, J., Yamamuro. T., Kotoura, Y., Takada, N., Kawaguchi,N., Kaneko, Y., Nakamura, Y, Sasaki, M. S., and Ishizaki, K. Allelotype analysis inosteosarcomas: frequent alÃeleloss on 3q, 13q, 17p and 18q. Cancer Res., 52:2419-2423, 1992.

19. Scheck, A. C., and Coons, S. W. Expression of the tumor suppressor gene DCC inhuman gliomas. Cancer Res., 53: 5605-5609, 1993.

20. Klingelhutz, A. J., Smith, P. P., Garrelt, L. R., and McDougall, J. K. Alteration of theDCC tumor-suppressor gene in tumorigenic HPV-18 immortalized human keratino-cytes transformed by nitrosomethylurea. Oncogene, 8: 95-99, 1993.

21. Porfiri, E., Seeker-Walker, L. M., Hofflirand, A. V., and Hancock, J. F. DCC tumorsuppressor gene is inactivated in hÃ©matologiemalignancies showing monosomy 18.Blood, 81: 2696-2701, 1993.

22. Miyake, K., Inokuchi, K., Dan, K., and Nomura, T. Expression of the DCC gene inmyelodysplastic syndromes and overt leukemia. Leukemia Res., 17: 785â€”788,1993.

23. Hedrick, L., Cho, K. R., Fearon, E. R., Wu, T-C, Kinzler, K. W., and Vogelstein, B.

The DCC gene product in cellular differentiation and colorectal tumorigenesis. Genes& Dev., 8: 1174-1183, 1994.

24. Hedrick, L., Cho, K. R., Boyd, J., Risinger, J., and Vogelstein, B. DCC: a tumorsuppressor gene expressed on the cell surface. Cold Spring Harbor Symp. Quant.Biol., 57: 345-351, 1992.

25. Migro, J. M., Cho K. R., Fearon, E. R., Kern, S. E., Ruppert, J. M., Oliner, J. D.,Kinzler, K. W., and Vogelstein, B. Scrambled exons. Cell, 64: 607-613, 1991.

26. Cunningham, B. A., Hemperly, J. J.. Murray, B. A., Prediger, E. A., Brackenbury, R.,and Edelman, G. M. Neural cell adhesion molecule: structure, immunoglobulin-likedomains, cell surface modulation, and alternative RNA splicing. Science (WashingtonDC), 236: 799-805, 1987.

27. Edelman, G. M., and Crossin, K. L. Cell adhesion molecules. Implications for amolecular histology. Annu. Rev. Biochem., 60: 155-190, 1991.

28. Vogelstein, B., Fearon, E. R., Hamilton, S. R., Preisinger, A. C., Willard, H. F.,Michelson, A. M., Riggs, A. D., and Orkin, S. H. Clonal analysis using recombinantDNA probes from the X-chromosome. Cancer Res., 47: 4806-4813, 1987.

29. Sanger, F., Nicklen, S., and Coulsen, A. R. DNA sequencing with chain terminatinginhibitors. Proc. Nati. Acad. Sci. USA. 74: 5463-5467, 1977.

30. Tabor, S., and Richardson, C. C. DNA sequence analysis with a modified bacteri-ophage T7 DNA polymcrase. Proc. Nati. Acad. Sci. USA, 84: 4767-4771, 1987.

31. Kraft, R., Tardiff, J. Krauter, K. S., and Leinwand, L. A. Using miniprcp plasmidDNA for sequencing double-stranded templates with sequenase. Biotechniques, 6:544-546, 1988.

32. Prober, J. M., Traninor, G. L., Dam, R. J., Hobbs, F. W., Robertson, C. W., Zagursky,R. J., Cocuzza, A. J., Jensen, M. A., and Baumesticr, K. A system for rapid DNAsequencing with fluorescent chain-terminating dideoxynucleotides. Science (Washington DC), 2.?Â«.'336-341, 1987.

33. Brumbaugh, J. A., Middendorf, L. R., Grone, D. L., and Ruth, J. L. Continuous,

on-line DNA sequencing oligodeoxynucleotide primers with multiple fluorophores.Proc. Nati. Acad. Sci. USA, 85: 5610-5614, 1988.

34. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem., 762:156-159, 1987.

35. Noonan, K. E., and Roninson, I. B. mRNA phenotyping by enzymatic amplificationof randomly primed cDNA. Nucleic Acids Res., 16: 10366. 1988.

36. Saiki, R. K., Gelfano, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., MullÃs,K. B., and Erlich, H. A. Primer-directed enzymatic amplification of DNA with athermostable DNA polymerase. Science (Washington DC), 239: 487-491, 1988.

37. Stuber. D., Matile, H., and Garotta. G. System for high-level production in

Escherichia coli and rapid purification of recombinant proteins: application ofepitope mapping, preparation of antibodies and structure-function analysis. In: I.Lefkovits and B. Pernis (eds.), Immunological Methods, Vol. IV, pp. 121-152.

New York: Academic Press, 1990.38. Pierceall, W. E., Cho, K. R., Getzenberg. R. H., Reale, M. A., Hedrick. L.. Vogelstcin,

B., and Fearon, E. R.. The deleted in colorectal cancer (DCC) tumor suppressor geneproduct stimulates neurite outgrowth in rat PCI 2 pheochromocytoma cells. J. CellBiol., 124: 1017-1027, 1994.

39. Kornfield, R., and Kornfield, S. Assembly of asparagine-linkcd oligosaccharides.Annu. Rev. Biochem., 50: 631-664, 1985.

40. Kemp, B. E., and Pearson. R. B. Protein kinasc recognition sequence motifs. TrendsBiochem. Sci., 15: 342-346, 1990.

41. Kuenzel, E. A., Mulligan, J. A., Sommercorn, J., and Krebs, E. G. Substrate specificity determinants for casein kinasc II as deduced from studies with syntheticpeptides. J. Biol. Chem., 262: 9136-9140, 1987.

42. Elbein, A. D. Inhibitors of the biosynthesis and processing of /V-linked oligosaccha-ride chains. Annu. Rev. Biochem., 56: 497-534, 1987.

43. Dickson, G., Cower, H. J., Barton, C. H., Prentice, H. M., Elsom, V. L., Moore, S. E.,Cox, R. D., Ouinn. C., Putt, W., and Walsh, F. S. Human muscle neural cell adhesionmolecule (N-CAM): identification of a muscle-specific sequence in the extracellulardomain. Cell, JO: 1119-1130, 1987.

44. Hamshere, M., Dickson, G., and Eperon, I. The muscle specific domain of mouseN-CAM: structure and alternative splicing patterns. Nucleic Acids Res., 19:4709-4716, 1991.

45. Walsh, F. S., Parekh. R. B., Moore, S. E., Dickson, G., Barton, C. H., Gower, H. J.,Dwek, R. A., and Rademacher, T. W. Tissue specific O-linked glycosylation of theneural cell adhesion molecule (N-CAM). Development (Camb.), 105: 803-811,

1989.46. Miura, M., Kobayashi, M., Asou, H., and Uyemura, K. Molecular cloning of cDNA

encoding the rat neural cell adhesion molecule LI. Two LI isoforms in the cytoplas-mic region are produced by differential splicing. FEBS Lett., 289: 91-95, 1991.

47. Sorkin, B. C., Hoffman, S., Edelman, G. M., and Cunningham, B. A. Sulfation andphosphorylation of the neural cell adhesion molecule, N-CAM. Science (WashingtonDC), 225: 1476-1478, 1984.

48. Sadoul, R., Kirchhoff, F., and Schachner, M. A protein kinase activity is associatedwith and specifically phosphorylates the neural cell adhesion molecule LI. J.Neurochem., 53: 1471-1478, 1989.

4501



1994;54:4493-4501. Cancer Res Michael A. Reale, Gang Hu, Abrahim I. Zafar, et al.

) Gene in Normal and Malignant TissuesDCCCancer (Expression and Alternative Splicing of the Deleted in Colorectal

Updated version

http://cancerres.aacrjournals.org/content/54/16/4493

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/54/16/4493To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Expression and Alternative Splicing of the Deleted in ... · The DCC (deleted in colorectal cancer) gene was identified because it is affected by somatic mutations in colorectal tumors,