Epithelial mucins are large, secreted and cell surface glycoproteins involved in epithelial cell protection, adhesion modulation, and signaling. Using differential display, we have identified two novel mucin cDNAs (dd34 and dd29), hereafter designated MUC11 and MUC12, respectively, that are down-regulated in colorectal cancers. Northern blots demonstrated polydisperse signals characteristic of mucin transcripts in RNA from normal colon that were absent in colorectal cancer. Both cDNAs were mapped by fluorescence in situ hybridization to chromosome band 7q22, the location of the MUC3 mucin gene, thus suggesting that there may be a cluster of mucin genes at this locus. The sequences of both differential display clones were extended by a combination of screening libraries and PCR. The 2.8-kb MUC11 cDNA composite encoded 35 serine/threonine-rich, mucin-like degenerate 28 amino acid tandem repeats. The MUC12 cDNA composite encoded a putative transmembrane mucin containing two extracellular cysteine-rich, EGF-like domains, a coiled-coil region, and a mucin-like domain consisting of 28 amino acid degenerate tandem repeats. Distinct patterns of expression of MUC11, MUC12, and MUC3 mRNAs were observed in a range of normal human tissues. MUC12 mRNA was not expressed in any of six colorectal cancer cell lines examined and was down-regulated or absent in 6 of 15 (40%) tumors compared with matched normal colonic tissue. In contrast, MUC11 showed a different pattern of mRNA expression, with four of these lines showing low levels and the other two lines showing relatively high levels of MUC11 transcripts. Expression of MUC11 was down-regulated in the tumors of 12 of 15 (80%) paired samples. Structural homology of MUC12 with rat, mouse, and human MUC3 and human and rat MUC4/ASGP2 indicate that there is a distinct subfamily of transmembrane mucins with conserved epidermal growth factor domains. The homology of MUC12 with epidermal growth factor-like growth factors and its down-regulation in colorectal cancers, together with known interactions between rat MUC4 and c-erbB-2 growth factor receptors, suggests that MUC12 may be involved in epithelial cell growth regulation.

Epithelial mucins are a family of secreted and cell surface glycoproteins expressed by epithelial tissues. They are characterized by a central polymorphic tandem repeat structure, which comprises most of the protein backbone, and a large number of O-linked carbohydrate side chains. Ten human epithelial mucin genes have been identified, although full-length cDNA clones are available for only five of these genes, primarily because of their large size (up to 33 kb) and the presence of tandem repeats (1).

It is emerging that there are several classes of human mucins, each with distinct functions. Many human mucin genes appear to encode secreted mucins that protect and lubricate epithelial tissues by forming a layer of viscoelastic gel. These gel-forming mucins are encoded by a cluster of four mucin genes (MUC6, MUC2, MUC5AC, and MUC5B) on chromosome band 11p15.5 (2). The MUC1 mucin is expressed by almost all human glandular epithelial cells and is one of only two human membrane-anchored, epithelial mucins to be identified thus far (3). It has a distinct role in adhesion modulation and cell signaling. The second transmembrane mucin, MUC4, is also widely distributed in human epithelial tissues. MUC4 and its rat orthologue, ASGP1/2 (hereafter designated rMuc4), contain two extracellular cysteine-rich, EGF3-like domains adjacent to a large mucin-like tandem repeat domain (4, 5).

MUC3, located on chromosome 7q22, encodes an intestinal mucin expressed by goblet and absorptive cells. MUC3 is similar to MUC4 because, in contrast to a previous report (6), we have determined recently that it contains two cysteine-rich, EGF-like domains, a transmembrane region, and a cytoplasmic tail at its COOH terminus, in addition to mucin-like repeat domains.4 EGF-like motifs are found in several growth factors and in numerous extracellular proteins involved in formation of the extracellular matrix, cell adhesion, chemotaxis, and wound healing. These motifs may allow exposure of ligand-binding sites on the exterior regions of a trilobed structure formed by disulfide bridging of the characteristic core six cysteine residues. Although no ligands for MUC3 have been identified, association of rMuc4 with the c-erbB-2 growth factor receptor and consequent modulation of receptor function has been demonstrated (7), implicating the EGF-like, domain-containing transmembrane mucins in growth modulation.

Altered mucin expression has long been associated with the pathology of epithelial diseases such as inflammatory bowel disease (8) and respiratory diseases, including cystic fibrosis (9). Additional evidence has implicated a role in adenocarcinomas, including colorectal carcinoma. Using a combination of in situ hybridization and immunohistochemistry, Chang et al.(10) found that MUC2 and MUC3 were down-regulated in colorectal cancer. A more recent in situ hybridization study found that expression of these genes was markedly reduced in nonmucinous colorectal cancers but was maintained in mucinous tumors (11). Expression of the mucin carbohydrate structure, sialosyl-Tn, has been found to be an independent predictor of poor prognosis in colon cancer, suggesting that carbohydrate alterations are also important (12). More recently, LS-C cells, which produce mucin that contains truncated carbohydrates enriched for sialosyl-Tn, have been shown to bind to basement membrane matrix and metastasize to a greater extent than another clone from the same parent line, LS-B cells, which produce more fully glycosylated mucins (13), thus suggesting a role for mucins in invasion and metastasis. Expression of MUC1 at the deepest invasive portion of colorectal cancers showing submucosal invasion has been shown to be a strong predictor of lymph node metastasis, implicating this transmembrane mucin in colorectal cancer metastasis (14).

In the present study, differential display was used to compare gene expression between paired normal colon and primary tumor tissues to identify new genes involved in the development of colorectal cancer. We identified two novel colonic mucin cDNAs that were found to be commonly down-regulated in colorectal cancer. The corresponding genes (designated MUC11 and MUC12) were localized to chromosome band 7q22, suggesting the presence of a mucin gene cluster at this locus. These genes show differential tissue expression to each other and MUC3 and share little DNA sequence homology. MUC12 is a transmembrane mucin with two EGF-like domains, one of which is homologous with EGF receptor ligands.

Differential Display.

The differential display method was devised from the original technique described by Liang and Pardee (15). Total RNA was isolated by a method described previously (16). Reverse transcription was carried out using one of four anchored primers, T12MG, T12MC, T12MA, and T12MT (Operon Technologies Inc., Alameda, CA) and Superscript RNase H- reverse transcriptase (Life Technologies, Inc., Gaithersburg, MD). One arbitrary 10-mer primer (Operon Technologies, Inc.) was selected at random to be used in a PCR with the appropriate anchored primer. Two patients, 101 and 112, were analyzed simultaneously, and duplicates of two separate reverse transcription reactions electrophoresed on each gel. Gels were put down wet and autoradiographed for 1–3 days. DNA was removed from gel slices by boiling and reamplified by PCR. Bands were then cloned into pGEM-T (Promega Corp., Madison, WI) and sequenced. Sequences were analyzed by multiple sequence similarity searches using BLAST algorithms (17) accessed through the National Center of Biotechnology Information.5

Northern Blot Analysis.

Northern blot analysis was performed on paired normal and tumor total RNA extracted from the same patients used in the differential display experiment. dd29 (MUC12) and dd34 (MUC11) were random primer-labeled using a Megaprime DNA labeling system (Amersham, Aylesbury, United Kingdom), and hybridization was performed at 65°C in buffer containing 7% SDS, 0.26 m Na2HPO4, 1 mm EDTA, and 1% BSA (18).

Multiplex Semiquantitative RT-PCR.

Multiplex semiquantitative RT-PCR was performed on total RNA isolated from six colorectal cancer cell lines and from paired normal colonic mucosa and tumor colorectal cancer tissues from 20 patients, five of each Dukes’ stage. Informed consent was obtained from each subject after approval by the appropriate hospital Ethics Committee. PCR products were quantitated relative to a β2-microglobulin cDNA amplification control using densitometry. First-strand cDNA synthesis was accomplished using 1 μg of total RNA. PCR amplification of cDNA was performed in a total volume of 25 μl containing 1 μl of the first-strand cDNA synthesis reaction products, 2.5 μl of 10× Taq polymerase buffer [25 mm Tris-(hydroxymethyl)-methyl-amino-propane-sulfonic acid, sodium salt (pH 9.3), and 50 mm KCl], 2 mm deoxynucleotide triphosphates, 25 mm MgCl2, 20 pmol each of the forward and reverse primers, and 2.5 units of Taq polymerase. Gene-specific forward and reverse primers for dd29 and dd34 were designed to produce PCR products of 510 and 169 bp, respectively. Primers for β2-microglobulin generated a PCR product of 247 bp (19). Primers were: dd29F1, 5′-TGAAGGGCGACAATCTTCCTC-3′; dd29R1, 5′-TACACGAGGCTCTTGGCGATGTTG-3′; dd34F1, 5′-CAGGCGTCAGTCAGGAATCTACAG-3′; dd34R1, 5′-GAGGCTGTGGTGTTGTCAGGTAAG-3′; β-21F, 5′-TGAATTGCTATGTGTCTGGGT-3′; and β-21R, 5′-CCTCCATGATGCTGCTTACAT-3′.

After an initial denaturation step of 94°C for 5 min, the amplification conditions were: 21 cycles of denaturation at 94°C (30 s) for dd29 [24 cycles of denaturation at 94°C (30 s) for dd34], annealing at 60°C (30 s), and extension at 72°C (30 s). PCR products were electrophoresed on 1.2% 1× TBE gels and photographed.

Expression Analysis.

A human RNA “master blot” (Clontech, Palo Alto, CA) with RNA from 50 different tissues and controls was used to examine mucin gene expression. DNA fragments encoding dd29, dd34, and MUC3 (GenBank accession no. M55405, a gift from Dr Sandra Gendler, Mayo Clinic, Scotsdale, Arizona) were excised from vector and radiolabeled as described above. Hybridization was performed as per the manufacturer’s instructions. The master blot was reprobed with a radiolabeled β-actin cDNA as a loading control.

Extending the Sequences of dd29 and dd34.

For library screening, a λgt11 human fetal brain 5′-STRETCH PLUS cDNA library (Clontech) was screened using radiolabeled dd29 and dd34. λ DNA was extracted, and inserts were excised, cloned into pBSK-, and sequenced. Screening of the fetal brain library with clone dd34 yielded two new cDNA clones: clone 2 (1043 bp) and clone Ii5 (1045 bp). Clone dd34 was a perfect match to the middle of the larger clone 2. cDNA from clone Ii5, however, was highly homologous but not identical to the cDNA from clone dd34. To ascertain whether these partial cDNAs arose from a single mRNA transcript, RT-PCR was carried out using combinations of forward and reverse primers specific for each cDNA in an attempt to link them. RT-PCR was performed on total RNA extracted from normal colon in a stringent touchdown PCR using high fidelity DyNAzyme DNA polymerase (Finnzymes, Espoo, Finland). Primer combination dd34F1 and Ii5R (5′-GGGAACACTGTGGTTTCAGTTGAG-3′) yielded a PCR product of 2 kb, demonstrating that these two cDNAs were derived from a single transcript. This product was cloned into pGEM-T and sequenced. For PCR library screening to extend the sequence of dd29, forward and reverse primers for dd29 (dd29F1 and dd29R1) were used in combination with a T7 vector-derived primer in a stringent touchdown PCR to screen an ulcerative colitis plasmid library (a gift from Dr Jonathon Fawcett, Queensland Institute of Medical Research, Brisbane, Australia). Amplified products were purified, cloned into pGEM-T, and sequenced.

FISH.

DNA fragments excised from dd29 (720 bp) and dd34 (530 bp) were nick translated with biotin-14-dATP and hybridized in situ at a final concentration of 10 ng/μl to metaphases from two normal males. The FISH method was modified from that described previously (20) in that chromosomes were stained before analysis with both propidium iodide as counterstain and 4′,6-diamidino-2-phenylindole for chromosome identification. Images of metaphase preparations were captured by a cooled CCD camera using the CyroVision Ultra image collection and enhancement system (Applied Imaging International, Ltd., Sunderland, United Kingdom).

Identification by Differential Display of Two cDNAs Encoding Mucins Down-Regulated in Colorectal cancer.

Differential display was performed on RNA from paired normal colonic mucosa and primary colorectal cancers. Using a PCR primer combination of T12MG and 10-mer 5′-ACTTCGCCAC-3′, bands dd29 (MUC12) and dd34 (MUC11) were both amplified from normal colonic mucosal RNA of two patients and were consistently down-regulated in the tumors from these patients in multiple PCR reactions (Fig. 1 A). After reamplification PCR, discrete bands of ∼720 bp for dd29 and 530 bp for dd34 were isolated and cloned into pGEM-T. Sequence analysis showed that both cDNAs were novel, with no match in any database accessed through the National Center of Biotechnology Information. Repetitive segments typical of mucin tandem repeats were observed in dd34.

Northern blot analyses of dd29 (Fig. 1,B) and dd34 (Fig. 1 C) with colonic total RNA used for the differential display reactions revealed a polydisperse signal beginning near the top of the gel for RNA isolated from normal colonic mucosa and no signal in tumor-derived RNA. Probe dd29 showed some cross-hybridization to rRNA. Polydispersity of signal is a hallmark of mucin RNA blots because of shearing of very high molecular weight transcripts.

Multiplex Semiquantitative RT-PCR.

Because of the polydisperse signals obtained by Northern analysis, expression of dd29 and dd34 was examined in a range of colorectal cancer cell lines and tissue mRNAs by multiplex semiquantitative RT-PCR. dd29 was not expressed in any of six colorectal cancer cell lines examined (Fig. 1,D). In contrast, dd34 showed a different pattern of expression, with HT29, LIM1215, LIM1899, and LIM1863 lines revealing very faint PCR products and SW620 and SW480 lines showing relatively high levels of expression (Fig. 1,E). For tumor tissue-derived RNA, down-regulation was defined as amplified band intensity <30% of that observed from paired normal colon tissue. dd29 was found to be down-regulated or absent in 6 of 15 (40%) tumors with paired normal samples and at low levels in three of five (60%) Dukes’ stage D samples (where normal colon was not available for comparison; Fig. 1,F). dd34 was down-regulated in the tumors of 12 of 15 (80%) paired samples and expressed at low levels in four of five (80%) Dukes’ stage D samples. One of five Dukes’ stage D samples showed relatively high levels of expression of dd34 (Fig. 1 G). Significantly, 13 of 15 (87%) colorectal cancers showed down-regulation of at least one of these mucin genes, with 5 of 15 (33%) showing down-regulation of both genes.

Sequence Analysis of dd29 (MUC12).

The sequence of dd29 revealed that it was amplified as a result of priming of random 10-mer at both ends of the PCR product and that it did not contain a 3′ untranslated region or poly(A) tail. Screening of an UC cDNA library with dd29-specific primers extended the sequence 840 bp in the 5′ direction and 800 bp in the 3′ direction to the poly(A) tail (GenBank accession number AF147790). Conceptual translation of the composite MUC12 cDNA revealed the presence of serine/threonine and proline-rich degenerate tandem repeats (Fig. 2), consistent with this protein being a member of the epithelial mucin family. The deduced 28-amino acid tandem repeat structure is shown in Fig. 2. Following the mucin-repeat domain, MUC12 contains two cysteine-rich, EGF-like domains separated by a 150-amino acid, nonmucin-like sequence (amino acids 261–410) containing five N-glycosylation sites and a potential coiled-coil domain. The second cysteine-rich, EGF-like domain is immediately followed by a putative transmembrane domain containing 26 hydrophobic or uncharged amino acids and a cytoplasmic tail of 75 amino acids at the COOH terminus.

Sequence alignment of MUC12, human MUC3 (hMUC3), rat Muc3 (rMuc3), mouse Muc3 (mMuc3), human MUC4 (hMUC4), and rMuc4 is shown in Fig. 3. When aligned by the transmembrane amino acid sequences, MUC12 was found to have areas of significant homology to rMuc3, mMuc3, and hMUC3, including perfect conservation of eight cysteine residues in the second EGF-like domain. With inclusion of three small gaps, each of these cysteines also align with those in rat and human MUC4. Interestingly, all six mucins contain a conserved EGF-like sequence of Cx(5)GPxCxCx(9)GExC. Furthermore, there is some (four of eight) conservation of the cysteine residues between MUC12 and the human and rodent MUC3 and MUC4 mucins in the first EGF-like domain.

Sequence Analysis of dd34 (MUC11).

Clone dd34 (544 bp) was also obtained as a result of priming of random 10-mers at both ends of the PCR product. Screening of a λgt11 human fetal brain library yielded two positive plaques that hybridized to dd34, clone Ii5 (1045 bp) and clone 2 (1043 bp). These two clones represented opposite ends of a 2.8-kb partial MUC11 cDNA sequence (GenBank accession number AF147791), the linking of which was established by PCR (see “Materials and Methods”). Conceptual translation of the MUC11 composite is shown in Fig. 4. The entire 957-amino acid sequence consisted of serine, threonine, and proline-rich tandem repeats of 28 amino acids in length, consistent with it being derived from a large epithelial mucin. The deduced tandem repeat structure and consensus repeat sequence for MUC11 is shown in Fig. 4.

Differential Tissue Distribution of MUC11, MUC12, and MUC3 mRNAs.

Analysis of the tissue distribution of MUC11, MUC12, and MUC3 transcripts in RNA isolated from 50 different normal tissues showed a distinct pattern of expression for each gene (Fig. 5). MUC12 and MUC11 showed highest expression in colon but had different patterns in other organs, mainly restricted to those of epithelial type. MUC11 had a wider epithelial distribution than MUC12, which was restricted to expression in the colon and weakly in the pancreas, prostate, and uterus. Consistent with published findings (21), MUC3 was found to be predominantly expressed in the small intestine and at much lower levels in the colon. Interestingly, it was also present in the thymus.

Chromosomal Localization of MUC11 and MUC12.

Twenty metaphases from a normal male were examined for hybridization to dd29 and dd34 probes. For both genes, all of the metaphases showed strong signal on one or both chromatids of chromosome 7, at band 7q22 (data not shown). A similar result was obtained using metaphases from a second normal male.

Differential display has been used to identify two partial cDNAs (designated MUC11 and MUC12 by the Human Gene Nomenclature Committee), which encode novel colonic mucin-like proteins. Expression of both cDNAs was commonly down-regulated in colorectal cancers.

MUC11 and MUC12 were mapped by FISH to chromosome band 7q22. The location of another mucin gene, MUC3, at 7q22 suggests the presence of a new cluster of mucin genes at this locus. Interestingly, four genes encoding gel-forming mucins are found in a cluster on chromosome 11, and these genes appear to have originated from a common ancestral gene. Although the mucin cDNAs mapped to 7q22 most likely represent separate genes, it is also possible that they are produced as a result of alternative mRNA splicing from a single, large mucin gene. Northern blot analysis for MUC11, MUC12, and MUC3 shows that these encode large transcripts, estimated to be >12 kb.

Our multiple tissue RNA analysis shows no cross-reactivity between MUC11, MUC12, or MUC3. MUC11 and MUC12 showed highest expression in the colon, and in confirmation of a previous report (21), MUC3 was most highly expressed in the small intestine and at very low levels in the colon. The sequences of MUC11 and MUC12 are not homologous with any other human mucin genes but show some degree of similarity within their variable tandem repeat regions to each other (71% over 653 bp). However, their clear differential expression patterns in normal and tumor tissues as well as tumor cell lines show that they are distinct from each other and from MUC3. In addition, we have identified that the COOH terminus of hMUC3 contains two cysteine-rich, EGF-like domains, a transmembrane region, and a cytoplasmic tail, similar to but distinct from that of MUC12.

Although both MUC11 and MUC12 contain variable repeat regions typical of mucins, MUC12 is putatively a transmembrane mucin with features suggesting an involvement in growth regulation, a largely unrecognized function in human mucins. MUC12 is only the fourth human membrane-anchored epithelial mucin to be described to date, along with MUC1, MUC3, and MUC4. MUC1 has been shown to be involved in cell signaling via multiple tyrosine phosphorylation sites on its highly conserved cytoplasmic tail (22). At its COOH terminus, MUC12 possesses a cytoplasmic tail containing a YNNF sequence (amino acids 557–560 in Fig. 2), which is similar to motifs recognized by SH2 domain-containing proteins (23), suggesting that MUC12, like MUC1, may be involved in signal transduction.

The COOH terminus of MUC12 shows areas of high homology to the equivalent regions of both rat and mouse Muc3 proteins (24, 25, 26). The rodent genes have been named Muc3 because of their weak homology with hMUC3, their location on the syntenic regions to human chromosome 7, and their expression in intestine. The identification of a gene cluster on 7q22 reported here raises the possibility that the rodent genes have not been appropriately named. The COOH termini of MUC12 and MUC3 have overall 34 and 38% amino acid identity, respectively, to the rodent proteins. Undoubtedly, rat and mouse Muc3 are orthologues, with 82% amino acid identity, as well as similar consensus tandem repeat sequences of TTTADV/TTTVVV for mMuc3 and TTTPDV for rMuc3. Interestingly, MUC12 shows high homology (45 and 48% at the amino acid level) to the rodent proteins in the second EGF-like cysteine-rich domain, including alignment of all eight cysteines. The preservation of a CQLQRSGPxCLCxxTxTHWYxGExC sequence in all three proteins suggests that it is particularly important. Such an exposed potentially bilobular domain may function in ligand binding. The first cysteine-rich EGF-like domain of MUC12 and the rodent Muc3 proteins contain four cysteine residues that are precisely conserved, although this domain contains four additional cysteines in the rodent mucins. The conservation of parts of the COOH termini of rMuc3, mMuc3, and MUC12 suggests they are related to each other, but given the very high degree of conservation between the rodent proteins, it is more likely that MUC12 represents a closely related protein family member rather than the human orthologue.

These data highlight the emergence of a distinct subfamily of transmembrane epithelial mucins containing conserved cysteine-rich, EGF-like domains, that are widely expressed across human glandular epithelia. MUC12 and the rodent Muc3 mucins show a similar domain organization to hMUC4 (4) and its rat orthologue, rMuc4 (Fig. 6). All of these proteins contain an N-glycosylated domain of a similar size separating two EGF-like domains. hMUC4 and rMuc4 contain an additional large N-glycosylated domain between the first EGF domain and the mucin domain and also have a much smaller cytoplasmic tail than the other mucins. Interestingly, the rMuc4 isoform that contains the EGF-like domains but lacks a mucin domain has been shown in transfection studies to bind the c-erbB-2 growth factor receptor and promote signaling, both in the presence and absence of the c-erbB-2 ligand, resulting in increased mitogenesis (7). Whereas the EGF-like domain in rMuc4 shows homology with the c-erbB-2 ligands (heregulins), the first EGF-like domain in MUC12 shows homology to a number of EGF receptor-binding growth factors (Fig. 7). Many EGF-like factors are capable of binding the EGF receptor, and it is the differences in the sequences of these factors that seem to account for differential receptor dimerization and activation of downstream effectors (27). Interaction of the MUC12 EGF-like domain with the EGF receptor may be a critical function of MUC12 in normal colonic epithelium, and diminished expression of MUC12 may contribute to the aberrant growth properties of colonic tumors.

The deduced amino acid sequence of the partial MUC11 cDNA was composed entirely of serine/threonine-rich tandem repeats. There is a similarity between the tandem repeat consensus sequences of MUC11 (Fig. 4) and MUC12 (Fig. 2), and these also show limited homology to the MUC3 repeat (ITTETTSHSTPSFTSS). These similarities are consistent with evolution from a common ancestral gene. MUC11 is more widely expressed than MUC12 and MUC3, however, with RNA detected in gastrointestinal, respiratory, reproductive, and urinary tracts and unexpectedly in the liver and thymus.

The physiological roles of MUC11 and MUC12 in colonic epithelium are unknown. MUC11 and MUC12 are commonly down-regulated in colorectal cancer, suggesting they may play a role in epithelial cell growth modulation and/or differentiation. At present, it is not possible to comment on whether down-regulation of these genes is related to stage of tumor progression, because only 20 patients were analyzed in this study. However, down-regulation appears to be so frequent that it may be an early event in tumorigenesis. The down-regulation of these new 7q22 mucin genes in colorectal cancer is consistent with the demonstrated reduced expression of MUC3 in colorectal cancers (11); however, the mechanism(s) controlling their down-regulation remains unclear. Given the colocalization of these three mucin genes on chromosome 7q22, it is possible that their expression is coordinately regulated; hence, they are simultaneously down-regulated in a large proportion of colorectal cancers. The effect of down-regulation of these mucins on normal colonic epithelial cells could be substantial. Mucins are believed to protect epithelial cells from attack by pathogenic organisms and from mechanical and chemical damage. Therefore, reduced expression of these mucins could expose colonic epithelial cells to the harsh environment of the intestinal lumen. Furthermore, loss of a transmembrane mucin such as MUC12 may also contribute to loss of critical cell signaling.

The location of these two novel mucin genes on chromosome 7q22 may have significance for two nonmalignant epithelial diseases where aberrant mucin expression and/or function is a recognized component of pathology, i.e., inflammatory bowel disease and cystic fibrosis. Susceptibility genes for inflammatory bowel disease have been located on chromosomes 3, 12, and 7q22 (28), and MUC3 is considered a candidate gene for the 7q22 inflammatory bowel disease locus. Thus, MUC11 and MUC12 must also be considered candidates for involvement in inflammatory bowel disease, given their chromosomal localization, expression in normal colon, and the documented alterations in mucins in this disease (8). Mucins may also play a role in cystic fibrosis because patients with the same CFTR gene mutation do not demonstrate exactly the same phenotype in terms of mucus obstruction. The existence of modifier genes has been postulated, and mucin genes are obvious candidates (9). The CFTR gene lies in the adjacent chromosome band (7q31) to the MUC3, MUC11 and MUC12 genes. Although the significance of this finding is not clear, MUC11 and MUC12, which are expressed in many of the tissues affected by cystic fibrosis, should be considered as candidate modifier genes involved in the etiology of this disease.

Mucins are encoded by large genes that have proved difficult to clone by conventional methods because of the repetitive nature of their tandem repeat regions. Here, we have identified by differential display two partial cDNAs that represent novel mucin genes that are highly expressed in colonic epithelium, both of which are down-regulated in colorectal cancer. These findings, together with the sequence homology between the MUC12 EGF-like domain and EGF receptor-binding growth factors, suggest that these mucins may function as growth regulators in colonic epithelium. Down-regulation of these two novel mucin genes could be an important and previously unrecognized step in colorectal carcinogenesis.

The costs 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 with 18 U.S.C. Section 1734 solely to indicate this fact.

      
1

This research was supported by grants from the Queensland Cancer Fund, the Royal Children’s Hospital Foundation, Brisbane, Queensland, and the National Health and Medical Research Council. M. A. M. is supported by National Health and Medical Research Council Grant 98/1291.

            
3

The abbreviations used are: EGF, epidermal growth factor; RT-PCR, reverse transcription-PCR; FISH, fluorescence in situ hybridization.

      
4

S. J. Williams, D. J. Munster, R. J. Quin, D. C. Gotley, and M. A. McGuckin. The MUC3 gene encodes a transmembrane mucin and is alternatively spliced. Biochem. Biophys. Res. Commun., in press, 1999.

      
5

Internet address: http://www.ncbi.nlm.nih.gov.

Fig. 1.

A, autoradiograph of a differential display gel showing amplified products from RNA isolated from matched normal colon (N) and primary colorectal tumor (P) tissues. Arrows, differentially expressed bands dd29 (MUC12) and dd34 (MUC11). B, Northern blot analysis of total RNA from patient 101 hybridized with the dd29 probe. C, Northern blot analysis of RNA from patient 112 hybridized with the dd34 probe. Signal corresponding to 18S rRNA is shown as a loading control. D, multiplex, semiquantitative RT-PCR showing amplification of dd29 mRNA transcripts from matched normal colonic mucosa and primary tumor no. 40, normal mucosa from patient 81, and six colorectal cancer cell lines. E, multiplex, semiquantitative RT-PCR showing dd34 mRNA transcripts in matched normal colonic mucosa and primary tumors of patients 40, 164, and 97 and six colorectal cancer cell lines. Amplification of β2-microglobulin (β2-MG) is included as a measure of total RNA. F, multiplex semiquantitative RT-PCR showing amplification of dd29 mRNA transcripts from matched normal colonic mucosa and primary tumors 346, 84, 128, 97, and 316 and from five unpaired Dukes’ stage D tumors (M) 93, 361, 107, 357, and 367. G, multiplex semiquantitative RT-PCR showing dd34 mRNA transcripts in matched normal colonic mucosa and primary tumors of patients 110, 346, 84, 128, and 348 and from five unpaired Dukes’ stage D tumors (M) 93, 107, 361, 367, and 357. Ma, molecular size markers.

Fig. 1.

A, autoradiograph of a differential display gel showing amplified products from RNA isolated from matched normal colon (N) and primary colorectal tumor (P) tissues. Arrows, differentially expressed bands dd29 (MUC12) and dd34 (MUC11). B, Northern blot analysis of total RNA from patient 101 hybridized with the dd29 probe. C, Northern blot analysis of RNA from patient 112 hybridized with the dd34 probe. Signal corresponding to 18S rRNA is shown as a loading control. D, multiplex, semiquantitative RT-PCR showing amplification of dd29 mRNA transcripts from matched normal colonic mucosa and primary tumor no. 40, normal mucosa from patient 81, and six colorectal cancer cell lines. E, multiplex, semiquantitative RT-PCR showing dd34 mRNA transcripts in matched normal colonic mucosa and primary tumors of patients 40, 164, and 97 and six colorectal cancer cell lines. Amplification of β2-microglobulin (β2-MG) is included as a measure of total RNA. F, multiplex semiquantitative RT-PCR showing amplification of dd29 mRNA transcripts from matched normal colonic mucosa and primary tumors 346, 84, 128, 97, and 316 and from five unpaired Dukes’ stage D tumors (M) 93, 361, 107, 357, and 367. G, multiplex semiquantitative RT-PCR showing dd34 mRNA transcripts in matched normal colonic mucosa and primary tumors of patients 110, 346, 84, 128, and 348 and from five unpaired Dukes’ stage D tumors (M) 93, 107, 361, 367, and 357. Ma, molecular size markers.

Close modal
Fig. 2.

Predicted amino acid sequence of MUC12.Right, numbering of amino acids. Top, the consensus sequence of the degenerate tandem repeat structure. The two cysteine-rich, EGF-like domains are double underlined, a potential coiled-coil domain is in bold, the hydrophobic domain is singly underlined, and potential N-glycosylation sites are shaded. ∗, stop codon.

Fig. 2.

Predicted amino acid sequence of MUC12.Right, numbering of amino acids. Top, the consensus sequence of the degenerate tandem repeat structure. The two cysteine-rich, EGF-like domains are double underlined, a potential coiled-coil domain is in bold, the hydrophobic domain is singly underlined, and potential N-glycosylation sites are shaded. ∗, stop codon.

Close modal
Fig. 3.

Amino acid sequence alignment of the COOH termini of MUC12, hMUC3 (amino acids 1–366),4 mMuc3 (Ref. 26; amino acids 637-1015), rMuc3 (Refs. 24 and 25; amino acids 356–447 and 1–379, respectively), hMUC4 (Ref. 5; amino acids 861-1156), and rMuc4 (Ref. 4; amino acids 451–744). Light shading demonstrates identity with MUC12, and dark shading highlights all cysteine residues. Hyphens, gaps inserted to optimize the alignment.

Fig. 3.

Amino acid sequence alignment of the COOH termini of MUC12, hMUC3 (amino acids 1–366),4 mMuc3 (Ref. 26; amino acids 637-1015), rMuc3 (Refs. 24 and 25; amino acids 356–447 and 1–379, respectively), hMUC4 (Ref. 5; amino acids 861-1156), and rMuc4 (Ref. 4; amino acids 451–744). Light shading demonstrates identity with MUC12, and dark shading highlights all cysteine residues. Hyphens, gaps inserted to optimize the alignment.

Close modal
Fig. 4.

Predicted amino acid sequence of MUC11 showing the degenerate tandem repeat structure. The consensus sequence is shown at the top, and amino acids not consistent with this sequence are shown in bold. Hyphens, gaps placed to optimize the amino acid alignment. A potential N-glycosylation site is shaded.

Fig. 4.

Predicted amino acid sequence of MUC11 showing the degenerate tandem repeat structure. The consensus sequence is shown at the top, and amino acids not consistent with this sequence are shown in bold. Hyphens, gaps placed to optimize the amino acid alignment. A potential N-glycosylation site is shaded.

Close modal
Fig. 5.

mRNA tissue distribution of the 7q22 mucin gene family. Only those tissues showing a positive signal by Northern blot analysis are represented in the histogram. Sixteen tissues of neural origin, heart, aorta, skeletal muscle, bladder, stomach, testis, ovary, spleen, pituitary gland, adrenal gland, thyroid gland, salivary gland, and mammary gland were negative for mucin mRNA expression. Expression was quantified by densitometry and is shown as a proportion of the tissue showing highest expression.

Fig. 5.

mRNA tissue distribution of the 7q22 mucin gene family. Only those tissues showing a positive signal by Northern blot analysis are represented in the histogram. Sixteen tissues of neural origin, heart, aorta, skeletal muscle, bladder, stomach, testis, ovary, spleen, pituitary gland, adrenal gland, thyroid gland, salivary gland, and mammary gland were negative for mucin mRNA expression. Expression was quantified by densitometry and is shown as a proportion of the tissue showing highest expression.

Close modal
Fig. 6.

Domain organization of the COOH termini of human MUC12, hMUC3, the rodent Muc3 mucins, and the rat and human MUC4 mucins. The relative size of domains is accurate except that the N-glycosylated domain adjacent to the mucin domain in MUC4 is shown at approximately one-fifth of its actual size. Only the beginning of the large mucin domains are shown.

Fig. 6.

Domain organization of the COOH termini of human MUC12, hMUC3, the rodent Muc3 mucins, and the rat and human MUC4 mucins. The relative size of domains is accurate except that the N-glycosylated domain adjacent to the mucin domain in MUC4 is shown at approximately one-fifth of its actual size. Only the beginning of the large mucin domains are shown.

Close modal
Fig. 7.

Alignment of the first extracellular EGF-like domain of MUC12 with human EGF-like growth factors. Dark shading, identical amino acids; light shading, conservative amino acid substitutions.

Fig. 7.

Alignment of the first extracellular EGF-like domain of MUC12 with human EGF-like growth factors. Dark shading, identical amino acids; light shading, conservative amino acid substitutions.

Close modal
1
Seregni E., Botti C., Massaron S., Lombardo C., Capobianco A., Bogni A., Bombardieri E. Structure, function and gene expression of epithelial mucins.
Tumori
,
83
:
625
-632,  
1997
.
2
Dessyn J. L., Guyonnet-Duperat V., Porchet N., Aubert J. P., Laine A. A. Human mucin gene MUC5B. The 10.7 kb large central exon encodes various alternate subdomains, resulting in a super-repeat. Structural evidence for a 11p15.5 gene family.
J. Biol. Chem.
,
272
:
3168
-3178,  
1997
.
3
Gendler S. J., Spicer A. P., Lalani E. N., Duhig T., Peat N., Burchell J., Pemberton L., Boshell M., Taylor-Papadimitriou J. Structure and biology of a carcinoma-associated mucin, MUC1.
Am. Rev. Respir. Dis.
,
144
:
S42
-S47,  
1991
.
4
Sheng S., Wu K., Carraway K. L., Fregien N. Molecular cloning of the transmembrane component of the 13762 mammary adenocarcinoma sialomucin complex: a new member of the epidermal growth factor superfamily.
J. Biol. Chem.
,
267
:
16341
-16346,  
1992
.
5
Moniaux N., Nollet S., Porchet N., Degand P., Laine A., Aubert J. Complete sequence of the human mucin MUC4: a putative cell membrane-associated mucin.
Biochem. J.
,
338
:
325
-333,  
1998
.
6
Gum J. R., Ho J. J. L., Pratt W. S., Hicks J. W., Hill A. S., Vinall L. E., Roberton A. M., Swallow D. M., Kim Y. S. MUC3 human intestinal mucin. Analysis of gene structure, the carboxyl terminus, and a novel upstream repetitive region.
J. Biol. Chem.
,
272
:
26678
-26686,  
1997
.
7
Mcneer R. R., Price-Schiavi S. A., Komatsu M., Fregien N., Carraway C. A. C., Carraway K. L. Sialomucin complex in tumors and tissues.
Front. Biosci.
,
2
:
449
-459,  
1998
.
8
Rhodes J. M. Mucins and inflammatory bowel disease.
Q. J. Med.
,
90
:
79
-82,  
1997
.
9
Harris A., Reid C. Cystic fibrosis and mucins.
J. Med. Genet.
,
35
:
82
-83,  
1997
.
10
Chang S., Dohrman A. F., Basbaum C. B., Ho S. B., Tsuda T., Toribara N. W., Gum J. R., Kim Y. S. Localization of mucin (MUC2 and MUC3) messenger RNA and peptide expression in human normal intestine and colon cancer.
Gastroenterology
,
107
:
28
-36,  
1994
.
11
Weiss A. A., Babyatsky M. W., Ogata S., Chen A., Itzkowitz S. H. Expression of MUC2 and MUC3 mRNA in human normal, malignant, and inflammatory intestinal tissues.
J. Histochem. Cytochem.
,
44
:
1161
-1166,  
1996
.
12
Itskowitz S. H., Bloom E. J., Kokal W. A., Modin G., Hakomori S., Kim Y. S. Sialosyl-Tn. A novel mucin antigen associated with prognosis in colorectal cancer patients.
Cancer (Phila.)
,
66
:
1960
-1966,  
1990
.
13
Bresalier R. S., Byrd J. C., Brodt P., Ogata S., Itzkowitz S. H., Yunker C. K. Liver metastasis and adhesion to the sinusoidal endothelium by human colon cancer cells is related to mucin carbohydrate chain length.
Int. J. Cancer
,
76
:
556
-562,  
1998
.
14
Aoki R., Tanaka S., Haruma K., Yoshihara M., Sumii K., Kajiyama G., Shimamoto F., Kohno N. MUC-1 expression as a predictor of the curative endoscopic treatment of submucosally invasive colorectal carcinoma.
Dis. Colon Rectum
,
41
:
1262
-1272,  
1998
.
15
Liang P., Pardee A. B. Differential display of eukaryotic messenger RNA by means of the polymerase chain reaction.
Science (Washington DC)
,
257
:
967
-971,  
1992
.
16
Chomczynski P., Sacchi N. Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction.
Anal. Biochem.
,
162
:
156
-159,  
1987
.
17
Altschul S. F., Gish W., Miller W., Myers E. W., Lipman D. J. Basic local alignment search tool.
J. Mol. Biol.
,
215
:
403
-410,  
1990
.
18
Church G. M., Gilbert W. Genomic sequencing.
Proc. Natl. Acad. Sci. USA
,
81
:
1991
-1995,  
1984
.
19
Gussow D., Rein R., Ginjaar I., Hochstenbach F., Seemann G., Kottman A., Ploegh H. L. The human β2-microglobulin gene. Primary structure and definition of the transcriptional unit.
J. Immunol.
,
139
:
3132
-3138,  
1987
.
20
Callen D. F., Baker E., Eyre H. J., Chernos J. E., Bell J. A., Sutherland G. R. Reassessment of two apparent deletions of chromosome 16p to an ins(11;16) and a t(1;16) by chromosome painting.
Ann. Genet.
,
33
:
219
-221,  
1990
.
21
Van Klinken B. J., Van Dijken T. C., Oussoren E., Buller H. A., Dekker J., Einerhand A. W. Molecular cloning of human MUC3 cDNA reveals a novel amino acid tandem repeat region.
Biochem. Biophys. Res. Commun.
,
238
:
143
-148,  
1997
.
22
Zrihan-Licht S., Baruch A., Elroy-Stein O., Keydar I., Wreschner D. H. Tyrosine phosphorylation of the MUC1 breast cancer membrane proteins cytokine receptor-like molecules.
FEBS Lett.
,
356
:
130
-136,  
1994
.
23
Songyang Z., Shoelson S. E., McGlade J., Olivier P., Pawson T., Bustello X. R., Barbacid M., Sabe H., Hanafusa H., Yi T., Ren R., Baltimore D., Ratnofsky S., Feldman R. A., Cantley L. C. Specific motifs recognized by the SH2 domains of Csk, 3PB2, fps/fes, GRB2, HCP, SHC, Syk and Vav.
Mol. Cell. Biol.
,
14
:
2777
-2785,  
1994
.
24
Gum J. R., Hicks J. W., Lagace R. L., Byrd J. C., Toribara N. W., Siddiki B., Fearney F. J., Lamport D. T., Kim Y. S. Molecular cloning of rat intestinal mucin. Lack of conservation between mammalian species.
J. Biol. Chem.
,
266
:
22733
-22738,  
1991
.
25
Khatri I. A., Forstner G. G., Forstner J. F. The carboxyl-terminal sequence of rat intestinal mucin rMuc3 contains a putative transmembrane region and two EGF-like motifs.
Biochim. Biophys. Acta
,
1326
:
7
-11,  
1997
.
26
Shekels L. L., Hunninghake D. A., Tisdale A. S., Gipson I. K., Kieliszewski M., Kozak C. A, Ho S. B. Cloning and characterization of the mouse intestinal MUC3 mucin: 3′ sequence contains epidermal-growth-factor-like domains.
Biochem. J.
,
330
:
1301
-1308,  
1998
.
27
Burgess A. W. Growth control mechanisms in normal and transformed intestinal cells.
Phil. Trans. R. Soc. Lond.
,
353
:
903
-909,  
1998
.
28
Satsangi J., Parkes M., Louis E., Hashimoto L., Kato N., Welsh K., Terwilliger J. D., Lathrop G. M., Bell J. I., Jewell D. P. Two stage genome-wide search in inflammatory bowel disease provides evidence for susceptibility loci on chromosomes 3, 7 and 12.
Nat. Genet.
,
14
:
199
-202,  
1996
.