Loss of beta-catenin regulation by the APC tumor suppressor protein correlates with loss of structure due to common somatic mutations of the gene.

The mutation cluster region in the APC gene defines a region of approximately 660 bp, in which the vast majority of its somatic mutations are found. These mutations disrupt the polypeptide chain, typically eliminating five of the seven repeated sequences of 20 amino acids (aa) each in the central region of the APC protein. To examine the relationship between loss of this structure and loss of function, we constructed APC deletion mutants that progressively truncated the protein across the mutation cluster region. The mutants were tested for their association with beta-catenin and their ability to down-regulate it in SW480 cells. The binding of beta-catenin to APC fragments required the inclusion of only a single 20-aa repeat sequence, whereas down-regulation required the presence of at least three of these repeat sequences, and those including the second repeat exhibited the highest activity. The mutation of three conserved serine residues in the second repeat greatly reduced the activity of an otherwise highly active APC fragment. Thus, the repeated 20-aa sequence is directly implicated in beta-catenin turnover. The elimination of at least five of these seven repeats due to somatic mutations suggests that loss of beta-catenin regulation by APC is selected for during tumor progression.

Deletion of an amino-terminal sequence beta-catenin in vivo and promotes hyperphosporylation of the adenomatous polyposis coli tumor suppressor protein.

Regulation of cell adhesion and cell signaling by beta-catenin occurs through a mechanism likely involving the targeted degradation of the protein. Deletional analysis was used to generate a beta-catenin refractory to rapid turnover and to examine its effects on complexes containing either cadherin or the adenomatous polyposis coli (APC) protein. The results show that amino-terminal deletion of beta-catenin results in a protein with increased stability that acts in a dominant fashion with respect to wild-type beta-catenin. Constitutive expression in AtT20 cells of a beta-catenin lacking 89 N-terminal amino acids (deltaN89beta-catenin) resulted in severely reduced levels of the more labile wild-type beta-catenin. The mutant beta-catenin was expressed at endogenous levels but displaced the vast majority of wild-type beta-catenin associated with N-cadherin. The deltaN89beta-catenin accumulated on the APC protein to a level 10-fold over that of wild-type beta-catenin and recruited a kinase into the APC complex. The kinase was highly active toward APC in vitro and promoted a sodium dodecyl sulfate gel band shift that was also evident for endogenous APC from cells expressing the mutant beta-catenin. Unlike wild-type beta-catenin, which partitions solely as part of a high-molecular-weight complex, the deltaN89 mutant protein also fractionated as a stable monomer, indicating that it had escaped the requirement to associate with other proteins. That similar N-terminal mutants of beta-catenin have been implicated in cellular transformation suggests that their abnormal association with APC may, in part, be responsible for this phenotype.

Binding of GSK3beta to the APC-beta-catenin complex and regulation of complex assembly.

The adenomatous polyposis coli gene (APC) is mutated in most colon cancers. The APC protein binds to the cellular adhesion molecule beta-catenin, which is a mammalian homolog of ARMADILLO, a component of the WINGLESS signaling pathway in Drosophila development. Here it is shown that when beta-catenin is present in excess, APC binds to another component of the WINGLESS pathway, glycogen synthase kinase 3beta (GSK3beta), a mammalian homolog of Drosophila ZESTE WHITE 3. APC was a good substrate for GSK3 beta in vitro, and the phosphorylation sites were mapped to the central region of APC. Binding of beta-catenin to this region was dependent on phosphorylation by GSK3 beta.

Functional wild-type and carboxyl-terminal-tagged rat substance P receptors expressed in baculovirus-infected insect Sf9 cells.

The rat substance P (SP) receptor (SPR) was expressed in insect Sf9 cells by infection with recombinant baculovirus. The receptor bound SP with high affinity (KD = 360 pM) and had a rank order of affinity of SP > neurokinin A > neurokinin B. Ligand activation of the receptor resulted in an increase in both inositol lipid hydrolysis and intracellular Ca2+ concentration ([Ca2+]i). However, high-level expression of the receptor, in the absence of ligand, was correlated with increased basal turnover of inositol lipids and an elevated rate of Ca2+ influx. These results demonstrate that the Sf9 cells provide a suitable environment for the high-level expression of a functionally active SPR. Two carboxy-terminal epitope-tagged receptors (SPR-KT3 = SPR-TPPPEPET, COOH; SPR-Glu = SPR-EEEEYMPME, COOH) were also expressed. The affinity of the KT3-tagged receptor for ligand was similar to that of the wild-type receptor (KD = 405 pM), and that of the Glu-tagged receptor was slightly lower (KD = 1,082 pM). The high-affinity SP binding site of all three receptors was sensitive to guanosine 5'-O-(3-thiotriphosphate) pretreatment. The maximal signal-transducing ability of the epitope-tagged receptors was comparable to that of the wild-type receptor ([Ca2+]i rise as a percentage of wild-type: SPR-KT3, 80-100%; SPR-Glu, 88-100%). These data show that heterologous expression in the baculovirus system results in high expression of functional wild-type and tagged receptors.

The APC protein and E-cadherin form similar but independent complexes with alpha-catenin, beta-catenin, and plakoglobin.

The tumor suppressor APC protein associates with the cadherin-binding proteins alpha- and beta-catenin. To examine the relationship between cadherin, catenins, and APC, we have tested combinatorial protein-protein interactions in vivo, using a yeast two-hybrid system, and in vitro, using purified proteins. beta-Catenin directly binds to APC at high and low affinity sites. alpha-Catenin cannot directly bind APC but associates with it by binding to beta-catenin. Plakoglobin, also known as gamma-catenin, directly binds to both APC and alpha-catenin and also to the APC-beta-catenin complex, but not directly to beta-catenin. beta-Catenin binds to multiple independent regions of APC, some of which include a previously identified consensus motif and others which contain the centrally located 20 amino acid repeat sequences. The APC binding site on beta-catenin may be discontinuous since neither the carboxyl- nor amino-terminal halves of beta-catenin will independently associate with APC, although the amino-terminal half independently binds alpha-catenin. The catenins bind to APC and E-cadherin in a similar fashion, but APC and E-cadherin do not associate with each other either in the presence or absence of catenins. Thus, APC forms distinct heteromeric complexes containing combinations of alpha-catenin, beta-catenin, and plakoglobin which are independent from the cadherin-catenin complexes.

Regulation of intracellular beta-catenin levels by the adenomatous polyposis coli (APC) tumor-suppressor protein.

The APC tumor-suppressor protein associates with beta-catenin, a cell adhesion protein that is upregulated by the WNT1 oncogene. We examined the effects of exogenous APC expression on the distribution and amount of beta-catenin in a colorectal cancer cell containing only mutant APC. Expression of wild-type APC caused a pronounced reduction in total beta-catenin levels by eliminating an excessive supply of cytoplasmic beta-catenin indigenous to the SW480 colorectal cancer cell line. This reduction was due to an enhanced rate of beta-catenin protein degradation. Truncated mutant APC proteins, characteristic of those associated with cancer, lacked this activity. Mutational analysis revealed that the central region of the APC protein, which is typically deleted or severely truncated in tumors, was responsible for the down-regulation of beta-catenin. These results suggest that the tumor-suppressor activity of mutant APC may be compromised due to a defect in its ability to regulate beta-catenin.

Oncogenic activation of v-kit involves deletion of a putative tyrosine-substrate interaction site.

The transforming gene of the Hardy-Zuckerman-4 strain of feline sarcoma virus, v-kit, arose by transduction of the cellular c-kit gene, which encodes the receptor tyrosine kinase (RTK) p145c-kit. To gain insight into the molecular basis of the v-kit transforming potential, we characterized the feline c-kit by cDNA cloning. Comparison of the feline v-kit and c-kit sequences revealed, in addition to deletions of the extracellular and transmembrane domains, three additional mutations in the v-kit oncogene product: deletion of tyrosine-569 and valine-570, the exchange of aspartate at position 761 to glycine, and replacement of the C-terminal 50 amino acids by five unrelated residues. Examinations of individual v-kit mutations in the context of chimeric receptors yielded inhibitory effects for some mutants on both autophosphorylation and substrate phosphorylation functions. In contrast, deletion of tyrosine-569 and valine-570 significantly enhanced transforming and mitogenic activities of p145c-kit, while the other mutations had no significant effects. Conservation in subclass III RTKs and the identification of the corresponding residue in beta PDGF-R, Y579, as a binding site for src family tyrosine kinases suggests an important role for Y568 in kit signal regulation and the definition of its oncogenic potential. Repositioning of Y571 by an inframe two codon deletion may be the crucial alteration resulting in enhancement of v-kit oncogenic activity.

The APC gene product associates with microtubules in vivo and promotes their assembly in vitro.

Defects in the APC gene occur frequently in patients with familial adenomatous polyposis coli and are associated with the progression of sporadic tumors of the colon and stomach. We examined the subcellular location of adenomatous polyposis coli (APC) protein resulting from transient expression of full length and partial APC complementary DNAs in epithelial cells. Immunofluorescent detection revealed an association of APC with cytoplasmic microtubules. Expression of partial complementary DNA constructs indicated that the carboxy-terminal region of the APC protein, typically deleted in cancers, is essential for this association. The same APC polypeptides that associated with microtubules in vivo also dramatically promoted their assembly in vitro. These results suggest that wild-type APC protein binds to and affects the assembly of microtubules, whereas the mutants identified in tumors have lost this activity.

Association of the APC gene product with beta-catenin.

Mutations in the human APC gene are linked to familial adenomatous polyposis and to the progression of sporadic colorectal and gastric tumors. To gain insight into APC function, APC-associated proteins were identified by immunoprecipitation experiments. Antibodies to APC precipitated a 95-kilodalton protein that was purified and identified by sequencing as beta-catenin, a protein that binds to the cell adhesion molecule E-cadherin. An antibody specific to beta-catenin also recognized the 95-kilodalton protein in the immunoprecipitates. These results suggest that APC is involved in cell adhesion.

Molecular cloning of a GTPase activating protein specific for the Krev-1 protein p21rap1.

The rap1/Krev-1 gene encodes a ras-related protein that suppresses transformation by ras oncogenes. We have purified an 88 kd GTPase activating protein (GAP), specific for the rap1/Krev-1 gene product, from bovine brain. Based on partial amino acid sequences obtained from this protein, a 3.3 kb cDNA was isolated from a human brain library. Expression of the cDNA in insect Sf9 cells resulted in high level production of an 85-95 kd rap1GAP that specifically stimulated the GTPase activity of p21rap1. The complete deduced amino acid sequence is not homologous to any known protein sequences, including GAPs specific for p21ras. Northern and Western blotting analysis indicate that rap1GAP is not ubiquitously expressed and appears most abundant in fetal tissues and certain tumor cell lines, particularly the Wilms' kidney tumor, SK-NEP-1, and the melanoma, SK-MEL-3, cell lines.

Inositol 1,4,5-trisphosphate receptors: distinct neuronal and nonneuronal forms derived by alternative splicing differ in phosphorylation.

We have identified two distinct transcripts of inositol 1,4,5-trisphosphate receptor by using the PCR on first-strand cDNAs from various rat tissues. The longer form, corresponding to the previously cloned adult rat brain inositol 1,4,5-trisphosphate receptor, contains a 120-nucleotide insert between the two cAMP-dependent protein kinase phosphorylation consensus sequences. The shorter form (lacking the insert) predominates in fetal brain and peripheral tissues and appears to represent a nonneuronal receptor, whereas the longer form is found in adult brain and appears to be exclusively neuronal. The phosphorylation kinetics by cAMP-dependent protein kinase and the phosphopeptide maps differ for inositol 1,4,5-trisphosphate receptors purified from tissues predominantly expressing different forms of the transcript.

Expression of the hst-1 and c-kit protooncogenes in human testicular germ cell tumors.

Seventy testicular germ cell tumors were analyzed at the DNA and RNA levels for the c-kit, hst-1, and int-2 oncogenes using Northern and Southern blot analyses, respectively. There were significant differences in oncogene expression between seminomas and nonseminomas with c-kit being expressed in 24 of 30 (80%) seminomas but in only 3 of 40 (7%) nonseminomatous tumors (P = 0.0001, chi 2 test) and hst-1 being expressed in 24 of 38 (63%) nonseminomas but only 1 of 24 (4%) of seminomas (P = 0.0001, chi 2 test), demonstrating an inverse relationship in the expression pattern of these 2 oncogenes in human testicular germ cell tumors. A significant association between tumor stage and hst-1 expression in the nonseminoma group was found (P = 0.0002, chi 2 test). No gross alterations in the c-kit, hst-1, and int-2 loci were found at the DNA level and no int-2 mRNA expression was detected in any of the germ cell tumors examined.

Molecular cloning and expression of a G25K cDNA, the human homolog of the yeast cell cycle gene CDC42.

G25K is a low-molecular-mass GTP-binding protein with a broad distribution in mammalian tissues. A cDNA clone was isolated by using oligonucleotides corresponding to the partial amino acid sequence of purified human G25K. The cDNA encodes an 191-amino-acid polypeptide containing GTP-binding consensus sequences and a putative farnesylation site at the C terminus. The sequence exhibits 50 and 70% identities to the mammalian rho and rac proteins, respectively, and an 80% identity to the Saccharomyces cerevisiae CDC42 gene product. Insect Sf9 cells infected with recombinant baculovirus vectors expressing the G25K cDNA produced a 25-kDa protein that bound GTP and was recognized by antibodies specifically reactive to G25K. G25K appears to be the human homolog of the CDC42 gene product, since expression of the G25K cDNA in S. cerevisiae suppressed both cdc42-1 and cdc24-4 temperature-sensitive lethal mutations.

Human kidney amiloride-binding protein: cDNA structure and functional expression.

Phenamil, an analog of amiloride, is a potent blocker of the epithelial Na+ channel. It has been used to purify the porcine kidney amiloride-binding protein. Synthetic oligonucleotides derived from partial sequences have been used to screen a human kidney cDNA library and to isolate the cDNA encoding the human amiloride-binding protein. The primary structure was deduced from the DNA sequence analysis. The protein is 713 residues long, with a 19-amino acid signal peptide. The mRNA was expressed in 293-S and NIH 3T3 cells, yielding a glycoprotein (i) that binds amiloride and amiloride analogs with affinities similar to the amiloride receptor associated with the apical Na+ channel in pig kidney membranes and (ii) that is immunoprecipitated with monoclonal antibodies raised against pig kidney amiloride-binding protein.

Epidermal growth factor receptor cytoplasmic domain mutations trigger ligand-independent transformation.

The transforming gene product of avian erythroblastosis virus, v-erbB, is derived from the epidermal growth factor (EGF) receptor but has lost its extracellular ligand-binding domain and was mutated in its cytoplasmic portion, which is thought to be responsible for biological signal generation. We have repaired the deletion of extracellular EGF-binding sequences and investigated the functional consequences of cytoplasmic erbB mutations. Within the resulting EGF receptors, the autophosphorylation activities of the cytoplasmic domains of v-erbB-H and v-erbB-ES4 were fully ligand dependent in intact cells. However, the mitogenic and transforming signaling activities of an EGF receptor carrying v-erbB-ES4 (but not v-erbB-H) cytoplasmic sequences remained ligand independent, whereas those of a receptor with a v-erbB-H cytoplasmic domain were regulated by EGF or transforming growth factor alpha. Thus, structural alterations in the cytoplasmic domain of growth factor receptor tyrosine kinases may induce constitutive signaling activity without autophosphorylation. These findings provide new insight into the mechanism of receptor-mediated signal transduction and suggest a novel alternative for subversion of cellular control mechanisms and proto-oncogene activation.

Biosynthesis of reovirus-specified polypeptides: effect of point mutation of the sequences flanking the 5'-proximal AUG initiator codons of the reovirus S1 and S4 genes on the efficiency of mRNA translation.

The effect on translation of site-directed nucleotide substitutions around the 5'-proximal AUG initiation codon of the reovirus s1 mRNA specifying polypeptide sigma 1 and the reovirus s4 mRNA specifying polypeptide sigma 3 was examined. The efficiency of synthesis of the S1-encoded sigma 1 polypeptide and the S4-encoded sigma 3 polypeptide was analyzed in transfected simian COS cells. Mutant s1 mRNAs possessing either GCU AUG G or GCA AUG G sequences surrounding the 5'-proximal sigma 1 AUG were translated with an efficiency comparable to that of the wild-type s1 mRNA which possesses the flanking sequence CCU AUG G. Mutant s4 mRNAs possessing either CCU AUG G or CCA AUG G sequences surrounding the 5'-proximal sigma 3 AUG were translated with an efficiency comparable to that of wild-type s4 mRNA which possesses the flanking sequence GCA AUG G. The s4 mRNAs, both wild-type and mutant, were translated in vivo about five times more efficiently than the s1 mRNAs, both wild-type and mutant. These results suggest that nucleotide positions other than the -3, -2, -1, and +4 positions relative to the 5'-proximal initiator AUG, where the A is +1, play a dominant role in determining the efficiency of translation of these two reovirus mRNAs in vivo.

Human proto-oncogene c-kit: a new cell surface receptor tyrosine kinase for an unidentified ligand.

Structural features of v-kit, the oncogene of HZ4 feline sarcoma virus, suggested that this gene arose by transduction and truncation of cellular sequences. Complementary DNA cloning of the human proto-oncogene coding for a receptor tyrosine kinase confirmed this possibility: c-kit encodes a transmembrane glycoprotein that is structurally related to the receptor for macrophage growth factor (CSF-1) and the receptor for platelet-derived growth factor. The c-kit gene is widely expressed as a single, 5-kb transcript, and it is localized to human chromosome 4 and to mouse chromosome 5. A c-kit peptide antibody permitted the identification of a 145,000 dalton c-kit gene product that is inserted in the cellular plasma membrane and is capable of self-phosphorylation on tyrosine residues in both human glioblastoma cells and transfected mouse fibroblasts. Our results suggest that p145c-kit functions as a cell surface receptor for an as yet unidentified ligand. Furthermore, carboxy- and amino-terminal truncations that occurred during the viral transduction process are likely to have generated the transformation potential of v-kit.

Biosynthesis of reovirus-specified polypeptides. Molecular cDNA cloning and nucleotide sequence of the reovirus serotype 1 Lang strain s2 mRNA which encodes the virion core polypeptide sigma 2.

Human reovirus serotype 1 Lang strain s2 mRNA, which encodes the virion inner capsid core polypeptide sigma 2, was cloned as a cDNA:mRNA heteroduplex in Escherichia coli using phage M13. A complete consensus nucleotide sequence was determined. The Lang strain s2 mRNA is 1331 nucleotides in length and possesses an open reading frame with a coding capacity of 335 amino acids, sufficient to account for a sigma 2 polypeptide of 37,682 daltons. Comparison of the serotype 1 Lang s2 sequence derived from cDNA clones of s2 mRNA with the serotype 3 Dearing S2 sequence derived from cDNA clones of the S2 dsRNA genome segment reveals 86 percent homology at the nucleotide level. The predicted sigma 2 polypeptides of the Lang and Dearing strains display 98 percent homology at the amino acid level. Of 147 silent nt differences in the translated region, 136 were in the third base position of codons.

Biosynthesis of reovirus-specified polypeptides. Efficiency of expression of cDNAs of the reovirus S1 and S4 genes in transfected animal cells differs at the level of translation.

Full-length cDNAs of the reovirus serotype 1 Lang strain S1 and S4 genes were cloned in Escherichia coli using bacteriophage M13 and expressed in monkey COS cells under the control of the SV40 late promoter using the eukaryotic expression vector pJC119. The s1-encoded sigma 1 and s4-encoded sigma 3 gene products were expressed in transfected COS cells and were indistinguishable from the authentic sigma 1 and sigma 3 polypeptides synthesized in reovirion-infected COS cells. The relative translational efficiencies of the s1 and s4 mRNAs in transfected COS cells were similar to the efficiencies observed in virion-infected cells; the s4 mRNA was translated approximately five times more efficiently than the s1 mRNA. Our results suggest that the differential translation of the reovirus s1 and s4 mRNAs in vivo may be attributed to intrinsic structural properties of the individual mRNAs and is independent of competition with other viral mRNAs.

Biosynthesis of reovirus-specified polypeptides. Molecular cDNA cloning and nucleotide sequence of the reovirus serotype 1 Lang strain bicistronic s1 mRNA which encodes the minor capsid polypeptide sigma 1a and the nonstructural polypeptide sigma 1bNS.

Human reovirus serotype 1 Lang strain s1 mRNA, which encodes the minor capsid cell attachment protein sigma 1a and the nonstructural protein sigma 1bNS, was cloned as a cDNA:mRNA heteroduplex in Escherichia coli using phage M13. The Lang strain s1 mRNA is 1462 nucleotides in length and possesses two open reading frames. The first begins at nt 14 and has a coding capacity of 418 amino acids, sufficient to account for sigma 1a; the second begins at nt 75 and has a coding capacity of 119 amino acids, sufficient to account for sigma 1bNS. Comparison of the Lang serotype s1 sequence derived from cDNA clones of s1 mRNA with the Lang S1 sequence derived from cDNA clones of the S1 dsRNA genome segment definitively establishes that reovirus plus-strand mRNA is structurally equivalent to the plus-strand of the dsRNA genome segment.