Faculty Bibliography

We report the first intracellular characterization of an endogenous nontransmembrane protein tyrosine phosphatase (PTP). Using affinity-purified polyclonal antibodies, we have identified PTP-1B as a 50 kd serine phosphoprotein in immunoprecipitation and immunoblotting assays. Surprisingly, indirect immunofluorescence experiments indicate that PTP-1B is localized predominantly in the endoplasmic reticulum (ER). Subcellular fractionation is consistent with this localization and establishes that PTP-1B is tightly associated with microsomal membranes, with its phosphatase domain oriented towards the cytoplasm. The C-terminal 35 amino acids of PTP-1B are both necessary and sufficient for targeting to the ER. The finding of a tyrosine phosphatase on the ER suggests new possibilities for cellular events controlled by tyrosine phosphorylation.

Tyrosine phosphorylation is controlled by the opposing actions of tyrosine kinases and phosphotyrosine phosphatases (PTPs). src homology 2 domains (SH2) are found in several types of signaling proteins, including some tyrosine kinases. These domains bind phosphotyrosyl proteins and thus help promote signal transduction. Using mixed oligonucleotide-directed polymerase chain reactions, two previously undescribed rat PTP cDNA fragments were generated. Through subsequent screening of rat megakaryocyte and human erythroleukemia libraries, we obtained a full-length coding sequence for one of these fragments. This cDNA, SH-PTP1, encodes a tyrosine phosphatase containing two highly conserved SH2 domains. SH-PTP1, with a 2.4-kilobase mRNA, a predicted open reading frame of 595 amino acids, and a structure suggesting a nontransmembrane protein, is expressed primarily in hematopoietic and epithelial cells. When expressed in Escherichia coli, SH-PTP1 possesses PTP activity. The structure of SH-PTP1 establishes an additional branch of the tyrosine phosphatase family and suggests mechanisms through which tyrosine phosphatases might participate in signal transduction pathways.

One of the three human retinoic acid receptors, RAR-beta, maps to a region on the short arm of chromosome 3 frequently deleted in lung cancer. Because retinoic acid is required for normal epithelial cell growth and regulation, and loss of a retinoic acid receptor might be expected to contribute to oncogenesis, we examined RAR-beta RNA and DNA in normal lung, 33 lung cancer cell lines and nine primary lung tumors. Normally, RAR-beta is expressed as two transcripts, of sizes 3.1 kb and 2.8 kb, which are strongly induced by retinoic acid. At least 50% of the cell lines and 30% of the tumor samples show altered RAR-beta expression and/or inducibility, including examples of absence or specific loss of one of the RAR-beta transcripts. Abnormalities in the expression patterns of RAR-alpha and RAR-gamma also are found, but at a lower frequency than RAR-beta abnormalities. Southern analysis reveals alteration of the RAR-beta gene in three of the cell lines. Our data suggest that abnormalities in structure and expression of the RAR-beta gene may be involved in the pathogenesis of lung cancer.

We have isolated a cDNA clone encoding the major protein-tyrosine-phosphatase (protein-tyrosine-phosphate phosphohydrolase, EC 3.1.3.48) of human placenta. Degenerate oligonucleotides, based on the amino acid sequence of the protein, were used to amplify an internal fragment of the gene from human placental cDNA by the polymerase chain reaction. This fragment was then used to probe a human placental cDNA library. A 3.3-kilobase (kb) insert was isolated and sequenced. The insert has a single extended open reading frame that predicts a 435 amino acid protein of Mr approximately 50,000. From the amino terminus to residue 321, the deduced amino acid sequence is identical to that previously determined by peptide sequencing [Charbonneau, H., Tonks, N. K., Kumar, S., Diltz, C. D., Harrylock, M., Cool, D. E., Krebs, E. G., Fischer, E. H. & Walsh, K. A. (1989) Proc. Natl. Acad. Sci. USA 86, 5252-5256]; however, the sequence predicts that the protein contains an additional 114 amino acids not present in the reported peptide sequence. In vitro translation of the 3.3-kb insert produces a protein of Mr 56,000, in general agreement with the predicted size. The phosphatase gene appears to be present as a single copy in human genomic DNA and is transcribed into a 3.5-kb message in a variety of tissues.

The human germ-line positions of the oncogenes ABL, SIS, and FES, the cellular counterparts of the v-onc genes of Abelson murine leukemia virus, simian sarcoma virus, and feline sarcoma virus, respectively, have been determined by in situ molecular hybridization of 3H-labeled v-onc gene probes to meiotic pachytene chromosomes. The position of ABL at 9q34.1 corresponds to the breakpoint in chromosome 9 in the translocation that gives rise to the Philadelphia chromosome, t(9;22) (q34; q11); the position of SIS at 22q13.1 is distal to the breakpoint in this chromosome. FES at 15q26.1 is also distal to the breakpoint in chromosome 15 in the translocation commonly seen in acute promyelocytic leukemia, t(15;17) (q24;q22).

The c-ras family is a set of c-onc genes that are highly conserved in vertebrates. The genes in this family are homologous to the transforming genes of Harvey and Kirsten murine sarcoma viruses (v-Ha-ras and v-Ki-ras, respectively). Using an in situ molecular hybridization method, we detected three sites on the human pachytene chromosomes that exhibited significant hybridization to v-Ki-ras and v-Ha-ras probes. These were chromomere positions that corresponded to bands 11p14.1, 12p12.1, and 12q24.2 of somatic chromosomes. The relationship between these chromosomal sites and previously defined members of the human c-ras gene family is discussed. These chromosomal sites are known to be involved in specific chromosome changes in a variety of tumors and in several congenital disorders that predispose to neoplastic disease.

We have used in situ chromosome hybridization techniques to map the human cellular counterparts (c-onc genes) of the transforming genes of two RNA tumor viruses on human meiotic pachytene and somatic metaphase chromosomes. We find that the human c-mos gene is located on chromosome 8 at a position corresponding to band 8q22 on the somatic map. The human c-myc gene is found on chromosome 8 at position 8q24. These regions on the long arm of chromosome 8 have been previously reported to be involved in specific translocations found in the M-2 subset of acute nonlymphoblastic leukemias. Burkitt lymphoma, and other forms of non-Hodgkin lymphoma, and a familial abnormality that predisposes to renal cell carcinoma. These results suggest that translocations of the human c-mos or c-myc genes may be causally related to neoplastic transformation.

We isolated molecular clones of the provirus-host cell junctions (tumor junction fragments) from two avian leukosis virus-induced lymphomas and compared the structures of these clones with a clone of the normal c-myc gene. Restriction mapping and DNA sequencing demonstrated that normal proviral integration events occurred adjacent to c-myc in both tumors, without gross structural alteration of c-myc. The right long terminal repeat of an avian leukosis virus provirus is integrated upstream from the bulk of the c-myc coding sequences and oriented such that transcription can initiate within the long terminal repeat and proceed downstream into c-myc. A comparison of a tumor junction fragment with the v-myc gene showed that there are two regions of v-myc-related sequences (which are probably exons) separated by 1 kilobase of sequences unrelated to v-myc (probably an intron). A DNA sequence analysis of the tumor junction fragments suggested that integration had occurred in exons adjacent to splice donor sites. This suggests that there are additional exons and introns in c-myc. Based on these findings, a model is proposed for the genesis of the tumor-specific RNAs containing viral-5' and c-myc information in avian leukosis virus-induced lymphomas.

We have isolated a replication-defective rapidly transforming sarcoma virus (designated 16L virus) from a fibro-sarcoma in a chicken infected with td107A, a transformation-defective deletion mutant of subgroup A Schmidt-Ruppin Rous sarcoma virus. 16L virus transforms fibroblasts and causes sarcomas in infected chickens within 2 wk. Its genomic RNA is 6.0 kilobases and contains sequences homologous to the transforming gene (fps) of Fujinami sarcoma virus (FSV). RNase T1 oligonucleotide analysis shows that the 5' and 3' terminal sequences of 16L virus are indistinguishable from (and presumably derived from) td107A RNA. The central part of 16L viral RNA consists of fps-related sequences. These oligonucleotides fall into four classes: (i) oligonucleotides common to the putative transforming regions of FSV and another fps-containing avian sarcoma virus, UR1; (ii) an oligonucleotide also present in FSV but not in UR1; (iii) an oligonucleotide also present in UR1 but not in FSV; and (iv) an oligonucleotide not present in either FSV, UR1, or td107A. Cells infected with 16L virus synthesize a protein of Mr 142,000 that is immunoprecipitated with anti-gag antiserum. This protein has protein kinase activity. These results suggest that 16L virus arose by recombination between td107A and the cellular fps gene.