Isolation and characterization of four type-1 ribosome-inactivating proteins, with polynucleotide:adenosine glycosidase activity, from leaves of Phytolacca dioica L.

Four type-1 (single-chain) ribosome-inactivating proteins (RIPs), with isoelectric points between 9.5 and 9.7, were isolated from leaves of Phytolacca dioica L. The purification procedure furnished the four proteins with an overall yield of about 16% and separated them from a protein of 29,407 +/- 2 Da, as determined by electrospray mass spectrometry, whose N-terminal amino acid sequence differed from that of pokeweed (Phytolacca americana L.) leaf chitinase (PLC-B) by only one amino acid (R17I). The four RIPs (PD-L1 to PD-L4) inhibited protein synthesis by a rabbit reticulocyte lysate with 50% inhibition at the picomolar level, and produced the beta-fragment, diagnostic of the specific enzymatic action of RIPs, on yeast ribosomes. Comparison of their N-terminal sequences, up to residue 45, showed that PD-L1 is identical to PD-L2 [designated the isoleucine (Ile) form from the N-terminal residue] and PD-L3 is identical to PD-L4 [designated the valine (Val) form from the N-terminal residue] and that there are 35 identical residues between the two forms. Furthermore, the Val form presents the same number of identical residues as PD-S2, an RIP isolated from the seeds of the same plant. With the exception of PD-L4, the purified RIPs gave a positive reaction when stained for sugars on SDS-PAGE gels and, when analyzed by electrospray mass spectrometry, had M(r) values of 32,715 +/- 1 (PD-L1), 31,542 +/- 1 (PD-L2), 30,356 +/- 1 (PD-L3) and 29,185 +/- 1 Da (PD-L4). The 1171 kDa difference in M(r), within the same RIP form, could be due to glycosylation. Like leaf saporins and many other RIPs, the four RIPs released several adenines from poly(A), herring sperm DNA and rRNA 16S + 23S, thus acting as polynucleotide:adenosine glycosidases. This property was less pronounced in PD-L1 and PD-L3 than in PD-L2 and PD-L4, respectively. The proteins PD-L1 and PD-L4 showed 3.7% reactivity with the antiserum anti-dianthin 32 and no reactivity with antisera to PAP-R saporin-S6, momordin 1 and even PD-S2, an RIP isolated from the seeds of the same plant. Protein PD-L4 showed 12.5% cross-reactivity with anti-PD-L1, while the opposite cross-reactivity was 100%.

Mitogenic and dedifferentiating effect of the K-fgf/hst oncogene on rat thyroid PC clone 3 epithelial cells.

Transfection of rat thyroid differentiated epithelial cells, PC clone 3, with the K-fgf/hst oncogene results in alleviation of thyrotropin dependency for growth and suppression of the differentiated phenotype without the acquisition of the fully transformed phenotype. An autocrine mechanism is responsible for these effects, since the proliferation of PC clone 3 cells, induced by K-FGF-conditioned medium, is blocked by anti-K-FGF-specific antibodies. Moreover, addition of K-FGF-conditioned medium inhibits the thyroid differentiated functions. Also, such other members of the fibroblast growth factor family as basic and acidic fibroblast growth factor are able to induce thyroid cell growth and to block expression of the differentiated functions.

Isolation of a human placenta cDNA coding for a protein related to the vascular permeability factor.

A human cDNA coding for a protein related to the vascular permeability factor (VPF) was isolated from a term placenta cDNA library; we therefore named its product placenta growth factor (PlGF). PlGF is a 149-amino-acid-long protein and is highly homologous (53% identity) to the platelet-derived growth factor-like region of human VPF. Computer analyses reveal a putative signal peptide and two probable N-glycosylation sites in the PlGF protein, one of which is also conserved in human VPF. By using N-glycosidase F, tunicamycin, and specific antibodies produced in both chicken and rabbit, we demonstrate that PlGF, derived from transfected COS-1 cells, is actually N-glycosylated and secreted into the medium. In addition, PlGF, like VPF, proves to be a dimeric protein. Finally, a conditioned medium from COS-1 cells containing PlGF is capable of stimulating specifically the growth of CPA, a line of endothelial cells, in vitro.

Protection of mice against tumor growth by immunization with an oncogene-encoded growth factor.

The K-fgf/hst oncogene encodes a growth factor of the fibroblast growth factor (FGF) family that is secreted and transforms cells through a mechanism of autocrine cell proliferation. K-fgf-transformed cells are highly tumorigenic in immunocompetent allogeneic and syngeneic animals. BALB/c mice were immunized with a bacterial fusion protein consisting of a portion of the MS2 polymerase and of the human K-FGF precursor lacking only the first 4 amino acids or with a recombinant protein corresponding to the mature, secreted form of K-FGF (176 amino acids). They were then challenged with syngeneic K-fgf- or H-ras-transformed cells. Vaccinated animals exhibited a significant degree of protection against tumor induction, which was specific for K-fgf-transformed cells and correlated with the ability of the immunized mice to produce high titers of anti-K-FGF antibodies. Thus immunization with a single oncogene product can protect animals against tumor cells expressing this oncogene.

Autocrine growth stimulation by secreted Kaposi fibroblast growth factor but not by endogenous basic fibroblast growth factor.

We studied the different potentials of a secreted and a nonsecreted member of the fibroblast growth factor (FGF) family to induce autocrine growth stimulation in human adrenal cortex carcinoma cells (SW-13). These epithelial cells express basic FGF (bFGF) cell surface receptors, and picomolar concentrations of bFGF suffice to induce anchorage-independent growth. The requirement for exogenously added bFGF contrasts with the intracellular storage of biologically active bFGF in SW-13 cells greater than 10,000-fold in excess of the concentration needed to stimulate anchorage independent growth. To study whether the expression of a secreted FGF would alter the growth phenotype of these cells, we transfected them with an expression vector coding for the Kaposi-fgf (K-fgf) oncogene. In contrast to controls, K-fgf-transfected cells secrete significant amounts of biologically active K-fgf protein into the growth media, show up to 50-fold increased colony formation in soft agar, and grow into rapidly progressing, highly vascularized tumors in athymic nude mice. A reversible inhibition of the autocrine growth stimulation in vitro is brought about by the polyanionic compound suramin. We conclude that FGF has to be released from SW-13 cells to function fully as a growth stimulator in vitro and in vivo.

Expression of the K-fgf protooncogene is repressed during differentiation of F9 cells.

Utilizing F9 embryonal carcinoma cells as a model system for early mammalian development, we have studied the pattern of expression of the endogenous murine homolog of the human K-fgf/hst oncogene, which encodes a new member of the fibroblast growth factors (FGFs) family. The K-fgf mRNA is expressed in undifferentiated F9 cells and its level becomes undetectable upon the induction of differentiation. Furthermore, a growth-promoting activity with properties identical to those of K-FGF is present in the conditioned medium of F9 cells, but absent in that of differentiated cells. Shut-off of K-fgf expression is mediated at the transcriptional level. The acidic FGF gene is also expressed in undifferentiated F9 cells and down modulated once differentiation is induced. In contrast, int-2, another member of the FGF gene family, is transcriptionally induced in differentiated F9 cells. Our data suggest that single members of the FGF gene family may perform distinct functions in vivo, and that the physiological role of K-FGF may be related to early development.

bFGF as an autocrine growth factor for human melanomas.

Normal human melanocytes in culture require specific additives such as basic fibroblast growth factor (bFGF) and dibutyryl cyclic adenosine monophosphate (dbcAMP) in order to proliferate in defined or serum-containing medium (Halaban et al., 1987). This stringent requirement is absent in cells derived from metastatic melanomas which not only proliferate in regular culture medium, but also produce a substance immunologically related to bFGF (Halaban et al., 1987). We show here that the mitogenic activity necessary for normal human melanocytes is constitutively present in several lines of human metastatic melanomas and that this activity is inactivated by anti-bFGF antibodies. Melanoma cells, but not normal melanocytes, express bFGF gene transcripts. Although the molecular mechanism underlying the abnormal expression of bFGF in melanomas is not known, the results suggest that bFGF acts as an autocrine growth factor in melanomas.

Processing, secretion, and biological properties of a novel growth factor of the fibroblast growth factor family with oncogenic potential.

We recently reported that the protein encoded in a novel human oncogene isolated from Kaposi sarcoma DNA was a growth factor with significant homology to basic and acidic fibroblast growth factors (FGFs). To study the properties of this growth factor (referred to as K-FGF) and the mechanism by which the K-fgf oncogene transforms cells, we have studied the production and processing of K-FGF in COS-1 cells transfected with a plasmid encoding the K-fgf cDNA. The results show that, unlike basic and acidic FGFs, the K-FGF protein is cleaved after a signal peptide, glycosylated, and efficiently secreted as a mature protein of 176 or 175 amino acids. Inhibition of glycosylation impaired secretion, and the stability of the secreted K-FGF was greatly enhanced by the presence of heparin in the cultured medium. We have used the conditioned medium from transfected COS-1 cells to test K-FGF biological activity. Similar to basic FGF, the K-FGF protein was mitogenic for fibroblasts and endothelial cells and induced the growth of NIH 3T3 mouse cells in serum-free medium. Accordingly, K-fgf-transformed NIH 3T3 cells grew in serum-free medium, consistent with an autocrine mechanism of growth. We have also expressed the protein encoded in the K-fgf protooncogene in COS-1 cells, and it was indistinguishable in its molecular weight, glycosylation, secretion, and biological activity from K-FGF. Taken together, these results suggest that the mechanism of activation of this oncogene is due to overexpression rather than to mutations in the coding sequences.

The FGF-related oncogene, K-FGF, maps to human chromosome region 11q13, possibly near int-2.

The protein encoded in a novel human oncogene isolated by transfection of Kaposi's sarcoma DNA is a growth factor with significant homology to basic and acidic FGFs. The genomic structure of this oncogene (designated K-FGF), as originally isolated, carried DNA rearrangements upstream and downstream of the coding region. The normally discontinuous sequence upstream of the K-FGF coding region derived from the 3' end of the c-fms gene and thus originated from human chromosome 5. In order to determine the normal chromosomal location of the K-FGF gene and of the DNA sequences adjacent to its 3' end, we have correlated the presence of these sequences with retention of specific human chromosome regions in rodent-human somatic cell hybrids. These experiments mapped the K-FGF gene to human chromosome region 11q13----11q23, and in situ hybridization localized it more precisely to region 11q13 near int-2, which also belongs to the FGF family. The sequence downstream of the gene in transfectants and discontinuous with K-FGF in normal human DNA derives from chromosome region 12p12----12q13, possibly near the int-1 locus.

Isolation of a rearranged human transforming gene following transfection of Kaposi sarcoma DNA.

By transfecting high molecular weight DNA from a Kaposi sarcoma lesion into murine NIH 3T3 cells, we have identified and molecularly cloned a set of human DNA sequences capable of inducing focus formation, growth in agar, and tumorigenicity in these cells. The human DNA sequences present in primary, secondary, and tertiary NIH 3T3 transformants encompass about 32 kilobases (kb) and contain four rearrangements with respect to normal human DNA and a portion of the c-fms protooncogene (FMS in human gene nomenclature). However, the minimal transforming region (6.6 kb) identified in our cloned DNA borders on the c-fms DNA region but does not contain c-fms coding sequences. The fms sequences are also not represented in the two transcripts (approximately equal to 1.2 and 3.5 kb) detected in NIH 3T3 transformants; however, they might provide elements regulating expression. Hybridization to several known oncogene probes and preliminary sequencing data indicate that we have identified a previously unrecognized "activated" oncogene. Since the rearrangements present in our cloned DNA sequences are not detectable in the original Kaposi tumor DNA used for transfection, it is possible that this oncogene was generated during gene transfer.

An oncogene isolated by transfection of Kaposi's sarcoma DNA encodes a growth factor that is a member of the FGF family.

We recently reported the cloning of a rearranged human oncogene following transfection of DNA from Kaposi's sarcoma into NIH 3T3 cells. To identify the protein(s) encoded in two novel mRNAs of 3.5 and 1.2 kb expressed in NIH 3T3 transformants, we constructed a cDNA library. One of the cDNA clones isolated (KS3) corresponded to the 1.2 kb mRNA and transformed NIH 3T3 cell when inserted into a mammalian expression vector. The 1152 nucleotide KS3 cDNA encodes a protein of 206 amino acids with significant homology to the growth factors basic FGF and acidic FGF. Expression of the KS3 product as a bacterial fusion protein or in COS cells allowed us to determine that both proteins had significant growth-promoting activity and that the COS cell protein was glycosylated. Thus one of the mRNAs transcribed from the KS oncogene encodes a growth factor that could transform cells by an autocrine mechanism and appears to represent a new member of the FGF family.

Presence of chromosomal abnormalities and lack of AIDS retrovirus DNA sequences in AIDS-associated Kaposi's sarcoma.

The frequent occurrence of Kaposi's sarcoma (KS) in association with the acquired immune deficiency syndrome (AIDS) could be due to the fact that the etiological agent of this tumor is the same retrovirus causing AIDS, to another oncogenic virus frequently found in AIDS patients, or to the unmasking of the tumorigenic potential of KS cells by immunosuppression. We have therefore investigated the presence of DNA sequences homologous to the AIDS retrovirus, cytomegalovirus (CMV), and hepatitis B virus in 13 KS necropsies and biopsies from AIDS patients. All KS DNA samples were negative for AIDS retrovirus or hepatitis B DNA sequences. Two DNAs from necropsies contained CMV DNA, but the data suggested the presence of replicating CMV DNA due to generalized infection. We have also studied cell cultures derived from KS skin biopsies of AIDS patients. These cultures had a short lifetime in vitro and expressed some markers of endothelial cells. The cells were not tumorigenic in nude mice but contained a number of chromosomal rearrangements which were often monoclonal within the same culture. However, these abnormalities were different from culture to culture and even in cultures from the same biopsy. The presence of these chromosomal abnormalities seemed to correlate with the cell positivity for endothelial markers. Taken together these results indicate that neither the AIDS retrovirus, CMV, or hepatitis B virus is directly responsible for the altered growth of KS cells, that KS may be polyclonal even within the same lesion, and that KS cells have a tendency to karyotypic rearrangements.

Selection of mouse neuroblastoma cell-specific polyoma virus mutants with stage differentiative advantages of replication.

Two mouse neuroblastoma cell lines were analyzed for their permissivity for polyoma virus growth. One (N18) is fully permissive for polyoma replication, the other (41A3) shows limited permissivity and the viral genome persists, without noticeable cell death. Virus persistence does not seem to alter the cells' ability to differentiate in vitro and leads to selection of viral mutants altered in the untranscribed regulatory region of the genome. The mutant types obtained appear to be related to the degree of host cell differentiation. Nucleotide sequence analysis of the restriction fragment covering the regulatory region shows that duplications are present in all mutants, while deletions in the non-duplicated segment are only present in mutants selected from less differentiated cells. These alterations involve both domains of the regulatory region that are considered to be essential for DNA replication and for enhancer activity. Mixed infections with polyoma wild type show that the selected mutants have cis-advantage in replication in neuroblastoma cells and not in 3T6 cells. Mutants carrying the deletion in the non-duplicated segment of the enhancer show a selective advantage in replication over the undeleted one in mixed infection. This advantage is much stronger in neuroblastoma cells in suspension (less-differentiated stage) than in monolayer cells (more-differentiated stage). An interpretation of the overall structure of the regulatory enhancer region, based on the observed differences between the mutants selected at different stages of differentiation in neuroblastoma and previously described mutants selected in undifferentiated cells, is discussed.

Polyomavirus growth and persistence in Friend erythroleukemic cells.

Infection of Friend erythroleukemic (FL) cells by polyomavirus (Py) invariably results in the selection of persistently infected FL-Py cell lines and clones. Anti-Py serum treatment of FL-Py lines and clones leads to the loss of Py genome and consequent cell cure. Conversely, cure has not been obtained in FL-PytsA cell lines (isolated after infection by a Py thermosensitive early mutant) and their derivative clones cultivated for a long time at nonpermissive temperature (39 degrees C), where viral large-T protein is inactive. Rescue of viral particles has always been obtained after shifting cells to 32 degrees C. Integrated viral genomes were detected by blot hybridization in an FL-PytsA clone at 39 degrees C. Long-term observation of FL-Py cell lines and their derivative clones reveals a reciprocal selection mechanism (coevolution) between the viral and the cellular populations, resulting in either a completely virus-free Py-resistant FL cell line (cure) or in a continuously Py-shedding line bearing Py genome variants. Structural analysis of these viral populations has been carried out, and some viral variants have been isolated and characterized. On the basis of the results obtained, the possible mechanisms of Py persistence in FL cells will be discussed.

Ultrastructural observations of polyoma infected Friend erythroleukemic cells.

Friend erythroleukemic cells (FLC) infected by Polyoma (Py) virus were studied at the electron microscope both during their lytic cycle, which occurs within the first 96 hr after infection, and in the surviving stabilized cells 30 days after infection. The results showed that the Polyoma virus and Friend leukemic virus (FLV) coexist and mature together in the same cell. During the lytic cycle, the Py particles were scattered in the nucleoplasm, and sometimes they were also found aggregated into crystals. Instead crystalline aggregates of Py particles were always present in the few surviving virus-carrying cells 30 days after infection. A possible interpretation could be that the crystal aggregation serves to maintain the viruses in a latency that allows the survival of the host cells.