A cell cycle and mutational analysis of anchorage-independent growth: cell adhesion and TGF-beta 1 control G1/S transit specifically.

We have examined cell cycle control of anchorage-independent growth in nontransformed fibroblasts. In previous studies using G0-synchronized NRK and NIH-3T3 cells, we showed that anchorage-independent growth is regulated by an attachment-dependent transition at G1/S that resembles the START control point in the cell cycle of Saccharomyces cerevisiae. In the studies reported here, we have synchronized NRK and NIH-3T3 fibroblasts immediately after this attachment-dependent transition to determine if other portions of the fibroblast cell cycle are similarly regulated by adhesion. Our results show that S-, G2-, and M-phase progression proceed in the absence of attachment. Thus, we conclude that the adhesion requirement for proliferation of these cells can be explained in terms of the single START-like transition. In related studies, we show that TGF-beta 1 overrides the attachment-dependent transition in NRK and AKR-2B fibroblasts (lines in which TGF-beta 1 induces anchorage-independent growth), but not in NIH-3T3 or Balb/c 3T3 fibroblasts (lines in which TGF-beta 1 fails to induce anchorage-independent growth). These results show that (a) adhesion and TGF-beta 1 can have similar effects in stimulating cell cycle progression from G1 to S and (b) the differential effects of TGF-beta 1 on anchorage-independent growth of various fibroblast lines are directly reflected in the differential effects of the growth factor at G1/S. Finally, we have randomly mutagenized NRK fibroblasts to generate mutant lines that have lost their attachment/TGF-beta 1 requirement for G1/S transit while retaining their normal mitogen requirements for proliferation. These clones, which readily proliferate in mitogen-supplemented soft agar, appear non-transformed in monolayer: they are well spread, nonrefractile, and contact inhibited. The existence of this new fibroblast phenotype demonstrates (a) that the growth factor and adhesion/TGF-beta 1 requirements for cell cycle progression are genetically separable, (b) that the two major control points in the fibroblast cell cycle (G0/G1 and G1/S) are regulated by distinct extracellular signals, and (c) that the genes regulating anchorage-independent growth need not be involved in regulating contact inhibition, focus formation, or growth factor dependence.

Akt inhibitor A-443654 induces rapid Akt Ser-473 phosphorylation independent of mTORC1 inhibition.

Rapamycin, a natural product inhibitor of the Raptor-mammalian target of rapamycin complex (mTORC1), is known to induce Protein kinase B (Akt/PKB) Ser-473 phosphorylation in a subset of human cancer cell lines through inactivation of S6K1, stabilization of insulin receptor substrate (IRS)-1, and increased signaling through the insulin/insulin-like growth factor-I/phosphatidylinositol 3-kinase (PI3K) axis. We report that A-443654, a potent small-molecule inhibitor of Akt serine/threonine kinases, induces Akt Ser-473 phosphorylation in all human cancer cell lines tested, including PTEN- and TSC2-deficient lines. This phenomenon is dose-dependent, manifests coincident with Akt inhibition and likely represents an alternative, rapid-feedback pathway that can be functionally dissociated from mTORC1 inhibition. Experiments performed in TSC2-/- cells indicate that TSC2 and IRS-1 cooperate with, but are dispensable for, A-443654-mediated Akt phosphorylation. This feedback event does require PI3K activity, however, as it can be inhibited by LY294002 or wortmannin. Small interfering RNA-mediated knockdown of mTOR or Rictor, components of the rapamycin-insensitive mTORC2 complex, but not the mTORC1 component Raptor, also inhibited Akt Ser-473 phosphorylation induced by A-443654. Our data thus indicate that Akt phosphorylation and activity are coupled in a manner not previously appreciated and provide a novel mode of Akt regulation that is distinct from the previously described rapamycin-induced IRS-1 stabilization mechanism.

Prevalence of narcolepsy-cataplexy in Korean adolescents.

BACKGROUND: Narcolepsy typically begins between adolescence and early adulthood causing severe neuropsychiatric impairments, but few prevalence studies are available on adolescent narcoleptics. In the present study, we investigated the prevalence of narcolepsy-cataplexy in adolescents. METHODS: In total 20,407 students, aged 14-19 years, participated in this study. Ullanlinna Narcolepsy Scale (UNS) was applied to all subjects and those with a UNS score of > or =14 were contacted by phone for semi-structured interview. Subjects then suspected of having narcolepsy participated in a laboratory investigation, which included polysomnography and HLA typing, or were interviewed in detail by telephone. RESULTS: Three subjects were finally diagnosed as narcolepsy with cataplexy and seven subjects might be diagnosed as narcolepsy without cataplexy. Among three narcoleptics with cataplexy, two subjects were HLA-DQB1*0602 and DRB1*1501 positive, but one subject had no test of HLA typing. The prevalence of narcolepsy with cataplexy in Korean adolescence was thus determined to be 0.015% (95% confidence interval = 0.0-0.0313%). CONCLUSION: This epidemiologic study is the first of its type on adolescent narcolepsy to use the International Classification of Sleep Disorders, 2nd edition (ICSD-2) diagnostic criteria. Considering those cases with an onset after adolescence were not included, the prevalence of narcolepsy with cataplexy determined in the present study is comparable with that of other studies in adults.

Overexpression of cyclin E in the HC11 mouse mammary epithelial cell line is associated with growth inhibition and increased expression of p27(Kip1).

To elucidate the role of cyclin E in cell growth and tumorigenesis in mammary epithelial cells, we have used retrovirus-mediated transduction to generate derivatives of the nontransformed HC11 mouse mammary epithelial cell line that stably express a human cyclin E cDNA (HU4). These derivatives expressed two distinct forms of the exogenous cyclin E protein, which were about M(r) 50,000 and M(r) 42,000, thus corresponding to endogenous cyclin E proteins found in human cells. In contrast to results obtained previously in fibroblasts, overexpression of the HU4 cyclin E cDNA in HC11 cells was associated with an increase in cell size, lengthening of G(1), and inhibition of both anchorage-dependent and independent growth. Furthermore, when quiescent serum-starved cells were restimulated with serum, entry into the S-phase was delayed in the overexpressor cells. Under these conditions, there was also delayed induction in the expression of the endogenous cyclin E protein and in other events involved in the G(1) transition. Despite the high level of expression of the exogenous cyclin E, the derivatives did not display increased cyclin E-associated in vitro kinase activity. The HC11 cells that overexpressed the exogenous cyclin E displayed an increase in the cyclin/cyclin-dependent kinase inhibitor p27(Kip1) in both asynchronous exponentially dividing and synchronous cell populations. These findings indicate that increased expression of this cyclin E cDNA in HC11 cells inhibits rather than stimulates growth and that this may be due to increased expression of the inhibitor p27(Kip1).

Antisense to cyclin D1 inhibits the growth and tumorigenicity of human colon cancer cells.

Cyclin D1 plays an important role in regulating the progression of cells through the G1 phase of the cell cycle. This gene is frequently overexpressed in human colon cancer. To address the role of cyclin D1 in growth control and tumorigenesis in this disease, we have overexpressed an antisense cyclin D1 cDNA construct in the human colon cancer cell line SW480E8, which expresses high levels of cyclin D1. The integration and expression of the antisense construct was verified by Southern and Northern blot analyses, respectively, and resulted in decreased expression of the cyclin D1 protein. This was associated with decreased levels of the Rb and p27Kip1 proteins. In addition, the hypophosphorylated form of Rb was increased in these cells. The SW480E8 antisense cyclin D1 cells displayed an increased doubling time, a decrease in saturation density, decreased plating efficiency and anchorage-independent growth, and a loss of tumorigenicity in nude mice. These findings provide direct evidence that increased expression of cyclin D1 in colon tumor cells contributes to their abnormal growth and tumorigenicity. The ability to revert the transformed phenotype of these cells with antisense cyclin D1 suggests that cyclin D1 or its associated cyclin-dependent kinase 4 may be useful targets in the therapy of colon cancer.

Identification and characterization of A-105972, an antineoplastic agent.

A high-throughput screening assay was designed to select compounds that inhibit the growth of cultured mammalian cells. After screening more than 60,000 compounds, A-105972 was identified and selected for further testing. A-105972 is a small molecule that inhibits the growth of breast, central nervous system, colon, liver, lung, and prostate cancer cell lines, including multidrug-resistant cells. The cytotoxic IC50 values of A-105972 were between 20 and 200 nM, depending on the specific cell type. The potency of A-105972 is similar in cells expressing wild-type or mutant p53. A majority of cells treated with A-105972 were trapped in the G2-M phases, suggesting that A-105972 inhibits the progression of the cell cycle. Using [3H]A-105972, we found that A-105972 bound to purified tubulin. Unlabeled A-105972 competed with [3H]A-105972 binding with an IC50 value of 3.6 microL. Colchicine partially inhibited [3H]A-105972 binding with an IC50 value of approximately 90 microM, whereas paclitaxel and vinblastine had no significant effect. Tumor cells treated with A-105972 were observed to contain abnormal microtubule arrangement and apoptotic bodies. DNA ladder studies also indicated that A-105972 induced apoptosis. A-105972 caused a mobility shift of bcl-2 on SDS-PAGE, suggesting that A-105972 induced bcl-2 phosphorylation. A-105972 treatment increased the life span of mice inoculated with B16 melanoma, P388 leukemia, and Adriamycin-resistant P388. These results suggest that A-105972 is a small molecule that interacts with microtubules, arrests cells in G2-M phases, and induces apoptosis in both multidrug resistance-negative and multidrug resistance-positive cancer cells. A-105972 and its analogues may be useful for treating cell proliferative disorders such as cancer.

A-204197, a new tubulin-binding agent with antimitotic activity in tumor cell lines resistant to known microtubule inhibitors.

Drug resistance is a prevalent problem in the treatment of neoplastic disease, and the effectiveness of many clinically useful drugs is limited by the fact that they are substrates for the efflux pump, P-glycoprotein. Because there is a need for new compounds that are effective in treating drug-resistant tumors, we tested A-204197 (4-[4-acetyl-4,5-dihydro-5-(3,4,5-trimethoxyphenyl)-1,3,4-oxadiazol-2-yl]-N,N-dimethylbenzeneamine), a novel oxadiazoline derivative with antiproliferative properties, on cell lines that were either sensitive or resistant to known microtubule inhibitors. Cell lines that were resistant to paclitaxel, vinblastine, or colchicine were equally sensitive to A-204197 (proliferation IC50s ranging from 36 to 48 nM) despite their expression levels of P-glycoprotein. The effect of A-204197 on cell growth was associated with cell cycle arrest in G2-M, increased phosphorylation of select G2-M checkpoint proteins, and apoptosis. In competition-binding assays, A-204197 competed with [3H]-labeled colchicine for binding to tubulin (K(i) = 0.75 microM); however, it did not compete with [3H]-labeled paclitaxel. A-204197 prevented tubulin polymerization in a dose-dependent manner (IC50 = 4.5 microM) in vitro and depolymerized microtubules in a time-dependent manner in cultured cells. These findings indicate A-204197 is a promising new tubulin-binding compound with antimitotic activity that has potential for treating neoplastic diseases with greater efficacy than currently used antimitotic agents.

PKC epsilon functions as an oncogene by enhancing activation of the Raf kinase.

Previous studies have indicated that PKCepsilon behaves as an oncogene when overproduced in rodent fibroblasts (Cacace et al., 1993; Mishak et al., 1993). In the present study, Western blot analysis revealed that the hyperphosphorylated form of Raf kinase was present at a high level in PKCepsilon overproducing R6 rat fibroblasts but not in R6 fibroblasts overproducing PKCalpha or beta1. Extracts from the PKCepsilon overproducing cells also exhibited a marked increase in Raf-1 kinase and MAP-kinase activity. To investigate the significance of these findings, dominant negative mutants of ras (N17) or raf (301-1) were stably expressed in early passage control and PKCepsilon-transformed R6 fibroblasts, by transduction using retrovirus-derived constructs. Dominant negative raf expressing clones exhibited a flat morphology, a decreased saturation density, and decreased growth in soft agar. In addition, these reverted clones exhibited decreased Raf kinase activity. In contrast, dominant negative ras expressing clones remained highly transformed. In addition, PKCepsilon was detected in Raf-1 immunoprecipitates indicating that PKCepsilon forms a complex with Raf-1 in vivo. Taken together, these results suggest that PKCepsilon functions as an oncogene in R6 cells by enhancing activation of the Raf-1 kinase.

The alpha isoform of protein kinase C mediates phorbol ester-induced growth inhibition and p21cip1 induction in HC11 mammary epithelial cells.

To clarify the roles of specific isoforms of PKC in regulating growth and cell cycle progression of the HC11 mammary epithelial cell line, we investigated the effects of activating endogenous PKC isoforms with the phorbol ester tumor promoter TPA, and also the effects of TPA on genetically engineered cells containing increased levels of individual PKC isoforms. We found that TPA treatment of HC11 cells induced a transient cell cycle arrest in G0/G1. Western blot analyses of the TPA treated cells provided evidence that the endogenous PKC alpha present in these cells mediated these effects. Indeed, derivatives of the HC11 cell line that inducibly overexpress an exogenous PKC alpha or ectopic PKC beta 1 exhibited more marked growth inhibition by TPA than control cells. Immunohistochemical staining of cells following treatment with TPA revealed selective translocation of PKC alpha into the nucleus, whereas PKC beta 1 remained in the cytoplasm. The transient arrest of HC11 cells following treatment with TPA was associated with marked induction of both p21cip1 mRNA and protein. This induction was exaggerated in the derivatives that overexpressed either PKC alpha or PKC beta 1. Therefore, in mouse mammary epithelial cells activation of the endogenous PKC alpha can transiently arrest cells in G0/G1 which may be due, at least in part, to induction of the transcription of p21cip1.

Stable overexpression of cyclin D1 in a human mammary epithelial cell line prolongs the S-phase and inhibits growth.

Amplification and/or increased expression of cyclin D1 occurs in an appreciable fraction of primary human breast carcinomas and several other types of human cancer. In addition, overexpression of cyclin D1 in rodent fibroblasts enhances growth and malignant transformation. The present study demonstrates that the extent of amplification and expression of cyclin D1 varies widely amongst a series of cell lines established from normal human mammary epithelium or human breast carcinomas. The HBL-100 mammary epithelial cell line did not display amplification or increased expression of cyclin D1. We used retrovirus-mediated transduction to obtain derivatives of this cell line that stably expressed relatively high levels of an exogenous cyclin D1 cDNA. These derivatives displayed an increased doubling time, decreased saturation density, decreased cloning efficiency, decreased anchorage-independent growth, an increased fraction of cells in the S-phase, and decreased tumorigenicity. Thus, increased expression of cyclin D1 in this cell line markedly inhibits rather than enhances growth, which may be due to the prolongation of S-phase.

Disorders in cell circuitry associated with multistage carcinogenesis: exploitable targets for cancer prevention and therapy.

The development of a malignant tumor involves the progressive acquisition of mutations and epigenetic abnormalities in multiple genes that have highly diverse functions. Some of these genes code for pathways of signal transduction that mediate the action of growth factors. The enzyme protein kinase C plays an important role in these events and in the process of tumor promotion. Therefore, we examined the effects of three inhibitors of protein kinase C, CGP 41251, RO 31-8220, and calphostin C, on human glioblastoma cells. These compounds inhibited growth and induced apoptosis; these activities were associated with a decrease in the level of CDC2 and cyclin B1/CDC2-associated kinase activity. This may explain why the treated cells accumulated in G2-M. In a separate series of studies, we examined abnormalities in cell cycle control genes in human cancer. We have found that cyclin D1 is frequently overexpressed in a variety of human cancers. Mechanistic studies indicate that cyclin D1 can play a critical role in carcinogenesis because: overexpression enhances cell transformation and tumorigenesis; introduction of an antisense cyclin D1 cDNA into either human esophageal or colon cancer cells reverts their malignant phenotype; and overexpression of cyclin D1 can enhance the amplification of other genes. The latter finding suggests that cyclin D1 can enhance genomic instability and, thereby, the process of tumor progression. Therefore, inhibitors of the function of cyclin D1 may be useful in both cancer chemoprevention and therapy. We obtained evidence for the existence of homeostatic feedback loops between cyclins D1 or E and the cell cycle inhibitory protein p27Kip1. On the basis of these and other findings, we hypothesize that, because of their disordered circuitry, cancer cells suffer from "gene addiction" and "gene hypersensitivity," disorders that might be exploited in both cancer prevention and therapy.

Blocking Chk1 expression induces apoptosis and abrogates the G2 checkpoint mechanism.

Checkpoint kinase 1 (Chk1) is a checkpoint gene that is activated after DNA damage. It phosphorylates and inactivates the Cdc2 activating phosphatase Cdc25C. This in turn inactivates Cdc2, which leads to G2/M arrest. We report that blocking Chk1 expression by antisense or ribozymes in mammalian cells induces apoptosis and interferes with the G2/M arrest induced by adriamycin. The Chk1 inhibitor UCN-01 also blocks the G2 arrest after DNA damage and renders cells more susceptible to adriamycin. These results indicate that Chk1 is an essential gene for the checkpoint mechanism during normal cell proliferation as well as in the DNA damage response.

Downregulation of Akt1 inhibits anchorage-independent cell growth and induces apoptosis in cancer cells.

The serine/threonine kinases, Akt1/PKBalpha, Akt2/PKBbeta, and Akt3/PKBgamma, play a critical role in preventing cancer cells from undergoing apoptosis. However, the function of individual Akt isoforms in the tumorigenicity of cancer cells is still not well defined. In the current study, we used an Akt1 antisense oligonucleotide (AS) to specifically downregulate Akt1 protein in both cancer and normal cells. Our data indicate that Akt1 AS treatment inhibits the ability of MiaPaCa-2, H460, HCT-15, and HT1080 cells to grow in soft agar. The treatment also induces apoptosis in these cancer cells as demonstrated by FACS analysis and a caspase activity assay. Conversely, Akt1 AS treatment has little effect on the cell growth and survival of normal human cells including normal human fibroblast (NHF), fibroblast from muscle (FBM), and mammary gland epithelial 184B5 cells. In addition, Akt1 AS specifically sensitizes cancer cells to typical chemotherapeutic agents. Thus, Akt1 is indispensable for maintaining the tumorigenicity of cancer cells. Inhibition of Akt1 may provide a powerful sensitization agent for chemotherapy specifically in cancer cells.

Modulation of drug resistance by alpha-tubulin in paclitaxel-resistant human lung cancer cell lines.

Beta(beta)-tubulin isotype variation has recently been implicated in the modulation of resistance to paclitaxel in human lung cancer cells and in primary human ovarian tumour samples. Whether alpha-tubulin is involved in drug resistance has not been reported. We have generated a paclitaxel-resistant cell line (H460/T800) from the sensitive human lung carcinoma parental cell line NCI-H460. The resistant cells are more than 1000-fold resistant to taxol and overexpress P-glycoprotein. Interestingly, H460/T800 cells also overexpress alpha- and beta-tubulin as detected by Western blot analysis. From Northern blot analysis, the mechanism of tubulin overexpression appears to be post-transcriptional. To understand whether alpha-tubulin plays a role in drug resistance, we transfected antisense human kalpha1 cDNA construct into the H460/T800 paclitaxel-resistant cells. The antisense clones displayed a reduced alpha-tubulin expression, and the cells were 45-51% more sensitive to paclitaxel and other known antimitotic drugs, compared with vector transfected controls. Complementary experiments of transfecting the sense kalpha1 cDNA into H460 cells conferred a 1.8- to 3.3-fold increase in the IC(50) of several antimitotic agents. Our study suggests that alpha-tubulin is one of the factors that contributes to drug resistance.

Biochemical and genetic characterization of the multidrug resistance phenotype in murine macrophage-like J774.2 cells.

The development of multidrug resistance (MDR) in malignant tumors is a major obstacle to the treatment of many cancers. MDR sublines have been derived from the J774.2 mouse macrophage-like cell line and utilized to characterize the phenotype at the biochemical and genetic level. Two isoforms of the drug resistance-associated P-glycoprotein are present and distinguishable both electrophoretically and pharmacologically. Genetic analysis has revealed the presence of a three-member gene family; expression of two of these genes, mdr1a and mdr1b, is associated with MDR whereas the expression of the third, mdr2, is not. Studies of these three genes have revealed similarities and differences in the manner in which they are regulated at the transcriptional level, and have suggested that post-transcriptional effects may also be important.

Sulindac derivatives inhibit growth and induce apoptosis in human prostate cancer cell lines.

We examined the activity of two metabolites of sulindac (a nonsteroidal anti-inflammatory drug), sulindac sulfide and sulindac sulfone (exisulind, Prevatec), and a novel highly potent analog of exisulind (CP248) on a series of human prostate epithelial cell lines. Marked growth inhibition was seen with the BPH-1, LNCaP, and PC3 cell lines with IC50 values of about 66 microM, 137 microM, and 64 nM for sulindac sulfide, exisulind, and CP248, respectively. DNA flow cytometry and 4',6'-diamido-2-phenylindole (DAPI) staining indicated that these three compounds also induced apoptosis in all of these cell lines. Similar growth inhibition also was seen with the PrEC normal human prostate epithelial cell line, but these cells were resistant to induction of apoptosis at concentrations up to 300 microM, 1 mM, and 750 nM of sulindac sulfide, exisulind, and CP248, respectively. Derivatives of LNCaP cells that stably overexpress bcl-2 remained sensitive to growth inhibition and induction of apoptosis by these compounds. In vitro enzyme assays indicated that despite its high potency in inhibiting growth and inducing apoptosis, CP248, like exisulind, lacked cyclooxygenase (COX-1 and COX-2) inhibitory activity even at concentrations up to 10 mM. Moreover, despite variations of COX-1 and COX-2 expression, the three benign and malignant prostate cell lines showed similar sensitivity to growth inhibition and induction of apoptosis by these three compounds. Therefore, sulindac derivatives can cause growth inhibition and induce apoptosis in human prostate cancer cells by a COX-1 and -2 independent mechanism, and this occurs irrespective of androgen sensitivity or increased expression of bcl-2. These compounds may be useful in the prevention and treatment of human prostate cancer.

Abrogation of G2 checkpoint specifically sensitize p53 defective cells to cancer chemotherapeutic agents.

BACKGROUND: Chkl is a checkpoint gene that is activated after DNA damage. It phosphorylates and inactivates Cdc25C at the late G2 phase. The inactivation of Cdc25C and consequently, the inactivation of Cdc2, are required for the G2 arrest induced by DNA damage. METHODS: We treated 184B5 cell line and its E6 transformed cell lines with adriamycin in the presence of staurosporine or UCNO1 and examined G2 arrest and cell death. RESULTS: We found that adriamycin induced a p53 and p21 response as well as a G1 arrest in 184B5 cells, but not in its E6 transformed cells. Staurosporine or UCNO1 abrogated the G2 arrest induced by adriamycin in both cell lines. In addition, staurosporine or UCNO1 specifically sensitized p53 incompetent cells to adriamycin. CONCLUSION: G2/M checkpoint abrogators can potentially enhance the cytotoxic effect of conventional chemotherapeutic reagents specifically to tumor cells.

Histological studies of surgically resected hepatocellular carcinoma following combined radiotherapy and hyperthermia.

Four cases of hepatocellular carcinoma (HCC) were surgically resected following combined radiotherapy (RT) and hyperthermia (HT). Complete necrosis of the tumor without viable tumor cell was found in one case and extensive tumor necrosis was observed in the other three cases; the percentage of necrosis in the specimens were 40%, 70%, and 80%, respectively. Histologic assessment showed mainly coagulative necrosis in the tumor with focal liquefactive necrosis. Cystic dilatation of sinusoids was observed in both tumor and nontumorous normal liver tissue. Other changes in normal liver tissue were unremarkable except for infiltration of inflammatory cells, fatty change, and proliferation of the bile ducts which can usually be seen beyond the area where any space occupying lesions are present. It is concluded that combined radiotherapy and hyperthermia can significantly induce coagulative necrosis of hepatocellular carcinoma with nonsignificant minimal histologic changes in adjacent nontumorous liver tissue.

Differential repair and recombination of psoralen damaged plasmid DNA in Saccharomyces cerevisiae.

Psoralen photoreaction with DNA produces interstrand crosslinks, which require the activity of excision and recombinational pathways for repair. Yeast replicating plasmids, carrying the HIS3, TRP1, and URA3 genes, were photoreacted with psoralen in vitro and transfected into Saccharomyces cerevisiae cells. Repair was assayed as the relative transformation efficiency. A recombination-deficient rad52 strain was the least efficient in the repair of psoralen-damaged plasmids; excision repair-deficient rad1 and rad3 strains had repair efficiencies intermediate between those of rad52 and RAD cells. The level of repair also depended on the conditions of transformant selection; repair was more efficient in medium lacking tryptophan than in medium from which either histidine or uracil was omitted. The plasmid repair differential between these selective media was greatest in rad1 cells, and depended on RAD52. Plasmid-chromosome recombination was stimulated by psoralen damage, and required RAD52 function. Chromosome to plasmid gene conversion was seen most frequently at the HIS3 locus. In RAD and rad3 cells, the majority of the conversions were associated with plasmid integration, while in rad1 cells most were non-crossover events. Plasmid to chromosome gene conversion was observed most frequently at the TRP1 locus, and was accompanied by plasmid loss.