Inhibition of proteolysis and cell cycle progression in a multiubiquitination-deficient yeast mutant.

The degradation of many proteins requires their prior attachment to ubiquitin. Proteolytic substrates are characteristically multiubiquitinated through the formation of ubiquitin-ubiquitin linkages. Lys-48 of ubiquitin can serve as a linkage site in the formation of such chains and is required for the degradation of some substrates of this pathway in vitro. We have characterized the recessive and dominant effects of a Lys-48-to-Arg mutant of ubiquitin (UbK48R) in Saccharomyces cerevisiae. Although UbK48R is expected to terminate the growth of Lys-48 multiubiquitin chains and thus to exert a dominant negative effect on protein turnover, overproduction of UbK48R in wild-type cells results in only a weak inhibition of protein turnover, apparently because the mutant ubiquitin can be removed from multiubiquitin chains. Surprisingly, expression of UbK48R complements several phenotypes of polyubiquitin gene (UB14) deletion mutants. However, UbK48R cannot serve as a sole source of ubiquitin in S. cerevisiae, as evidenced by its inability to rescue the growth of ubi1 ubi2 ubi3 ubi4 quadruple mutants. When provided solely with UbK48R, cells undergo cell cycle arrest with a terminal phenotype characterized by replicated DNA, mitotic spindles, and two-lobed nuclei. Under these conditions, degradation of amino acid analog-containing proteins is severely inhibited. Thus, multiubiquitin chains containing Lys-48 linkages play a critical role in protein degradation in vivo.

Effects of structural modifications of antitumor antibiotics (luzopeptins) on the interactions with deoxyribonucleic acid.

Luzopeptins consist of two identical, substituted quinolines linked to a cyclic decadepsipeptide, with a 2-fold symmetry. Luzopeptin A, with two acetylated sites in its peptide ring, is active against several experimental animal tumor systems. Luzopeptin B (one acetylated site) is less active, and luzopeptin C (no acetylation) is inactive. Our studies showed that all three luzopeptins and a half-molecule of luzopeptin C exhibited similar fluorescence (400 to 700 nm, with a peak at 490 to 496 nm) with an excitation spectrum in the 200 to 450 nm range (with a peak at 250 to 252, 337, and 385 nm). The half-molecule had the strongest fluorescence, followed in order by luzopeptins A, B, and C. DNA binding quenched both fluorescence and absorption of luzopeptins. Studies of the DNA-induced fluorescence and absorption quenching and the drug-induced viscosity and gel electrophoretic mobility changes of DNA suggested that luzopeptin C was slightly more effective than luzopeptins B and A in both the bifunctional DNA intercalation and the drug-induced DNA-DNA intermolecular cross-linking. Thus, the lack of antitumor activity of luzopeptin C is not the result of the lack of interactions with DNA. The half-molecule of luzopeptin C (quinoline with a pentapeptide) and smaller fragments (quinoline alone or with one to four peptide residues) did not react with DNA. Thus, the planar quinoline chromophore alone is unable to intercalate with DNA. The peptidic cyclic structure of luzopeptins is essential for the bifunctional intercalation of the twin chromophores, probably by providing proper conformational orientations of the chromophores.

Use of PM-2 DNA degradation as a pharmacokinetic assay for bleomycin.

The PM-2 DNA fluorescence assay has been shown to be a rapid, sensitive, and reproducible assay for bleomycin biochemical activity. The assay can detect bleomycin in human serum in the nmol range. The method measures DNA degradative activity of bleomycin and could be used to determine activity of bleomycin analogs and metabolites. The usefulness of the assay to perform bleomycin pharmacokinetic studies in cancer patients has been demonstrated. Linear regression analyses of parallel bleomycin assays with the radioimmunoassay gave a coefficient of correlation of 0.98 to 0.78 with trichloroacetic acid-treated serum. These results indicate excellent agreement between the two assays.

Comparison of DNA damage and single- and double-strand breakage activities on PM-2 DNA by talisomycin and bleomycin analogs.

Single- and double-strand breakage of isolated PM-2 DNA by structural analogs of the glycopeptide antitumor antibiotics bleomycin (BLM) and talisomycin (TLM) was investigated. Breakage of PM-2 DNA was determined by two systems: an ethidium bromide fluorescence assay; and agarose gel electrophoresis. The fluorescence assay, which measures total breakage of DNA including single- and double-strand breakage and alkaline labile damage, showed that the BLM's, A2 and B2 induced more total DNA breakage than did the TLM's A, B, S2b, and S10b. As measured by the comparison of the concentration of analog required to cause 50% breakage of superhelical DNA, BLM's A2 and B2 were 10 times more active than TLM's S2b and S10b and 25 times more active than TLM's A and B. Gel electrophoresis, which measures the extent of both single- and double-strand breakage of DNA, showed that at equivalent levels of breakage of superhelical DNA each of the TLM's caused more double-strand breakage of DNA than did the BLM's. Thus, the structural alterations near the bithiazole in the TLM's, which distinguish them structurally from the BLM's, result in a reduction of the total PM-2 DNA breakage activity and enhanced production of double-strand breaks relative to single-strand breaks by TLM when compared to BLM.

Comparative ultrastructural studies of nucleoli of tumor cells treated with adriamycin and the newer anthracyclines, carminomycin and marcellomycin.

This study was designed to determine the effects of several antitimor anthracyclines, including Adriamycin and its analogs, carminomycin and marcellomycin, on the ultrastructure of nucleoli of Novikoff hepatoma cells. Adriamycin and carminomycin, which are structurally related, induce nucleolar segregation following the formation of conspicuous fibrillar centers. Marcellomycin did not induce formation of nucleolar fibrillar centers. Instead, numerous microspherules formed following treatment with marcellomycin; later complete nucleolar segregation developed. The microspherules were observed to be in various stages of extrusion from the nucleolar body. This microspherule "migration" appeared to be both time and drug concentration dependent. These results show that the rate and extent of nucleolar ultrastructural aberration may be related to structural differences of the various anthracyclines.

Purification and mechanism of action of macromomycin.

Macromomycin (MCR), a polypeptide antibiotic previously shown to have antitumor activity in experimental tumors, has been purified into an electrophoretically homogeneous component with an approximate molecular weight of 12,500. MCR has alanine as an NH2-terminal amino acid, 4 cysteine residues, and no arginine or methionine residues. With a fluorescence assay and agarose gel electrophoresis, MCR was shown to induce strand breaks in PM2 DNA in vitro. 2-Mercaptoethanol inhibited the DNA cleavage activity of MCR. When incubated with Novikoff hepatoma ascites cells in tissue culture, MCR caused Novikoff hepatoma ascites cell DNA degradation as observed by the slower sedimentation of DNA on alkaline sucrose density gradient centrifugation when compared to untreated cell DNA. DNA synthesis in Novikoff hepatoma ascites cells was inhibited by 80% after a two-hr treatment with MCR (0.03 microgram/ml). RNA and protein syntheses were inhibited by 25 and less than 10%, respectively, at this concentration of drug. At a concentration of MCR (1.0 microgram/ml), syntheses of DNA and RNA in Novikoff hepatoma ascites cells were totally inhibited. The results of this study suggest that MCR may inhibit tumor cell growth by causing DNA breakage with subsequent inhibition of DNA and other macromolecule syntheses.

Ultrastructural study of the effect of bleomycin A2 on the nucleolus and its possibly related cytoplasmic constituents in Novikoff hepatoma cells.

Treatment of Novikoff hepatoma ascites cells with bleomycin A2, in vivo as well as in vitro, in varying doses produced marked alterations in nucleolar and cytoplasmic ultrastructural organization. A series of changes occur including formation of fibrillar centers, fragmentation of fibrillar nucleolar elements, appearance of microspherules, and the loss of fibrillar elements from these nucleoli. A bleomycin concentration of 10 mug/ml in vitro produced an increased number of fibrillar centers with well-defined nucleolonemas. At a concentration of 50 mug/ml, there was an increase in number and fragmentation of these fibrillar centers and many microspherules were found throughout the nucleolus. Approximately one-fourth of the cells contained cytoplasmic fibrillar bodies and amorphous fibrous tufts around the nuclear envelope. At a bleomycin concentration of 100 mug/ml, the nucleoli contained more granular elements and numerous microspherules. Almost 90% of the cells contained cytoplasmic fibrillar bodies. The effects of bleomycin in vivo (10 mg/kg) closely resemble those found in vitro with concentrations of 50 and 100 mug/ml.

In vivo resistance towards anthracyclines, etoposide, and cis-diamminedichloroplatinum(II).

From a single wild-type strain of Ehrlich ascites tumor, sublines resistant to daunorubicin, etoposide, and cis-diamminedichloroplatinum(II) have been developed in vivo. Different levels of resistance were achieved after 4 to 8 months for anthracyclines (greater than 32-fold), cis-diamminedichloroplatinum(II) (4-fold), and etoposide (greater than 6-fold). Anthracycline resistance was associated with decreased nuclear steady-state concentration of anthracyclines, increased content of high-molecular-weight membrane glycoproteins, and glucose-dependent drug extrusion after metabolic blockade with sodium azide. A similar "pump" system which was apparently not drug specific was also documented in etoposide resistance. Resistance towards cis-diamminedichloroplatinum(II) was accompanied by decreased cis-diamminedichloroplatinum(II)-induced DNA damage in vitro when proteinase K-resistant interstrand cross-links were measured by alkaline elution. Parallel in vivo studies revealed cross-resistance of various degrees among a number of anthracycline analogs, complete cross-resistance among daunorubicin, doxorubicin, and 4'-(9-acridinylamino)methanesulfon-M-anisidine (amsacrine), and partial cross-resistance between daunorubicin and etoposide. However, cis-diamminedichloroplatinum(II) was curative in anthracycline- and etoposide-resistant cells, as daunorubicin and etoposide were curative in acquired resistance towards cis-diamminedichloroplatinum(II). cis-Diamminedichloroplatinum(II) resistance was also overcome by the derivative 1,2,diaminocyclohexylplatinum malonate. The Vinca alkaloid vindesine, although only marginally active in the control tumor, was highly active in cells selected for cis-diamminedichloroplatinum(II) resistance. These in vivo patterns of cross-resistance and collateral sensitivity may be related to observations in clinical chemotherapy.

Effects of second-generation platinum analogs on isolated PM-2 DNA and their cytotoxicity in vitro and in vivo.

Using PM-2 DNA and cytotoxicity assay systems, we have studied several second-generation platinum analogs and compared these to the parent compound cis-diamminedichloroplatinum(II). These results indicate that all planar platinum(II) congeners induced similar effects upon interaction with PM-2 DNA, i.e., alteration of the tertiary DNA conformations. The reactivity of the analogs with DNA depended on the activity of the leaving groups. Octahedral platinum(IV) compounds, however, induced breakage of covalently closed circular PM-2 DNA, and the effects were not inhibited by either chloride or ethylenediaminetetraacetate. This suggests that breakage of isolated PM-2 DNA may be related to the axial trans bonds rather than the equatorial cis bonds of the solvated platinum(IV) compounds, since the activity of the dichloroplatinum(II) compounds has been shown to be inhibited by chloride ions. Studies on the in vitro and in vivo cytotoxicity of the platinum analogs demonstrated that the reactivity of analogs against PM-2 DNA correlated with in vitro and in vivo potencies. The reactivity with PM-2 DNA appeared to depend on the characteristics of the leaving group.

Interaction of cis-diamminedichloroplatinum(II) with PM-2 DNA.

The interaction of cis-diamminedichloroplatinum(II) (CDDP) with PM-2 DNA was studied using two techniques: (a) agarose gel electrophoresis of PM-2 DNA conformation isomers after CDDP binding; and (b) viscometric measurement of different forms of CDDP-bound PM-2 DNA. In both systems, the results indicated that the DNA isomers interacted differently with CDDP. CDDP induced a decrease of viscosity upon interacting with single-strand broken relaxed circular (Form II) and double-strand broken linear (Form III) PM-2 DNA's. These observations are consistent with a "DNA shortening effect" proposed by Cohen et al. [Science (Wash. D. C.), 203: 1014-1016, 1979] and Macquet et al. [Biochimie (Paris), 60: 901-914, 1978] When covalently closed circular (Form I) PM-2 DNA was used, increasing concentrations of CDDP induced an initial slight increase and then decrease of electrophoretic mobility to the degree that it comigrated with CDDP-bound Form II DNA. Further addition of CDDP restored the electrophoretic mobility of Form I DNA. Corresponding changes in the viscosity of CDDP-bound Form I DNA showed an initial decrease, then an increase, and a final prolonged decrease of viscosity. These effects are similar but not identical to those induced by either DNA intercalators (e.g., ethidium bromide) or certain DNA denaturating agents (e.g., formaldehyde, ultraviolet light, alkali trichloroacetate, methylmercuric hydroxide, and carbodiimide). Thus, CDP may induce a DNA superhelix-unwinding process followed either by rewinding or a denaturation process or both. Quantitative analysis of the agarose gel electrophoretic pattern plus sucrose density gradient centrifugation studies also indicated that there was little DNA strand breakage induced by CDDP treatment.

Acute ultrastructural effects of the antitumor antibiotic carminomycin on nucleoli of rat tissues.

Male Sprague-Dawley rats were treated with carminomycin i.v. in doses ranging from 1 to 40 mg/kg. Within 1 hr after the administration of carminomycin, 20 mg/kg, nucleoli of cardiac and skeletal muscle cells were segregated, while nucleoli of liver parenchyma cells were unaffected. Three and one-half hr after drug administration, cardiac muscle nucleoli reverted to normal ultrastructure. However, some skeletal muscle cell nucleoli were still segregated. Following treatment with carminomycin, 10 mg/kg, no significant ultrastructural changes were observed. These results demonstrate that at sufficiently high doses carminomycin induces ultrastructural lesions in nucleoli of both cardiac and muscle cells. The dose of carminomycin required to produce nucleolar segregation in cardiac and skeletal muscle is 6 times greater than the dose of Adriamycin (3.5 mg/kg) required to induce equivalent alterations.

Effects of DNA superhelical changes induced by ethidium bromide on the DNA-degrading activity of two antitumor antibiotics, bleomycin and phleomycin.

The effects of changes in the conformational state of DNA on the single-strand and double-strand breakage activity of two antitumor antibiotics, bleomycin (BLM) A2 and phleomycin D1, have been studied by the gel electrophoretic analysis of the drug-degraded PM2 phage superhelical DNA pretreated with an intercalating agent, ethidium bromide (EB). Both the single-strand and double-strand breakage activities of BLM A2 increased as the negatively superhelical turns of native PM2 DNA were gradually removed by intercalation with increasing EB concentrations. The activities peaked when DNA was completely relaxed and gradually decreased as the higher concentrations of EB twisted DNA into the positively superhelical form. The decrease in breakage activity was not due to any inhibitory effect of EB at higher concentrations, since treatment of the relaxed Form I0 DNA with low EB concentrations also reduced the activity. In contrast to BLM A2, phleomycin D1 responded minimally to DNA conformational changes, which suggested further that the two drugs may react with DNA differently. The differential responses of BLM A2 activity towards different DNA conformational states may have biological implications, since DNA in cells may exist in different conformational states relating to various gene functions. The current study may serve as a model for studying combined effects of intercalative and nonintercalative antitumor antibiotics which are used frequently in combination treatments of cancer.

Intermolecular cross-linking of DNA through bifunctional intercalation of an antitumor antibiotic, luzopeptin A (BBM-928A).

A bifunctional intercalator may intercalate with DNA in at least two ways. Both intercalating moieties may intercalate with the same DNA molecule (type I, intramolecular cross-linking) or with two separate DNA molecules (type II, intermolecular cross-linking). Production of type I is often assumed. Type II biintercalation has been suggested, but no direct evidence has been reported. In the present study, endonuclease-restricted PM2 phage or pBR322 plasmid DNA fragments were treated with the bifunctional intercalative antitumor antibiotics, luzopeptin A (BBM-928A) and echinomycin, and analyzed by agarose gel electrophoresis. Luzopeptin A treatment produced additional DNA bands which were the products of type II biintercalation. The types of restriction fragments involved were identified. Maximal type II biintercalation occurred at a luzopeptin A/DNA range of 0.14 to 0.18, at which more than 50% of the total DNA molecules were involved. Type II products were converted gradually to type I products upon prolonged incubation at 37 degrees, probably due to the tendency for intermolecular bonds to disrupt. Echinomycin treatment failed to produce type II products, probably because of a DNA-binding affinity weaker than that of luzopeptin A. Thus, it is possible to use the present gel system to demonstrate the type II biintercalation for strong biintercalators, but milder systems are needed for weak biintercalators.

Morphological effects of mitomycin C administered intravesically to normal mice and mice with N-[4-(5-nitro-2-furyl)-2-thiazolyl]-formamide-induce bladder neoplasms.

The intravesical administration of mitomycin C to normal and N-[4-(5-nitro-2-furyl)-2-thiozolyl]formamide-induced bladder tumor-carrying mice at initial concentrations of 2 and 5 mg/ml for 2 hr resulted in marked morphological effects including cytoplasmic and nuclear abnormalities and disruption of surface architecture. Tumors cells and normal urothelial cells were sensitive to the effects of mitomycin C. These effects were limited to epithelial cells.

DNA supercoiling, shortening, and induction of single-strand regions by cis-diamminedichloroplatinum(II).

cis-Diamminedichloroplatinum(II) was found to bind to covalently closed circular supercoiled, covalently closed circular nonsupercoiled, and single-strand broken relaxed PM2 DNA and induce different types of DNA conformational changes. Using Kleinschmidt's technique, it was found that binding of cis-diamminedichloroplatinum(II) decreased the DNA length to 75% of the original single-strand broken relaxed DNA without inducing superhelical conformational changes. cis-Diamminedichloroplatinum(II) also shortened the length of covalently closed circular nonsupercoiled DNA before a supercoiled conformation was generated. Single strand-specific nucleases were used to detect drug-induced DNA structural alterations. Local single-strand regions or regions of denaturation were detected by S1 nuclease from Aspergillus oryzae and by BAL-31 nuclease from Alteromonas espejiana BAL-31. The single-strand regions or local denaturation regions do not seem to be related to or caused by DNA superhelical conformational changes since they were detected at drug concentrations at which no significant DNA superhelical turns were found. DNA shortening, superhelical conformational changes, and local denaturation regions can be explained by the previously proposed "DNA intrastrand cross-linking" model (Stone et al., J. Mol. Biol., 104: 793-801, 1976).

Comparison of the sequences at specific sites on DNA cleaved by the antitumor antibiotics talisomycin and bleomycin.

We have investigated the site-specific cleavage of DNA by the antitumor antibiotics talisomycin and bleomycin by using 5'- or 3'-terminal 32P-labeled restriction fragments of pBR 322 DNA. Both drugs cleaved DNA preferentially at G-C and G-T sequences. However, the relative amounts of cleavage at particular cleavage sites differed between talisomycin and bleomycin at concentrations of the drugs which produced similar extents of total cleavage. In addition, talisomycin produced specific cleavages at G-A sequences which were relatively resistant to cleavage by bleomycin. Within a preferred sequence group (i.e., G-C sequences), some sites were cleaved to a greater extent relative to others by both talisomycin and bleomycin, suggesting that a greater degree of specificity than that provided by only two nucleotides is involved in the site-specific recognition and cleavage of DNA by these drugs.

Inhibition of PM-2 DNA degradation by a human serum protein.

A unique DNA-binding protein was detected that inhibited DNA degradation induced by bleomycin and was decreased in sera of cancer patients. The protein from normal human serum was purified to homogeneity by ammonium sulfate precipitation and DEAE-cellulose and DNA-cellulose column chromatography. Two-dimensional isoelectric focusing gel electrophoresis revealed a single protein spot with a molecular weight of 64,000 and a pI at pH 5.9. The NH2 terminus was lysine, and the ratio of acidic to basic residues was 1.2. DNA binding was demonstrated by column chromatography, agarose gel electrophoresis, fluorescence quenching, and circular dichroism. The inhibitory activity was abolished by treatment with Pronase but not by RNase or DNase I. FeCl2 caused a partial loss of inhibitory activity. The inhibition of DNA degradation was more effective for breakage induced by bleomycin than neocarzinostatin, macromomycin, or DNase I. Evidence from DNA-binding studies suggests the inhibition is due to binding of the protein to sites on DNA preferred by bleomycin. Thus, the protein could be useful for studies on the mechanism of action of bleomycin and other antitumor agents, the cytotoxic effects of which are due primarily to damage of cellular DNA. The protein was decreased significantly in sera of cancer patients, and its potential use as a diagnostic tool for neoplasias is being investigated further.

Bleomycin and talisomycin sequence-specific strand scission of DNA: a mechanism of double-strand cleavage.

Computer analyses of DNA sequencing data obtained using various restriction fragments of pBR 322 DNA indicate that a trinucleotide sequence (-Pyr-G-C-) is the most preferred site for cleavage by the antitumor antibiotic bleomycin A2. Talisomycin A, a structurally related bleomycin analog, cleaved at the sequences -G-T/A- most preferentially. However, the presence of a pyrimidine at the 5' side of guanine at the cleavage site did not increase the probability of that site being cleaved by talisomycin. Using denaturing and nondenaturing polyacrylamide gel analyses of the drug-DNA reaction products. The sites of both single- and double-strand breaks have been localized and differentiated. The results indicate that a major determinant for location of a site-specific double-strand break is the production of two closely spaced sequence-specific single-strand breaks by the drugs on opposite strands of the DNA. A four-base pair sequence is proposed for the optimal sequence for bleomycin-induced double-strand breaks.

Inhibition of bleomycin-induced DNA breakage by superoxide dismutase.

Inhibition of bleomycin (BLM)-induced DNA breakage by superoxide dismutase (SOD) has been reported and presumed to be due to its removal of the superoxide free radicals generated by BLM in the presence of iron(II). We have studied the possibility that the inhibitory effect might result from DNA-binding of SOD. The effect of copper-zinc SOD on BLM-induced DNA degradation was investigated using the PM-2 DNA fluorescence technique. PM-2 DNA was incubated with BLM in the presence or absence of native and heat-inactivated copper-zinc SOD as determined by the epinephrine autoxidation method. The concentrations of SOD required to inhibit 50% PM-2 DNA degradation for the native and the inactivated SOD were 100 and 120 microgram/ml, respectively. Analysis of the reaction mixture by agarose gel electrophoresis confirmed the absence of DNA degradation by BLM in the presence of either form of SOD. PM-2 DNA was shown to bind native or inactivated SOD by Sephadex G-100 column chromatography, fluorescence-quenching studies, and agarose gel electrophoresis. Thus, these results indicate that SOD is able to bind to PM-2 DNA and inhibit BLM-induced degradation independently of its free radical-scavenging activity. The inhibitor was more effective against BLM than other compounds which degrade PM-2 DNA. This suggests that SOD may bind to BLM-binding and/or BLM degradation sites in PM-2 DNA, and the observed inhibition is unrelated to its effects on free radicals.

Evaluation of the in vivo antitumor activity and in vitro cytotoxic properties of auranofin, a coordinated gold compound, in murine tumor models.

The coordinated gold compound, 2,3,4,6-tetra-O-acetyl-1-thio-beta-D-glucopyranosato-S-triethyl phosphine gold (auranofin; Ridaura), was evaluated for antitumor activity in a variety of mouse tumor models. Of the 15 tumor models evaluated, auranofin was found to be active only against i.p. P388 leukemia. A number of dose schedules was used to measure activity against P388 with optimal activity observed at 12 mg/kg given daily, i.p., on Days 1 to 5. Auranofin was active against i.p. P388 leukemia only when administered i.p.; the drug was completely inactive when administered i.v., s.c., or p.o. on Days 1 to 5. Evaluation of the effects of auranofin in vitro demonstrated that survival curves for B16 melanoma cells as measured by the clongenic and dye exclusion assays were exponential and monophasic; cell cycle distribution was not altered, and auranofin displayed no preferential cytotoxicity to logarithmic or plateau growth phase cell populations; auranofin inhibited DNA, RNA, and protein synthesis at cytotoxic concentrations but showed no selective effect; the cytotoxic activity and cellular association of gold from auranofin were dose, time, and temperature dependent; and binding of auranofin gold to serum proteins markedly decreased cellular uptake of gold and cytotoxicity of auranofin in vitro.