Contrasts in the actions of protein antibiotics on deoxyribonucleic acid structure and function.

The protein antibiotics neocarzinostain (NCS), macromomycin (MCR), and auromomycin (AUR), which is closely related to MCR, have been compared for their in vitro and in vivo actions on deoxyribonucleic acid (DNA). NCS, markedly stimulated by 2-mercaptoethanol, is much more active in inducing strand scissions in superhelical pMB9 and linear duplex lambda DNA than AUR, which is slightly inhibited by 2-mercaptoethanol. Purified MCR, even at very high levels, does not give any significant amount of cutting with either DNA substrate. 2-Propanol stimulates the activity of NCS but inhibits that of AUR. On the other hand, the antioxidant alpha-tocopherol strongly inhibits DNA breakage by both drugs. The intercalating drugs ethidium bromide, daunorubicin, proflavin, and actinomycin D at low concentrations inhibit DNA scission by AUR. The levels of intercalators required to inhibit NCS activity to comparable levels are about 10 times higher than those for AUR. Although MCR has virtually no in vitro DNA cutting activity, it is, like AUR and NCS, cytotoxic, as measured by the inhibition of DNA synthesis and induction of DNA strand breakage in HeLa cells.     

Antitumor proteins of Streptomyces macromomyceticus: purification and characterization of auromomycin, macromomycin A, and macromomycin D

Macromomycin A and the two related proteins auromomycin and macromomycin D were isolated from the culture filtrates of Streptomyces macromomyceticus by chromatography on columns of DEAE-cellulose, Amberlite XAD-7, and decylagarose. Antibodies prepared against macromomycin A showed antigenic identity by Ouchterlony double diffusion between the three purified proteins. This similarity was further demonstrated by their behavior on disc gel electrophoresis, the amino acid compositions, and comparative peptide mapping of the aminoethylated derivatives. They differed, however, in other chemical and biological properties. Auromomycin and macromomycin A, pI 5.4, have antibiotic activity, which is absent in macromomycin D, pI 5.2. This antibiotic activity was associated with chromophore groups that were extractable by methanol. High-pressure liquid chromatography of the methanol extracts gave difference profiles for each of the purified proteins. The differences in the three proteins extended to their ultraviolet-visible spectra, fluorescence and circular dichroism, and the changes of these properties with heating. The heat denaturation, with auromomycin and macromomycin melting at 70.5 degrees C and macromomycin D at 57.0 degrees C, was reversible. Changes were noted in the spectra both during and following heating at 80 degrees C; the antibacterial activity was lost in auromomycin and only partially reduced in macromomycin A. The properties of the three proteins support the general similarities in their polypeptide structures, modifications in the properties of which are endowed by the differences in the associated nonprotein chromophores.     

Strand scission of DNA in KB cells by macromomycin.

Macromomycin, an antitumor protein, produces strand scission of DNA in KB cells, which might appear to correlate with its concommitant inhibition of thymidine and uridine incorporation into nucleic acids. However, an acetyl derivative of macromomycin does not cause the same extent of strand breakage, yet it shows inhibitory and cytotoxic properties similar to the native protein. This is also seen in a derivative of macromomycin resulting from its reaction with the Bolton-Hunter reagent. Cesalin, another antitumor protein, does not possess DNA cleavage activity.     

Inhibition of mammalian deoxyribonucleic acid synthesis by neocarzinostatin: selective effect on replicon initiation in CHO cells and resistant synthesis in ataxia telangiectasia fibroblasts

Treatment of CHO cells with low doses of the protein antibiotic neocarzinostatin severely inhibited DNA replicon initiation but had no effect on chain elongation. The selectivity of the effect on initiation, which was greater than that seen with other chemical agents and comparable to that seen with X-rays, explains the biphasic dose response seen for DNA synthesis inhibition by this drug. Parallel experiments employing the nucleoid sedimentation technique indicated that half-maximal relaxation of domains of DNA supercoiling and half-maximal inhibition of replicon initiation required the same dose of neocarzinostatin, approximately 0.03 micrograms/mL. These results, similar results obtained with the protein antibiotic auromomycin, and previous results obtained with X-rays suggest a quantitative correlation between inhibition of replicon initiation and induction of sufficient strand breakage to relax domains of supercoiling in DNA of mammalian cells. Results in human ataxia telangiectasia fibroblasts indicated that neocarzinostatin, like X-rays, is much less effective in inhibiting DNA synthesis in these cells than in normal human fibroblasts. This finding is consistent with the hypothesis that the genetic defect in ataxia telangiectasia involves a failure to recognize the presence of strand breaks in cellular DNA.     

Structural study of crystalline auromomycin: Preliminary X-ray diffraction study of auromomycin and macromomycin

Auromomycin and macromomycin from the organism Streptomyces macromomyceticus have been crystallized. The X-ray diffraction pattern of crystals of each molecule is consistent with space group P2₁2₁2 with cell parameters $\text{a = 46.45}\text{A}\text{?}\text{, b = 54.34}\text{A}\text{?}\text{and}\text{c = 42.03}\text{A}\text{?}$ for auromomycin, and $\text{a = 46.45}\text{A}\text{?}\text{, b = 54.52}\text{A}\text{?}\text{and}\text{c = 41.54}\text{A}\text{?}$ for macromomycin. Diffraction analysis of auromomycin is in progress.     

Neocarzinostatin and auromomycin preferentially cleave simian virus 40 DNA and chromosomes at a number of discrete locations

Neocarzinostatin and auromomycin were shown to cleave simian virus 40 (SV40) DNA with preference for distinct regions of the viral genome. The positions cut by neocarzinostatin and auromomycin were similar, while micrococcal nuclease cleaved at positions other than those recognized by neocarzinostatin and auromomycin. Breaks were distributed throughout the viral genome and were not associated with any single type of genetic element. The limited number of locations in SV40 DNA that were preferentially cut by neocarzinostatin and auromomycin suggests that drug attack is directed by DNA structures other than the known trinucleotide sequence specificity of the drugs. Neocarzinostatin and auromomycin cut purified, cell-free, nuclear and intracellular chromosomal SV40 DNA at similar regions. The data indicate that there are regions in DNA that are hypersensitive to the drugs; the hypersensitivity may be determined by the microstructure of the DNA. The conformational change associated with the packing of the DNA into nucleosomes did not affect the microstructure of the sensitive region, nor did the shielding effect of nuclear proteins affect the drug's access to it. In addition, intracellular drug metabolism or other cellular factors did not alter the ability of drugs to interact at these sensitive regions.     

Cellular DNA damage by the antitumor protein macromomycin and its relationship to cell growth inhibition

Macromomycin, an antitumor protein, after purification inhibited HeLa S₃ cell growth at low nanomolar concentrations. DNA synthesis was inhibited at drug levels that left RNA and protein synthesis unaffected. Incubation of macromomycin with HeLa S₃ cells resulted in a rapid fragmentation of cellular DNA that was detectable at low nanomolar drug levels. Heat denaturation of macromomycin showed that both its ability to inhibit cell growth and to fragment cellular DNA were lost at similar rates.     

DNA strand breaks: the DNA template alterations that trigger p53-dependent DNA damage response pathways.

The tumor suppressor protein p53 serves as a critical regulator of a G1 cell cycle checkpoint and of apoptosis following exposure of cells to DNA-damaging agents. The mechanism by which DNA-damaging agents elevate p53 protein levels to trigger G1/S arrest or cell death remains to be elucidated. In fact, whether damage to the DNA template itself participates in transducing the signal leading to p53 induction has not yet been demonstrated. We exposed human cell lines containing wild-type p53 alleles to several different DNA-damaging agents and found that agents which rapidly induce DNA strand breaks, such as ionizing radiation, bleomycin, and DNA topoisomerase-targeted drugs, rapidly triggered p53 protein elevations. In addition, we determined that camptothecin-stimulated trapping of topoisomerase I-DNA complexes was not sufficient to elevate p53 protein levels; rather, replication-associated DNA strand breaks were required. Furthermore, treatment of cells with the antimetabolite N(phosphonoacetyl)-L-aspartate (PALA) did not cause rapid p53 protein increases but resulted in delayed increases in p53 protein levels temporally correlated with the appearance of DNA strand breaks. Finally, we concluded that DNA strand breaks were sufficient for initiating p53-dependent signal transduction after finding that introduction of nucleases into cells by electroporation stimulated rapid p53 protein elevations. While DNA strand breaks appeared to be capable of triggering p53 induction, DNA lesions other than strand breaks did not. Exposure of normal cells and excision repair-deficient xeroderma pigmentosum cells to low doses of UV light, under conditions in which thymine dimers appear but DNA replication-associated strand breaks were prevented, resulted in p53 induction attributable to DNA strand breaks associated with excision repair. Our data indicate that DNA strand breaks are sufficient and probably necessary for p53 induction in cells with wild-type p53 alleles exposed to DNA-damaging agents.     

Modulation of glycoprotein hormone alpha-subunit levels, alkaline phosphatase activity, and DNA replication by antimetabolites in HeLa cultures

Synthesis of the glycoprotein hormone common alpha-subunit and alkaline phosphatase (placental isozyme) has been examined in HeLa S3 cells. A variety of compounds that inhibit DNA synthesis lead to the increased production of both proteins. Experiments presented in this communication were undertaken to determine whether protein induction and DNA synthesis inhibition are coordinated. In general, nucleoside analogs and compounds that alter deoxynucleotide metabolism were good inducers of these ectopic products, whereas agents that altered DNA by intercalation, crosslinking, and covalent modification were poor inducers. The former class of effectors includes 5-bromo-2'-deoxyuridine, 5-fluoro-2'-deoxyuridine, 2'-deoxythymidine, 1-beta-D-arabinofuranosylcytosine, methotrexate, hydroxyurea, N-phosphonoacetyl-L-aspartic acid, and sodium butyrate; and the latter class of compounds includes ethidium bromide, acridine, bleomycin, mitomycin C, cesalin, macromomycin, and cis-diamminedichloroplatinum(II). A direct correlation between protein induction and DNA synthesis inhibition is unlikely based on the following observations: (i) for some effectors, the concentrations required to induce alpha-subunit and PAP were significantly different from those necessary to inhibit DNA synthesis; (ii) several agents inhibit DNA replication but do not enhance hormone or enzyme production; (iii) the kinetics of ectopic protein induction were similar for a number of inducers whereas the kinetics of DNA synthesis inhibition elicited by the same compounds were quite different. It is difficult from the data obtained, however, to rule out the possibility that inhibition of DNA synthesis may be required but is not sufficient for protein induction.     

(Na (+) + K (+)-stimulated ATPase of human kidney, normal and adenocarcinoma. Phosphorylation and inhibition by antitumor proteins

(Na⁺+K⁺)-ATPase was purified from human kidney of normal and tumor tissue with specific activities of 100.0 and 16.6 μmol Pi/mg/h, respectively. The antitumor proteins, macromomycin, largomycin, and NSC 327459 (50 μg/ml each) caused 70 to 90% inhibition of (Na⁺+K⁺)-ATPase from tumor tissue, whereas auromomycin had no effect. (Na⁺+K⁺)-ATPase from both sources could be phosphorylated by rabbit muscle protein kinase; there was 3 to 6-fold stimulation of phosphorylation by cyclic AMP. Phosphorylation resulted in 70 to 80% decrease in (Na⁺+K⁺)-ATPase activity, and caused the normal enzyme to become sensitive to inhibition by macromomycin.     

Qualitative and quantitative aspects of intercalator-induced DNA strand breaks.

The intercalating agents, adriamycin and ellipticine, were previously found to produce DNA strand breaks associated with DNA-protein crosslinks in mouse leukemia L1210 cells. The current work explores the nature of the agents that produce this effect and the quantitative relationship between the breaks and crosslinks. The protein-associated DNA breaks were produced by a wide variety of intercalators in addition to the above-mentioned compounds: actinomycin D, daunoycin, ethidium and lucanthone (miracil D). Treatment with several drugs that bind to DNA without intercalation, or that inhibit DNA synthesis without binding to DNA, did not cause DNA breaks. The strand break and crosslink frequencies were quantitated by means of alkaline elution methods. The strand break and crosslink frequencies were found to be within a factor of 2 of each other over a range of concentrations of adriamycin and ellipticine. It is proposed that intercalation-induced distortion of the DNA helix leads to strand scission by a nuclease which becomes bound to one terminus of the break so as to form a DNA-protein crosslink.     

Defective repair of DNA single- and double-strand breaks in the bleomycin- and X-ray-sensitive Chinese hamster ovary cell mutant, BLM-2

Using filter elution techniques, we have measured the level of induced single- and double-strand DNA breaks and the rate of strand break rejoining following exposure of two Chinese hamster ovary (CHO) cell mutants to bleomycin or neocarzinostatin. These mutants, designated BLM-1 and BLM-2, were isolated on the basis of hypersensitivity to bleomycin and are cross-sensitive to a range of other free radical-generating agents, but exhibit enhanced resistance to neocarzinostatin. A 1-h exposure to equimolar doses of bleomycin induces a similar level of DNA strand breaks in parental CHO-K1 and mutant BLM-1 cells, but a consistently higher level is accumulated by BLM-2 cells. The rate of rejoining of bleomycin-induced single- and double-strand DNA breaks is slower in BLM-2 cells than in CHO-K1 cells. BLM-1 cells show normal strand break repair kinetics. The level of single- and double-strand breaks induced by neocarzinostatin is lower in both BLM-1 and BLM-2 cells than in CHO-K1 cells. The rate of repair of neocarzinostatin-induced strand breaks is normal in BLM-1 cells but retarded somewhat in BLM-2 cells. Thus, there is a correlation between the level of drug-induced DNA damage in BLM-2 cells and the bleomycin-sensitive, neocarzinostatin resistant phenotype of this mutant. Strand breaks induced by both of these agents are also repaired with reduced efficiency by BLM-2 cells. The neocarzinostatin resistance of BLM-1 cells appears to be a consequence of a reduced accumulation of DNA damage. However, the bleomycin-sensitive phenotype of BLM-1 cells does not apparently correlate with any alteration in DNA strand break induction or repair, as analysed by filter elution techniques, suggesting an alternative mechanism of cell killing.     

The contribution of alkylation to the activity of quinone antitumor agents

Studies have shown that the quinone group can produce tumor cell kill by a mechanism involving active oxygen species. This cytotoxic activity can be correlated with the induction of DNA double strand breaks and is enhanced by the ability of the quinone compound to bind to DNA by alkylation. The cytotoxic activity and the production of DNA damage by model quinone antitumor agents were compared in L5178Y cells, sensitive and resistant to alkylating agents, to assess the contribution of alkylation to the activity of these agents. The resistant L5178Y/HN2 cells were found to be two fold and six fold more resistant to the alkylating quinones, benzoquinone mustard and benzoquinone dimustard, respectively, than parent L5178Y cells. In contrast, the L5178Y/HN2 cells showed no resistance to the nonalkylating quinones, hydrolyzed benzoquinone mustard and bis(dimethylamino)benzoquinone. The alkylating quinones produced approximately two fold less cross-linking in L5178Y/HN2 cells compared with L5178Y sensitive cells. DNA double strand break formation by hydrolyzed benzoquinone mustard and bis(dimethylamino)benzoquinone was not significantly different in sensitive and resistant cells. However, the induction of double strand breaks by the alkylating quinones benzoquinone mustard and benzoquinone dimustard was reduced by 5-fold and 15-fold, respectively, in L5178Y/HN2 cells. These results show that the alkylating activity of the alkylating quinones cannot directly explain all of the enhanced cytotoxic activity of these agents. Furthermore, they provide strong evidence that the enhanced formation of DNA double strand breaks by alkylating quinone agents is directly related to the ability of these agents to bind to DNA. This increased formation of strand breaks may account for the enhanced cytotoxic activity of the alkylating quinones.     

Degradation of HeLa S3 cell chromatin by auromomycin and its chromophore

The antitumor protein agent auromomycin was found to degrade chromatin structure primarily by inducing strand scissions in linker regions. The reaction was stimulated by dithiothreitol. The chromophore form of the drug caused similar effects on chromatin, but it appeared to function at a more rapid rate. There was no evidence that auromomycin could cause breakage in core regions of chromatin.     

Metal-induced DNA damage and repair in human diploid fibroblasts and Chinese hamster ovary cells

Cloning efficiency and DNA strand breaks induction were compared in human diploid fibroblasts (HSBP) and chinese hamster ovary (CHO) cells treated with various metal salts. Cadmium (Cd2+), nickel (Ni2+) and chromate (Cr2O7) reduced the cloning efficiency of HSBP cells more than that of CHO cells whereas the reverse was true after treatment with mercury (Hg2+), manganese (Mn2+) and cobalt (Co2+). The effects on cloning efficiency did not consistently correlate with DNA strand breaking activity as all metals except Cr(VI) were more effective at producing DNA strand breaks in CHO cells than in human cells. The differential responses of the two cell types was shown to be only partially due to differences in cellular uptake of metals. DNA breaks induced in human cells by Hg2+ and Cr2O7 were shown most likely to be alkaline labile sites rather than true strand breaks since no damage was detected in a nick translation assay which measures the amount of free 3'-OH terminals. Damage induced by Mn2+ and Co2+, however, appeared to be comprised at least in part by true DNA strand breaks. DNA damage was also induced in HSBP cells following treatment with selenium but only in the presence of reduced glutathione. These studies indicate that DNA damage is not as major a consequence following some metal treatments in human cells as it appears to be in rodent cells. This suggests that rodent models for risk estimation of metal-induced tumorigenesis may not always be appropriate for extrapolation to humans.     

In vivo non-enzymatic repair of DNA oxidative damage by polyphenols

The non-enzymatic repair of DNA oxidative damage can occur in a purely chemical system, but data show that it might also occur in cells. Human hepatoma cells (SMMC-7721) and human hepatocyte cells (LO2) were treated with 200 μM H₂O₂ for 30 min to induce oxidative DNA damage quantified by amount of 8-OHdG and degree of DNA strand breaks, without inducing enzymatic repair. The dynamics of enzymatic repair activity quantified by unscheduled DNA synthesis, within 30 min after removal of H₂O₂ enzymatic repair mechanism has not been initiated. However, pre-incubation with low micromolar level polyphenols, quercetin or rutin can significantly attenuate DNA damage in both cell lines, indicating that the polyphenols did not work through an enzymatic mechanism. Unscheduled DNA synthesis after removal of H₂O₂ was also markedly decreased by quercetin and rutin. Combined with our previous studies of fast reaction chemistry, the inhibitory effect of polyphenols have to be assigned to non-enzymatic repair mechanism rather than to enzymatic repair mechanism or antioxidant mechanism.     

Preliminary crystallographic study of macromomycin: The apoprotein of the antitumor antibiotic auromomycin

Macromomycin (M_r 12,000) is the apoprotein of the antitumor drug auromomycin, which inhibits DNA synthesis by causing single-strand breaks in DNA. Two orthorhombic crystal forms of macromomycin have been observed. The platelike crystals of one form belong to space group P2₁2₁2₁, with cell dimensions $\text{a = 48·92}\text{A}\text{?}\text{, b = 54·71}\text{A}\text{?}\text{, c = 103·31}\text{A}\text{?}\text{, Z = 8}$. The crystals of the second form are needle-shaped, and belong to space group P2₁2₁2, with cell dimensions $\text{a = 46·1}\text{A}\text{?}\text{, b = 54·4}\text{A}\text{?}\text{, c = 41·2}\text{A}\text{?}\text{, Z = 4}$. At this point in time, the platelike crystals appear the most suitable for continued crystallographic studies.     

Aldehyde oxidase generates deoxyribonucleic acid single strand nicks in vitro

SUMMARY

We purified aldehyde oxidase (AO) from rabbit livers and found that AO produced deoxyribonucleic acid (DNA) single strand nicks in vitro. Acetaldehyde, benzaldehyde, and certain purine bases were effective substrates for AO catalyzed DNA strand nicking. DNA strand nicking did not occur with the reducing substrates nicotinamide-adenine dinucleotide or dithionite that produce superoxide anion (O_2′^?). Inclusion of electron transport inhibitors, potassium cyanide, ferricyanide or menadione, prevented AO catalyzed nicking. AO induced DNA strand nicking was dependent upon hydrogen peroxide (H₂O₂) formation and most likely generation of hydroxyl radical (HO'). The present observations may be pertinent to the recently proposed involvement of AO in inherited juvenile familial amyotrophic lateral sclerosis (JFALS) and other oxygen radical mediated diseases.     

Different effects of genistein and resveratrol on oxidative DNA damage in vitro

Previous studies have demonstrated that phenolic compounds, including genistein (4′,5,7-trihydroxyisoflavone) and resveratrol (3,4′,5-trihydroxystilbene), are able to protect against carcinogenesis in animal models. This study was undertaken to examine the ability of genistein and resveratrol to inhibit reactive oxygen species (ROS)-mediated strand breaks in φX-174 plasmid DNA. H₂O₂/Cu(II) and hydroquinone/Cu(II) were used to cause oxidative DNA strand breaks in the plasmid DNA. We demonstrated that the presence of genistein at micromolar concentrations resulted in a marked inhibition of DNA strand breaks induced by either H₂O₂/Cu(II) or hydroquinone/Cu(II). Genistein neither affected the Cu(II)/Cu(I) redox cycle nor reacted with H₂O₂ suggest that genistein may directly scavenge the ROS that participate in the induction of DNA strand breaks. In contrast to the inhibitory effects of genistein, the presence of resveratrol at similar concentrations led to increased DNA strand breaks induced by H₂O₂/Cu(II). Further studies showed that in the presence of Cu(II), resveratrol, but not genistein was able to cause DNA strand breaks. Moreover, both Cu(II)/Cu(I) redox cycle and H₂O₂ were shown to be critically involved in resveratrol/copper-mediated DNA strand breaks. The above results indicate that despite their similar in vivo anticarcinogenic effects, genistein and resveratrol appear to exert different effects on oxidative DNA damage in vitro.     

Involvement of Hydroxyl Radical Formation and Dna Strand Breakage in the Cytotoxicity of Anthraquinone Antitumour Agents

Four 9,10-anthraquinones (AQ) mono- or bis-substituted with the -NH(CH₂)₂ NH(CH₂)₂OH group were studied. 1-AQ, 1,5-AQ and 1,8-AQ but not 1,4-AQ (100°M) generated pBR322 plasmid DNA single strand breaks in the presence of purified NADPH dependent cytochrome P450 reductase. 1-AQ, 1,5-AQ and 1,8-AQ (at 100 °M) stimulated hydroxyl radical formation in MCF-7 S9 cell fraction (as measured by dimethyl pyrolline N-oxide spin trapping) and MCF-7 DNA strand breaks as measured by alkaline filter elution. In contrast 1,4-AQ did not stimulate hydroxyl radical formation and produced considerably less strand breaks in MCF-7 cells compared to the other AQ's. It would appear that the position of the -NH(CH₂)₂ NH(CH₂)₂OH groups on the chromophore is an important determinant in the metabolic activation of cytotoxic anthraquinones. This may contribute to the cytotoxicity (ID₅₀ values) of 1-AQ (0.06 °M), 1-8-AQ (0.5 °M) and 1,5-AQ (12.3 °M) but not the 1,4-AQ (1.2 °M).