- Email: [email protected]
The mutagenic effects of diacridines and diquinolines in microbial systems
Mutation Research, 232 (1990) 337-343 Elsevier
337
MUT 04914
The mutagenic effects of diacridines and diquinolines in microbial systems Lynnette R. Ferguson, Pamela M. Turner and William A. Denny Cancer Research Laboratory, University of Auckland Medical School, Private Bag, Auckland (New Zealand) (Received 5 January 1990) (Revision received 15 May 1990) (Accepted 21 May 1990)
Keywords: Frameshift mutagenesis; Petite mutagenesis; Diacridines; Diquinolines
Summary Two series of difunctional DNA-intercalating agents (diacridines and diquinolines) were tested for mutagenic properties in Salmonella typhimurium strain TA1537, and for 'petite' mutagenesis activity in Saccharomyces cereoisiae, and also compared in terms of their structural, lipophilic and DNA-binding properties. Diacridines with only a short chain length were monointercalators, while those with an alkyl linker chain longer than C6 were bisintercalators. Although the bisintercalators especially bound very tightly to DNA, none of these compounds was as effective a frameshift mutagen in TA1537 as the parent chromophore 9-aminoacridine. However, the two (monointercalating) diacridines of shortest chain length were still able to cause frameshifts, and this ability returned (albeit weakly) in the bisintercalators of longest chain length. Although 9-aminoacridine showed no ability for 'petite' mutagenesis, the diacridines of longer chain length were very effective in causing this mitochondrial event. In the quinoline series, both the parent chromophore (4-aminoquinoline) and all the diquinolines were weak monointercalators. None of these compounds showed any ability for frameshift mutagenesis, although some were very weak mitochondrial mutagens. It is concluded that linking two acridines produces compounds whose mutagenic properties might have been predicted from our current knowledge of the parent molecules. However, despite a similar ability to intercalate DNA, the diquinolines show no resemblance to acridines in their mutagenic properties.
Simple acridines such as 9-aminoacridine possess three fused aromatic rings, which are thought to constitute a minimum requirement for efficient intercalative DNA binding (Albert, 1960; Atwell et al., 1988). In contrast, quinolines such as chloroquin are inefficient intercalators (Waring, 1970). Several studies have considered the effects
Correspondence: Dr. L.R. Ferguson, Cancer Research Laboratory, University of Auckland Medical School, Private Bag, Auckland (New Zealand).
on DNA binding of linking two acridine or two quinoline molecules using either flexible or inflexible linker chains. Such changes substantially alter the physical binding properties of the molecules. Many such diacridines of sufficiently long chain length have the ability to bisintercalate DNA inserting both of the aromatic chromophores between the basepairs at different sites on the DNA (Wakelin et al., 1978; Denny et al., 1985). However, the corresponding aikyl-linked diquinolines have been shown to bind only by monointercalation (insertion of only one of the two aromatic
0027-5107/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
338
DNA binding
1
2-10 (n = 2 Io 10)
I1
12-20 (n = 2 to 10) Fig. 1.
chromophores of the molecule between the basepairs) (MacFadyen et al., 1988). Such structural changes have an equally profound effect on biological properties. Thus while simple 9-alkylaminoacridines are poorly cytotoxic and do not show in vivo antitumour activity, some diacridines, particularly those with rigid linking chains, are potent cytotoxins and in vivo antitumour agents. A suggested reason is the much longer residence times shown by such compounds at individual DNA-binding sites (Denny et al., 1985). Acridines are well characterised as frameshift mutagens, and this effect is thought to relate to their ability to intercalate into DNA (Drake, 1989 and refs. therein), but no systematic mutagenicity studies on diacridines exist, to show how this mutagenicity is affected by bisintercalation. The mutagenic behaviour of diquinolines has also not been characterised. For this reason, we have collated information on the relevant physicochemical properties of two series of alkyMinked diacridines (2-10) and diquinolines (12-20), and the respective monomeric parent compounds 9-aminoacridine (1) and 4-aminoquinoline (11), and have compared these data with the mutagenic properties of the compounds in test systems using both the bacterium Salmonella typhimurium and the yeast Saecharomyces cerevisiae.
The relative binding affinities of the compounds to the DNA homopolymers poly(dA-dT) • poly(dA-dT) and poly(dG-dC) • poly(dG-dC) were measured by the ethidium displacement assay (Baguley and Falkenhaug, 1978; Denny et al., 1979). The concentration of ligand needed to reduce the fluorescence of DNA-bound ethidium by 50% (C50 value) is a measure of DNA binding affinity, being reciprocally correlated with DNA association constants for mono-intercalating agents (Baguley and Falkenhaug, 1978). Although such a strict relationship may not hold for compounds binding by other modes, C50 is still an adequate indication of relative DNA-binding affinity for the present compounds.
Lipophilicity The overall lipophilicities of the compounds were measured by thin-layer liquid-liquid chromatography as described previously (Cain et al., 1976). Relative chromatographic mobilities (Rm values) measured by this technique have been shown (Cain et al., 1976) to correlate directly with octanol-water partition coefficients which are widely used for estimation of the relative lipophilicity of drugs.
Saccharomyces cerevisiae The diploid strain D5 (Zimmermann, 1973) was kindly provided by Dr. B.S. Cox (Botany School, University of Oxford). For each experiment, a single-colony isolate was inoculated into liquid yeast complete medium (YC; Cox and Bevan, 1962) and grown to stationary phase for 24 h. DMSO was added to 10%, and 1-ml aliquots were frozen to - 7 0 °C and stored at this temperature before use. For all experiments, the 1-ml sample was thawed, added to 10 ml of fresh medium, and grown for exactly 2 h before use.
Salmonella typhimurium Materials and methods
Diacridines and diquinolines These were synthesized by published methods (Singh, 1975; Denny et al., 1985), and were formulated as the dihydrochloride salts.
Strain TA1537 was kindly supplied by Dr. B.N. Ames (Biochemistry Dept., University of California, Berkeley, CA, U.S.A.). Since we have found that the use of aliquots of frozen stock is necessary for reproducibility of experiments, the bacteria were initially grown to stationary phase
339 in L-broth (Luria and Burrous, 1957), and frozen (with 10% DMSO) in 1-ml aliquots at - 8 0 ° C.
Microtitre assay for petite mutagenesis This assay has been described in detail by Ferguson (1984). Briefly, a log-phase culture was diluted into fresh YC medium. A 96-well microtitre tray ( A / S Nunc, Denmark) was inoculated with (usually) 100-#1 aliquots of the diluted yeast culture, and drugs added at various dilutions to the wells, to a maximum of 1000 /~g/ml. Drugs were dissolved in 50% ethanol and dilutions made so that there was no more than 1% ethanol in each well. Trays were incubated for 20 h at 30 o C, an appropriate dilution made from each well into saline, and 100 #1 plated onto each of 3 YC plates. Cell numbers were calculated so that the dilutions at this point were at least 1 in 104, thereby effectively washing drugs from cells by dilution (Ferguson, 1984). Light was excluded from the assay at all points. Plates were incubated at 3 0 ° C for 2 days. Plates were scored for total surviving colonies as well as for the percentage of 'petite' colonies, following a tetrazolium overlay (Nagai, 1959). All experiments were performed at least twice, and the data compared for reproducibility. The presented data have been pooled and petite mutagenesis frequencies given as induced petites per 100 colonies. Spontaneous levels were around 0.5% for this strain. Bacterial mutagenicity assays Solutions of the drugs were dissolved in 50% ethanol for use, and freshly prepared for each experiment. Care was taken to exclude light from both the chemical and the assay plates. For each experiment, a 1-ml vial of bacteria was removed from the - 8 0 ° C freezer, inoculated into 20 ml of fresh bacterial complete medium, and grown for 4 h. Optical density was checked at that time and at intervals thereafter until a 1 / 1 0 dilution into fresh bacterial complete medium gave a reading of between 0.10 and 0.12 at 654 nm. (This was to ensure that all cultures were at the same stage of growth when used.) All of the compounds were tested quantitatively in strain TA1537. The S. typhimurium plate-incorporation assay was carried out as described by Maron and Ames (1983). Drug solu-
tions (in 50% ethanol) were added to 2 ml of soft agar containing 5 m M histidine-biotin, maintained at 4 2 ° C in a temperature block. 100 ~1 of the bacterial suspension was added, with or without 200/~1 of $9 mix, the tube was mixed and quickly poured over the surface of agar plates containing 20 ml of minimal medium (Vogel and Bonner, 1956). Plates were allowed to harden and then incubated at 37 ° C for 3 days before scoring colonies for reversion to histidine independence. A typical dose-response incorporated dose levels of 10, 15, 50, 100, 200, 300 and 400/~g/plate. Subsequent experiments raised the concentration as necessary, up to as high as 3 mg/plate. Where drug precipitated out before toxicity was reached (compounds 11-14) this is indicated by a figure for the minimum inhibitory concentration. Each experimental point was performed in triplicate on at least two separate occasions. Reversion characteristics of the strain were routinely tested in each experiment using the disc method described previously (Zeiger et al., 1981). Colony counts were determined on an Artek Model 880 automatic counter, calibrated with plates counted manually. High and low counts in a series of plates were checked periodically by manual counting. Each set of assays was performed in triplicate on two separate occasions. Data for the two experiments were combined and subjected to linear regression analysis. All data are presented as induced revertant colonies per ~g drug. Negative control values for TA1537 were in the range of 7-11 colonies (average) per plate.
Mammalian cytotoxicity studies P388 leukemia cells were maintained in exponential growth phase by subculturing in RPMI 1640 medium containing 10% fetal calf serum as previously described (Palmer et al., 1990). IC50 values were determined using log-phase cultures in 96-well microculture plates, and are calculated as the nominal drug concentration required to reduce the cell density to 50% of control values, using 8 control cultures on each microplate. Drug was present throughout the growth period (72 h) and final cell densities were determined using a minor modification of the M T T method of Mossman (Mossman, 1983).
Chain
0.08 0.40 0.29 0.21 0.02 0.15 0.21 0.28 0.40 0.45
- 0.39 -0.70 -0.57 --0.50 -0.39 -0.17 -0.02 0.11 0.22 0.39
-
Rm a
143 3.6 2.0 3.9 4.2 4.3 1.65 3.5 1.1 2.4
9.6 0.82 0.50 0.39 0.80 0.13 0.26 0.18 0.12 0.28
633 18.7 >100 k ND 108 >10 k 7.16 > 10 k 35 > 10 k
10.0 0.98 1.63 0.87 1.5 0.29 0.39 0.22 0.06 0.28
GC
D N A binding Cso b AT
M M M M M M M M M M
M M M M M/B B B B B B
mode ¢
> 25 9.5 18 15 6.5 0.83 0.68 0.66 0.12 0.35
3.0 3.7 0.11 0.06 0.17 0.16 0.28
> > > >
5.55 2.58 2.49 2.41 2.33 2.25 1.31 0.57 0.41 0.40
2.61 1.21 0.97 1.07 1.84 1.08 0.30 0.23 0.18 0.17
(#M/plate)
(#M)
1.9 4.9
MICe
S. typhimurium
iC50 d
P388
0 0 0 0 0 0 0 0 0 0
38.5 0.05 0.01 0 0 0 0 0.07 0.10 0.05
Mutagenicity f
TA1537
75.5 0.62 0.65 0.58 0.42 0.45 0.33 0.16 0.16 0.24
0.43 0.29 0.3 0.28 0.19 0.25 0.13 0.13 0.23 0.25
(#M)
Ds0
0 0 0 5 4 0 0 2 3 4
0 0 0 64 99 100 100 100 100 83
Pmax h
-
0.12 0.05 0.05 0.02 0.03 0.01 0.03
-
/50 i
' Petite' mutagenesis
S. cerevisiae
N N N W W N N W W W
N N N P S S S S S P
Class J
Rm value for relative lipophilicity of drug cations, measured by l i q u i d - l i q u i d thin-layer chromatography against an internal standard (see Cain, AtweU and Denny, 1976). C5° is the concentration (in micromolar) of ligand needed to displace 50% of D N A - b o u n d ethidium (see Baguley and Falkenhaug, 1978; D en n y et. al., 1979). Binding modes: M = monointercalation, B = bisintercalation. As determined by hydrodynamic assays (Wakelin et al., 1978; M c F a d y e n et al., 1988). iCs0 is the concentration of drug (in micromolar) to inhibit the growth of P388 leukemia cells by 50% measured in 96-well cultures (Finlay et al., 1986). MIC: M in im um inhibitory concentration is the lowest concentration (in micromoles per plate) at which the drug shows signs of causing toxicity, as ju d g ed by a thinning of the background colonies on the minimal media plates. Mutagenicity in S. typhimurium TA1537, expressed as revertant c o l o n i e s / # g drug. I)50: drug concentration (micromolar) causing a 50% cell kill compared with untreated control plates. Maximum percentage of 'petites' observed (as a percentage of viable colonies). Concentration of drug needed (in micromolar) for 50% of colonies to be 'petites'.
Parent 2 3 4 5 6 7 8 9 10
Parent 2 3 4 5 6 7 8 9 10
length
J Classification of 'petite'-inducing properties as follows: N = not active, W = weakly active ( < 10% ' pe t i t e s ' induced), P = proflavine-like, S = stronger 'p etite' mutagens than proflavine. For fuller definition of this classification see Ferguson and Baguley (1981b). k Some values could not be determined due to c ompound insolubility, or precipitation of the d r u g / D N A complex.
f g h i
a b c d
11 12 13 14 15 16 17 18 19 20
Diquinolines
1 2 3 4 5 6 7 8 9 10
Diacridines
Compd.
DNA BINDING, MAMMALIAN TOXICITY A N D M U T A G E N I C PROPERTIES OF M E T H Y L E N E D I Q U I N O L I N E S A N D DIACR1DINES
TABLE 1
341
Results The relative lipophilicities of the compounds, as determined by liquid-liquid chromatography, are recorded in Table 1. The values are for the drug cations or dications, which accounts for the apparent decrease in lipophilicity between the parent (monocation) and the early members of each bis series (dications). Within each series, lipophilicity increases with chain length, with an average increase of about 0.2 R m units per methylene. The most interesting point, seen by intercomparing the lipophilicities of diquinolines and diacridines with equal chain lengths, is that the two extra aromatic rings of the diacridines count for only about 0.2-0.3 R m units (not much more than one methylene unit). The DNA-binding characteristics of the compounds are also summarised in Table 1. Within each series all of the bis compounds bind more tightly to D N A than the parent chromophore, primarily because of the additional cationic charge. For the diacridines, there is a change in binding mode at C5, the compounds with linker chains longer than this being able to bisintercalate D N A (Wakelin et al., 1978), a property reflected in their considerably higher DNA-binding constants in the ethidium displacement assay (Table 1). There is almost no difference in binding preference between AT and G C binding sites. The parent 4aminoquinoline bound slightly less well to D N A than 9-aminoacridine, but all the diquinolines bound more tightly than either of the parent compounds. Previous work (McFadyen et al., 1988) has shown that these compounds all bind by monointercalation, even when the linker chain is long enough (C6 and greater) to span a bisintercalation site, and the similar binding constants (as estimated by C50 values) of all the diquinolines reflect this. The DNA-binding data presented here confirm a previous report (McFadyen et al., 1988) that, in contrast to the diacridines, the diquinolines have a high preference for binding to AT compared with G C sites. This may be due to the presence of a non-intercalated chromophore, which can interact more favourably with the narrow groove of AT-rich DNA, as do other non-intercalative binders (McFadyen et al., 1988). Within both series, toxicity towards mam-
marian cells increases with chain length (Table 1). In the diacridines, the sharp increase in mammarian toxicity between the C5 and C6 compounds (5 and 6) has been related to the change in DNA-binding mode which occurs at this point in the series (Denny et al., 1985). However, it seems likely that increasing lipophilicity also contributes to increased mammalian toxicity. In the diquinolines there is a similar increase in potency down the series, in the absence of a change in DNAbinding mode. The toxicities of both series in both bacteria and yeast seem to relate more to changes in lipophilicity, with a reasonably smooth increase in potency down each series. In initial experiments, the compounds were assayed, using well test methodology (Ferguson and MacPhee, 1984) in the four Ames test bacterial strains TA98, TA100, TA102 and TA1537. No positive results were seen in any strain except TA1537, and the quantitative experiments reported here then utilized only that strain. Comparative mutagenicity data for the diacridines are summarized in Table 1, and show a weak inverse relationship between the ability of the drugs to cause either frameshift mutagenesis in S. typhimurium or 'petite' mutagenesis in S. cerevisiae. Although none of the diacridines (2-10) causes such a high mutation frequency as the parent drug some compounds (particularly the more lipophilic ones) cause frameshift mutagenesis in TA1537. The relative effectiveness of the 6 frameshift mutagen drugs can be seen as induced revertants//~g on Table 1. Additionally the maximum numbers of mutant colonies differed very substantially, with drug 1 treatment leading to a maximum of 6864 induced colonies/plate, drug 2 to 45, drug 3 to 13, drug 8 to 35, drug 9 to 49 and drug 10 to 20. The ability to cause 'petite' mutagenesis is confined to the higher members of the series. However, the available data do not show strong relationships between mutagenic activity and either lipophiricity or D N A binding. In contrast, neither 4-aminoquinoline (11) nor any of the derived diquinolines (12-20) showed any evidence of frameshift mutagenesis properties in TA1537. While some showed reproducible 'petite' mutagenesis in the yeast model, the effects were very weak compared with the diacridines. For each of these drugs, numbers of petites on
342 treated plates were higher than on the control plates, suggesting the effect was due to induction rather than selection of 'petite' mutants.
Discussion This study shows differences between bacterial toxicity and mutagenic properties in two series of diacridines (2-10) and diquinolines (12-20) with methylene linker chains from C2 to C10 in length. Roos et al. (1985) studied the same series of diacridines for their ability to induce intracellular DNA damage, determined as single-strand D N A breakage. They showed that this ability (which may be mediated via topoisomerases) altered with chain length, with those diacridines of chain < C5 (DNA monointercalators) being less effective than 9-aminoacridine, and those of chain > C5 (DNA bisintercalators) being more effective. There was no similar discontinuity seen in either the toxicity or the mutagenic properties of the diacridines in the present study. Both of these properties instead correlate more closely with the strength of DNA binding rather than binding mode per se, in general agreement with previous results with acridines from this laboratory. Thus, Ferguson and Baguley (1981a) found a window for frameshift mutagenesis activity in Salmonella typhimurium TA1537, with those compounds of moderate DNA-binding activity being the most effective, weakly binding compounds inactive and the tightly binding drugs also being reduced in ability. Subsequent studies (Ferguson and Baguley, 1981b, 1985) showed that the tightest binding drugs became 'petite' mutagens in yeast (and weak frameshift mutagens at best), a pattern seen also with the present series of diacridines. The diquinolines show a very different pattern of mutagenic activity from the diacridines. They show no ability for frameshift or any other type of mutagenic activity in the bacteria tested. We note, however, that it is possible that the diquinolines show activity at sites other than the CCCC sequences detected in TA1537. Several of the compounds show weak but significant 'petite' mutagenesis in yeast, but this does not appear to be related to either lipophilicity or D N A binding (Table 1). In contrast to the diacridines, the diquinolines do show very selective binding to AT
regions in DNA. Many compounds with this DNA-binding pattern are 'petite' mutagens (Ferguson et al., 1989). However, it should be noted that the high AT-selectivity of the diquinolines is not due to increased binding to AT sites, but rather a complete loss of binding ability to GC regions (Table 1). The compounds thus bind much more weakly to AT regions than do many classical AT-selective 'petite' mutagens (Ferguson et al., 1989). The fact that the diquinolines are not frameshift mutagens in this test system despite binding to D N A by monointercalation even more tightly (in many cases) than 9-aminoacridine suggests that intercalative binding per se is not a 'critical' factor in this phenomenon. In fact both helix extension and unwinding assays show that 4-aminoquinoline and the diquinolines are much poorer intercalators than 9-aminoacridine, with unwinding angles of only 5-10 °, compared with 17 ° (Wright et al., 1980). It is still not resolved whether such low unwinding angles are due to a reduced degree of helix unwinding by each intercalated molecule (perhaps only partly-inserted between the basepairs), or whether only a proportion of the ligand fully intercalates with the rest binding on the outside of the helix, thus reducing the average unwinding per bound molecule.
Acknowledgements This work was supported by the Auckland Division of the Cancer Society of New Zealand and the Medical Research Council of New Zealand.
References Albert, A. (1960) The Acridines, Arnold, London, 1960. Atwell, G.J., C.D. Bos, B.C. Baguley and W.A. Denny (1988) Potential antitumor agents, 56. 'Minimal' DNA-intercalating ligands as antitumor drugs: phenylquinoline-8carboxamides, J. Med. Chem., 31, 1048-1052. Baguley, B.C., and E.-M. Falkenhaug (1978). The interaction of ethidium with synthetic double-strandedpolynucleotides at low ionic strength, Nucleic Acids Res., 5, 161-171. Cain, B.F., G.J. Atwell and W.A. Denny (1976) Potential antitumor agents, 17. 9-Anilino-10-methylacridiniumsalts, J. Med. Chem., 19, 772-777. Cox, B.S., and E.A. Bevan (1962) Aneuploidy in yeast, New Phytol., 61, 342-351.
343 Denny, W.A., G.J. Atwell, B.C. Baguley and B.F. Cain (1979) Potential antitumor agents, 29. Quantitative structure-activity relationships for the antileukemic bisquaternary ammonium heterocycles, J. Med. Chem., 22, 134-150. Denny, W.A., G.J. Atwell, B.C. Baguley and L.P.G. Wakelin (1985) Potential antitumor agents, 44. Synthesis and antitumor activity of new classes of diacridines: importance of linker chain rigidity for DNA-binding kinetics and biological activity, J. Med. Chem., 28, 1568-1574. Drake, J.W. (1989) Mechanism of mutagenesis, Environ. Mol. Mutagen., 14 (S16), 11-15. Ferguson, L.R. (1984) Apparent changes in structure-activity relationships for antimitochondrial effects of 9-anilinoacridines according to Saccharomvces cerevisiae strain and methodology, Mutation Res., 136, 223-234. Ferguson, L.R., and B.C. Baguley (1981a) Induction of 'petite' formation in Saccharomyces cerevisiae by experimental antitumor agents: structure-activity relationships for 9anilinoacridines, Mutation Res., 121, 263-269. Ferguson, L.R., and B.C. Baguley (1981b) The relationship between frameshift mutagenicity and DNA binding affinity in a series of acridine-substituted derivatives of the experimental anticancer drug 4'-(9-acridinylamino)methanesulfonanilide (AMSA), Mutation Res., 82, 31-39. Ferguson, L.R., and B.C. Baguley (1985) Inverse correlation between bacterial frameshift mutagenicity and yeast mitochondrial effects of antitumour acridines, Chem.-Biol. Interact., 56, 145-155. Ferguson, L.R., and D.G. MacPhee (1984) Frameshift mutagenesis by acridines in wild-type, uvrB and polA strains of Salmonella typhimurium with and without plasmid pKM101, Mutation Res., 141, 83-88. Ferguson, L.R., P.M. Turner, T.A. Gourdie, K.K. Valu and W.A. Denny (1989) 'Petite' mutagenesis and mitotic crossing-over in yeast by DNA-targeted alkylating agents, Mutation Res., 215, 213-222. Finlay, G.J., W.R. Wilson and B.C. Baguley (1986) Comparison of in vitro activity of cytotoxic drugs towards human carcinoma and leukemia cell lines, Eur. J. Cancer Clin. Oncol., 22, 655-662. Luria, S.E., and J.W. Burrous (1957) Hybridization between E. coli and Shigella, J. Bacteriol., 74, 461-467. Maron, D.M., and B.N. Ames (1983) Revised methods for the Salmonella mutagenicity test, Mutation Res., 113, 173-215.
McFadyen, W.D., W.A. Denny and L.P.G. Wakelin (1988) Alkyl-linked diquinolines are monofunctional AT-selective DNA-intercalating agents, FEBS Lett., 228, 235-240. Mossman, T. (1983) Rapid colourimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods, 65, 55. Nagai, S. (1959) Introduction of the respiration deficient mutation in yeast by various synthetic dyes, Science, 130, 11881189. Palmer B.D., W.R. Wilson, S.M. Pullen and W.A. Denny (1990) Hypoxia-selective antitumor agents. 3. Relationships between structure and cytotoxicity against cultured tumor cells for substituted N, N-bis(2-chloroethyl)anilines, J. Med. Chem., 33, 112. Roos I.A.G., L.P.G. Wakelin and S.J. Henry (1985) Intracellular damage produced by a series of diacridines, Biochem. J., 226, 175-182. Singh, A.K. (1975) Studies on antibacterial agents, III. Synthesis of N, N '-bis(4-quinolyl-, 4-quinaldyl-, 4-quinazolinyl- or 9-acridinyl)-polymethylenediamines and their sulfonamides, Chem. Pharm. Bull., 23, 1869-1873. Vogel, H.J., and D.M. Bonner (1956) Acetylornithinase of E. coli: partial purification and some properties, J. Biol. Chem., 281, 97-106. Wakelin, L.P.G., M. Romanos, T.-K. Chen, D. Glaubiger, E.S. Canellakis and M.J. Waring (1978) Structural limitations on the bifunctional intercalation of diacridines into DNA, Biochemistry, 17, 5057-5063. Waring, M.J. (1970) Variation of the supercoils in closed circular DNA by binding of antibiotics and drugs: evidence for molecular models involving intercalation, J. Mol. Biol., 54, 247-279. Wright, R.G. McR., L.P.G. Wakelin, A. Fieldes, R.M. Acheson and M.J. Waring (1980) Effects of ring substituents and linker chains on the bifunctional intercalation of diacridines into deoxyribonucleic acid, Biochemistry, 17, 5825-5836. Zeiger, E., D. Pagano and I.G.C. Robertson (1981) A rapid and simple scheme for confirmation of Salmonella tester strain phenotype, Environ. Mutagen., 3, 205-209. Zimmermann, F.K. (1973) A yeast strain for visual screening for the two reciprocal products of mitotic crossing-over, Mutation Res., 21,263-269.
337
MUT 04914
The mutagenic effects of diacridines and diquinolines in microbial systems Lynnette R. Ferguson, Pamela M. Turner and William A. Denny Cancer Research Laboratory, University of Auckland Medical School, Private Bag, Auckland (New Zealand) (Received 5 January 1990) (Revision received 15 May 1990) (Accepted 21 May 1990)
Keywords: Frameshift mutagenesis; Petite mutagenesis; Diacridines; Diquinolines
Summary Two series of difunctional DNA-intercalating agents (diacridines and diquinolines) were tested for mutagenic properties in Salmonella typhimurium strain TA1537, and for 'petite' mutagenesis activity in Saccharomyces cereoisiae, and also compared in terms of their structural, lipophilic and DNA-binding properties. Diacridines with only a short chain length were monointercalators, while those with an alkyl linker chain longer than C6 were bisintercalators. Although the bisintercalators especially bound very tightly to DNA, none of these compounds was as effective a frameshift mutagen in TA1537 as the parent chromophore 9-aminoacridine. However, the two (monointercalating) diacridines of shortest chain length were still able to cause frameshifts, and this ability returned (albeit weakly) in the bisintercalators of longest chain length. Although 9-aminoacridine showed no ability for 'petite' mutagenesis, the diacridines of longer chain length were very effective in causing this mitochondrial event. In the quinoline series, both the parent chromophore (4-aminoquinoline) and all the diquinolines were weak monointercalators. None of these compounds showed any ability for frameshift mutagenesis, although some were very weak mitochondrial mutagens. It is concluded that linking two acridines produces compounds whose mutagenic properties might have been predicted from our current knowledge of the parent molecules. However, despite a similar ability to intercalate DNA, the diquinolines show no resemblance to acridines in their mutagenic properties.
Simple acridines such as 9-aminoacridine possess three fused aromatic rings, which are thought to constitute a minimum requirement for efficient intercalative DNA binding (Albert, 1960; Atwell et al., 1988). In contrast, quinolines such as chloroquin are inefficient intercalators (Waring, 1970). Several studies have considered the effects
Correspondence: Dr. L.R. Ferguson, Cancer Research Laboratory, University of Auckland Medical School, Private Bag, Auckland (New Zealand).
on DNA binding of linking two acridine or two quinoline molecules using either flexible or inflexible linker chains. Such changes substantially alter the physical binding properties of the molecules. Many such diacridines of sufficiently long chain length have the ability to bisintercalate DNA inserting both of the aromatic chromophores between the basepairs at different sites on the DNA (Wakelin et al., 1978; Denny et al., 1985). However, the corresponding aikyl-linked diquinolines have been shown to bind only by monointercalation (insertion of only one of the two aromatic
0027-5107/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
338
DNA binding
1
2-10 (n = 2 Io 10)
I1
12-20 (n = 2 to 10) Fig. 1.
chromophores of the molecule between the basepairs) (MacFadyen et al., 1988). Such structural changes have an equally profound effect on biological properties. Thus while simple 9-alkylaminoacridines are poorly cytotoxic and do not show in vivo antitumour activity, some diacridines, particularly those with rigid linking chains, are potent cytotoxins and in vivo antitumour agents. A suggested reason is the much longer residence times shown by such compounds at individual DNA-binding sites (Denny et al., 1985). Acridines are well characterised as frameshift mutagens, and this effect is thought to relate to their ability to intercalate into DNA (Drake, 1989 and refs. therein), but no systematic mutagenicity studies on diacridines exist, to show how this mutagenicity is affected by bisintercalation. The mutagenic behaviour of diquinolines has also not been characterised. For this reason, we have collated information on the relevant physicochemical properties of two series of alkyMinked diacridines (2-10) and diquinolines (12-20), and the respective monomeric parent compounds 9-aminoacridine (1) and 4-aminoquinoline (11), and have compared these data with the mutagenic properties of the compounds in test systems using both the bacterium Salmonella typhimurium and the yeast Saecharomyces cerevisiae.
The relative binding affinities of the compounds to the DNA homopolymers poly(dA-dT) • poly(dA-dT) and poly(dG-dC) • poly(dG-dC) were measured by the ethidium displacement assay (Baguley and Falkenhaug, 1978; Denny et al., 1979). The concentration of ligand needed to reduce the fluorescence of DNA-bound ethidium by 50% (C50 value) is a measure of DNA binding affinity, being reciprocally correlated with DNA association constants for mono-intercalating agents (Baguley and Falkenhaug, 1978). Although such a strict relationship may not hold for compounds binding by other modes, C50 is still an adequate indication of relative DNA-binding affinity for the present compounds.
Lipophilicity The overall lipophilicities of the compounds were measured by thin-layer liquid-liquid chromatography as described previously (Cain et al., 1976). Relative chromatographic mobilities (Rm values) measured by this technique have been shown (Cain et al., 1976) to correlate directly with octanol-water partition coefficients which are widely used for estimation of the relative lipophilicity of drugs.
Saccharomyces cerevisiae The diploid strain D5 (Zimmermann, 1973) was kindly provided by Dr. B.S. Cox (Botany School, University of Oxford). For each experiment, a single-colony isolate was inoculated into liquid yeast complete medium (YC; Cox and Bevan, 1962) and grown to stationary phase for 24 h. DMSO was added to 10%, and 1-ml aliquots were frozen to - 7 0 °C and stored at this temperature before use. For all experiments, the 1-ml sample was thawed, added to 10 ml of fresh medium, and grown for exactly 2 h before use.
Salmonella typhimurium Materials and methods
Diacridines and diquinolines These were synthesized by published methods (Singh, 1975; Denny et al., 1985), and were formulated as the dihydrochloride salts.
Strain TA1537 was kindly supplied by Dr. B.N. Ames (Biochemistry Dept., University of California, Berkeley, CA, U.S.A.). Since we have found that the use of aliquots of frozen stock is necessary for reproducibility of experiments, the bacteria were initially grown to stationary phase
339 in L-broth (Luria and Burrous, 1957), and frozen (with 10% DMSO) in 1-ml aliquots at - 8 0 ° C.
Microtitre assay for petite mutagenesis This assay has been described in detail by Ferguson (1984). Briefly, a log-phase culture was diluted into fresh YC medium. A 96-well microtitre tray ( A / S Nunc, Denmark) was inoculated with (usually) 100-#1 aliquots of the diluted yeast culture, and drugs added at various dilutions to the wells, to a maximum of 1000 /~g/ml. Drugs were dissolved in 50% ethanol and dilutions made so that there was no more than 1% ethanol in each well. Trays were incubated for 20 h at 30 o C, an appropriate dilution made from each well into saline, and 100 #1 plated onto each of 3 YC plates. Cell numbers were calculated so that the dilutions at this point were at least 1 in 104, thereby effectively washing drugs from cells by dilution (Ferguson, 1984). Light was excluded from the assay at all points. Plates were incubated at 3 0 ° C for 2 days. Plates were scored for total surviving colonies as well as for the percentage of 'petite' colonies, following a tetrazolium overlay (Nagai, 1959). All experiments were performed at least twice, and the data compared for reproducibility. The presented data have been pooled and petite mutagenesis frequencies given as induced petites per 100 colonies. Spontaneous levels were around 0.5% for this strain. Bacterial mutagenicity assays Solutions of the drugs were dissolved in 50% ethanol for use, and freshly prepared for each experiment. Care was taken to exclude light from both the chemical and the assay plates. For each experiment, a 1-ml vial of bacteria was removed from the - 8 0 ° C freezer, inoculated into 20 ml of fresh bacterial complete medium, and grown for 4 h. Optical density was checked at that time and at intervals thereafter until a 1 / 1 0 dilution into fresh bacterial complete medium gave a reading of between 0.10 and 0.12 at 654 nm. (This was to ensure that all cultures were at the same stage of growth when used.) All of the compounds were tested quantitatively in strain TA1537. The S. typhimurium plate-incorporation assay was carried out as described by Maron and Ames (1983). Drug solu-
tions (in 50% ethanol) were added to 2 ml of soft agar containing 5 m M histidine-biotin, maintained at 4 2 ° C in a temperature block. 100 ~1 of the bacterial suspension was added, with or without 200/~1 of $9 mix, the tube was mixed and quickly poured over the surface of agar plates containing 20 ml of minimal medium (Vogel and Bonner, 1956). Plates were allowed to harden and then incubated at 37 ° C for 3 days before scoring colonies for reversion to histidine independence. A typical dose-response incorporated dose levels of 10, 15, 50, 100, 200, 300 and 400/~g/plate. Subsequent experiments raised the concentration as necessary, up to as high as 3 mg/plate. Where drug precipitated out before toxicity was reached (compounds 11-14) this is indicated by a figure for the minimum inhibitory concentration. Each experimental point was performed in triplicate on at least two separate occasions. Reversion characteristics of the strain were routinely tested in each experiment using the disc method described previously (Zeiger et al., 1981). Colony counts were determined on an Artek Model 880 automatic counter, calibrated with plates counted manually. High and low counts in a series of plates were checked periodically by manual counting. Each set of assays was performed in triplicate on two separate occasions. Data for the two experiments were combined and subjected to linear regression analysis. All data are presented as induced revertant colonies per ~g drug. Negative control values for TA1537 were in the range of 7-11 colonies (average) per plate.
Mammalian cytotoxicity studies P388 leukemia cells were maintained in exponential growth phase by subculturing in RPMI 1640 medium containing 10% fetal calf serum as previously described (Palmer et al., 1990). IC50 values were determined using log-phase cultures in 96-well microculture plates, and are calculated as the nominal drug concentration required to reduce the cell density to 50% of control values, using 8 control cultures on each microplate. Drug was present throughout the growth period (72 h) and final cell densities were determined using a minor modification of the M T T method of Mossman (Mossman, 1983).
Chain
0.08 0.40 0.29 0.21 0.02 0.15 0.21 0.28 0.40 0.45
- 0.39 -0.70 -0.57 --0.50 -0.39 -0.17 -0.02 0.11 0.22 0.39
-
Rm a
143 3.6 2.0 3.9 4.2 4.3 1.65 3.5 1.1 2.4
9.6 0.82 0.50 0.39 0.80 0.13 0.26 0.18 0.12 0.28
633 18.7 >100 k ND 108 >10 k 7.16 > 10 k 35 > 10 k
10.0 0.98 1.63 0.87 1.5 0.29 0.39 0.22 0.06 0.28
GC
D N A binding Cso b AT
M M M M M M M M M M
M M M M M/B B B B B B
mode ¢
> 25 9.5 18 15 6.5 0.83 0.68 0.66 0.12 0.35
3.0 3.7 0.11 0.06 0.17 0.16 0.28
> > > >
5.55 2.58 2.49 2.41 2.33 2.25 1.31 0.57 0.41 0.40
2.61 1.21 0.97 1.07 1.84 1.08 0.30 0.23 0.18 0.17
(#M/plate)
(#M)
1.9 4.9
MICe
S. typhimurium
iC50 d
P388
0 0 0 0 0 0 0 0 0 0
38.5 0.05 0.01 0 0 0 0 0.07 0.10 0.05
Mutagenicity f
TA1537
75.5 0.62 0.65 0.58 0.42 0.45 0.33 0.16 0.16 0.24
0.43 0.29 0.3 0.28 0.19 0.25 0.13 0.13 0.23 0.25
(#M)
Ds0
0 0 0 5 4 0 0 2 3 4
0 0 0 64 99 100 100 100 100 83
Pmax h
-
0.12 0.05 0.05 0.02 0.03 0.01 0.03
-
/50 i
' Petite' mutagenesis
S. cerevisiae
N N N W W N N W W W
N N N P S S S S S P
Class J
Rm value for relative lipophilicity of drug cations, measured by l i q u i d - l i q u i d thin-layer chromatography against an internal standard (see Cain, AtweU and Denny, 1976). C5° is the concentration (in micromolar) of ligand needed to displace 50% of D N A - b o u n d ethidium (see Baguley and Falkenhaug, 1978; D en n y et. al., 1979). Binding modes: M = monointercalation, B = bisintercalation. As determined by hydrodynamic assays (Wakelin et al., 1978; M c F a d y e n et al., 1988). iCs0 is the concentration of drug (in micromolar) to inhibit the growth of P388 leukemia cells by 50% measured in 96-well cultures (Finlay et al., 1986). MIC: M in im um inhibitory concentration is the lowest concentration (in micromoles per plate) at which the drug shows signs of causing toxicity, as ju d g ed by a thinning of the background colonies on the minimal media plates. Mutagenicity in S. typhimurium TA1537, expressed as revertant c o l o n i e s / # g drug. I)50: drug concentration (micromolar) causing a 50% cell kill compared with untreated control plates. Maximum percentage of 'petites' observed (as a percentage of viable colonies). Concentration of drug needed (in micromolar) for 50% of colonies to be 'petites'.
Parent 2 3 4 5 6 7 8 9 10
Parent 2 3 4 5 6 7 8 9 10
length
J Classification of 'petite'-inducing properties as follows: N = not active, W = weakly active ( < 10% ' pe t i t e s ' induced), P = proflavine-like, S = stronger 'p etite' mutagens than proflavine. For fuller definition of this classification see Ferguson and Baguley (1981b). k Some values could not be determined due to c ompound insolubility, or precipitation of the d r u g / D N A complex.
f g h i
a b c d
11 12 13 14 15 16 17 18 19 20
Diquinolines
1 2 3 4 5 6 7 8 9 10
Diacridines
Compd.
DNA BINDING, MAMMALIAN TOXICITY A N D M U T A G E N I C PROPERTIES OF M E T H Y L E N E D I Q U I N O L I N E S A N D DIACR1DINES
TABLE 1
341
Results The relative lipophilicities of the compounds, as determined by liquid-liquid chromatography, are recorded in Table 1. The values are for the drug cations or dications, which accounts for the apparent decrease in lipophilicity between the parent (monocation) and the early members of each bis series (dications). Within each series, lipophilicity increases with chain length, with an average increase of about 0.2 R m units per methylene. The most interesting point, seen by intercomparing the lipophilicities of diquinolines and diacridines with equal chain lengths, is that the two extra aromatic rings of the diacridines count for only about 0.2-0.3 R m units (not much more than one methylene unit). The DNA-binding characteristics of the compounds are also summarised in Table 1. Within each series all of the bis compounds bind more tightly to D N A than the parent chromophore, primarily because of the additional cationic charge. For the diacridines, there is a change in binding mode at C5, the compounds with linker chains longer than this being able to bisintercalate D N A (Wakelin et al., 1978), a property reflected in their considerably higher DNA-binding constants in the ethidium displacement assay (Table 1). There is almost no difference in binding preference between AT and G C binding sites. The parent 4aminoquinoline bound slightly less well to D N A than 9-aminoacridine, but all the diquinolines bound more tightly than either of the parent compounds. Previous work (McFadyen et al., 1988) has shown that these compounds all bind by monointercalation, even when the linker chain is long enough (C6 and greater) to span a bisintercalation site, and the similar binding constants (as estimated by C50 values) of all the diquinolines reflect this. The DNA-binding data presented here confirm a previous report (McFadyen et al., 1988) that, in contrast to the diacridines, the diquinolines have a high preference for binding to AT compared with G C sites. This may be due to the presence of a non-intercalated chromophore, which can interact more favourably with the narrow groove of AT-rich DNA, as do other non-intercalative binders (McFadyen et al., 1988). Within both series, toxicity towards mam-
marian cells increases with chain length (Table 1). In the diacridines, the sharp increase in mammarian toxicity between the C5 and C6 compounds (5 and 6) has been related to the change in DNA-binding mode which occurs at this point in the series (Denny et al., 1985). However, it seems likely that increasing lipophilicity also contributes to increased mammalian toxicity. In the diquinolines there is a similar increase in potency down the series, in the absence of a change in DNAbinding mode. The toxicities of both series in both bacteria and yeast seem to relate more to changes in lipophilicity, with a reasonably smooth increase in potency down each series. In initial experiments, the compounds were assayed, using well test methodology (Ferguson and MacPhee, 1984) in the four Ames test bacterial strains TA98, TA100, TA102 and TA1537. No positive results were seen in any strain except TA1537, and the quantitative experiments reported here then utilized only that strain. Comparative mutagenicity data for the diacridines are summarized in Table 1, and show a weak inverse relationship between the ability of the drugs to cause either frameshift mutagenesis in S. typhimurium or 'petite' mutagenesis in S. cerevisiae. Although none of the diacridines (2-10) causes such a high mutation frequency as the parent drug some compounds (particularly the more lipophilic ones) cause frameshift mutagenesis in TA1537. The relative effectiveness of the 6 frameshift mutagen drugs can be seen as induced revertants//~g on Table 1. Additionally the maximum numbers of mutant colonies differed very substantially, with drug 1 treatment leading to a maximum of 6864 induced colonies/plate, drug 2 to 45, drug 3 to 13, drug 8 to 35, drug 9 to 49 and drug 10 to 20. The ability to cause 'petite' mutagenesis is confined to the higher members of the series. However, the available data do not show strong relationships between mutagenic activity and either lipophiricity or D N A binding. In contrast, neither 4-aminoquinoline (11) nor any of the derived diquinolines (12-20) showed any evidence of frameshift mutagenesis properties in TA1537. While some showed reproducible 'petite' mutagenesis in the yeast model, the effects were very weak compared with the diacridines. For each of these drugs, numbers of petites on
342 treated plates were higher than on the control plates, suggesting the effect was due to induction rather than selection of 'petite' mutants.
Discussion This study shows differences between bacterial toxicity and mutagenic properties in two series of diacridines (2-10) and diquinolines (12-20) with methylene linker chains from C2 to C10 in length. Roos et al. (1985) studied the same series of diacridines for their ability to induce intracellular DNA damage, determined as single-strand D N A breakage. They showed that this ability (which may be mediated via topoisomerases) altered with chain length, with those diacridines of chain < C5 (DNA monointercalators) being less effective than 9-aminoacridine, and those of chain > C5 (DNA bisintercalators) being more effective. There was no similar discontinuity seen in either the toxicity or the mutagenic properties of the diacridines in the present study. Both of these properties instead correlate more closely with the strength of DNA binding rather than binding mode per se, in general agreement with previous results with acridines from this laboratory. Thus, Ferguson and Baguley (1981a) found a window for frameshift mutagenesis activity in Salmonella typhimurium TA1537, with those compounds of moderate DNA-binding activity being the most effective, weakly binding compounds inactive and the tightly binding drugs also being reduced in ability. Subsequent studies (Ferguson and Baguley, 1981b, 1985) showed that the tightest binding drugs became 'petite' mutagens in yeast (and weak frameshift mutagens at best), a pattern seen also with the present series of diacridines. The diquinolines show a very different pattern of mutagenic activity from the diacridines. They show no ability for frameshift or any other type of mutagenic activity in the bacteria tested. We note, however, that it is possible that the diquinolines show activity at sites other than the CCCC sequences detected in TA1537. Several of the compounds show weak but significant 'petite' mutagenesis in yeast, but this does not appear to be related to either lipophilicity or D N A binding (Table 1). In contrast to the diacridines, the diquinolines do show very selective binding to AT
regions in DNA. Many compounds with this DNA-binding pattern are 'petite' mutagens (Ferguson et al., 1989). However, it should be noted that the high AT-selectivity of the diquinolines is not due to increased binding to AT sites, but rather a complete loss of binding ability to GC regions (Table 1). The compounds thus bind much more weakly to AT regions than do many classical AT-selective 'petite' mutagens (Ferguson et al., 1989). The fact that the diquinolines are not frameshift mutagens in this test system despite binding to D N A by monointercalation even more tightly (in many cases) than 9-aminoacridine suggests that intercalative binding per se is not a 'critical' factor in this phenomenon. In fact both helix extension and unwinding assays show that 4-aminoquinoline and the diquinolines are much poorer intercalators than 9-aminoacridine, with unwinding angles of only 5-10 °, compared with 17 ° (Wright et al., 1980). It is still not resolved whether such low unwinding angles are due to a reduced degree of helix unwinding by each intercalated molecule (perhaps only partly-inserted between the basepairs), or whether only a proportion of the ligand fully intercalates with the rest binding on the outside of the helix, thus reducing the average unwinding per bound molecule.
Acknowledgements This work was supported by the Auckland Division of the Cancer Society of New Zealand and the Medical Research Council of New Zealand.
References Albert, A. (1960) The Acridines, Arnold, London, 1960. Atwell, G.J., C.D. Bos, B.C. Baguley and W.A. Denny (1988) Potential antitumor agents, 56. 'Minimal' DNA-intercalating ligands as antitumor drugs: phenylquinoline-8carboxamides, J. Med. Chem., 31, 1048-1052. Baguley, B.C., and E.-M. Falkenhaug (1978). The interaction of ethidium with synthetic double-strandedpolynucleotides at low ionic strength, Nucleic Acids Res., 5, 161-171. Cain, B.F., G.J. Atwell and W.A. Denny (1976) Potential antitumor agents, 17. 9-Anilino-10-methylacridiniumsalts, J. Med. Chem., 19, 772-777. Cox, B.S., and E.A. Bevan (1962) Aneuploidy in yeast, New Phytol., 61, 342-351.
343 Denny, W.A., G.J. Atwell, B.C. Baguley and B.F. Cain (1979) Potential antitumor agents, 29. Quantitative structure-activity relationships for the antileukemic bisquaternary ammonium heterocycles, J. Med. Chem., 22, 134-150. Denny, W.A., G.J. Atwell, B.C. Baguley and L.P.G. Wakelin (1985) Potential antitumor agents, 44. Synthesis and antitumor activity of new classes of diacridines: importance of linker chain rigidity for DNA-binding kinetics and biological activity, J. Med. Chem., 28, 1568-1574. Drake, J.W. (1989) Mechanism of mutagenesis, Environ. Mol. Mutagen., 14 (S16), 11-15. Ferguson, L.R. (1984) Apparent changes in structure-activity relationships for antimitochondrial effects of 9-anilinoacridines according to Saccharomvces cerevisiae strain and methodology, Mutation Res., 136, 223-234. Ferguson, L.R., and B.C. Baguley (1981a) Induction of 'petite' formation in Saccharomyces cerevisiae by experimental antitumor agents: structure-activity relationships for 9anilinoacridines, Mutation Res., 121, 263-269. Ferguson, L.R., and B.C. Baguley (1981b) The relationship between frameshift mutagenicity and DNA binding affinity in a series of acridine-substituted derivatives of the experimental anticancer drug 4'-(9-acridinylamino)methanesulfonanilide (AMSA), Mutation Res., 82, 31-39. Ferguson, L.R., and B.C. Baguley (1985) Inverse correlation between bacterial frameshift mutagenicity and yeast mitochondrial effects of antitumour acridines, Chem.-Biol. Interact., 56, 145-155. Ferguson, L.R., and D.G. MacPhee (1984) Frameshift mutagenesis by acridines in wild-type, uvrB and polA strains of Salmonella typhimurium with and without plasmid pKM101, Mutation Res., 141, 83-88. Ferguson, L.R., P.M. Turner, T.A. Gourdie, K.K. Valu and W.A. Denny (1989) 'Petite' mutagenesis and mitotic crossing-over in yeast by DNA-targeted alkylating agents, Mutation Res., 215, 213-222. Finlay, G.J., W.R. Wilson and B.C. Baguley (1986) Comparison of in vitro activity of cytotoxic drugs towards human carcinoma and leukemia cell lines, Eur. J. Cancer Clin. Oncol., 22, 655-662. Luria, S.E., and J.W. Burrous (1957) Hybridization between E. coli and Shigella, J. Bacteriol., 74, 461-467. Maron, D.M., and B.N. Ames (1983) Revised methods for the Salmonella mutagenicity test, Mutation Res., 113, 173-215.
McFadyen, W.D., W.A. Denny and L.P.G. Wakelin (1988) Alkyl-linked diquinolines are monofunctional AT-selective DNA-intercalating agents, FEBS Lett., 228, 235-240. Mossman, T. (1983) Rapid colourimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods, 65, 55. Nagai, S. (1959) Introduction of the respiration deficient mutation in yeast by various synthetic dyes, Science, 130, 11881189. Palmer B.D., W.R. Wilson, S.M. Pullen and W.A. Denny (1990) Hypoxia-selective antitumor agents. 3. Relationships between structure and cytotoxicity against cultured tumor cells for substituted N, N-bis(2-chloroethyl)anilines, J. Med. Chem., 33, 112. Roos I.A.G., L.P.G. Wakelin and S.J. Henry (1985) Intracellular damage produced by a series of diacridines, Biochem. J., 226, 175-182. Singh, A.K. (1975) Studies on antibacterial agents, III. Synthesis of N, N '-bis(4-quinolyl-, 4-quinaldyl-, 4-quinazolinyl- or 9-acridinyl)-polymethylenediamines and their sulfonamides, Chem. Pharm. Bull., 23, 1869-1873. Vogel, H.J., and D.M. Bonner (1956) Acetylornithinase of E. coli: partial purification and some properties, J. Biol. Chem., 281, 97-106. Wakelin, L.P.G., M. Romanos, T.-K. Chen, D. Glaubiger, E.S. Canellakis and M.J. Waring (1978) Structural limitations on the bifunctional intercalation of diacridines into DNA, Biochemistry, 17, 5057-5063. Waring, M.J. (1970) Variation of the supercoils in closed circular DNA by binding of antibiotics and drugs: evidence for molecular models involving intercalation, J. Mol. Biol., 54, 247-279. Wright, R.G. McR., L.P.G. Wakelin, A. Fieldes, R.M. Acheson and M.J. Waring (1980) Effects of ring substituents and linker chains on the bifunctional intercalation of diacridines into deoxyribonucleic acid, Biochemistry, 17, 5825-5836. Zeiger, E., D. Pagano and I.G.C. Robertson (1981) A rapid and simple scheme for confirmation of Salmonella tester strain phenotype, Environ. Mutagen., 3, 205-209. Zimmermann, F.K. (1973) A yeast strain for visual screening for the two reciprocal products of mitotic crossing-over, Mutation Res., 21,263-269.