Interaction of mithramycin with metal ions and DNA

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

April 28, 1989

Interaction

Pages

of mithramycin Benjamin

M.G.

with

metal

Cons and Keith

ions

517-524

and DNA

R. Fox

Dept. Physiology & Pharmacology University of Southampton Bassett Crescent East Southampton SO9 3TU U.K. Received February 2, 1989 The interaction of mithramycin with metal ions has been studied by absorbance and fluorescence spectroscopy. Magnesium shifts the drug absorbance spectrum to longer wavelengths and displays a weak binding constant (&=lmM); no interaction with calcium was detected. The drug requires magnesium for binding to DNA and this is characterised by small additional hypochromic and bathochromic changes. Mithramycin does not bind to DNA in the presence of calcium. With 1OmM magnesium the drug binds to DNA with an association constant of 9.2x104 M-l. The inability of calcium to substitute for magnesium has been confirmed by 'footprinting' I and hydroxyl radicals. 0 1989Academrc experiments using both DNase PreS.5,

Inc.

Mithramycin is an antitumour antibiotic of the aureolic acid group which is active against a variety of experimental and It is useful for treating disseminated human tumours [II. testicular carcinomas [2], Paget's disease and hypercalcaemia The mechanism of its action on calcium metabolism has not [3,41. been clearly defined although it appears to inhibit bone resorption t51, either by a direct interaction with calcium or by an antivitamin D action [4]. Its antitumour activity is due to inhibition of DNA-directed RNA synthesis, an effect which is achieved by binding to doublestranded DNA [6,7]. This interaction with DNA requires divalent metal ions [8] and the presence of guanine bases L9.101. DNA footprinting studies have shown that mithramycin recognizes the sequence GpG [ll-131. The interaction of these antibiotics with magnesium has previously been studied by several workers [13-161. The absorbance spectrum of mithramycin shifts to longer wavelengths on adding millimolar concentrations of magnesium. Despite this observation most studies on the binding of mithramycin to DNA have used buffers in which little attention has been paid to the 0006-291X/89 $1.50 517

Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYBICAL RESEARCH COMMUNICATIONS

magnesium concentration. Prasad and Nayak [14] presented spectral shifts on adding DNA to mithramycin, yet used only 15OuM Mga+, so that it was not clear whether the changes resulted from drug binding to DNA or a shift in the amount of mithramycin complexed to magnesium. In previous footprinting studies (12,131 magnesium was either added with the digesting enzyme (DNase I) or not at all (DNase II and micrococcal nuclease). The purpose of this work was therefore to examine the binding of mithramycin to divalent metal ions and to assess the interaction of the drug-metal ion complex with DNA under conditions in which their concentration is carefully controlled. Materials and Methods Chemicals and enzymes Mithramycin was a gift from Pfizer Inc U.S.A. and was dissolved in 1OmM Tris-HCl pH 8.0 containing 1OmM NaCl. It was stored as a 1mM stock solution in the dark at 4°C. DNase I was purchased from Sigma and stored as a 7200 units/ml solution in the dark at -20°C. All restriction enzymes were purchased from Northumbria Biologicals Ltd. Calf thymus DNA was purchased from Sigma and dissolved in 1OmM Tris-HCl pHa.Ocontaining 1OmM NaCl. DNA concentrations are expressed with respect to nucleotides and determined from an E(p),,, of 6600. DNA fragments. The 135 base pair DNA fragment (Figure 1) was obtained from plasmid pXbs1 [la] by cutting with Hind111 and Sau3Al and was labelled at the 3'-end of the Hind111 end using reverse transcriptase and a-[32PldATP. Spectroscopic studies. Absorbance spectra were recorded on a Beckman Model 25 spectrophotometer using lcm path length optical cuvettes. Fluorescence measurements were determined using a Perkin-Elmer 3000 fluorescence spectrometer. In all spectroscopic studies drug complexes were left to equilibrate for at least 15 min since a portion of the reaction has been shown to occur slowly [lo]. DNA binding data were analysed according to equation (10) of McGhee and von Hippel [20] and are presented as a Scatchard plot of r/c against r where r is the bound drug per DNA nucleotide and c is the free drug concentration. DNA footprinting DNase I footprinting was performed as omitting previously described [12] except that the experiments magnesium contained 5mM CaCl2 instead of 2mM MgCl-L, 2mM Hydroxyl radical footprinting [191 MnC12 in the digestion buffer. was performed by mixing 10~1 of the drug-DNA complex with 10~1 of a freshly prepared mixture containing 0.05mM ferrous ammonium sulphate, O.lmM EDTA, 2.5mM ascorbic acid and 0.07% (v/v) hydrogen peroxide. The reaction was stopped after 3 mins by adding 10~1 1OOmM thiourea The products of footprinting and the DNA precipitated with ethanol. reactions were separated on 8% (w/v) polyacrylamide gels containing 3'-~CCGTCCTG~CCCGTCGAGACGmGACATmGGCCTGmCCGAAAGGGGACCGAATGTGCG 10 20 30 40 50 ~CCCTTCCCGGAAAGGACTCCTCCACTCGCCG~GGACCTGAGCCCCTACCGCGACC~CA-5' 70 80 90 100 110

Figure strand

60 120

130

Sequence of the 135 base pair DNA fragment. Only the bearing the 3'-radioactive label (circled) is shown. 1

518

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

8M urea. Gels were run at 1500V for 2hr, then fixed in 10% (v/v) acetic acid, transferred to Whatman 3MM paper, dried under vacuum BO°C and subjected to autoradiography at -70°C with an intensifying screen. Bands were assigned by comparison with dimethylsulphatepiperidine markers specific for guanine. Results The visible absorbance spectrum of mithramycin at pH 8.0 has a peak at 402nm which shifts to 422nm on addition of excess magnesium chloride (Figure 2). Both the magnesium-free and bound forms of the drug obeyed Beer's law at 420nm over the range of concentrations used. The interaction with magnesium was examined in more detail by carefully titrating magnesium into a solution of mithramycin and monitoring the absorbance changes. All the spectra generated passed through a common isosbestic point at 411nm, indicating that only one spectrally distinct mithramycin-magnesium complex is formed. In each complex the fraction of drug bound to the clivalent metal ion (Cb) was estimated from the absorbance at 420nm according to Cb=(EIxCT-OD)/(EI-Eb) where Ef and Eb are the extinction coefficients for the magnesium-free and -bound forms of the drug respectively, CT is the total drug concentration and OD is the measured optical density of the drug-magnesium mixture. A binding curve generated from such an analysis is presented in Figure 3A and yields an affinity constant of 0.63 +/- 0.15 x 103 M-l. Such a low value suggests that at 1.5mM MgCl 2 only 50% of the drug will be complexed with magnesium. Even at Mg2+ concentrations as

X,lm

Fiuure 2 Absorbance spectra of mithramycin (50pM) in the presence and absence of magnesium and DNA. Curve a free mithamycin; curve b in the presence of 1OmMMgCl=; curve c in the presence of 1OmMMgCla and 5OOp DNA. 519

at

Vol. 160, No. 2, 1969

BIOCHEMICAL

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

0.06 '/c $tM-' I O.OL

0.02

I 0.5

I l.0

I 1.5

Mg

I 2.0

-

I 2.5

I

1

0.05

I

0.10

I

0.15

0.20

,

0

0.25

ImM)

Fiqure 3 A1 Binding curve for the interaction of mithramycin with magnesium. The drug concentration was 75p. The curve drawn to the points is a least squares binding curve characterized by K-0.63x10= M-l. B) Scatchard plot for the binding of mithramycin to calf thymus DNA. The curve drawn to the points is a non-linear least squares fit calculated according to equation (10) of McGhee & von Hippel I201 and has the parameters K=9.2x104 M-l, n=3.0 bases.

high

as 1OmM only

implications spectrum

for

about the

produced

sufficient

to

form

90% will

proper

be complexed.

analysis

by DNA, since a single

divalent

drug

This

of any changes metal

species

has important in the

drug

ion concentrations

must first

be present.

In contrast to the results with magnesium, calcium at concentrations as high as 50mM failed to provoke any changes the drug absorbance spectrum. Barium also failed changes although zinc, manganese and nickel all longer

wavelengths Figure

addition

of

similar

2 displays a large

excess

to the of

those spectral

seen with changes

DNA to a solution

in

to induce any caused shifts to

magnesium. that

occur

on

of mithramycin

in

1OroM MgC12. There is only a small bathochromic and hypochromic shift in the spectra, which on careful analysis do not pass through a single isosbestic point. These changes were too small to permit an accurate determination of a binding constant and are in contrast to the large working at much lower represent the binding

spectral changes claimed by other authors magnesium concentrations 114,151. These must of more magnesium to the drug according to

the law of mass action as metal-bound drug Addition of DNA to a mixture of mithramycin 520

is absorbed by the DNA. and calcium (1OmM)

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

caused calcium

no spectral perturbations. consistent with the view that can not substitute for magnesium. In order to further examine the interaction of the drug with DNA we have also measured its fluorescence properties. On addition of magnesium to a solution of mithramycin its fluorescence (excitation at 412nm emission at 547nml is totally Addition of DNA to this mixture causes an increase in quenched. fluorescence up to about 80% of the original value. These data were analysed according to the modified Scatchard equation of McGhee and van Hippel [20] and yielded the plot shown in Figure 2b characterized by the parameters K=9.2x104 M-l n=3.0. These results demonstrate that magnesium is required for the interaction of mithramycin with DNA and that it can not be replaced by calcium. In order to confirm these effects we have performed footprinting experiments with DNA fragments of known sequence. Figure 4A displays the results of experiments using hydroxyl radicals as the footprinting probe in the presence and absence of various concentrations of mithramycin. This cleavage agent generates a fairly even ladder of bands in the control lanes and with 1OmM magnesium mithramycin produces clear protection around positions 38, 47, 54, 58, 72, 78, 94 and 100. At concentrations of 10uM and below no protection is afforded. Each of these sites of protection corresponds to the dinucleotide step A more detailed analysis of the precise sequence GpG (CpCl. selectivity will be presented elsewhere. When calcium is used in place of nlag3leSiUln (lanes i,jj no drug-induced protection is produced confirming that binding to DNA can not occur. Figure 48 presents the results of similar experiments using DNase I as the footprinting probe. In this case all the drug treated lanes contained 50uM mithramycin with varying amounts of magnesium. With 1 and 1OmM magnesium large blockages are clearly visible around positions 50, 75 and 100 corresponding to the protections seen with hydroxyl radicals, but are larger due to the size of DNase I and the close proximity of several of the binding sites. At lower magnesium concentrations blockages are still evident but are less well defined. When calcium is used as the divalent cation no protections are seen. Discussion The results presented mithramycin has no significant Calcium does not affect either

in

this paper demonstrate that interaction with calcium ions. the absorbance or fluorescence 521

Vol. 160, No. 2, 1989

BIOCHEMICAL

Mg

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Ca

A

IOO-

80.

60,

LO

Figure 4 Footprinting of mithramycin on the 135 base pair DNA fragment. A) Hydroxyl radical footprinting; lanes a-h in the presence of 1OmMMgClz, lanes i,j in the presence of 1OmM CaClz. The drug concentrations were a,i OF; b, 0.5uM; c, ~JJM; d, 5p; e, lop; f, 20)1M; g, 3OuM; h,j, 50pM. B) DNAase I footprinting in the absence (con) and presence of 50).&l mithramycin. Each pair of lanes represents digestion by the enzyme for 1 and 5 mins. The figures above each lane refer to the MgCla concentration. Lanes labelled Ca were in the presence of 1OmMCaCla. Tracks labelled "G" correspond to Maxam Gilbert dimethylsulphate-piperidine markers specific for guanine. spectra

of

These results with

the

drug are

and does not

in contrast

magnesium and several

may be the mechanism of the we can confidently rule out

to other

induce those divalent

any

interaction

seen for metal

hypocalcaemic effect a direct interaction

the ions.

with

DNA.

interaction Whatever

of mithramycin with calcium.

The apparent binding constant of magnesium to mithramycin is low (less than lo3 ?Pz), so that high concentrations are required before the true spectral changes caused by addition of DNA can be determined. The similarity of the absorbance spectrum of free and DNA-bound mithramycin may suggest that the drug conformational changes induced by magnesium are retained on binding to DNA. It has previously been suggested that the metal ion binds across the hydrophilic face of the drug molecule shielding the negative charges of mithramycin (pK=5.0). While such

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYSICAL RESEARCH COMMUNICATIONS

a model is plausible it does not readily account for the ability of mithramycin to discriminate between calcium and magnesium, and would obscure those groups on the chromophore which are likely to be responsible for the pronounced GC-selectivity. We are of the opinion that the metal ion is involved in orienting the sugar residues around the hydrophobic edge of the chromophore. consistent with recent NMR experiments [211. Such an interaction and may be able to discriminate between the is likely to be weak, two metal ions. The inability of calcium to substitute for magnesium in sequence selective binding is confirmed by the footprinting experiments. The observation that with magnesium concentrations below 1mM mithramycin produces footprints which are less well defined is consistent with the weak interaction with the divalent metal ion. This underlines the importance of using sufficient magnesium since the formation of the drug-metal ion complex is limited by its low affinity constant. The lack of an isosbestic point in the small absorbance changes produced on binding to DNA is unusual, though consistent with a recent study with chromomycin [22]. We can rule out any complications arising from the effect of magnesium on the free and bound drug so that the most likely explanation is that more than one type of DNA-bound drug species is formed. If the absorbance of mithramycin varies according to the DNA sequence to which it is bound then no clear isosbestic point will be detected. The dissociation constant from calf thymus DNA (1OpM) agrees well with the footprinting data for which no protections are evident at drug concentrations below 2O)l.M. Acknowledcrments Medical Research studentship.

This work was supported by a grant from Council. BMGC is supported by an SERC

the

References Kennedy, B.J.. Yarbo, J.W., Kickertz, V. Fu Sandberg[ll. Wollbeim, M. (1968) Cancer Res. 28, 91-97. Kennedy, B.J. (19701 Am. J. Med. 49, 494-503. 121. Perlia, C.P., Gubisch, N.J.. Wolter, J., Edelberg. D., 131. Dedderick, M.M., & Taylor, S.G. (1970) Cancer 25, 389-394. Slayton, R.E., Snider, B.I., Elias, E., Horton, J. & [41. Perlia, C.P. (1971) Clin. Pharmacol. Ther. 12, 833-837. D.T. Loken, M.K. & Kennedy, B.J. (1979) J. Clin. I51. Kiang. Endocrin.

161. t71. [al.

& Metab.

48,

341-344.

Kersten, W. Kersten, H. & Szybalski, (1966) Biochemistry 5. 236-244. Fok, J. & Waring, M.J. (1972) Mol. Pharmacol. 8, 65-74. Behr, W. & Hartmann, G. (19651 Biochem.2. 343, 519-527. 523

Vol. 160, No. 2, 1989

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH COMMUNICATIONS

Ward, D.C. Reich. E. & Goldberg, I.H. (1965) Science 149, 1259-1263. Behr, W. Honikel, K., & Hartmenn, G. (1969) Eur,. J. [lOI. Biochem. 9, 82-92. Van Dyke, M.W. & Dervan. P.B. (1983) Biochemistry 22, [ill. 2373-2377. [121. Fox, K.R. & Howarth. N.R. (1985) Nucl. Acids Res. 13, 8695-8714. M.J. (1987) Biochim. Biophys. Acta 909, [131. Fox, K.R. & Waring, 145-155. K.S. & Nayak, R. (1976) FEBS Lett. 71, 171-174. [141. Prasad. M. & Podder, S.K. (1973) FEBS Lett. 30, [151. Nayak, R. Sirsi, 157-162. [16]. Nayak, R. Sirsi. S.K. & Podder, (1975) Biochim. Biophys. dcta 378, 195-204. D., Shashiprabha, B.K. & Podder, S.K. (1979) [171. Dasgupta, Ind. J. Biochem. Biophys. 16. 18-21. R.C., Doering, J.L. & Brown. D.D. (1980) Cell 20. [181. Peterson, 131-141. T.D. & Dombroski, B.A. (1986) Proc. Natl. Acad. 1191. Tullius. Sci. U.S.A. 83, 5469-5473. P.H. (1974) J. Mol. Biol. [201. McGhee, J.D & von Hippel. 86,469-489. [211. Keniry, M.A., Brown, S.C., Berman. E. & Shafer, R.H. (1987) Biochemistry 26, 1058-1067. R.H., Roques, B.P., LePecq, J.-B. E; Delepierre, M. [221. Shafer, (1988) Eur. J. Biochem. 173, 377-382. [91.

524