Metal-polynucleotide interactions A comparison of carcinogenic and non-carcinogenic metals in vitro

256

Biochimica et Biophysica Acta, 4 2 5 ( 1 9 7 6 ) 2 5 6 - - 2 6 1 @) E l s e v i e r S c i e n t i f i c P u b l i s h i n g C o m p a n y , A m s t e r d a m - - P r i n t e d in T h e N e t h e r l a n d s

BBA 98528

METAL-POLYNUCLEOTIDE INTERACTIONS A COMPARISON OF CARCINOGENIC AND NON-CARCINOGENIC METALS IN VITRO

M A R K J. M U R R A Y

a n d C. P E T E R F L E S S E L

Department of Biology, University of San Francisco, San Francisco, Calif. 94117 (U.S.A.) (Received July 25th, 1975)

Summary The effects of divalent cations on mixing curves of synthetic polyribonucleotides in solution are described. Manganese and cadmium chlorides at 10 -3 M induce changes suggestive of base mispairing. MnC12 induces mispairing in complexes formed between both poly(I) and poly(C,U) and poly(I) and poly(C,A) while CdC12 affects base pairing between poly(I) and poly(C,U) only. By contrast, the chlorides of magnesium and zinc show no mispairing effects with either polymer pair. Manganese and cadmium are both reported carcinogens in animals while magnesium and zinc are not. The possibility that direct metal-nucleic acid interaction may be involved in metal carcinogenesis is discussed.

Introduction This investigation concerns the in vitro interaction of carcinogenic and noncarcinogenic metals with synthetic polyribonucleotides in solution. Our findings indicate that manganese(II) and cadmium(II), both reported carcinogens (refs. 1--3 and Furst, A., personal communication) induce mispairing between nucleic acid bases whereas two chemically related but non-carcinogenic cations, magnesium(II) and zinc(II), do not. This work confirms and extends the optical studies of Eichhorn and co-workers [4,5]. Materials and Methods

Polymers. The polynucleotides * (Miles Laboratories) were supplied as lyo* Poly(I), polyriboinosinic acid, a h o m o p o l y m e r containing h y p o x a n t h i n e as the base; poly(C,U), polyribocytidylic-uridyilc acid, a c o p o l y m e r c o n t a i n i n g cytosine and uracil in u n k n o w n sequence: poly(C,A): polyribocytidylic-adenylic acid, a c o p o l y m e r containing cytosine and adenine in unk n o w n sequence.

257 philized potassium (poly(I)) or sodium (poly(C,U), poly(C,A)) salts and stored at --60° C until used. Metal solutions. Stock solutions of the metals (0.1 M) were stored at approximately l°C. The zinc solution was acidified slightly with HCI to prevent hydroxide formation. All experiments were performed in a standard buffer containing 0.1 M NaC1 and 0.005 M N-methylmorpholine (Aldrich Chemical Company), titrated to pH 7.0 with 0.6 M HC1. Optical measurements. The optical measurements were made with a Beckman DU Spectrophotometer model 2400 using standard 1 cm quartz curvettes. The cell compartment was thermojacketed and kept at a constant temperature of 15°C +- 0.5°C by cold water circulation. Continuous variation experiments. Mixtures were prepared by combining aliquots of equimolar stock solutions of the two polymers in varying proportions. Stock solutions of each polymer were made by dissolving the lyophilized polymers in the buffer at room temperature. Concentrations were calculated from the absorbances at 260 nm using the molar extinction coefficients supplied by the manufacturer. They were: poly(C,A), 8 . 2 0 . 1 0 s ; poly(C,U), 7.33 • 103; poly(I), 5.28.103. Metals were added to-the samples immediately after the polymer mixtures were prepared to give a final concentration of 10 -3 M. The mixtures were stored overnight (16--20 h) at 0--2°C. Before reading, each sample was equilibrated briefly * at 15°C in the spectrophotometer cell compartment. The absorbance readings were made near the isosbestic wavelength, by definition the wavelength at which the extinction coefficients of both absorbing species are equal. These were determined from spectra to be 252--256 nm for poly(I) + poly(C,U) and 254 nm for poly(I) + poly(C,A) in the standard buffer. Several experiments done in presence of selected divalent cations gave similar values. The relative absorbance was calculated from the relationship Ar ~" Aobs] {XHoAHo + (1 -- XHo)Aco), where Aobs is the observed absorbance, AHo the mole fraction of the homopolymer (poly(I)) and Aco the absorbance and 1 -- XHO the mole fraction of the copolymer (either poly(C,U) or poly(C,A) [6,7]. For each mixture, Ar was plotted as a function of the mole fraction of the homopolymer. Standard deviations in Ar between replicate experiments were less than 0.015 units. The mole fractions were subject to experimental errors in sample preparation estimated to be less than 5%. Results

Continuous variation experiments. Mixing curves for poly(I) and poly(C,U) in the presence of various metal chlorides are shown in Fig. 1. The comparable curves for the poly(I) and poly(C,A) are shown in Fig. 2. Examination of these figures reveals curves of two types, those showing little or no effect of metal addition (Mg, Zn) and those showing significant effects, (Mn, Cd). Furthermore, the changes induced by the addition of manganese are different than the w e equilibrated samples for several minutes in the cell compartment. Samples occasionally incubated longer g a v e c o n s t a n t readings for several hours.

* Routinely,

258 1.01

~'.

A

,~,

'~,. " ~ 0.95 -

.

,a_?

~1.0

.a--':'g'~ ,,,,

0.95

~,, u Z

B

_o,-~ "°

O.9O

090

m

ZnCI 2

MgCI 2 I

1.0

I

I

I

I

I

I

I

l

l

I

I

I

I

I

I

d.0

C

~u

Q95

0.95

/o

OJi~

J

,. 02

l

o

d

, 0.4

,

MOLE

MnCI2 i 0.6

I

J

0.90

,

CdCI 2 I

i

8

FRACTION

I

I

1.00

02

I

I

0.4

I

I

0.6

HOMOPOLYMER

I

I

1.0

0.8

(Polyl)

Fig. I . M i x i n g c u r v e s o f p o l y ( I ) + p o l y ( C , U ) . M e t a l c u r v e s ( . . . . . . ); A, I 0 -3 M MgCI2; B, 10 - 3 M Z n C l 2 ; C, 10 -3 M M n C l 2 ; D, 10 -3 M C d C I 2 a n d c o r r e s p o n d i n g c o n t r o l c u r v e s (o . . . . o) c o n t a i n i n g n o a d d e d d i v a l e n t c a t i o n s d o n e in 0.1 M N a C l a n d 0 . 0 0 5 M N - m e t h y l m o r p h o l i n e b u f f e r at p H 7.0 a n d 1 5 ° C . C u r v e s A , C, a n d D w e r e d o n e a t 2 5 2 n m a n d c u r v e B a t 2 5 6 n m . E a c h is t h e a v e r a g e o f at l e a s t t w o s e p a r a t e e x p e r i ments.

changes induced by cadmium, whose effects are significant only in the poly(I) + poly(C,U) system (Figs. 1D; 2D). Magnesium and zinc. The mixing curves done in the presence of 10 -3 M MgC12 or ZnC12 were essentially indistinguishable from control curves, within 1.0~

B

,o~-

0.95

q.O

0.95

ua U z 0.9O ,¢

0.90 MgCI 2 I

I

l

I

I

I

I

I

ZnCI 2 I

I

I

I

I

I

I

I

I

I

m

'~

9.0

1.Od

',

/

"b,'~"°/°"°/'°'/°.."'d'!

> ~-

0.95

0.95

~, /a" "a----a----a/"

0.90

0.90

MnCI 2 I

0

i

02

i

I

0.4 MOLE

i

i

0.6

J

i

0.8

FRACTION

CdCI 2 I

I

1.00

1

i

0.2

HOMOPOLYMER

I

O4

I

i

i

O6

l

O8

I

1.0

( Poly I )

Fig. 2. Mixing curves of poly(I) + poly(C,A). Metal curves (...... ); A, 10 -3 M MgCl2; B, 10 -3 M ZnCl2; C, 10 -3 M MnCl2; D, 10 -3 M CdCl 2 and corresponding control curves (o .... o) d o n e in 0.1 M N a C l and 0.005 M N-methylmorpholine buffer at p H 7.0 and 15°C. All curves were d o n e at 2 5 4 n m . Each curve is the average of at least t w o separate experiments.

259 experimental error (see Methods and Materials). This was true for both polymer-copolymer systems (Figs. 1A,B; 2A,B). The poly(I) + poly(C,A) curve with MgC12 showed a small b u t reproducible increase in relative absorbance at the poly(I)-rich end of the series (Fig. 2A). All of the curves done in magnesium or zinc (as well as the controls) had points of minimum absorbance (maximum hypochromicity) within the range XI = 0.3--0.5. This reflects the formation of p o l y ( I ) , poly(C,U) or poly(I) • poly(C,A) complexes with stoichiometric ratios between 1 : 2 and 1 : 1. Manganese and cadmium. In contrast to the minimal effects obtained with MgC12 and ZnCl2, addition of MnCI2 or CdC12 produced much more noticeable differences. As shown in Figs. 1C and 2C, MnC12 caused a substantial (50%) increase in hypochromicity relative to the control. In both cases, the point of maximum hypochromicity also appeared to be shifted to the right. This indicates that in the presence of MnC12 both copolymers interact more extensively with poly(I) forming complexes that are more nearly equimolar. In the case of CdCI2, the results were similar b u t n o t identical. Addition of CdC12 caused a significant decrease in hypochromicity (i.e. an increase in Ar) with poly(I) + poly(C,U) (Fig. 1D) b u t n o t with poly(I) + poly(C,A) (Fig. 2D). The cadmium-induced changes reflect a decrease in extent of interaction of the polynucleotides chains. In the case of the poly(I) + poly(C,U) interaction, cadmium also induced a small shift in the overall stoichiometry of complex formation (Fig. 1D). This shift, though small, was reproducible and implies the formation of more equimolar poly(I) • poly(C,U) complexes. In contrast, no change in the stoichiometry of the poly(I) • poly(C,A) complex was observed (Fig. 2D). Poly(I) precipitation by nickel(II) and cobalt(II) salts. Attempts to obtain mixing curve data for other metals were thwarted by the inherent insolubility of several salts at neutral pH and also by the fact that some metal chlorides were found to precipitate with poly(I). Both NiC12 and COC12 at 10 -3 M caused the precipitation of poly(I). In the absence of the polymer, no precipitation was observed. This result is quantitatively displayed in Table I. Incubation of poly(I) in the absence of added metal salts, or in the presence of 10 -3 M MgCI2, MnC12, ZnC12, or CdC12 for 24 h at 15°C resulted in no loss of ultraviolet-

TABLE I POLY(I) PRECIPITATION

Metal chloride added*

BY N I C K E L A N D C O B A L T S A L T S

A b s o r b a n c e at 2 4 8 n m

Percentag e absorbance

Time = 0

remaining =

Time = 24 h

(A t = 2 4 / A t = 0 ) × 1 0 0

None MgC12 MnC12

ZnC12 CdC12 NiC12 CoCI 2

0.820 0.818 0.755 0.755 0.731 0.850 0.892

0,810 0.802 0.802 0,762 0,718 0,104 0,158

98 98 106 101 98 12 18

* T h e f i n a l c o n c e n t r a t i o n w a s 10 - 3 M in all cases.

260

absorbing material. Addition of the same concentration of NiC12 or CoCl~ led to precipitation which was observable within 5 min and which, after 24 h of incubation had resulted in the loss of 80--90% of the ultraviolet absorbing material from the solution. Discussion The primary conclusion to be drawn from this work is that MnCI2 and CdC12 alter the interactions of selected pairs of synthetic polyribonucleotides, while MgC12 and ZnCl2 do not. These results extend and partially confirm the studies of Eichhorn [5]. In that work, mispairing of nucleotide bases was reported to be induced by mangenese(II), magnesium(II) and nickel(II) at a concentration, (0.01 M), ten-fold higher than used here. Because genetic studies showed manganese(II) and nickel(II) to be optimally mutagenic at 10 -3 M, we began the optical measurements at this concentration. At 10 -3 M it was possible to distinguish between two carcinogenic metals Mn{II) and Cd(II) and two noncarcinogenic metals Mg(II) and Zn(II). The two metal carcinogens altered the salient characteristics of the mixing curves in ways suggesting that these "divalent ions indeed produce mispairing of bases" [5]. The two non-carcinogens, however, induced no such changes. This conclusion follows from the continuous variation experiments described above. In the absence of any added metal salts, mixtures of poly(I) with either poly(C,U) or poly(C,A) form complexes containing approximately one mole of the h o m o p o l y m e r to two moles of the copolymer. Complex formation presumably reflects Watson-Crick base pairing between the hypoxanthine residues in the poly(I) and the cytosine residues in the poly(C,U) or poly(C,A). In the presence of MnC12, poly(I) forms complexes with both copolymers which are more nearly equimolar when compared with the controls (Figs. 1C and 2C). This shift from a (1 : 2) towards a (1 : 1) complex suggests more extensive interaction between hypoxanthine and either uracil or adenine. In contrast to Mn(II), Cd(II) facilitates this shift only with the poly(I) • poly(C,U) complex. No detectable shift in the stoichiometry of the poly(I) • poly(C,A) complex was observed (compare Figs. 1D and 2D), even though some destabilization of the complex occurred. This suggests that Cd{II) can stabilize a helical array involving hypoxanthine and uracil more readily than one involving hypoxanthine and adenine. Presumably this is related to the fact that it is easier to a c c o m m o d a t e base pairing between a purine (hypoxanthine) and a pyrimidine (uracil) than between t w o purines (hypoxanthine and adenine) in a helix of constant diameter. In contrast to Mn(II) and Cd(II), Mg(II) and Zn(II) induce no optical changes suggesting mispairing. This is despite the fact that "almost all complexes of Mn(II) with ligands containing oxygen and nitrogen are isomorphous with corresponding Mg(II) complexes" [8]. In view of this it is interesting that Mn(II) and Mg(II) are clearly distinguishable under our assay conditions (Figs. 2A; 2C; 3A and 3C). This suggests that the larger ionic radius of Mn(II) v i s a vis Mg(II) may be important to its mispairing ability. Comparison of Zn and Cd leads to a similar conclusion. Both are members of Group I I B in the periodic table. Both have identical dl°s 2 electron configurations, with Zn occupying a position at the end of the first and Cd a position at the end of the second transition series. Salts of the t w o metals form

261

similar ligand co-ordination complexes, but just as with Mn and Mg, the metal ion with the larger ionic radius, (Cd), induces mispairing with the smaller ion, (Zn), does not. It may be significant that the two metals which induce changes consistent with some degree of mispairing, Mn and Cd, are both believed to be carcinogens in animals. This, coupled with the demonstration that manganese (ref. 9 and Sholla, C. and Flessel, C., manuscript in preparation) and nickel (Schibler, M. and Flessel, C., manuscript in preparation) are mutagens in bacteria, makes it appear plausible that metal carcinogenesis at least in some cases may involve the direct interaction of metal ions with the nucleic acids. Acknowledgements This study was made possible by a grant from the Lilly Drake Cancer Research Fund, University of San Francisco. The authors wish to thank Professor Arthur Furst and Ms. Norma Washington for their support, advice and assistance throughout this study. We also thank Theodore Jones for his thoughtful comments on the manuscript. References 1 Heath, J.C. and Webb, M. (1967) Br. J. Cancer 2 1 , 7 6 8 - - 7 7 9 2 Furst, A. and Haro, R.T. (1969) Jerusalem Symposia on Q u a n t u m Chemistry and Biochemistry (Bergman, E.D. and PulLman, B., eds.), pp. 310--320, The Israel A c a de my of Sciences and Humanities, Jerusalem 3 Furst, A. (1971) E n v i r o n m e n t a l Geochemistry in Health and Disease (Cannon, A.L. and Hopps, H.C., eds.), pp. 109--130, Geological Society of America, Boulder, Colorado 4 Eichhorn, G. and Shin, Y. (1968) J. Am. Chem. Soc. 90, 7323--7328 5 Eichhorn, G.L., Richardson, C. and Pitha, J. (1971) Am. Chem, Soc. Meet. Abstr., Washington, D.C Biol. 17 6 Blake, R.D., Massaoulie, J. and Fresco, J.R. (1967) J. Mol. Biol. 30, 291--308 7 L o m a n t , A. and Fresco, J. (1973) Biopolymers 12, 1889--1903 8 0 r g e l , A. and Orgel, L.E. (1965) J. Mol. Biol. 14, 453--457 9 Dernerec, M. and Hanson, J. (1951) Cold Spring Harbor Symp. Quant. Biol. 1 6 , 2 1 5 - - 2 2 8