378

Biochimica et Biophysica Acta, 479 (1977) 378--390 © Elsevier/North-Holland Biomedical Press

BBA 99066

THE INTERACTION OF PROPIDIUM DIIODIDE WITH SELFCOMPLEMENTARY DINUCLEOSIDE MONOPHOSPHATES

MICHAEL W. DAVIDSON, BILLY G. GRIGGS, IRENE G. LOPP and W. DAVID WILSON *

Department of Chemistry, Georgia State University, Atlanta, Ga. 30303 (U.S.A.) (Received May 2nd, 1977)

Summary The interactions of a quinacrine derivative, methylated at both the aromatic and aliphatic nitrogens, and propidium diiodide with the dinucleoside monophosphates CpG, GpC, UpA and ApU have been investigated using 13C-NMR (for the quinacrine derivative prepared with [~3C]methyl substituents and ~HNMR and ultraviolet-visible spectroscopy. The quinacrine derivative displayed negligible interaction with the dinucleosides at concentrations up to 5 . 1 0 -4 M. Propidium did form complexes with dinucleosides even at concentrations as low as 10 -4 M. Propidium displayed a pyrimidine-purine binding preference and gave especially large changes in ultraviolet-visible and 1H-NMR spectra in the presence of CpG. This suggests that propidium forms an intercalated complex with a Watson-Crick hydrogen-bonded CpG dimer. At higher concentrations UpA and GpC gave similar spectral changes indicating that they could also form significant amounts of an intercalated complex with propidium under appropriate conditions. The changes caused by ApU were small under all conditions and were more similar to the effects caused by mononucleotides. These results indicate that, at least for phenanthridines, cationic side chains do not greatly inhibit complex formation with dinucleoside monophosphates, and suggest that the weak interaction of the quinacrine derivative with dinucleosides is due to weaker interactions of the acridine ring system with nucleoside bases relative to the phenanthridine ring system.

* To w h o m correspondence should be addressed. Abbreviations: CpG, cytidylyl(3t-~of)guanosine; GpC, guanylyl(3t-->Sr)cytidine; UpA, uridylyl(3~--->Sr)adenosine; ApU, adenyIyl(3~-'>St)uridine; pG, guanosine 5r-monophosphoric acid; propidium diiodide, 3,8-diarnlno-5-(diethylmethylaminopropyl)-6-phenylphenanthridinium diiodide; TSP, sodium 3-trirnethylsilylpropionate-2,2,3,3-d4.

379 Introduction

Phenanthridine derivatives, especially ethidium bromide (Fig. 1) have been known for many years as trypanocidal drugs which inhibit nucleic acid synthesis in .several in vivo and in vitro systems [1]. Henry and coworkers [2] have found that phenanthridine derivatives also exhibit significant antineoplastic activity and inhibit both DNA and RNA synthesis in L-1210 t u m o r cell culture. Acridine derivatives also have a variety of biological effects, and considerable information has been accumulated on the interaction of the antimalarial acridine, quinacrine, with DNA [3]. Several acridine derivatives related to quinacrine have also been found to have antineoplastic activity [4]. Both quinacrine [5] and ethidium [6] bind strongly to DNA through formation of an intercalated complex. One of the most important aspects for characterizing intercalation of small molecules into double helical nucleic acids is base pair specificity in binding. There are ten different base pair intercalation sites in double helical nucleic acids [8] and each site, in theory, can exhibit different interactions with bound molecules. General preferences for binding sites containing A - T or G . C base pairs can be determined through binding experiments with DNA samples of varying base pair composition [9,10], but this does not determine sequence specificity. Direct determination of sequence specificity using synthetic double helical polynucleotides, while possible, is limited due to the lack of availability of the appropriate polynucleotides. Miiller et al. [11], using a three-chamber equilibrium dialysis cell and two DNA samples of different base pair composition have developed an accurate m e t h o d for determining general base pair binding preferences. Using this technique, however, no significant specificity in the binding of either ethidium or quinacrine to DNA was found [12]. Short self-complementary nucleotide sequences, down to the dinucleotide level, can form Watson-Crick hydrogen-bonded complexes in solution [ 13--15]. These samples have been used to characterize binding sequence specificity for drugs such as ethidium bromide [16--18]. These smaller nucleic acid segments can also tumble much faster in solution than native DNA or RNA molecules allowing NMR studies of both the association of complementary nucleotide segments to form miniature double helices and their interaction with intercalating molecules. A model for the small molecule-nucleic acid complex can thus be developed in combination with other spectral studies (fluorescence, circular dichroism, ultraviolet-visible absorption) which can also be conducted with high molecular weight DNA and RNA. Extrapolation of these studies with short nucleotide sequences to double helical DNA or RNA requires the assumption that pronounced solvent interactions of the base pairs, which are usually negligible in large polymers, do not greatly influence complex structure or binding specificity. With dinucleoside mon ophosphates, in particular, this could be a problem since they generally form significant amounts of the miniature double helices in solution only in the presence of intercalating molecules. The fact that the dinucleosides can form Watson-Crick hydrogen-bonded dimers in the absence and presence of intercalating molecules is strongly supported by spectroscopic evidence and by several X-ray structures of crystallized complexes [19--21].

380 With actinomycin the binding specificity with DNA and complementary dinucleoside monophosphate segments is in good agreement and indicates strong binding to G • C sequences [22--24]. With ethidium the results are more complex. As discussed above, studies with high molecular weight DNA indicated no pronounced base pair binding preference [12]. With dinucleosides, however, stronger interactions are found with pyrimidine-purine sequences regardless of the composition [16]. The molecular basis for this binding specificity is not known at present. Because of the importance of determining binding specificity when analyzing small molecule-nucleic acid interactions, we have extended our studies with acridine and phenanthridine derivatives to include interactions with the self complementary dinucleoside monophosphates CpG, GpC, UpA and ApU. To facilitate studies by removing complications due to water and overlapping peaks between drug and dinucleoside, we have synthesized a 13C-labelled methylated quinacrine derivative (Fig. 1). This c o m p o u n d was found to interact negligibly with dinucleosides even at concentrations near its solubility limit. To help resolve the question of why this c o m p o u n d binds weakly to dinucleosides relative to ethidium, we analyzed the interaction of propidium diiodide (Fig. 1), a phenanthridine with a cationic side chain, with dinucleoside monophosphates. Propidium interacts strongly with some self-complementary dinucleoside monophosphates in a manner qualitatively similar to ethidium but with some important quantitative differences. Experimental section

Materials. Dinucleoside monophosphates were purchased from Sigma Chemical Co. and dissolved as concentrated stock solutions 0 . 8 . 1 0 - 2 - - 2 . 0 • 10 -2 M) in buffer (5 • 10 -3 M NaH2PO4, 10 -4 M EDTA, adjusted to pH 7.0 with NaOH). Concentrations were determined by diluting the above solutions in 0.1 M HC1 and using extinction coefficients listed in PL Biochemicals catalog No. 104. Propidium diiodide was a product of Calbiochem and ethidium bromide was purchased from Aldrich Chemical Co. Both samples were monitored for purity using thin-layer chromatography, spectroscopy, and elemental analysis. Samples of propidium and ethidium were dried to constant weight in a vacuum at 100°C before preparation of stock solutions (2 • 10-3--4 • 10 -3 M) in buffer. 2-Methoxy-6-chloro-9-(1-methyl-4-(diethylmethylammonium)-butylamino)-10methylacridinium diiodide was synthesized from quinacrine and characterized as described elsewhere (Wilson et al., unpublished data). Buffer salts were of the highest purity commerically available and deionized water was glass distilled from acidic permanganate. All 2H20 was purchased from Aldrich Chemical Co. NMR spectroscopy. For all experiments compounds were dissolved in a 99.8 atom % 2H20 solution of standard buffer adjusted to a pH meter reading of 7.0 with NaO2H. To remove exchangeable protons from the sample and reduce the intensity of the solvent resonance in 1H-NMR experiments, stock solutions were diluted into the proper concentration in buffer, lyophilized, a n d then redissolved under N2 atmosphere in 99.96 atom % 2H20. NMR experiments were performed on a JEOL FX-60 Fourier Transform Spectrometer. The 1H-NMR spectra were collected at a frequency of 59.79 MHz utilizing

381 5-mm diameter NMR tubes (Wilmad 507-pp). Data were accumulated in a Texas Instruments 980B computer using a 1000 Hz spectral width in 4096 data points to give 0.48 Hz data point resolution. Samples were irradiated using a 90 ° pulse width, a 2.5 s pulse repetition and normally were accumulated for approximately 104 pulses. Temperature control was achieved using the JEOL variable temperature module. The chemical shifts were referenced to TSP 0.005%) which was added as a 2H20 solution after the lyophilized samples were redissolved in 99.96 atom % 2H20. Inclusion of TSP in the samples before lyophilization led, in some cases, to precipitates which were difficult to redissolve and resulted in loss of the TSP signal. The 13C-NMR experiments were conducted at 15.04 MHz using 10-mm diameter NMR tubes. Samples were pulsed approximately 2 • 104 times using 8192 computer words with a 2500 Hz spectral width yielding 0.9 Hz data point resolution. A 45 ° pulse width and a 2.5 s repetition time were utilized. The samples were referenced to internal dioxane s e t at 67.4 ppm relative to Me4Si. All experiments utilized broad-band proton noise decoupling. Visible spectroscopy. The visible electronic absorption spectra of propidium and propidium-dinucleoside monophosphate solutions were recorded utilizing either 0.1- or 1.0-cm lightpath quartz cuvettes (Pyrocell) in a Cary 17Dspectrophotometer. Temperature was regulated to 0.01°C with a Haake {model No. FE 391) circulating water bath. During low temperature studies, the sample chamber of the spectrophotometer was purged with dry N~ to reduce moisture condensation on the cuvette windows. Temperature was monitored through a thermister immersed in the sample cell and coupled to a Beckman temperature module accessory (No. 754030). Spectrophotometric titrations were conducted by t w o methods: (i) successive aliquots of concentrated dinucleoside monophosphate solution were added with a calibrated microliter syringe (Hamilton} to a 10 -4 M solution of propidium (0.9--1.2 ml) in a reduced volume 1.0 cm cuvette, the resulting solution was carefully mixed and, after temperature equilibration, the spectrum was recorded; and (ii) aliquots of dinucleoside monophosphate solution were added as before to a 10 -3 M propidium solution {1.3--1.5 ml) in a circular 0.1-cm cuvette fitted with a teflon stopper, the solution was mixed by repeated inversion of the cuvette, and the spectrum was recorded. In both titration methods, the dinucleoside monophosphate stock solution contained propidium at the same concentration as in the sample cuvette to avoid dilution of the drug during the course of the titration. Results

NMR Methylated quinacrine. The 13C-NMR spectrum of [13C]methylated quinacrine {Fig. 1) displayed two peaks, a singlet at 36.8 ppm, and a triplet at 47.5 ppm. These were assigned by synthesizing a non-13C-labeled m o n o m e t h y l derivative. In the natural abundance 13C spectrum of this c o m p o u n d only the side chain carbon peaks shifted relative to quinacrine and a new peak at 47.5 ppm, corresponding to the added methyl group, was present. An unlabeled dimethyl quinacrine derivative, identical in structure to the labeled c o m p o u n d {Fig. 1), had shifts in both the aromatic and side-chain carbon peaks in the natural abun-

382 9

I0

I

2

H2N-~'~~NH2

ethidium, ..-c.2c. s propidium, .--CH2C.2C.2 (cHs)(c2He2

R

~

e

41 NHCH(CH3)(CH2 )3N(C2H5)2 el i OCH3

R

quinacrine, ,.-H dimethylquinacrine, R.-c. 3 Fig. 1. T h e d i m e t h y l q u i n a e r i n e d e r i v a t i v e w a s p r e p a r e d b o t h w i t h 13C.labele d a n d n o n - l a b e l e d m e t h y l groups at the acridine and side-chain q u a t e r n a r y nitrogens.

dance 13C-NMR spectrum in addition to the new methyl peaks at 36.8 and 47.5 ppm. The signal at 47.5 ppm is therefore assigned to the side-chain m e t h y l group and is split into a triplet due to coupling to the aliphatic side-chain nitrogen. The signal at 36.8 ppm is assigned to the m e t h y l substituent on the aromatic nitrogen and no detectable coupling to the acridine nitrogen is found. The 13C-labeled dimethyl quinacrine derivative (5 • 10 -4 M, 15°C) was titrated with up to a 3-fold excess of dinucleoside monophosphates and the methyl resonances were monitored. Within experimental error, no changes in the chemical shifts and no broadening of these peaks occurred. Titration of the methylated quinacrine (5 • 10 -4 M, 15°C) in the visible spectral region with CpG also indicated no detectable change in extinction coefficient or shift in the wavelength of m a x i m u m absorbance. Propidium diiodide. The 1H-NMR spectrum of propidium is shown in Fig. 2. The assignment of the aromatic protons follows directly from that for ethidium bromide [25]. The resonances for H-1 and H-10 are better resolved than is observed in the spectrum of ethidium, indicating less dimerization for propidium at similar concentrations [22]. For this reason, dimer-induced chemical shift changes for the propidium aromatic proton signals were negligible on complexation with dinucleoside monophosphates. The resonances for H-I, H-10 (doublet centered at 8.73 and 8.58 ppm), and H-7 (a doublet at 6.67 ppm) are easily distinguished from the remaining aromatic protons of propidium (at 10 -3 M) and most protons of the dinucleosides. They provide convenient signals for

383

CpG, PI

// !

r

2,1

I

i

ko.5,~ HDO

TSP

I HI, ~0,i I

I

I

I

I

9

8

,7

6

5

I

I

I

I

I

4

3

2

I

0

ppm Fig. 2. 1 H - N M R s p e c t r a o f p r o p i d i u m ( 1 0 -3 M) a n d p r o p i d i u m - C p G s o l u t i o n s a t 3 0 ° C . T h e c o n c e n t r a t i o n o f C p G is given as a m o l a r ratio t o t h e p r o p i d i u m c o n c e n t r a t i o n . T h e b o t t o m s p e c t r u m is f o r p r o p i d i u m a l o n e a n d t h e n e x t t h r e e s p e c t r a h a v e CpG c o n c e n t r a t i o n s of 0.5 • 10 - 3 , 1 • 10 -3 a n d 2.0 • 10 -3 M, res p e c t i v e l y . All s p e c t r a are r e f e r e n c e d t o TSP. PI, p r o p i d i u m diiodide.

monitoring complex formation with dinucleosides and will be referred to frequently. The triplet at 1.21 ppm arises from the methyl of the N-ethyl substit u e n t and a singlet due to the N-methyl substituent is obtained at 2.9 ppm. These signals are useful in monitoring the cationic side chain of propidium during complex formation. In Fig. 2, spectra for titration of propidium with CpG at 30°C are shown. Increasing the ratio of CpG to propidium causes drastic broadening of all the propidium and CpG signals. Ring current-induced shielding occurs for the aromatic protons resulting in upfield shifts for H-1 and H-10 of 0.5 ppm and of 0.75 ppm for H-7 at a 2 : 1 ratio of CpG to propidium. No shift of the resonances for the side-chain methyls is observed, however, the signals do broaden appreciably indicating restricted rotation and m o v e m e n t of the side chain in

311, c

i I

I

I

! I GH8 C H 6

| 9

I 8

I 7

I 6

I 5

ppm Fig. 3. 1 H - N M R s p e c t r a o f e q u i m o l a r m i x t u r e s o f p r o p i d i u m ( 1 0 -3 M) a n d C p G o r G p C . S p e c t r a o f C p G a n d (}pC ( 1 0 -3 M) a r e i n c l u d e d f o r r e f e r e n c e . PI, p r o p i d i u m d i i o d i d e .

the propidium-CpG complex. The guanine H-8 and cytosine H-6 signals shift upfield b u t are difficult to quantitate due to overlap with signals from the propidium phenyl, H-2, H-4, and H-9 protons which are all located between 7 and 8 ppm. Cytosine H-5 is also difficult to quantitate due to overlap with the ribose H-I' signal. In Fig. 3 the spectrum of GpC and propidium at 30°C and at a molar ratio of 1 : 1 is compared with the CpG propidium spectrum under the same conditions. The differences between the t w o spectra are striking in that GpC produces very little broadening or chemical shift changes of the propidium signals compared with the effects observed for CpG. The signals for H-l, H-10, and H-7 in the CpG-propidium sample are all shifted approximately 0.5 ppm upfield relative to the GpC-propidium sample. The side-chain methyl signals are n o t shifted relative to free propidium for either dinucleoside b u t are significantly broader in the CpG than in the GpC sample. The spectra for the dinucleoside monophosphates, CpG and GpC at 1.0 • 10 -3 M have been included for

385

•

i IO*C

1

!

'

20"C

I

HI'HIOIi i/~

=H? /

9

P

i

!

I

8

z

.

s

~

ppm

3

2

;

o

Fig. 4. 1 H - N M R s p e c t r a o f a n e q u i m o l a r m i x t u r e of propidium and GpC ( b o t h a t 1 0 -3 M) a t 30, 20, 10, and 5°C.

reference. As in Fig. 2 the guanine and cytosine ring proton shifts are difficult to quantitate b u t do move upfield on interaction with propidium. Broadening of these protons in CpG is significantly greater than in GpC (Fig. 3). To further investigate restraints on the binding of propidium with GpC, spectra for the 1 : 1 solution with propidium were obtained at varying temperatures .(Fig. 4). As the temperature is decreased, the aromatic peaks are broadened and shifted upfield. At 5°C the chemical shift changes and broadening of the aromatic signals are similar to those found for the 1 : 1 CpG complex at 30°C. The side-chain methyl groups of propidium do n o t shift relative to the free comp o u n d even at 5°C in the GpC complex b u t they do become broader as the temperature is lowered. Much smaller changes in the PMR spectrum of propidium are obtained on

386

AH8 I

UH5 AH2

n

UpA

UpA,PI 2,1

HI,~: Z If I' AH8 ' i

II

PI 2,1

f/

ApU'

TSP

UH5 a l

,t I ApU

7 6 5ppm 4

I

3

i

;

o

Fig. 5. I H - N M R s p e c t r a o f m i x t u r e s o f p r o p i d i u m ( 1 0 -3 M) w i t h U p A a n d A p U a t a 2 : 1 m o l a r r a t i o a n d 15°C. S p e c t r a o f U p A a n d A p U ( 1 0 -3 M) h a v e b e e n i n c l u d e d f o r r e f e r e n c e . PI, p r o p i d i u m diiodide.

addition of UpA and ApU. At 30°C and a 2 : 1 ratio of ApU to propidium there are upfield shifts for H-l, H-10 of approx. 0.1 ppm and less than 0.1 ppm for H-7. With UpA at 30°C and a 2 : 1 ratio, the H-1 and H-10 signals shift 0.25 ppm and H-7 shifts 0.1 ppm. No signal shift is obtained for either side-chain methyl group with ApU or UpA. As the temperature of these samples is lowered, larger upfield shifts for the propid]um aromatic signals are obtained. In Fig. 5, the spectra for propidium with UpA and ApU are shown at 15°C along with reference spectra for ApU and UpA. The upfield shifts induced by UpA remain greater than for ApU. The signals for propidium H-1 and H-10 are somewhat obscured by H-8 of adenine b u t are shifted farther upfield by UpA than ApU. The signals for H-7 are easily resolved and are shifted upfield 0.1 ppm in the ApU complex and over 0.3 ppm in the UpA complex (Fig. 5). In both complexes, H-8 of adenine shifts upfield less than 0.1 ppm. The H-2 proton of ade-

387

nine is more sensitive to complex formation and shifts 0.1 ppm for the ApU complex and 0.25 ppm in the UpA complex. The H-5 and H-6 protons of uracil shift slightly upfield on complex formation with both UpA and ApU. The propidium side-chain methyl signals broaden slightly but do not shift for either complex. At 5°C all propidium signals continue to broaden and signals in the aromatic region shift upfield with the effects for UpA larger than for ApU. Because of broadening and overlap with adenine and uracil protons, the changes could not be accurately determined at this temperature.

Visible absorption spectral changes Large hypochromic effects and shifts of absorption maxima to longer wavelengths are noted on titrating propidium (10 -4 M) with CpG but only marginal changes are produced with GpC (not illustrated). Isosbestic points occur in the CpG titration at 400 and 521 nm but no definite isobestic points are found with GpC. The isosbestic points found in the CpG-propidium titration curves correspond closely to those obtained when propidium is titrated with native duplex DNA (Wilson, W.D. and Davidson, M.W., unpublished data). Both ApU and UpA induce even smaller effects in the propidium absorption spectrum than does GpC at these concentrations. A much larger decrease in absorption

I

I.Ot;

E;o 0.7 0.65 0.6 0.55 0.5

I

2

3

4

5

6

7'

C n (X 10 4) F i g . 6. C h a n g e s i n p r o p i d i u m e x t i n c t i o n c o e f f i c i e n t at 4 9 2 n m as a f u n c t i o n o f d i n u c l e o s i d e p h a t e c o n c e n t r a t i o n . C o n d i t i o n s are similar t o t h o s e d e s c r i b e d i n F i g . 2. o, C p G 0 p r o p i d i u m t i o n = 9 . 9 1 • 1 0 --4 M ; e , U p A , p r o p i d i u m c o n c e n t r a t i o n = 9 . 9 5 • 1 0 --4 M ; ~, G p C , p r o p i d i u m t i o n 1 . 0 0 • 1 0 - 3 M ; A A p U , p r o p i d i u m c o n c e n t r a t i o n = 1 . 0 4 • 1 0 - 3 M : a n d A, p G , p r o p i d i u m t i o n = 9 . 9 2 • 1 0 - 4 M.

monophosconcentraconcentraconcentra-

388 occurs in the ethidium titration with UpA than observed with propidium, although this change is still small relative to that obtained with CpG (not shown). Absorption decreases are similar for ethidium and propidium when titrated with the other dinucleoside monophosphates. Perturbations in the propidium absorption spectrum induced by dinucleoside monophosphates were also monitored at 10-fold higher concentration (approx. 10 -3 M), the same concentration utilized in NMR experiments. ApU produces changes similar to those produced by GpC at the lower concentration (not shown). The UpA-propidium spectra are somewhat more complicated. At higher wavelengths, two sets of apparent isosbestic points occur at 516 and 522 nm, while at lower wavelengths no definite isosbestic points are noted. Absorbance changes produced by the interaction of propidium with all four dinucleoside monophosphates and with pG are illustrated in Fig. 6. CpG induces a greater hypochromic effect than GpC, consistent with previous results at lower concentrations. The titration curve for UpA with propidium shows much greater change than at lower concentrations. At high dinucleotide to propidium ratios, the changes produced by UpA are second only to those produced by CpG. ApU and pG have similar, quite small, effects on the propidium absorption spectrum. Discussion

The dimethyl quinacrine derivative, illustrated in Fig. 1, interacts strongly with DNA (Wilson, W.D. and Lopp, I.G., unpublished data) b u t no significant interaction was observed with up to a 3-fold molar excess of any of the dinucleoside monophosphates used in our studies. Strong interactions with all four dinucleosides have been found with ethidium under similar conditions [16,17, 26]. The dramatic differences in dinucleoside interactions between ethidium and the methylated quinacrine derivative could be due to (i) differences in stacking of the substituted acridine ring with the dinucleoside base pairs relative to the substituted phenanthridine ring, or (ii) the presence of a cationic side chain on the quinacrine derivative, or (iii) a combination of these effects. In an a t t e m p t to determine which of the above is correct, the interaction of propidium iodide (Fig. 1) with UpA, ApU, CpG, and GpC was investigated. The shielding and line broadening obtained for the propidium aromatic protons in NMR experiments with CpG and UpA at high concentration and low temperature are t o o large to arise simply from interaction of the phenanthridine ring with only a single dinucleoside m o n o p h o s p h a t e [16,17,25]. In addition, the visible spectral changes obtained for propidium at high ratios of CpG and UpA are quite similar to those obtained when propidium binds to DNA (Wilson, W.D. and Davidson, M.W., unpublished data). These results suggest the propidium can complex with self-complementary pyrimidine-purine dinucleosides in solution to form a Watson-Crick hydrogen-bonded dinucleoside dimer in which the phenanthridine ring is intercalated between base pairs. The interaction of propidium and ethidium is quite similar for the dinucleoside monophosphates: CpG, GpC, and ApU b u t propidium complexes more weakly with UpA than does ethidium. Sobell and coworkers [19] using X-ray crystallography have determined the structure of a double helical complex of ethi-

389 dium and 5-iodo-UpA. In this structure, the phenyl and ethyl substituents of ethidium are directed to what would be the minor groove of macromolecular double helical DNA. We have analyzed the interaction of ethidium and propidium with a Corey-Pauling-Koltun space-filling molecular model of a WatsonCrick hydrogen-bonded complex of UpA using the orientations found in the X-ray structure of ethidium and 5-iodo-UpA. Using these models, no obvious steric constraints are apparent in the propidium-UpA complex which do not also exist for the ethidium complex. These studies with propidium illustrate that compounds with cationic side chains can interact quite strongly with dinucleoside monophosphates. Although some quantitative differences in binding such as with UpA, can result when a cationic side chain is present, it seems highly unlikely that addition of a simple side chain would abolish binding completely. The lack of interaction of the methylated quinacrine derivative with dinucleoside monophosphates must be due, then, to differences between the substituted acridine and phenanthridine ring systems. It seems likely that the substituted phenanthridine ring has molecular structural features which optimize complex formation with dinucleoside monophosphates and in particular, self-complementary pyrimidine-purine dinucleosides. The acridine ring system and the antibiotic echinomycin [9] do not significantly interact with dinucleosides, suggesting that they lack these critical structural features even though they both bind quite strongly to double helical DNA. The elucidation of the binding differences among these molecules should prove helpful in understanding interaction specificities for nucleic acids in general. Acknowledgements The authors thank Professor David W. Boykin for helpful comments concerning this research and manuscript. This work was supported in part by grants from the Research Corporation, the Petroleum Research Fund, and the Georgia State University College of Arts and Sciences Research Fund. The authors wish to thank June M. Nicks for preparation of the manuscript. References 1 Waring, M.J. (1974) in Antibiotics I n (Corcoran, J.W. and Hahn, F.E., ed.), pp. 141--165, SpringerVerlag, New York 2 Mosher, C.W., Ku hlmann, K.F. and Henry, D.W. (1976) Abstracts 172nd ACS National Meeting, San Francisco, California August 29, September 3, MEDI 78 3 Blake, A. and Peacocke, A.R. (1968) Biopolymers 6, 1225--1253 4 Cain, B.F., Atweli, G.J. and Denny, W.A. (1976) J. Med. Chem. 19, 772--778 5 Saucier, J.M., Festy, B. and Le Pecq, J.-B. (1971) Biochimie 53, 973--980 6 Waling, M.J. (1970) J. Mol. Biol. 54, 247--279 7 Wilson, W.D., Gough, A.N., Doyle, J.J. and Davidson, M.W. (1976) J. Med. Chem. 19, 1261--1263 8 Gabbay, E.J., DeStefano, R. and Sanford, K. (1972) Biochem. Biophys. Res. C ommun. 46, 155--161 9 Wakelin, L.P.G. and Waring, M.J. (1976) Biochem. J. 1 5 7 , 7 2 1 - - 7 4 0 10 Wartell, R.M., Larson, J.E. and Wells, R.D. (1974) J. Biol. Chem. 249, 6719--6731 11 M/ilier, W., Crothers, D.M. and Waring, M.J. (1973) Eur. J. Biochem. 39, 223--234 12 M/iller, W. and Crother, D.M. (1975) Eur. J. Biochem. 54, 267--277 13 Borer, P.N., Kan, L.S. and Ts'O, P.O.P. (197,5) Biochemistry 14, 4847---4863 14 Patel, D.J. and Tonelll, A.E. (1975) Biochemistry 14, 39 90--3996 15 Krugh, T.R. and Young, M.A. (1975) Biochem. Biophys. Res. C o m m u n . 62, 1025--1031

390

16 17 18 19 20 21 22 23 24 25 26

K r u g h , T . R . , W i t t l i n , F . N . a n d C r a m e r , S.P. ( 1 9 7 5 ) B i o p o l y m e r s 1 4 , 1 9 7 - - 2 1 0 K r u g h , T . R . a n d R e i n h a r d t , C . G . ( 1 9 7 5 ) J. Mol. Biol. 9 7 , 1 3 3 - - 1 6 2 P a t e l , D . J . a n d C a n u e l , L . L . ( 1 9 7 6 ) P r o c . N a t l . A c a d . Sci. U.S. 7 3 , 3 3 4 3 - - 3 3 4 7 Tsai, C.-C., J a i n , S.C. a n d SobeU, H . M . ( 1 9 7 5 ) P r o c . N a t l . A c a d . Sci. U.S. 72° 6 2 8 - - 6 3 2 S e e m a n , N.C., R o s e n b e r g , J . M . , S u d d a t h , F.L.0 K i m , J . J . P . a n d R i c h , A. ( 1 9 7 6 ) J. Mol. Biol. 1 0 4 109--144 R o s e n b e r g , J . M . , S e e m a n , N.C., D a y , R . O . a n d R i c h , A. ( 1 9 7 6 ) J. Mol. Biol. 1 0 4 , 1 4 5 - - 1 6 7 K r u g h , T . R . a n d C h e n , Y.-U. ( 1 9 7 5 ) B i o c h e m i s t r y 1 4 , 4 9 1 2 - - 4 9 2 2 K r u g h 0 T . R . , M o o b e r y , E.S. a n d C h a i o , Y.-U.C. ( 1 9 7 7 ) B i o c h e m i s t r y 1 6 , 4 7 0 - - 4 7 8 Wells, R . D . a n d L a r s o n , J . E . ( 1 9 7 0 ) J. M o l . Biol. 4 9 , 3 1 9 - - 3 4 2 K r e i s h m a n 0 G . P . , C h a n , S.I. a n d B a u e r , W. ( 1 9 7 1 ) J . Mol. Biol. 6 1 , 4 5 - - 5 8 Patel, D.J. (1976) Biochim. Biophys. Acta 442, 98--108