RNA-diethylstilbestrol interaction studied by Fourier transform infrared difference spectroscopy.

Diethylstilbestrol (DES), a synthetic estrogen, is known to be a carcinogen in human and in animals. This study was designed to examine the interaction of DES with yeast RNA in aqueous solution at physiological pH with drug/RNA-phosphate (P) molar ratios of 1/80, 1/40, 1/20, 1/10, 1/4, and 1/2. Fourier transform infrared (FTIR) difference spectroscopy was used to determine the drug binding mode, the binding constant, the sequence selectivity, and RNA secondary structure in the RNA.DES complexes. Spectroscopic evidence showed that at low drug concentration (1/80 and 1/40), DES is intercalating through both Gua-Cyt and Ade-Urd base pairs with minor interaction with the backbone PO2 group (external binding). The calculated binding constant of K approximately 8.5 x 10(4) M-1 at a drug concentration of 3.12 x 10(-4) M shows that DES is a weaker intercalator than those of the methylene blue, acridine orange, and ethidium bromide. At high drug content (r > 1/40, where r represents the DES/RNA-phosphate molar ratio), a partial helix destabilization occurs with no alteration of RNA conformation upon drug complexation. However, a comparison with DNA.DES complexes showed that drug intercalation causes major reduction of the B-DNA structure in favor of A-DNA with no participation of the backbone PO2 group in the DES. DNA complexation.

DES 1 (see Structure 1), a synthetic estrogen, is known to be a carcinogen both in human and in animals (1,2). It was used world wide from the 1940s until 1970s to prevent miscarriages in women, while at the same time, it was found to cause cancer in experimental animals. Although DES application has been banned for pregnancy, it remains in use for other clinical purposes such as estrogen replacement therapy for hormone deficiency. The DES-induced carcinogenesis has been considered to be due to the high hormonal potency of this synthetic estrogen and possible DES⅐DNA adduct formation in vivo (3). Radioactively labeled DES was found to bind DNA in vivo (3)(4)(5) and in vitro (6), but the nature of complexation with nucleic acids could not be clarified in these investigations. The major problem in defining the nature of drug-DNA interaction was mainly related to the instability of DNA⅐DES and DNA⅐DESQ complexes (diethylstilbestrol-4Ј,4Љ-quinone, DESQ, a metabolic intermediate product derived from DES oxidation) formed in vivo and in vitro (6,7). Similarly, the possibility of DNA intercalate formation with both DES and DESQ was not included (7,8). It has also been suggested that DES-DNA adduct formation may occur under oxidative stress (9,10). However, a number of biological and biochemical effects of DES were noted, which depended on metabolic activation of stilben and are commonly associated with genotoxic activity (11). Since the DES-induced carcinogenesis is related to its complexation with nucleic acids (2), the structural analysis of DES complexes with DNA and RNA and their constituents has major biological importance. Our recent structural characterization of DES⅐DNA complexes showed that DES is a weak intercalator with affinity toward Ade-Thy-rich region (12). However, at high drug concentration, a partial helix opening occurs with major reduction of B-DNA structure in favor of A-DNA (12). The present study is also designed to investigate the interaction between RNA and DES in vitro and to establish correlations between spectral changes and drug binding mode, binding constant sequence selectivity, RNA stability, and conformation, using FTIR difference spectroscopy. In recent years, we have used vibrational spectroscopy (infrared and Raman) for the structural characterization of several DNA⅐drug (12)(13)(14), DNA⅐carbohydrate (15), DNA⅐cation (16,17), and protein complexes (18). We believe FTIR spectroscopy can also be used here to obtain structural information regarding the nature of DES-RNA interaction and the effect of drug complexation on the biopolymer secondary structure in aqueous solution.
In this work, FTIR difference spectroscopy is applied to study the interaction between yeast RNA and DES in water-ethanol solution at physiological pH, with drug/RNA-phosphate molar ratios of 1/80 to 1/2. Evidence for the drug binding mode, sequence preference, RNA secondary structure, and helix stability is provided. Furthermore, comparisons were made with those of the corresponding DNA⅐DES complexes and the results are reported here.

MATERIALS AND METHODS
Yeast RNA sodium salt was purchased from Sigma and used as supplied. DES was from Sigma and recrystallized from ethanol/water mixture. Other chemicals were reagent grade and used without further purification.
Preparation of Stock Solution-Sodium-RNA was dissolved to 2% (w/w, 0.05 M phosphate) in 0.05 M NaCl and 1 mM sodium cacodylate solution (pH 7.30) at 5°C with occasional stirring. DES solution of 0.3-12 mM was also prepared in ethanol (DES is not soluble in water). Mixtures of drug and RNA were prepared by adding DES solution dropwise to RNA solution with constant stirring to give the desired drug/RNA molar ratios of 1/80, 1/40, 1/20, 1/10, 1/4, and 1/2 at a final RNA concentration of 1% w/w or 0.025 M RNA-phosphate. Solution pH was 7.30 -6.80, and the infrared spectra were recorded 4 h after initial mixing of drug and RNA solution. The infrared spectra of DES⅐RNA complexes with r Ͼ 1/2 could not be recorded in solution due to precipitation.
Infrared spectra were recorded on a BOMEM DA3-0.02 Fourier transform infrared spectrometer equipped with nitrogen-cooled HgC-dTe detector and a KBr beam splitter. The solution spectra were taken on AgBr windows with spectral resolution of 2-4 cm Ϫ1 and 100 -500 scans. Water subtraction was carried out as in our previous report (16). A good water subtraction was achieved by a flat base line around 2200 cm Ϫ1 , where the H 2 O combination mode is located (19). The difference spectra ((RNA solution ϩ DES solution) -(RNA solution)) were produced, using the RNA band at 913 cm Ϫ1 as internal reference. This band, due to ribose-phosphate vibration exhibits no spectral changes (intensity or shifting) upon drug complexation, and it is cancelled on spectral subtraction. The difference spectra contain several positive and negative derivative features in the region of 1800 -600 cm Ϫ1 , whose amplitudes are less than 20% of the original peaks, with an estimated error of Ϯ 5% (absorbance). Several positive features in the difference spectra of DES⅐RNA complexes are coming from DES vibrations and are not due to RNA vibrational mode (properly labeled).
The intensity ratios of several RNA in-plane vibrations related to Ade-Urd, Gua-Cyt, and the backbone PO 2 groups were measured as a function of drug concentration with error of Ϯ 5%. These intensity ratio variations were used to quantitatively measure the amounts of drug base and drug-PO 2 bindings. The detailed intensity ratio measurements and spectral manipulations are described in our previous publication (16).
The calculation of the binding constant was also carried out as reported for other drug complexes with DNA, RNA, and mononucleotides (20 -23). By assuming that there is only one type of interaction (intercalation) between drug and RNA molecule, Equations 1 and 2 can be established.
The relative intensities for the band at 1698 cm Ϫ1 (mainly Gua), at 1608 cm Ϫ1 (mainly Ade), and at 1244 cm Ϫ1 (backbone PO 2 ) were calculated for each drug concentration. The calculated intensities were used as a function of drug concentration to estimate the K(G) for guanine, K(A) for adenine, and K(PO 2 ) for the phosphate group. The overall association constant (K) estimated for the above equation was 8.5 ϫ 10 4 M Ϫ1 at DES concentration of 3.12 ϫ 10 Ϫ4 M.

RESULTS AND DISCUSSION
RNA⅐DES Complexes-At low drug concentration (r ϭ 1/80 and 1/40), DES (Structure 1) intercalates in the Gua-Cyt and Ade-Urd-rich regions. Evidence for this comes from minor intensity increase (10%) of the RNA in-plane vibrations at 1698 cm Ϫ1 (Gua, Urd) and 1654 cm Ϫ1 (Urd, Gua, Ade, Cyt) (24 -29). The increase in the intensity was associated also with the shift of the band at 1698 cm Ϫ1 toward a lower frequency at 1695 cm -1 , while the band at 1654 cm Ϫ1 exhibited no shifting upon DES complexation (Figs. 1 and 2). Similarly, the mainly adenine band at 1608 cm Ϫ1 (19) gained intensity (20%) on drug interaction (Fig. 2). The positive derivative features at 1690 -1689 cm Ϫ1 (Gua, Urd), 1656 -1650 cm Ϫ1 (Urd, Gua, Ade, Cyt), 1610 -1608 cm Ϫ1 (Ade), and 1250 -1245 cm Ϫ1 (PO 2 ) in the difference spectra of DES⅐RNA complexes are arising from the increase in the intensity of the RNA bands at 1698, 1654, 1608, and 1244 cm Ϫ1 (Figs. 1 and 2, r ϭ 1/80 and 1/40). The observed spectral changes are related to drug intercalation with both Gua-Cyt and Ade-Urd base pairs. However, at this stage, the drug intercalation with the Gua-Cyt bases is larger than that of the Ade-Urd-rich region. The calculated binding constant K ϳ 8.5 ϫ 10 4 M Ϫ1 obtained at a drug concentration of 3.12 ϫ 10 Ϫ4 M is indicative of DES being a weak intercalator. The distributions of drug around DNA bases were about 35% with the Gua-Cyt and 25% with Ade-Urd base pairs (Fig. 3). The RNA backbone PO 2 anti-symmetric band at 1244 cm Ϫ1 (24, 26) also gained intensity (10%) upon drug complexation (Fig. 2). This is indicative of some degree of DES-PO 2 interaction (external binding) in these DES⅐RNA complexes. However, the amount of drug-phosphate binding is less than 20% (Fig. 3). Similar increases in the intensity of DNA in-plane vibrations were observed in the Raman spectra of several porphyrins and their metal derivatives, intercalated in the Ade-Thy-or Gua-Cyt-rich regions of the synthetic and native DNAs (30,31). DES intercalation with DNA resulted also in major spectral changes (intensity increase and shifting) of several DNA in-plane vibrations, related to the Ade-Thy and Gua-Cyt base pairs (12). The possibility of the external binding with the Ade-Thy, Gua-Cyt, or PO 2 donor groups have also been examined for several DNA intercalators (31)(32)(33)(34).
The band at 1488 cm Ϫ1 related mainly to the cytosine vibrations (24) exhibited no major intensity variations or shifting upon DES interaction (Fig. 2). This is indicative of less perturbation of cytosine bases in DES⅐RNA complex formation.
A comparison with the infrared spectra of other strong DNA and RNA intercalators, such as ethidium bromide, acridine orange, and methylene blue (recorded in our laboratory), showed that DES is a weaker intercalator than these pigments. The calculated binding constants for DNA⅐pigment and RNA- pigment complexes are ranging from 10 5 M Ϫ1 to 10 6 M Ϫ1 (21)(22)(23)(35)(36)(37)(38). These pigments are intercalated in both Gua-Cyt, Ade-Thy, or Ade-Urd-rich regions with no major sequence preference and with minor drug-phosphate interaction (external binding), similar to RNA⅐DES complexes investigated here.
At r Ͼ 1/40, DES interaction causes a partial helix destabilization. Evidence for this comes from major increase (20%) in the intensity of RNA bands at 1698, 1654, 1608, and 1244 cm Ϫ1 (Fig. 2, r ϭ 1/20). A similar increase of intensity was observed for several DNA or RNA in-plane vibrations upon thermal denaturation and acid fixation (16,23). The partial helix opening increases the chance of drug binding (externally) to different RNA donor sites that are available on local helix melting. At r Ͼ 1/20, some decrease in the intensity (20 -30%) of several RNA bands at 1698, 1654, 1488, and 1244 cm Ϫ1 are observed as results of minor perturbations of ribose-phosphate backbone geometry, upon drug complexation (Fig. 2, r ϭ 1/10). Evidence for such structural changes comes from the shift of the infrared marker bands at 1698 (Gua, Urd) to 1689 cm Ϫ1 , 1244 (PO 2 stretch) to 1247 cm Ϫ1 , and 861 (sugar-phosphate) to 870 cm Ϫ1 . However, since the infrared indicators at 810 and 870 cm Ϫ1 related to ribose in C3Ј-endo/anti-conformation (25, 39) exhibited no major shifting, the RNA remains in the A-family structure in these DES⅐RNA complexes (Fig. 1). No further drug-RNA interaction occurred at r Ͼ 1/10, due to the lack of spectral changes (intensity or shifting) observed for RNA and drug vibrational frequencies (Fig. 2). No major differences are also observed between the difference spectra obtained for the DES⅐RNA complexes with r ϭ 1/10 and those of the corresponding drug-RNA complexes with r ϭ 1/4 and 1/2 ( Figs. 1 and 2).
Additional evidence for DES⅐RNA interaction comes also from the major spectral shiftings of several DES vibrational frequencies (12). The shifts of the bands at 1612 cm Ϫ1 (DES CAC stretches) to 1610 cm Ϫ1 ; 1462 (DES COO stretches) to 1460 cm Ϫ1 ; 1045 (DES COC stretches) to 1040 cm Ϫ1 , and 836 (DES ring vibrations) to 838 cm Ϫ1 are related to the DES⅐RNA interaction (Fig. 1). However, it should be noted that at low drug concentration, most of the DES infrared vibrations were masked by the strong RNA vibrations, while at high drug content, some of the RNA vibrations were overlapped by drug vibrational frequencies (Fig. 1). However, these overlappings did not interfere with conclusive interpretation of our infrared results regarding the RNA⅐DES complexes. The positive derivative features observed at 1610, 1513, 1460, 1372, 1250, 1038, 850, and 838 cm Ϫ1 in the difference spectra of the DES⅐RNA complexes are coming from DES vibrations, and they are not related to the RNA molecule ( Fig. 1; r ϭ 1/10).
Comparison Between RNA⅐DES and DNA⅐DES Complexes-Recent spectroscopic results on the DES⅐DNA complexes showed that at low drug concentration, DES is a weak intercalator with major affinity toward Ade-Thy-rich region, whereas at high drug content, DES binding extended to Gua-Cyt base pairs with no drug-PO 2 interaction (12). At high DES content, a major helix destabilization was also observed with a partial reduction of B-DNA structure in favor of A-DNA (12). However, on the basis of the FTIR spectroscopic results on RNA⅐DES complexes presented here for the first time, it is clearly evident that at low drug concentration, DES intercalates in both Ade-Urd-and Gua-Cyt-rich regions of RNA with some degree of drug-PO 2 interaction (external binding). At high drug content, a partial helix opening occurs, while RNA remains in the A-family structure. The reason is mainly due to less flexibility of RNA structure (mainly A-conformation) (40) with respect to the polymorphic nature of DNA conformations, which readily adapt B, A, C, D, and Z structures (41,42). The local helix melting provides additional binding sites for the DES-biopolymer interaction (groove binding) via Gua-Cyt and Ade-Urd bases donor groups. A partial helix destabilization occurred for DNA at high DES content (r Ͼ 1/10), while RNA local helix opening was observed at lower drug concentration (r Ͼ 1/40). This can also be due to a complete DNA double helicity with respect to RNA structure, which contains a partial non-helical structure, and to minor differences in some of the H-bonding arrangements (40,43).