An Ultraviolet Resonance Raman Spectroscopic Study of Cisplatin

Sep 19, 2017 - Ultraviolet resonance Raman (UVRR) spectroscopy is a label-free method to define biomacromolecular interactions with anticancer compoun...
0 downloads 16 Views 2MB Size
Article pubs.acs.org/JPCB

An Ultraviolet Resonance Raman Spectroscopic Study of Cisplatin and Transplatin Interactions with Genomic DNA Jiafeng Geng,†,‡ Mena Aioub,†,‡,§ Mostafa A. El-Sayed,†,‡,§ and Bridgette A. Barry*,†,‡ †

School of Chemistry and Biochemistry, ‡Parker H. Petit Institute of Bioengineering and Bioscience, and §Laser Dynamics Laboratory, Georgia Institute of Technology, Atlanta, Georgia 30332, United States S Supporting Information *

ABSTRACT: Ultraviolet resonance Raman (UVRR) spectroscopy is a label-free method to define biomacromolecular interactions with anticancer compounds. Using UVRR, we describe the binding interactions of two Pt(II) compounds, cisplatin (cis-diamminedichloroplatinum(II)) and its isomer, transplatin, with nucleotides and genomic DNA. Cisplatin binds to DNA and other cellular components and triggers apoptosis, whereas transplatin is clinically ineffective. Here, a 244 nm UVRR study shows that purine UVRR bands are altered in frequency and intensity when mononucleotides are treated with cisplatin. This result is consistent with previous suggestions that purine N7 provides the cisplatin-binding site. The addition of cisplatin to DNA also causes changes in the UVRR spectrum, consistent with binding of platinum to purine N7 and disruption of hydrogen-bonding interactions between base pairs. Equally important is that transplatin treatment of DNA generates similar UVRR spectral changes, when compared to cisplatin-treated samples. Kinetic analysis, performed by monitoring decreases of the 1492 cm−1 band, reveals biphasic kinetics and is consistent with a two-step binding mechanism for both platinum compounds. For cisplatin−DNA, the rate constants (6.8 × 10−5 and 6.5 × 10−6 s−1) are assigned to the formation of monofunctional adducts and to bifunctional, intrastrand cross-linking, respectively. In transplatin−DNA, there is a 3.4-fold decrease in the rate constant of the slow phase, compared with the cisplatin samples. This change is attributed to generation of interstrand, rather than intrastrand, adducts. This longer reaction time may result in increased competition in the cellular environment and account, at least in part, for the lower pharmacological efficacy of transplatin.



bloodstream.4 In vitro, it is known that this water ligand can be displaced by the N-heterocyclic nucleobases of DNA, with a preference for guanine and adenine at the N7 position (Figure 1b and c).5 After the initial binding, the remaining chloride ligand can be displaced by an additional nucleobase.6 The most common cisplatin modification of DNA has been described as a 1,2-intrastrand cross-link of two adjacent purine bases (Figure 1d). The formation of such adducts distorts the structure of DNA, and this interaction has been postulated to activate DNA repair mechanisms, which lead to apoptosis.3 Other cellular components, such as RNA and protein, have also been identified as binding targets of cisplatin.7−9 Interestingly, the cisplatin stereoisomer, transplatin (transdiamminedichloroplatinum(II), Figure 1a) is clinically ineffective.10 One reason for the inactivity of transplatin may be the types of DNA adducts formed by this stereoisomer. Whereas the main adducts of cisplatin and DNA are intrastrand crosslinks, the majority of transplatin−DNA adducts identified thus

INTRODUCTION Although there have been many important developments in cancer treatment leading to increased survival rates, there remains a need to develop new therapeutic approaches, particularly for drug-resistant malignancies. A critical barrier in the development of new therapeutics is often an incomplete understanding of drug−biomacromolecule interactions. The development of a rapid and noninvasive method to characterize these interactions will provide rational design strategies for new therapeutic compounds. Here, we report the use of ultraviolet resonance Raman (UVRR) spectroscopy as a sensitive, labelfree, and specific probe of drug−biomacromolecule interactions. Cisplatin (cis-diamminedichloroplatinum(II), i.e., Platinol, shown in Figure 1a) is a DNA-targeting anticancer therapeutic, which received FDA approval in 1978.1 Today, cisplatin remains widely used in chemotherapy, especially in treatment for the early stage of testicular cancer,2,3 along with lymphoma and small cell lung, bladder, ovarian, and cervical cancers. Following the uptake of cisplatin by cancer cells, one of the chloride ligands of cisplatin is readily replaced by water due to the low chloride concentration in the cytoplasm relative to the © XXXX American Chemical Society

Received: August 15, 2017 Revised: August 30, 2017

A

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 1. Pt(II) drugs and their interactions with DNA. (a) Molecular structures of cisplatin and transplatin. (b) Watson−Crick base pairs. The hydrogen-bonding interactions between the nucleobases are shown as dotted lines. The arrows indicate the binding sites of Pt(II) drugs. (c) Adducts of cisplatin with A and G. (d) Illustration of the 1,2-intrastrand cross-link adduct of cisplatin and DNA.

Figure 2. Schematic of the UVRR setup. (a) Optical path of the Raman instrument (L, lens; M, mirror). The incident laser beam is shown in black and purple lines. The scattered light is shown in blue lines. (b) Diagram of the flow sample cell. A peristaltic pump was used to circulate the sample at ca. 0.8 m/s to avoid photodamage potentially caused by the UV probe beam.

far are either monofunctional or interstrand.11,12 A second reason for the inactivity of transplatin may be increased chemical reactivity, which renders the compound susceptible to deactivation in the cell.13,14 In previous work, visible Raman spectroscopy and nearinfrared Raman spectroscopy have been employed to characterize drug−biomacromolecule interactions, including the reaction between cisplatin and DNA.15−17 Compared to visible and near-infrared Raman spectroscopy, UVRR spectroscopy generates less biofluorescence and produces reliable spectral baselines.18 Under resonance Raman conditions, where the energy of the laser probe matches a specific electronic transition of the target molecule, Raman scattering associated with this chromophore is enhanced by as much as 106 compared to nonresonance Raman scattering (see refs 18−21). In applications to mechanistic studies of drug−DNA interactions, UVRR spectroscopy offers a significant advantage due to the fact that the molecular target, i.e., DNA, has intense optical absorption in the UV region. As a result, the Raman scattering from nucleobases is selectively enhanced, providing specific information on drug-induced structural changes of DNA. In this report, we describe the interactions of two Pt(II) compounds, cisplatin and transplatin, with DNA using UVRR spectroscopy. Spectral assignments are performed by comparison to a detailed set of model compounds. Notably, the use of a flow cell22−25 minimizes UV-induced photodamage in the samples and improves spectral fidelity and reproducibility. After incubation with the Pt(II) compounds, the binding of platinum to purine bases induces significant spectral changes. Our results suggest that platinum binding also disrupts the hydrogenbonding interactions between base pairs, altering the structure of DNA. Reaction mechanisms for cisplatin and transplatin are proposed on the basis of time-dependent UVRR measurements. These UVRR experiments provide new information

regarding the interactions of cisplatin and transplatin with genomic DNA. This approach offers an effective method to study the mechanism of DNA−drug interactions that can be used to validate or improve clinical drug candidates.



MATERIALS AND METHODS Materials. Dulbecco’s phosphate buffered saline (PBS, pH 7.4) was purchased from VWR (Radnor, PA). Cisplatin and transplatin were purchased from Alfa Aesar (Lancashire, U.K.) and dissolved in normal saline (0.9% NaCl) as stock solutions. 2′-Deoxyadenosine 5′-monophosphate (dAMP), 2′-deoxycytidine 5′-monophosphate (dCMP) sodium salt, and 2′deoxyguanosine 5′-monophosphate (dGMP) sodium salt hydrate were purchased from Sigma-Aldrich (St. Louis, MO). 2′-Deoxythymidine-5′-monophosphate (dTMP) disodium salt was purchased from Chem-Impex International (Wood Dale, IL). The nucleotides were dissolved in PBS. The calf-thymus DNA (CT-DNA, GC content ∼42% according to the supplier) was purchased from Sigma-Aldrich and dissolved in PBS. For spectroscopic measurements, samples were diluted to specified concentrations (see figure legends) in PBS. UVRR Spectroscopy. UVRR spectra were recorded at room temperature using methods described previously.22−26 The 244 nm probe beam was generated from an intracavity frequency-doubled Lexel 95 argon ion laser (Cambridge Lasers Laboratories, Inc., Fremont, CA). As shown in Figure 2a, the probe beam was coupled to a Renishaw inVia Raman microscope system equipped with a UV-coated, deep-depletion charge-coupled-device (CCD) detector. Backscattering from the samples was collected by a 15× UV objective with a numerical aperture (NA) value of 0.32 (Thorlabs, Newton, NJ), assembled in a Leica Microsystems microscope. The interval between adjacent data points was 3.8 cm−1. Spectral calibration was performed using the Raman band of diamond B

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B (1332 cm−1) as the reference standard. The full-width at halfheight of the diamond Raman band was 9 cm−1. The UVRR spectrum of a tyrosine solution (10 mM, pH 11) was acquired before each set of UVRR experiments and was also used to calibrate the Raman shift. To prevent UV-induced photodamage, the UVRR samples were circulated in a homemade flow cell composed of a peristaltic pump and a nozzle (ca. 120 μm inner diameter), as illustrated in Figure 2b. The flow cell generated a liquid jet of the sample solution, on which the 244 nm probe beam was focused using the microscope objective. Incubation of the nucleotides/CT-DNA with cisplatin/transplatin was performed at room temperature for the specified time and concentration (see the figure legends for details). The UVRR spectra were acquired with 3 mW laser power and 4−10 min data collection times. Each UVRR spectrum was the sum of two individual measurements. Experiments were replicated on two different sets of samples to establish the reproducibility of the kinetic parameters. The UVRR data were analyzed using OriginPro 2016 software (OriginLab Corporation, Northampton, MA). Baselines were fit with a cubic function and corrected using PeakFit software (Systat Software, San Jose, CA). A minimal background contribution from PBS alone was also measured and subtracted from each Raman spectrum. For kinetic analysis, data were normalized at 1700 cm−1, a region that was not significantly altered by cisplatin and transplatin treatment. Optical Absorption Spectroscopy. Optical absorption spectra were recorded on a Shimadzu UV-1700 spectrometer (Columbia, MD) at room temperature. The slit width and spectral resolution were 1 nm.

S1 shows the optical absorption spectra of the nucleotides prior to and after incubation with cisplatin. Incubation with cisplatin results in a decrease of the extinction coefficients of dGMP and dAMP at 244 nm (Figure S1a and b). In contrast, the optical absorption spectra of dCMP and dTMP are not significantly perturbed (Figure S1c and d). The result is consistent with the conclusion that cisplatin coordinates purines but not pyrimidines. Figure 4 shows the UVRR spectra of the nucleotides prior to and after incubation with cisplatin. Incubation with cisplatin

Figure 4. UVRR spectra of dNMPs (1 mM) before (black traces) and after (red traces) incubation with cisplatin (1 mM): (a) dGMP; (b) dAMP; (c) dCMP; (d) dTMP. The incubation time with cisplatin was 356 h. 244 nm laser probe power, 3 mW; data collection time, 10 min.



causes a decrease in the intensity of the UVRR signals of dGMP and dAMP (Figure 4a and b). In contrast, the UVRR spectra of dCMP and dTMP remain unaltered (Figure 4c and d). In the spectrum of dGMP, the most pronounced impact of cisplatin is an intensity decrease and frequency shift of the Raman bands at 1577, 1488, and 1329 cm−1 to 1598, 1503, and 1339/1360 cm−1, respectively. All three bands are assignable to aromatic ring modes involving the N7 atom (Table S1). These observations are consistent with previous Raman results in the literature.32,37 In the spectrum of dAMP, the major impact of cisplatin is an intensity decrease of the Raman bands at 1484 and 1341 cm−1. Again, the bands arise from aromatic ring modes involving the N7 atom (Table S1). UVRR Spectra of a dNMP Mixture and CT-DNA. Figure 5 shows the UVRR spectra of the calf-thymus DNA (CT-DNA) and a dNMP mixture. The dNMP mixture contains the same deoxynucleotide composition and concentration as the CTDNA sample. Table S1 shows UVRR band frequencies for the CT-DNA sample and normal mode assignments based on previous literature results.28,31,38 As expected from the results presented above, the 244 nm UVRR spectrum of the CT-DNA is dominated by Raman signals from dGMP and dAMP. When comparing the UVRR spectrum of the dNMP mixture (Figure 5, black trace) with that of the CT-DNA (Figure 5, blue trace), it is evident that all Raman bands display a lower signal intensity in duplex DNA. The observed UVRR spectral difference between the dNMP mixture and CT-DNA is due to the hypochromic effect,39 as demonstrated in the optical absorption spectra shown in Figure S2. In addition, these two samples exhibit bands with distinct frequencies. For example, the dNMP mixture exhibits two Raman bands at 1603 and 1653 cm−1. However, the former is not detected in the UVRR spectrum of the CT-DNA, and the latter shifts to 1646 cm−1 in

RESULTS UV Absorption and UVRR Spectra of Nucleotides. Figure 3 shows the optical absorption (Figure 3a) and UVRR

Figure 3. UV absorption (a) and UVRR (b) spectra of the nucleotides. The nucleotide concentrations used in the absorption and UVRR experiments were 100 μM and 5 mM, respectively. UVRR measurement condition: laser power at 244 nm, 3 mW; data collection time, 4 min. The dotted line in part a indicates the wavelength of the UV probe beam used in the UVRR experiments.

(Figure 3b) spectra of the individual monophosphate nucleotides: dGMP, dAMP, dCMP, and dTMP. Compared to the pyrimidine nucleotides, dGMP and dAMP display higher absorbance at 244 nm, the wavelength of the UVRR laser probe (Figure 3a). As a result, these two nucleotides exhibit more intense UVRR signals compared to those of the other two nucleotides (Figure 3b). Table S1 summarizes the major Raman signals of the nucleotides as well as their spectral assignments, based on the literature.27−36 UV Absorption and UVRR Characterization of the Interactions between Cisplatin and Nucleotides. Figure C

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Figure 5. UVRR spectra of the dNMP mixture (black trace) and calfthymus DNA (CT-DNA, blue trace). The GC content of the CTDNA was 42%. The DNA concentration was 1 mg/mL, corresponding to ca. 1.5 mM base pairs. The dNMP mixture was chosen to mimic the composition and concentration of the CT-DNA sample. 244 nm laser probe power, 3 mW; data collection time, 10 min.

the DNA spectrum. As indicated in Table S1, both Raman bands are derived from normal modes that are influenced by hydrogen-bonding interactions of the Watson−Crick base pairs. To assess whether the laser probe employed in the UVRR experiments can cause photodamage to DNA, the UVRR spectrum of the flowing CT-DNA sample was compared to one recorded from a nonflowing sample (Figure S3). The most prominent spectral feature indicative of UV-induced photodamage in the stationary sample is an intense, broad Raman band around 1600 cm−1. Such a Raman signal has been previously reported as a photodamage marker in UVRR studies on DNA and cells40−42 and can be assigned to oxidized derivatives of guanine.31,43−45 In contrast, the Raman bands of the CT-DNA recorded using the jet-flow system are sharp, and the broad band at 1600 cm−1 is not observed. Therefore, we conclude that flowing the samples minimizes UV-induced photodamage of DNA. This conclusion agrees with previous UVRR studies on other biomolecules.22−26,46,47 UVRR spectroscopy is sensitive to drug−DNA modifications, which involve the nucleobases.48−52 Figure 6a shows that incubation with cisplatin causes a pronounced decrease in the intensity of the UVRR signals over time in the dNMP mixture. This spectral change is attributed to the binding of cisplatin to dGMP and dAMP (Figure 4 and Figure S1). Figure 6b presents the time course of the UVRR signals of the cisplatin-treated CT-DNA. As shown, incubation with cisplatin leads to a noticeable decrease in the intensity of the majority of the UVRR bands, although the change is not as pronounced as that observed in the dNMP mixture. To highlight the spectral changes induced by cisplatin, UVRR difference spectra were generated and are presented in Figure 6c. The difference spectrum in black is derived from the dNMP mixture (Δ(dNMPs)cis); the difference spectrum in blue is derived from the CT-DNA (Δ(CT-DNA)cis). Both difference spectra are subtractions of the last incubation time point from the data obtained in the absence of cisplatin. A comparison of these two difference spectra confirms that the spectral change in the CT-DNA sample is less significant than that in the dNMP mixture. It should be noted that, in the CT-DNA sample, a band shift is detected at ∼1600 cm−1 after incubation with cisplatin, generating a differential band in the corresponding difference spectrum, Δ(CT-DNA)cis. This differential feature has a negative band at 1593 cm−1 and a positive band at 1627 cm−1 and is not present in the difference spectrum of the dNMP mixture, Δ(dNMPs)cis. We propose that this band

Figure 6. Time-dependent UVRR spectral changes, induced by incubation with cisplatin. (a) dNMP mixture incubated with cisplatin (1 mM). The dNMP mixture contained the same deoxynucleotide composition as that of the CT-DNA (i.e., 42% GC and 58% AT). The mixture was composed of 0.42 mM dGMP, 0.42 mM dCMP, 0.58 mM dAMP, and 0.58 mM dTMP, corresponding to 1 mM base pairs. The incubation periods for the UVRR spectra (from purple to red) were 0, 5, 11, 27, 50, 70, 140, 238, and 356 h, respectively. The arrow indicates the overall trend of the spectral change over time. 244 nm laser probe power, 3 mW; data collection time, 10 min. (b) CT-DNA (1 mg/mL, ca. 1.5 mM base pairs) incubated with cisplatin (1.5 mM). The incubation periods for the UVRR spectra (from purple to red) were 0, 2, 5, 11, 25, 49, 75, 147, and 243 h, respectively. The arrow indicates the overall trend of the spectral change over time. The UVRR measurement conditions and data processing methods are the same as those described in part a. (c) UVRR difference spectra, associated with cisplatin treatment. The black trace is the difference spectrum derived from the dNMP mixture, i.e., Δ(dNMPs)cis. It was obtained by subtracting the UVRR spectrum collected after incubation with cisplatin (356 h, the red trace in part a) from that collected before incubation (0 h, the purple trace in part a). The blue trace is the difference spectrum derived from CT-DNA, i.e., Δ(CT-DNA)cis. It was obtained by subtracting the UVRR spectrum collected after incubation with cisplatin (243 h, the red trace in part b) from that collected before the incubation (0 h, the purple trace in part b). The traces in part c were scaled to account for the small difference in base pair concentration.

shift is a marker for disruption of base pair interactions in duplex DNA. We conclude that cisplatin treatment alters hydrogen bonding interactions in the CT-DNA. In cisplatinmodified regions of duplex DNA, the Watson−Crick base pairs and associated molecular interactions are disrupted, rendering interactions in the biomacromolecule more like interactions in the free nucleotides. Optical absorption spectroscopy was also used to study the binding reactions of cisplatin. As shown in Figure S4a, incubation of the dNMP mixture with cisplatin causes a redshift (from 260 to 265 nm) and an intensity decrease of the optical absorption band. Incubation of the CT-DNA with cisplatin also causes a red-shift (from 258 to 263 nm) and an intensity decrease of the optical absorption band (Figure S4b). Figure S4c shows the difference spectra of the dNMP mixture and CT-DNA in the UV region. Both data sets are generated by D

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

data, the fast phase (k1) is assigned to the formation of monofunctional cisplatin−DNA adducts. The slow step (k2) is proposed to be the transition from the monofunctional adducts to bidentate adducts, such as the 1,2-intrastrand cross-links (Figure 7b). As shown in Figure S5b, kinetic analysis of the UVRR bands at 1340 and 1578 cm−1 also yields biphasic kinetics with similar rate constants, when compared to those obtained from Figure 7b. At 1340 cm−1, the rate constants were 7.4 × 10−5 (13%) and 5.9 × 10−6 (15%), while, at 1578 cm−1, the rate constants were 6.5 × 10−5 (22%) and 4.2 × 10−6 (17%). Interactions between Transplatin and the CT-DNA. The interactions between transplatin and the CT-DNA were also characterized using UVRR spectroscopy. As shown in Figure 8a, incubation of the CT-DNA with transplatin leads to

subtracting the spectrum obtained after cisplatin incubation from the original spectrum. As evident, a decrease in extinction coefficient at 244 nm is less pronounced in the CT-DNA sample compared to the dNMP mixture. This difference is attributed to disruption of the Watson−Crick base pairs by cisplatin, which attenuates the hypochromic effect in the DNA sample. Time-Dependent Interactions between Cisplatin and CT-DNA, as Assessed from UVRR Spectroscopy. As shown in Figure 7a, the time-dependent change in intensity of the

Figure 8. UVRR spectra associated with transplatin treatment of CTDNA. (a) Time-resolved UVRR spectra of CT-DNA (1 mg/mL, ca. 1.5 mM base pairs) incubated with transplatin (1.5 mM). The incubation periods for the UVRR spectra (from purple to red) were 0, 2, 5, 11, 25, 49, 75, 147, and 243 h, respectively. The arrow indicates the overall trend of the spectral change over time. 244 nm laser probe power, 3 mW; data collection time: 10 min. (b) Comparison of the spectral changes induced by cisplatin and transplatin treatment of DNA. The solid trace is the difference spectrum obtained by subtracting the UVRR spectrum collected after the incubation with transplatin (243 h, the red trace in part a) from that collected before the incubation (0 h, the purple trace in part a). The dotted trace is the corresponding difference spectrum obtained from the experiments with cisplatin, i.e., repeated from the blue trace in Figure 6c.

Figure 7. Cisplatin-induced decrease of the 1485/1492 cm−1 band in the UVRR spectra of the dNMP mixture (a) and CT-DNA (b). The error bars represent the range, calculated from two replicate experiments. The solid traces are fits of the data to a singleexponential or double-exponential function. The fitting parameters are summarized in Table S2. Proposed mechanisms are shown.

UVRR signal from the dNMP mixture at 1485 cm−1 can be fit by a single-exponential equation. The rate constant was determined to be 1.5 × 10−6 s−1 (Table S2). Kinetic analysis was also performed by following the intensity changes of the UVRR bands at 1340 and 1579 cm−1, and similar rate constants were obtained (Figure S5a). The mechanism proposed in Figure 7a is consistent with these data and proposes the formation of two types of Pt(II) adducts, containing either one or two nucleotide ligands. In the proposed mechanism, the two binding reactions are assigned with the same rate constant and are thus indistinguishable (Figure 7a). The time-dependent change of the UVRR signal from the CT-DNA at 1492 cm−1 can be satisfactorily fit by a doubleexponential equation but not by a single-exponential equation (Figure 7b), suggesting that the spectral change is derived from a two-step mechanism. The rate constants are determined to be 6.8 × 10−5 and 6.5 × 10−6 s−1 (Table S2). The fast phase corresponds to 27% of the total amplitude; the slow phase corresponds to 34% of the total amplitude. To explain these

a decrease in the signal intensity of many UVRR bands. Although smaller in magnitude compared to the cisplatininduced perturbation, the transplatin-induced spectral changes resemble those induced by cisplatin (Figure 8b). This suggests that transplatin coordinates the purine residues of the CT-DNA in solution and alters its double-helix structure. Similar to the cisplatin results, the transplatin-induced change of the UVRR signal at 1492 cm−1 can be fit by a double-exponential equation but not a single-exponential equation (Figure 9), indicating that the spectral change corresponds to a two-step mechanism. As shown in Table S2, the rate constants derived from the doubleexponential fitting are 5.7 × 10−5 s−1 (12% of total amplitude) and 1.9 × 10−6 s−1 (40% of total amplitude). While the rate constant for the fast phase (k1) is unaltered when the data of cisplatin and transplatin are compared, the percentage of the total amplitude attributed to this fast kinetic phase decreases in the transplatin case. In addition, the rate constant of the slow E

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

hypochromic effect that results from the interactions between the nucleobases. When the UVRR spectra of the dNMP mixture and CT-DNA are compared, the Raman bands, involving hydrogen-bonding interactions between the Watson−Crick base pairs, exhibit the most pronounced intensity changes and frequency shifts. As examples, bands at 1603 and 1653 cm−1 were present in the UVRR spectrum of the dNMP mixture but were absent (1603 cm−1) or shifted (1653 cm−1) in the UVRR spectrum of the CT-DNA (Figure 5). These two Raman bands involve hydrogen-bonding interactions of all four nucleotides. The band at 1603 cm−1 is assigned to dGMP (N1H b; C2N s) and dAMP (NH2 b; C5C6, C6N s); the band at 1653 cm−1 is assigned to dCMP (C2O, C2N3 s) and dTMP (C4O, C4C5 s) (Table S1). Cisplatin-Induced Changes in the UVRR Spectra of Nucleotides. The spectroscopic results presented here are consistent with the coordination of cisplatin to purines but not pyrimidines. Binding of cisplatin to dGMP and dAMP alters the electronic structures of the nucleobases. The resulting red-shift of the optical absorption bands of these two nucleotides causes a decrease of their UV absorption at 244 nm (Figure S1). As expected, we show that binding of cisplatin leads to an intensity loss as well as frequency shift of a number of Raman bands in the UVRR spectra of dGMP and dAMP (Figure 4). It should be noted that the Raman bands involving the N7 atom of these two nucleotides, i.e., 1598, 1503, and 1360 cm−1 for dGMP and 1484 and 1341 cm−1 for dAMP, display the most pronounced spectral changes, consistent with previous reports that the N7 atom of the purines is the binding site for cisplatin.7 Cisplatin-Induced Changes in the UVRR Spectrum of the CT-DNA. From previous model studies on duplex oligonucleotides, it was proposed that binding of cisplatin causes unwinding of the helical structure and disruption of the hydrogen-bonding interactions between Watson−Crick base pairs.53 Our UVRR results are consistent with this hypothesis and support the assignment of cisplatin-induced changes in the CT-DNA to two effects. First, direct binding of cisplatin to the purine residues decreases the intensity of Raman bands associated with the N7 atom of the purine residues. This change is observed both in the dNMP mixture sample and in the CT-DNA sample. Second, a disruption of Watson−Crick base pairs and a loss of hydrogen-bonding interactions occur. This change is specific to the DNA sample. In our experiments, the cisplatin-induced structural alteration of duplex DNA shifts the frequency of some Raman bands and attenuates the hypochromic effect (Figures 6 and S4). When evaluated either by UV absorption or UVRR signal, the intensity change due to cisplatin addition is larger in the dNMP mixture, compared to duplex DNA. Kinetic Analysis of Cisplatin-Induced Spectral Changes. Kinetic UVRR experiments provide additional information about the binding of cisplatin to the CT-DNA. Notably, the kinetic profile of the cisplatin-induced spectral change, as derived from CT-DNA, is different from that of the dNMP mixture. We model the former as a two-step mechanism, whereas the latter is modeled as a single-step mechanism. The two-step mechanism is proposed to involve the formation of monofunctional cisplatin adducts and the subsequent generation of bifunctional intrastand adducts (Figure 7b). The formation of the monofunctional adducts is relatively rapid, with a rate constant of 6.8 × 10−5 s−1, and is comparable to the previously reported values for the first hydrolysis step of cisplatin.7,54 In comparison, the binding of

Figure 9. Transplatin-induced decrease of the 1492 cm−1 band in the UVRR spectrum of CT-DNA. The error bars represent the range, calculated from two replicate experiments. The solid traces are fits to a single-exponential or double-exponential function. The fitting parameters are summarized in Table S2. The dotted line is the biexponential fit to the cisplatin data (Figure 7b), repeated for comparative purposes. A proposed mechanism is shown.

phase (k2) in the transplatin-treated CT-DNA decreases significantly, relative to the cisplatin case. The amplitude of the nondecaying component also increases from 39 to 48% in the transplatin-treated CT-DNA. The fast step in the mechanism is attributed to the formation of monofunctional DNA adducts. The rate constant of this step is similar to the k1 value of the cisplatin experiment. The slow step is attributed to the formation of bifunctional interstrand adducts, which are associated with a smaller rate constant in the transplatin case, when compared to the formation of the intrastrand adducts upon cisplatin treatment (Figure 9).



DISCUSSION Therapeutic drug design is most efficient when the basis of drug−biomacromolecule interactions is known. Platinum-based drugs, such as cisplatin, are widely used in anticancer therapies, but their mechanisms of action are still not completely understood. Compared with other Raman techniques, UVRR spectroscopy has the advantage that the spectral contributions of the nucleobases in DNA are selectively enhanced, allowing for clear spectral assignments without the need for additional labeling steps. In this work, UVRR spectroscopy is employed as a noninvasive, high-resolution method to study the interactions of platinum-based drugs with nucleobases and intact genomic DNA. Although the use of the 244 nm UV probe can potentially cause photodamage in the DNA samples, the flow cell22−26 employed in this work effectively prevents this complication (Figure S3). Here, changes in the intensity of UVRR signals are used to obtain new information about the mechanism of the Pt(II) compounds. UVRR Spectra of Individual Nucleotides and Genomic DNA. UVRR spectra of the nucleotides and CT-DNA are acquired in this study. Assignments of their Raman bands are summarized in Table S1, as derived from previous results.27,29,31,32,37 The UVRR spectra of the four nucleotides are distinguishable (Figure 3). Notably, the UVRR bands of dGMP and dAMP show the strongest resonance enhancement with the 244 nm excitation. Compared to the spectrum of the dNMP mixture with the same nucleobase composition and concentration, the UVRR spectrum of CT-DNA exhibits lower signal intensity (Figure 5). This decrease in signal intensity is due to a loss of UV absorption at 244 nm caused by the F

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

pronounced. Kinetic analysis is consistent with a slower formation of cross-linked bifunctional adducts in transplatintreated DNA, when compared to the cisplatin case.

cisplatin to the dNMP mixture is a single-step mechanism with a rate constant of 1.5 × 10−6 s−1 (Figure 7a), much slower than that determined with the CT-DNA samples. Binding of cisplatin to the CT-DNA may be facilitated by the negative charge on the phosphate groups, which can attract the positively charged hydrolysis product of cisplatin, i.e., [Pt(NH3)2Cl(H2O)]+.7,54 Transplatin-Induced Changes in the UVRR Spectrum of the CT-DNA. The influence of transplatin binding on the structure of DNA remains relatively unexplored compared to the cisplatin effects. Currently, there is no crystal structure of a transplatin−DNA adduct. An NMR solution structure of a transplatin−oligonucleotide adduct with a unique intrastrand G−C cross-link has been reported.55 However, the structure of this adduct is uncommon, and its physiological relevance remains to be established. In this work, the interaction between transplatin and CTDNA is characterized using UVRR spectroscopy. The transplatin-induced spectral change in the UVRR spectrum of CTDNA is similar, although less intense, when compared to the change induced by cisplatin. It should be noted that the major adducts of transplatin with DNA have been described as monofunctional and/or interstrand-cross-linked. This is distinct from the proposed cisplatin−DNA adducts, which are mainly intrastrand and involve neighboring purine residues.11,12 Despite the difference in the binding modes of these two Pt(II) compounds, we show for the first time that the binding of transplatin to intact genomic DNA generates similar hydrogen-bonding changes in the DNA structure, when compared to cisplatin (Figure 8b). The kinetic profile derived for transplatin−DNA is composed of two phases (Figure 9). The rate constant of the fast phase is similar to that derived for cisplatin treatment, suggesting that the formation rate of the monofunctional transplatin−DNA adducts is comparable to that of the monofunctional cisplatin−DNA adducts. The second phase, however, displayed a significantly slower rate constant when compared to the value derived from the cisplatin-treated samples. This temporal difference is likely due to the fact that the formation of interstrand cross-links requires a more significant rearrangement of the DNA structure. Importantly, the lower pharmacological efficacy of transplatin, compared to cisplatin, may be attributable to this change in the kinetics of bifunctional DNA adduct generation. In the cellular environment, this slower rate may permit competition from side reactions as well as give an increased probability of DNA repair due to the longer time scale involved.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b08156. Table S1 (normal mode assignments), Table S2 (kinetic constants), Figures S1, S2, and S5 (UV absorption spectra), Figure S3 (UVRR spectra of a flowing versus stationary sample), and Figure S4 (kinetic analysis of UVRR bands) (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Mena Aioub: 0000-0001-9799-0634 Mostafa A. El-Sayed: 0000-0002-7674-8424 Bridgette A. Barry: 0000-0003-3421-1407 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors are grateful for financial support from an American Heart Association Postdoctoral Fellowship (16POST30990053, J.G.) and NSF-DMR (1206637, M.A.E.).



REFERENCES

(1) Dasari, S.; Tchounwou, P. B. Cisplatin in cancer therapy: molecular mechanisms of action. Eur. J. Pharmacol. 2014, 740, 364− 78. (2) Jamieson, E. R.; Lippard, S. J. Structure, recognition, and processing of cisplatin-DNA adducts. Chem. Rev. 1999, 99, 2467−98. (3) Wang, D.; Lippard, S. J. Cellular processing of platinum anticancer drugs. Nat. Rev. Drug Discovery 2005, 4, 307−20. (4) Reedijk, J. New clues for platinum antitumor chemistry: kinetically controlled metal binding to DNA. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 3611−6. (5) Takahara, P. M.; Rosenzweig, A. C.; Frederick, C. A.; Lippard, S. J. Crystal structure of double-stranded DNA containing the major adduct of the anticancer drug cisplatin. Nature 1995, 377, 649−52. (6) Poklar, N.; Pilch, D. S.; Lippard, S. J.; Redding, E. A.; Dunham, S. U.; Breslauer, K. J. Influence of cisplatin intrastrand crosslinking on the conformation, thermal stability, and energetics of a 20-mer DNA duplex. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 7606−11. (7) Johnstone, T. C.; Suntharalingam, K.; Lippard, S. J. The next generation of platinum drugs: targeted Pt(II) agents, nanoparticle delivery, and Pt(IV) prodrugs. Chem. Rev. 2016, 116, 3436−86. (8) Hostetter, A. A.; Osborn, M. F.; DeRose, V. J. RNA-Pt adducts following cisplatin treatment of Saccharomyces cerevisiae. ACS Chem. Biol. 2012, 7, 218−25. (9) Chapman, E. G.; DeRose, V. J. Enzymatic processing of platinated RNAs. J. Am. Chem. Soc. 2010, 132, 1946−52. (10) Aris, S. M.; Farrell, N. P. Towards antitumor active transplatinum compounds. Eur. J. Inorg. Chem. 2009, 2009, 1293−1302. (11) Eastman, A.; Jennerwein, M. M.; Nagel, D. L. Characterization of bifunctional adducts produced in DNA by transdiamminedichloroplatinum(II). Chem.-Biol. Interact. 1988, 67, 71−80. (12) Brabec, V.; Leng, M. DNA interstrand cross-links of transdiamminedichloroplatinum(II) are preferentially formed between guanine and complementary cytosine residues. Proc. Natl. Acad. Sci. U. S. A. 1993, 90, 5345−9.



CONCLUSION UVRR spectroscopy presents a method to develop a molecular understanding of drug interactions with biomacromolecular targets. We present high quality UVRR spectra of nucleotide and genomic DNA solutions using a flow cell, allowing selective enhancement of the Raman signals of the purine bases. Upon incubation with the Pt(II) compounds, significant spectral changes are observed, owing to the molecular interactions of the Pt(II) compounds with nucleobases. The binding of the platinum drugs also disrupts the hydrogen-bonding interactions of the Watson−Crick base pairs and alters the double-helix DNA structure, as evident in the UVRR difference spectra. Although both cisplatin and transplatin induce similar spectral changes in the CT-DNA, the influence of cisplatin over the incubation time used for the kinetic experiments is more G

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B (13) Kasparkova, J.; Marini, V.; Bursova, V.; Brabec, V. Biophysical studies on the stability of DNA intrastrand cross-links of transplatin. Biophys. J. 2008, 95, 4361−71. (14) Florea, A. M.; Busselberg, D. Cisplatin as an anti-tumor drug: cellular mechanisms of activity, drug resistance and induced side effects. Cancers 2011, 3, 1351−71. (15) Nawaz, H.; Bonnier, F.; Knief, P.; Howe, O.; Lyng, F. M.; Meade, A. D.; Byrne, H. J. Evaluation of the potential of Raman microspectroscopy for prediction of chemotherapeutic response to cisplatin in lung adenocarcinoma. Analyst 2010, 135, 3070−3076. (16) Vrana, O.; Masek, V.; Drazan, V.; Brabec, V. Raman spectroscopy of DNA modified by intrastrand cross-links of antitumor cisplatin. J. Struct. Biol. 2007, 159, 1−8. (17) Masetti, M.; Xie, H. N.; Krpetic, Z.; Recanatini, M.; AlvarezPuebla, R. A.; Guerrini, L. Revealing DNA interactions with exogenous agents by surface-enhanced Raman scattering. J. Am. Chem. Soc. 2015, 137, 469−476. (18) Asher, S. A. UV resonance Raman spectroscopy for analytical, physical, and biophysical chemistry. Part 2. Anal. Chem. 1993, 65, 201A−210A. (19) Schmitt, M.; Popp, J. Raman spectroscopy at the beginning of the twenty-first century. J. Raman Spectrosc. 2006, 37, 20−28. (20) Mukerji, I. Resonance Raman spectroscopy. In Encyclopedia of Life Sciences; John Wiley & Sons, Ltd: Hoboken, NJ, 2012; DOI: 10.1002/9780470015902.a0003113.pub2. (21) Mathies, R. Biological applications of resonance Raman spectroscopy in the visible and ultraviolet: visual pigments, purple membrane, and nucleic acids. In Chemical and Biochemical Applications of Lasers; Moore, C. B., Ed. Academic Press: New York, 1979; Vol. 4, p 55. (22) Chen, J.; Barry, B. A. Ultraviolet resonance Raman microprobe spectroscopy of photosystem II. Photochem. Photobiol. 2008, 84, 815− 818. (23) Chen, J.; Bender, S. L.; Keough, J. M.; Barry, B. A. Tryptophan as a probe of photosystem I electron transfer reactions: a UV resonance Raman study. J. Phys. Chem. B 2009, 113, 11367−11370. (24) Chen, J.; Yao, M. D.; Pagba, C. V.; Zheng, Y.; Fei, L. P.; Feng, Z. C.; Barry, B. A. Directly probing redox-linked quinones in photosystem II membrane fragments via UV resonance Raman scattering. Biochim. Biophys. Acta, Bioenerg. 2015, 1847, 558−564. (25) Pagba, C. V.; Barry, B. A. Redox-induced conformational switching in photosystem-II-inspired biomimetic peptides: a UV resonance Raman study. J. Phys. Chem. B 2012, 116, 10590−10599. (26) Pagba, C. V.; McCaslin, T. G.; Veglia, G.; Porcelli, F.; Yohannan, J.; Guo, Z. J.; McDaniel, M.; Barry, B. A. A tyrosinetryptophan dyad and radical-based charge transfer in a ribonucleotide reductase-inspired maquette. Nat. Commun. 2015, 6, 10010. (27) Wen, Z. Q.; Thomas, G. J. UV resonance Raman spectroscopy of DNA and protein constituents of viruses: assignments and cross sections for excitations at 257, 244, 238, and 229 nm. Biopolymers 1998, 45, 247−256. (28) Russell, M. P.; Vohnik, S.; Thomas, G. J. Design and performance of an ultraviolet resonance Raman spectrometer for proteins and nucleic acids. Biophys. J. 1995, 68, 1607−1612. (29) Toyama, A.; Hanada, N.; Ono, J.; Yoshimitsu, E.; Takeuchi, H. Assignments of guanosine UV resonance Raman bands on the basis of 13 C, 15N and 18O substitution effects. J. Raman Spectrosc. 1999, 30, 623−630. (30) Shanmugasundaram, M.; Puranik, M. Computational prediction of vibrational spectra of normal and modified DNA nucleobases. J. Raman Spectrosc. 2009, 40, 1726−1748. (31) D’Amico, F.; Cammisuli, F.; Addobbati, R.; Rizzardi, C.; Gessini, A.; Masciovecchio, C.; Rossi, B.; Pascolo, L. Oxidative damage in DNA bases revealed by UV resonant Raman spectroscopy. Analyst 2015, 140, 1477−1485. (32) Perno, J. R.; Park, Y. D.; Reedijk, J.; Spiro, T. G. UV resonance Raman study of platinum binding to guanine in mononucleotides and dinucleotides. J. Raman Spectrosc. 1988, 19, 203−212.

(33) Shanmugasundaram, M.; Puranik, M. Vibrational markers of structural distortion in adenine nucleobases upon DNA damage. Phys. Chem. Chem. Phys. 2011, 13, 3851−3862. (34) Perno, J. R.; Grygon, C. A.; Spiro, T. G. Ultraviolet Raman excitation profiles for the nucleotides and for the nucleic acid duplexes poly(rA)-poly(rU) and poly(dG-dC). J. Phys. Chem. 1989, 93, 5672− 5678. (35) Mathlouthi, M.; Seuvre, A. M.; Koenig, J. L. FT-IR and laserRaman spectra of constituents of nucleic acids. Part II. FT-IR and laser-Raman spectra of adenine and adenosine. Carbohydr. Res. 1984, 131, 1−15. (36) Nishimura, Y.; Tsuboi, M.; Kubasek, W. L.; Bajdor, K.; Peticolas, W. L. Ultraviolet resonance Raman bands of guanosine and adenosine residues useful for the determination of nucleic acid conformation. J. Raman Spectrosc. 1987, 18, 221−227. (37) Perno, J. R.; Cwikel, D.; Spiro, T. G. Ultraviolet resonance Raman study of deoxyguanosine monophosphate and its complexes with cis-(NH3)Pt2+, Ni2+, and H+. Inorg. Chem. 1987, 26, 400−405. (38) Zhao, X. J.; Vinson, M.; Malins, D. C.; Spiro, T. G. Characterization of DNA isolated from normal and cancerous ovarian tissues by ultraviolet resonance Raman spectroscopy. Proc. SPIE 2000, 3918, 146−152. (39) D’Abramo, M.; Castellazzi, C. L.; Orozco, M.; Amadei, A. On the nature of DNA hyperchromic effect. J. Phys. Chem. B 2013, 117, 8697−704. (40) Kumamoto, Y.; Taguchi, A.; Smith, N. I.; Kawata, S. Deep UV resonant Raman spectroscopy for photodamage characterization in cells. Biomed. Opt. Express 2011, 2, 927−936. (41) Yazdi, Y.; Ramanujam, N.; Lotan, R.; Mitchell, M. F.; Hittelman, W.; Richards-Kortum, R. Resonance Raman spectroscopy at 257 nm excitation of normal and malignant cultured breast and cervical cells. Appl. Spectrosc. 1999, 53, 82−85. (42) Kumamoto, Y.; Taguchi, A.; Smith, N. I.; Kawata, S. Deep ultraviolet resonant Raman imaging of a cell. J. Biomed. Opt. 2012, 17, 0760011. (43) Douki, T.; Perdiz, D.; Grof, P.; Kuluncsics, Z.; Moustacchi, E.; Cadet, J.; Sage, E. Oxidation of guanine in cellular DNA by solar UV radiation: biological role. Photochem. Photobiol. 1999, 70, 184−90. (44) Zhang, Y. Y.; Dood, J.; Beckstead, A. A.; Li, X. B.; Nguyen, K. V.; Burrows, C. J.; Improta, R.; Kohler, B. Efficient UV-induced charge separation and recombination in an 8-oxoguanine-containing dinucleotide. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 11612−11617. (45) Zhang, Y. Y.; Dood, J.; Beckstead, A. A.; Li, X. B.; Nguyen, K. V.; Burrows, C. J.; Improta, R.; Kohler, B. Photoinduced electron transfer in DNA: charge shift dynamics between 8-oxo-guanine anion and adenine. J. Phys. Chem. B 2015, 119, 7491−7502. (46) Shreve, A. P.; Cherepy, N. J.; Franzen, S.; Boxer, S. G.; Mathies, R. A. Rapid-flow resonance Raman spectroscopy of bacterial photosynthetic reaction centers. Proc. Natl. Acad. Sci. U. S. A. 1991, 88, 11207−11211. (47) Mathies, R.; Oseroff, A. R.; Stryer, L. Rapid-flow resonance Raman spectroscopy of photolabile molecules - rhodopsin and isorhodopsin. Proc. Natl. Acad. Sci. U. S. A. 1976, 73, 1−5. (48) Chan, S. S.; Austin, R. H.; Mukerji, I.; Spiro, T. G. Temperaturedependent ultraviolet resonance Raman spectroscopy of the premelting state of dA.dT DNA. Biophys. J. 1997, 72, 1512−20. (49) Mukerji, I.; Williams, A. P. UV resonance Raman and circular dichroism studies of a DNA duplex containing an A3T3 tract: evidence for a premelting transition and three-centered H-bonds. Biochemistry 2002, 41, 69−77. (50) Mukerji, I.; Shiber, M. C.; Fresco, J. R.; Spiro, T. G. A UV resonance Raman study of hairpin dimer helices of d(A-G)10 at neutral pH containing intercalated dA residues and alternating dG tetrads. Nucleic Acids Res. 1996, 24, 5013−5020. (51) Knee, K. M.; Dixit, S. B.; Aitken, C. E.; Ponomarev, S.; Beveridge, D. L.; Mukerji, I. Spectroscopic and molecular dynamics evidence for a sequential mechanism for the A-to-B transition in DNA. Biophys. J. 2008, 95, 257−272. H

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B (52) Sokolov, L.; Wojtuszewski, K.; Tsukroff, E.; Mukerji, I. Nucleic acid structure investigated by UV resonance Raman spectroscopy: protonation effects and a tract structure. J. Biomol. Struct. Dyn. 2000, 17, 327−334. (53) Gelasco, A.; Lippard, S. J. NMR solution structure of a DNA dodecamer duplex containing a cis-diammineplatinum(II) d(GpG) intrastrand cross-link, the major adduct of the anticancer drug cisplatin. Biochemistry 1998, 37, 9230−9. (54) Choudhury, J. R.; Rao, L.; Bierbach, U. Rates of intercalatordriven platination of DNA determined by a restriction enzyme cleavage inhibition assay. JBIC, J. Biol. Inorg. Chem. 2011, 16, 373−80. (55) Paquet, F.; Boudvillain, M.; Lancelot, G.; Leng, M. NMR solution structure of a DNA dodecamer containing a transplatin interstrand GN7-CN3 cross-link. Nucleic Acids Res. 1999, 27, 4261− 4268.

I

DOI: 10.1021/acs.jpcb.7b08156 J. Phys. Chem. B XXXX, XXX, XXX−XXX