Photoinduced Single Strand Breaks and Intrastrand Cross-Links in an

Apr 25, 2014 - 5-Bromouracil (BrU) is photoreactive toward near UVB photons and can be introduced into genomic DNA during its biosynthesis in cells...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/JPCB

Photoinduced Single Strand Breaks and Intrastrand Cross-Links in an Oligonucleotide Labeled with 5‑Bromouracil Magdalena Zdrowowicz, Barbara Michalska, Agnieszka Zylicz-Stachula, and Janusz Rak* Faculty of Chemistry, University of Gdańsk, Wita Stwosza 63, 80-308 Gdańsk, Poland S Supporting Information *

ABSTRACT: 5-Bromouracil (BrU) is photoreactive toward near UVB photons and can be introduced into genomic DNA during its biosynthesis in cells. However, PCR seems to be a simpler approach, which can be used to obtain labeled DNA similar to that synthesized within the cell. In the current work, PCR has been employed and optimized in order to substitute all thymines (besides those present in starters) with BrU in the dsDNA fragment of 80 base pairs (bp) in length. The modified oligonucleotide was irradiated with 300 nm photons in a buffered aqueous solution (pH = 7) and digested with a cocktail of enzymes specific to the phosphodiester bond cleavage. Initially, the extent of damage in the intact photolyte was measured with DHPLC. Then, the digested reaction mixture was subjected to HPLC and MS analyses and, in addition to the formation of 5-bromo-2′deoxuyridine, which proves the occurrence of single strand breaks (SSBs) due to irradiation, U∧U and U∧C dimers were found, whose molecular structure was confirmed by MS/MS analysis. Although the abundance of such tandem lesions is lower than that of the SSB type, they pose a potent threat to genome integrity. Thus, our findings shed new light on the photosensitizing properties of BrU toward DNA.

1. INTRODUCTION The abilities of 5-bromo-2′-deoxyuridine (BrdU) to photo-1,2 and radiosensitize3−5 DNA have been studied since the late 1950s. Despite the fact that in vitro BrdU-labeled cells are 2−3 times more radiosensitive than nonlabeled ones,6 a most thorough phase III clinical trial did not demonstrate increased survival in patients to whom BrdU was administered during the radiotherapy of various astrocytomas and malignant gliomas.7 Nevertheless, recent years have witnessed renewed interest in halogenated uridine analogues as radiosensitizing agents, especially for poorly responding cancers.8 It is worth mentioning that in cells pretreated with BrdU or 5-iodo-2′deoxyuridine both ionizing and UV−C photons enhance the activation of DNA damage signaling.9,10 Indeed, for UV−C enhanced apoptosis was observed, whereas ionizing radiation mainly induced rapid senescence in cells.10 This renewed interest in the DNA sensitizing properties of halouracils resulted in a series of studies of the molecular mechanism leading to the UV-induced damage of DNA labeled with BrdU.11−18 For instance, on the basis of denaturing gel electrophoresis assays Greenberg et al.18−20 and Sanche et al.21 suggested that single strand breaks (SSBs) form as a result of the UV irradiation of double-stranded 20−30 base pair sequences point-labeled with BrdU.19 On the other hand, using HPLC Sugiyama et al.22 demonstrated the occurrence of UV-induced 2-ribonolactone rather than SSB formation for shorter BrdU-labeled oligonucleotides.23 Similarly, our group very recently published a DHPLC/MS study that argues, in agreement with the reports of Sugiyama et al.,24 for a secondary pathway of SSB formation in UVB-irradiated BrdU-labeled © 2014 American Chemical Society

DNA. A reasonable explanation for these apparently inconsistent findings might be the fact that 2-ribonolactone is a thermally unstable species that easily converts into a secondary strand break.25 Hence, the ultimate conclusion related to the mechanism of photochemical damage to BrdU-labeled DNA depends on the analytical method employed for damage detection. A common transient species presumed in all mechanistic proposals of photodamage to BrU-labeled DNA is the reactive uracil-5-yl radical,22 which forms from the unstable BrU anion.26,27 The latter species is a primary product of the UVinduced electron transfer between the electronically excited BrU and a distant guanine, or else it comes from the decay of the excited state of BrU if the above-mentioned long-range electron transfer is forbidden because of a disfavored DNA sequence.22 The resulting uracil-5-yl radical may become stabilized by the transfer of a hydrogen atom from the adjacent 2′-deoxyribose moiety, which produces 2′-deoxyuridine; this can be observed in HPLC coupled to the enzymatic digestion of UV-damaged DNA. Another pathway leading to the stabilization of the uracil-5-yl radical is the formation of socalled tandem lesions,11−15,28,29 which are produced via the attack of the uracil radical on a neighboring base and results in a covalent bond between the two bases. Such DNA intrastrand cross-links (ICL), in which two neighboring nucleobases are covalently tethered to each other, represent an important class Received: January 7, 2014 Revised: April 24, 2014 Published: April 25, 2014 5009

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

of radiation-induced DNA damage. Many details concerning ICL formation and their influence on the B-DNA conformation are still unclear. Nevertheless, it has already been shown that some ICLs block DNA replication in vitro.30,31 Similarly, Wang’s group demonstrated that intrastrand cross-link lesions result in significant destabilization to the DNA double helix.32 Finally, Basu et al. showed that ICLs cause a wide variety of mutations at or near the cross-link, including single- and multiple-base substitutions as well as small frameshifts.33 All these findings suggest that ICLs, if not repaired, can be highly cytotoxic and mutagenic. In this work, we employed completely denaturing HPLC (DHPLC), denaturing gel electrophoresis (DGE), enzymatic digestion and mass spectrometry (MS) to identify the products of UVB-damage to DNA multiply labeled with 5-bromouracil. The labeled oligonucleotide was prepared using PCR. DNA and its aqueous solutions were irradiated with photons of 300 nm. The main photodegradation products turned out to be 2′deoxuyridine as well as two intrastrand cross-links: the d(U∧U) and d(U∧C) dimers. Hence, we demonstrated that two parallel damage mechanisms are triggered by the UVB photons in the multiply labeled double-stranded oligonucleotides. One of them is associated with single strand breaks (SSBs), the other one with tandem lesion formation. We suggest that these two DNA photodamage mechanisms are also operative in the BrUlabeled cells.

Table 1. Cycling Conditions of the PCR Reaction with BrdUTP step

T [°C]

t [s]

Initial denaturation hot start−addition of the polymerase denaturation annealing of the forward primer elongation

95 85 95 55 72

30 60 30 1 2

addition of the reverse primer denaturation annealing elongation

95 95 56 72

30 5 1 2

1 cycle

30 cycles

2.3. UV Irradiation. Photolysis of purified DNA was carried out in quartz capillaries (3 mm × 3 mm) filled with a DNA solution (200 ng/μL, 4.1 × 10−6 M) containing Tris-HCl (pH = 8.3) in a total volume of 50 μL with a 500 W high-pressure mercury lamp for 1, 5, 10, and 30 min. The 300 nm wavelength of incident light (half-width 2.5 nm) was selected using a prismatic monochromator (SPM-2 Carl Zeiss, Jena). The incident light intensity for the above irradiation setup was c. 8.0 W/m2. This value was measured with a Maestro11 laser power/ energy meter (Standa), equipped with an 11PD photodetector (Standa). 2.4. Denaturing PAGE Electrophoresis. The 15% denaturing (7 M urea) polyacrylamide gels were prepared in 1 × TBE buffer.34 Samples were incubated at 95 °C for 20 min before loading onto the gel. The gels were visualized after staining with Sybr Green I using a 312 nm UV transilluminator and photographed with a SYBR Green gel stain photographic filter. 2.5. Enzymatic Digestion. Ten μg of photoirradiated DNA (concentration 200 ng/μL, 4.2 × 10−6 M) were digested to nucleosides with 0.2 U of P1 nuclease and 0.005 U of calf spleen phosphodiesterase. For this purpose, 10 μL of a solution containing both enzymes, 300 mM sodium acetate (pH 5.0) and 10 mM zinc acetate were added to 50 μL of the irradiated DNA and made up with water to a final volume of 100 μL. The digestion was carried out at 37 °C for 3 h. After the incubation, 40 U of DNase I, 10 mM CaCl2, 100 mM MgCl2, 100 mM Tris-HCl (pH 8.9), 400 μM DTPA, and water were added to the mixture to yield a final volume of 150 μL. The incubation was continued at 37 °C for 1 h. Then 1 U of BAP was added and the sample incubated at 37 °C for an additional hour. In the final step of the protocol 0.00004 U of SVP was added and the incubation was prolonged for one more hour. Finally, the solution was chloroform-extracted to remove the enzymes. The aqueous layer was lyophilized and the resulting pellet resuspended in water for liquid chromatography−mass spectrometry/mass spectrometry (LC-MS/MS) analysis. 2.6. Chromatography. The HPLC and DHPLC separation was performed on a Dionex UltiMate 3000 System with Diode Array Detector, which was set at 260 nm for monitoring the effluents. HPLC conditions. A Waters Atlantis reverse-phase dC18 column (4.6 mm × 150 mm; 5 μm in particle size and 100 L in pore size) with a mobile phase consisting of deionized water, acetonitrile (Sigma-Aldrich, Poland) and 1% formic acid (POCH S.A., Poland) (pH 2.55; 87.7:2:10.3, v/v/v) was used. The flow rate was set at 1 mL/min.

2. EXPERIMENTAL SECTION 2.1. Chemicals and Reagents. The deoxyribonucleosides, DNase I, snake venom phosphodiesterase (SVP), calf spleen phosphodiesterase (CSP), and nuclease P1 were purchased from Sigma-Aldrich, the bacterial alkaline phosphatase (BAP) and the DNA purification kit from Eurx Molecular Biology Products (Gdańsk, Poland), and Marathon DNA polymerase from A&A Biotechnology (Gdynia, Poland). The PCR primers and template were obtained from GeneSys (Wroclaw, Poland), HPLC grade acetonitrile and formic acid were purchased from PoCh (Gliwice, Poland). Ultrapure water was obtained using the Hydrolab system (Polska HLP). 2.2. PCR Reaction with BrdUTP. PCR reactions were performed using an Eppendorf thermocycler in 100 μL of a reaction mixture containing 50 mM Tris−HCl (pH 9.0), 20 mM (NH4)2SO4, 2.5 mM MgCl2, 5% DMSO, 4 ng of singlestranded DNA template (5′-ACGACAGGTTTCCCGACTGGAAAGCGGGCAGTGAGCGCAACGCAATTAATGTGAGTTAGCTCACTCATTAGGCACCCCAG-3′), 1 μM of each primer (the forward primer sequence: 5′-ACGACAGGTTTCCCGAC-3′ the reverse primer sequence: 5′CTGGGGTGCCTAATGAGTG-3′), 200 μM of each dNTP (either dATP, dGTP, dCTP, TTP or dATP, dGTP, dCTP, BrdUTP) and 1 unit of Marathon DNA polymerase, enriched with thermostable UTP-ase, which removes dUTP. A modified PCR protocol was used in order to increase the reaction yield and eliminate the formation of nonspecific products. A synthetic single-stranded oligonucleotide (80 nucleotide (nt) in length) was used as a DNA template in this protocol. In the first PCR cycle a complementary DNA strand was synthesized to create a double-stranded DNA template as specified in Table 1. Then 30 cycles of standard PCR were performed. The PCR conditions are specified in Table 1. The resulting double-stranded PCR products were 80 bp in length (see Figure 1). The amplified DNA fragments were purified with a PCR/DNA clean up purification kit. 5010

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

Figure 1. DNA sequence of nonlabeled and BrdU-labeled 80 bp PCR fragments. The 5′-bromo-2′-deoxyuridine incorporated in the labeled DNA fragment is marked in red B. The DNA sequences responsible for long-range electron transfer are shown in the yellow boxes. FragN_5′/FragN_3′, where N = 1, 2, or 3, stand for the fragments formed as a result of secondary single strand break; number of nucleotides in the praticular fragments given in brackets.

DHPLC conditions. An XBridge OST, reverse-phase C18 column (2.5 μm in particle size, 4.6 × 50 mm) was used. For the DHPLC separations a linear gradient from 10.5 to 11.5% (20 min) was employed in a buffer containing triethylamine acetate (120 mM, pH 7) at a flow rate of 0.5 mL/min, and carried out at a column temperature of 70 °C. 2.7. Mass Spectrometry. Electrospray ionization-mass spectrometry (ESI-MS) and tandem MS (MS/MS) experiments were conducted on an HCT Ultra ion-trap mass spectrometer (Bruker Daltonics). The spray voltage was −4.0 kV, the drying gas (N2) pressure was 50 psi, the flow rate was 11 l/min and the temperature was 360 °C. The spectra were registered in positive-ion mode. Each spectrum was obtained by averaging 3 scans, the time of each scan being 0.1 s. LC-MS/MS conditions. An Agilent 1200 Technologies HPLC System was employed for the LC-MS/MS experiments. A Waters Atlantis reverse-phase dC18 column (4.6 mm × 150 mm; 5 μm in particle size and 100 L in pore size) with a mobile phase consisting of deionized water, acetonitrile (SigmaAldrich, Poland) and 1% formic acid (POCH S.A., Poland) (pH 2.55; 87.7:2:10.3, v/v/v) was used. A 20 μL aliquot of digested sample was injected in each run. The effluent was coupled to the HCT Ultra ion-trap mass spectrometer, which was operated in positive-ion mode.

to be irradiated in a single experiment. On the one hand this complicates the analysis, but on the other, such an approach seems to be cost-effective, since otherwise all the sequences would have to be studied separately. In the following, we will demonstrate that due to large differences in photoreactivity only a few sequences are important for the damage. 3.2. DHPLC Analysis. The irradiated aqueous solutions of the synthesized oligonucleotide were initially analyzed by DHPLC since conventional HPLC is not suitable for the assay of damage to double-stranded DNA. Indeed, only a small broadening of the HPLC signal with the irradiation time is observed (cf. chromatograms A and B in Figure S2, Supporting Information). As double-stranded oligonucleotides longer than 100 bp are not appropriate for DHPLC (they cannot separate completely at 70−80 °C, the temperature range routinely employed with this technique36,37), the length of the amplified DNA fragment was set at 80 bp. Very recently, we demonstrated that DHPLC is quite convenient for studies of DNA damage and appears to be more accurate and more sensitive than gel electrophoresis.24 As a matter of fact, nondenaturing polyacrylamide gel electrophoresis employed against the irradiated samples shows no effect of irradiation (see Figure S1). This is in complete contrast to the DHPLC chromatograms depicted in Figure 2. In fact, the results presented in Figure 2 clearly demonstrate increasing damage to both strands of the labeled DNA fragment with increasing irradiation time (no effect of irradiation was observed with DHPLC in the nonlabeled PCR fragment). Two well-separated peaks corresponding to particular strands occur in the DHPLC chromatogram for the nonirradiated sample (see Figure 2a). After 10 min exposure both signals decrease and new multiple overlapping peaks in the 8−14 min range of retention time arise (see Figure 2b). Comparing the sum of integrated signals of nondamaged DNA in the irradiated sample to that of the nonirradiated one, one can estimate that ca. 70% of the labeled oligonucleotide was damaged. A 30 min exposure (Figure 2c) leads to further decomposition, which is associated with the almost complete disappearance of the substrate DNA; the chromatogram in Figure 2c indicates that about 90% of the substrate was converted into the products. Thus, with DHPLC one can quantitatively evaluate the damage for a given label (photosensitizer) and as a result one can tell which of the sensitizers under scrutiny is the most efficient, which again points to DHPLC as the method of choice for DNA damage studies.

3. RESULTS AND DISCUSSION 3.1. Model. In order to understand the impact of UVB photons on the sensitized cells, one should use model oligonucleotides, which resemble DNA labeled in cells. Most earlier studies on damage to Hal-NB-labeled DNA were carried out for short oligonucleotide models containing up to 2 labeled molecules per oligonucleotide21,22 as opposed to biosynthesized DNA, which is substituted to quite a large extent. For this reason, we decided to use the PCR reaction with a mixture of nucleotides containing BrdUTP (5-bromo-2′-deoxuridine-5′triphosphate) instead of dTTP (2′-deoxythymidine-5′-triphosphate) to synthesize a labeled oligonucleotide. For the first time this approach to labeling DNA with BrdU was employed by Sugiyama’s group.35 In this way, we obtained the doublestranded (ds) oligonucleotide in which all the thymines apart from those present in the starters were substituted with BrdU. Such material resembles cellular DNA that arises after several divisions of cells cultured in a medium containing BrdU. Nearly every BrdU label is surrounded by different nucleobases (as the template is a random part of the pUC19 plasmid−see the Experimental Section) in the PCR synthesized oligonucleotide (as in the biosynthesized DNA), which enables many sequences 5011

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

irradiation causes three new signals to be produced. The one with the retention time of 7.813 min (see Figure 3a) is due to 2′-deoxyuridine (dU), as revealed by the retention time of the dU standard under the HPLC conditions employed in our experiment. The occurrence of dU as the main photoproduct (see Figure 3a) suggests that the formation of secondary strand breaks is a major photodegradation pathway. At first glance it is not clear, however, if the formation of dU is due to secondary strand cleavage or is a result of the hydrogen atom transfer from Tris to the uracil-5-yl radical. Therefore, to resolve this unambiguity, we carried out additional irradiation of the studied solution in which Tris was substituted with the phosphate buffer (no external hydrogen donor). The HPLC analysis of the digested photolyte leads to the chromatogram very similar both qualitatively and quantitatively to that depicted in Figure 3. Regardless of the buffer type, irradiation leads to the formation of the same amount of 2′-deoxyuridine. The amount of dimers (a minor photoproduct) seems also to be identical in both buffers. If the source of hydrogen atoms were different in the studied buffers, different amounts of products shoud be formed (different source of hydrogen would result in the different rates of 2′-deoxyuridine and dimers formation, which would change the final yields of the products). Moreover, if the hydrogen atom originated from Tris not only the formation of 2′doxyribonolactone would be quenched due to H transfer from the Tris molecule to the uridine-5-yl radical but also the guanine radical cation (G●+) forming in the primary ET process would react with water leading to 8-oxo-2′doxyguanosine at the amount comparable to that of 2′deoxyuridne (G●+ reacts with water in dsDNA38). Another stabilization pathway of G●+ could be its fast deprotonation followed by the reaction with soluble oxygen that finally would lead to oxazolone.38 Here, it is worth mentioning that the HPLC analysis of the digested photolyte do not show any significant signals beside those for BrdU, five nucleosides and intrastrand dimers (see Figure 3). Thus, there is no indication for the presence of 8-oxo-G or oxazolone in the irradiated system buffered with Tris which should form in the amount comparable to that of dU if hydrogen came from the Tris molecule. Thus, all the above-mentioned facts suggest that in both studied buffers the hydrogen atoms originate from the same source, i.e. from DNA deoxyribose neighboring with BrdU. Since the presence of dU in the digested photolyte suggests SSBs formation we carried out additional experiments with the use of denaturing polyacrylamide gel electrophoresis. In contrast to native PAGE electrophoresis (see Supporting Information), its denaturing variant enables such a type of DNA damage to be assayed. Figure 4 compares a denaturing electropherogram for nonirradiated labeled DNA with those for the irradiated samples. Several bands are observed in lane 2 where the irradiated sample was loaded, which indicates that UV radiation produces shorter DNA fragments (proof of strand breaks). An equimolar mixture of oligonucleotides, that should form if photoinduced SSBs took place in the hot spots depicted in Figure 1, was used as a mass standard in the denaturing PAGE experiment shown in Figure 4. The locations of fragments resulting from the photodegradation match perfectly those corresponding to the mass standards which confirm the assumed cleavage sites. Note that the number of fragments is quite small despite the fact that the 80 bp oligonucleotide is completely labeled with BrU (there are as many as 29 BrUs

Figure 2. DHPLC chromatogram of (a) a nonirradiated DNA fragment labeled with 5-bromouracil and irradiated for (b) 10 min and (c) 30 min.

3.3. Enzymatic Digestion Coupled to LC-MS Analysis. In order to identify the photoproducts the photolyte was digested to nucleosides. The composition of the enzymatic cocktail was optimized prior to digestion (see Supporting Information) and we finally settled on a mixture of four enzymes (nuclease P1, calf spleen phosphodiesterase, alkaline phosphatase and snake venom phosphodiesterase). Figure 3 shows the HPLC chromatograms corresponding to the samples irradiated for 30 min and the nonirradiated labeled samples digested with the above-mentioned mixture of enzymes. Comparison of Figure 3a with 3b shows that UV

Figure 3. HPLC chromatogram of a DNA fragment labeled with 5bromouracil after enzymatic digestion (a) irradiated for 30 min and (b) nonirradiated. 5012

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

moiety and an H3PO4 component from m/z 517 and 516, respectively (see Table 2). Finally, the presence of the fragmentation ion of m/z 222 is consistent with the mass of two covalently bonded molecules of uracil, whereas the fragment ion of m/z 223 corresponds to the covalent dimer of uracil and cytosine. The above analysis allows us to conclude that the 3.897 and 6.373 min fractions (see Figure 5a) contain the d(U∧C) and d(U∧U) intrastrand cross-links, respectively (Figure 4c). A similar fragmentation pattern was observed for the intrastrand cross-links between pyrimidine and purine bases.10−13 3.4. Sequence Dependence of Damage. Taking into account the intensity of the incident light, the geometry of the irradiation setup and the concentration of DNA solution, we calculated the degradation quantum yield (QY) to be 8.4 × 10−3 for an irradiation time of 1 min (for such a short time of exposure, the error related to the fact that not only the intact DNA but also the photodegradation products absorb at 300 nm (the internal filter effect) is relatively small and can be neglected. Indeed, the degree of degradation versus time of irradiation displays almost perfect linearity (R2=0.9987) during the first 2 min of irradiation (see Figure S5). This value is between the QYs measured for singly labeled oligonucleotides (QY=0.07),21 where long-range electron transfer triggers the damage, and those characteristic of intrastrand dimer formation (QY=0.004).24 Indeed, as demonstrated in the previous section, irradiation of the studied DNA fragment leads to two types of damage: 2′deoxyuridine coupled to a strand break, and intrastrand dimers. The former is formed as a result of longrange electron transfer (ET) between the photoexcited 5bromuracil and a distant 5′-side guanine that ultimately converts BrdU into dU.22 On the other hand, the intrastrand pyrimidine dimers are produced via photocycloaddition induced by the electronic excitation of BrU initially leading to cyclobutapyrimidine.28,29 The latter reaction is followed by photochemical and thermal degradation resulting in the opening of the cyclobutane ring and formation of the ICL dimers29 that are assayed in the HPLC analysis. It is worth noting that the quantum yields for the formation of 2deoxyuridine (strand break) and intrastrand dimers are quite different and it seems reasonable to assume that the damage quantum yield measured in our experiments is a weighted mean of QYs specific to the two damage pathways, i.e., strand breaks and ICLs. Thus, in order to rationalize the measured QY value we analyzed the local surroundings of a particular label in the dsDNA fragment studied. There are only three sequences5′CABrU, which occurs in both strands, and 5′-CAABrU, present in only one of the strandsthat can be responsible for longrange electron transfer (see Figure 1). Sugiyama et al.’s results indicate that the photoreactivity of the former sequence is c. 71% of the latter one,35 while Sanche et al.21 estimate the absolute quantum yield for 5′-CAABrU damage at 6.8 × 10−2 (after correcting the value at λ = 300 nm for the absorption specific to BrU only). Hence, the quantum yield for the 5′CABrU sequence is equal to 0.048. These three “hot spots” should produce three pairs of fragments with various lengths: 30:49, 35:44, and 13:66, respectively (electron transfer leads to the heat-labile deoxyribonolactone, which easily decomposes, producing two fragments of total length shorter by one than the substrate oligonucleotide). This conclusion remains in full accordance with the results of denaturing PAGE electrophoresis (see Figure 4).

Figure 4. 15% denaturing PAGE in 1x TBE buffer. Lane 1: mass standard−an equimolar mixture of oligonucleotides that should occur if photoinduced single strand breaks concerned the hot spots presented in Figure 1 (for fragment labels see Figure 1). Lane 2: BrdU-labeled PCR fragment irradiated for 40 min. Lane 3: nonirradiated BrdU-labeled PCR fragment.

molecules present in the studied oligonucleotide). Thus, one may infer that only a few of the BrUs present in the oligonucleotide molecule lead to strand breaks. This remains in accordance with Sugiyama’s and our earlier observations that there are only a few hot spots in multiply labeled oligonucleotides. Two additional signals in the chromatogram of the digested oligonucleotide were analyzed using LC-MS and LC-MS/MS. The positive-ion ESI-MS of the products eluting at 3.897 and 6.373 min (see Figure 5a) shows ions of m/z equal to 516 and 517, respectively (see Figure 4a). These masses correspond to two dinucleotides, d(U∧C) and d(U∧U), in which the nucleobases are covalently bonded (see Figure 4c). Indeed, fragmentation (MS/MS spectra) of the m/z 516 and 517 ions leads to the formation of fragments with m/z equal to 498, 302, 222, 179 and 419, 303, 223, 180, respectively (see Figure 4b). The fragment ion of m/z 498 can be attributed to the loss of a H2O molecule from m/z 516. On the other hand, the ion of m/z 419 is due to the loss of a C5H6O2 moiety (the 2′deoxyribose component) from m/z 517, while the ions of m/z 303 and 302 are related to the eliminations of the C5H6O2 5013

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

Figure 5. (a) LC-MS analysis: positive-ion ESI-MS of the products eluting at 3.897 and 6.373 min. (b) LC-MS/MS analysis for monitoring the fragmentation pathway of m/z 516.1 and m/z. c) Chemical structure of intrastrand dimers.

arrangement (where Pyr stands for a pyrimidine base).29 Assuming the quantum yield for strand breaks of 6.8 × 10−2 and 4.8 × 10−2 (depending on the sequence) and that of ICL to be 4.0 × 10−3 (see the Discussion), one can estimate the QY of the overall damage as the weighted mean, which is then equal to 8.6 × 10−3. The latter value corresponds very well to the measured total quantum yield of 8.4 × 10−3, which on the one hand confirms the assumed mechanisms and on the other hand rationalizes the observed amount of damage. Finally, one could wonder if an efficient transfer of excitation energy (TEE), demonstrated recently for calf-thymus DNA in an elegant study employing time-resolved fluorescence spectroscopy39 may influence the above-described results. Because of the fact that our samples are excited at 300 nm, i.e. in the spectral range where the oligonucleotide absorption is determined mainly by BrdU the requirements of effective energy transfer are not fulfilled. Indeed, at the edge of absorption band the excitons are localized to a single base which slows down possible TEE.40 Moreover, as indicated by Markovitsi et al.,39 dark charge transfer (CT) states are involved in TEE in dsDNA. However, as we demonstrated41 300 nm photons are not capable of populating efficiently the CT states in B-DNA labeled with BrdU since these states are higher in energy (at least for the regular B-DNA geometry) than the first ππ* excited state localized to BrdU. Finally, the spectral overlap integral between BrdU and other nucleobases is rather small which should further diminish the probability of TEE. Hence, under the employed experimental conditions, excitation energy transfer should not influence the observed

Table 2. Mass Measurements of the [M + H]+ Ions and Product Ions Observed in the MS/MS of the [M + H]+ Ions of the two Intrastrand Cross-Links m/z

ion identities d(U∧U)

517.1 419.1 303.1 223.0 180.1 516.1 498.1 302.1 222.1 179.1

[M + H]+ −2′-deoxyribose −2′-deoxyribose, −phosphate [nucleobases + H]+ [nucleobases + H]+−HNCO d(U∧C) [M + H]+ −H2O −2′-deoxyribose, −phosphate [nucleobases + H]+ [nucleobases + H]+−HNCO

As indicated by the analysis of the studied 80 bp oligonucleotide sequence, 21 of the remaining 26 BrUs may produce a pyrimidine ICL due to the electronic excitation of 5bromouracil. Moreover, since we did not observe mixed purinepyrimidine ICLs (although we did attempt to identify them with the LC-MS/MS experiment), we assumed the quantum yield of their formation was close to zero. The above analysis leads to the conclusion that the 3 BrUs give rise to strand breaks while the other 21 lead to the U∧U or U∧C dimers. We assume that the HPLC signal corresponding to the uridinecytidine dimer is due to the U∧C isomer, since cycloaddition is strongly preferred for the 5′-BrU-Pyr (rather than 5′-Pyr-BrU) 5014

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

Society, Miami Beach, FL, April 7−12, 1957; American Chemical Society: Washington, DC, 1957; p 3C. (2) Greer, S. Studies on Ultraviolet Irradiation of Escherichia coli Containing 5-Bromouracil in its DNA. J. Gen. Microbiol. 1960, 22, 618−634. (3) Djordjevic, B.; Szybalski, W. Genetics of Human Cell Lines. III. Incorporation of 5-Bromo and 5-Iododeoxyuridine into the Deoxyribonucleic Acid of Human Cells and its Effects on Radiation Sensitivity. J. Exp. Med. 1960, 112, 509−531. (4) Kaplan, H. S.; Tomplin, P. A. Enhancement of X-ray Sensitivity of E-coli by 5-Bromouracil. Radiat. Res. 1960, 12, 447−448. (5) Bagshaw, M. A.; Doggett, R. L.; Smith, K. C.; Kaplan, H. S.; Nelsen, T. S. Intra-Arterial 5-Bromodeoxyuridine and X-ray Therapy. Am. J. Roentgenol. 1967, 99, 886−894. (6) Brust, D.; Feden, J.; Farnsworth, J.; Amir, C.; Broaddus, W. C.; Valerie, K. Radiosensitization of Rat Glioma with Bromodeoxycytidine and Adenovirus Expressing Herpes Simplex Virus-Thymidine Kinase Delivered by Slow, Rate-Controlled Positive Pressure Infusion. Cancer Gene Ther. 2000, 7, 778−788. (7) Prados, M. D.; Scott, C.; Sandler, H.; Buckner, J. C.; Phillips, T.; Schultz, Ch.; Urtasun, R.; Davis, R.; Gutin, P.; Cascino, T.; Greenberg, H. S.; Curran, W. J. A Phase 3 Randomized Study of Radiotherapy Plus Procarbazine, CCNU, and Vincristine (PCV) with or without BUdR for the Treatment of Anaplastic Astrocytoma: a Preliminary Report of RTOG 9404. Int. J. Radiat. Oncol. Biol. Phys. 1999, 45, 1109−1115. (8) Timothy, J.; Kinsella, T. J. Update on Radiosensitization by Halogenated Thymidine Analogs-Molecular Mechanisms of Drug Processing and Cell Death Signaling. Implications for Future Clinical Trials. Cancer Biol. Ther. 2008, 7, 1567−1569. (9) Rieber, M.; Rieber, M. S. Sensitization to Radiation-Induced DNA Damage Accelerates Loss of Bcl-2 and Increases Apoptosis and Autophagy. Cancer Biol. Ther. 2008, 7, 1561−1566. (10) Yan, T.; Seo, Y.; Schupp, J. E.; Zeng, X.; Desai, A. B.; Kinsella, T. J. Methoxyamine Potentiates Iododeoxyuridine-Induced Radiosensitization by Altering Cell Cycle Kinetics and Enhancing Senescence. Mol. Cancer Ther. 2006, 5, 893−902. (11) Michalska, B.; Sobolewski, I.; Polska, K.; Zielonka, J.; Ż yliczStachula, A.; Skowron, P.; Rak, J. PCR Synthesis of Double Stranded DNA Labeled with 5-Bromouridine. A Step Towards Finding a Bromonucleoside for Clinical Trials. J. Pharm. Biomed. Anal. 2011, 56, 671−677. (12) Zeng, Y.; Wang, Y. Facile Formation of an Intrastrand CrossLink Lesion Between Cytosine and Guanine upon Pyrex-Filtered UV Light Irradiation of d(BrCG) and Duplex DNA Containing 5Bromocytosine. J. Am. Chem. Soc. 2004, 126, 6552−6553. (13) Hong, H.; Wang, Y. Formation of Intrastrand Cross-Link Products Between Cytosine and Adenine from UV Irradiation of d(BrCA) and Duplex DNA Containing a 5-Bromocytosine. J. Am. Chem. Soc. 2005, 127, 13969−13977. (14) Zeng, Y.; Wang, Y. Sequence-Dependent Formation of Intrastrand Crosslink Products from the UVB Irradiation of Duplex DNA Containing a 5-Bromo-2′-Deoxyuridine or 5-Bromo-2′-Deoxycytidine. Nucleic Acids Res. 2006, 34, 6521−6529. (15) Zeng, Y.; Wang, Y. UVB-Induced Formation of Intrastrand Cross-Link Products of DNA in MCF-7 Cells Treated with 5-Bromo2′-Deoxyuridine. Biochemistry 2007, 46, 8189−8195. (16) Fujimoto, K.; Ikeda, Y.; Ishihara, S.; Saito, I. Deoxyribonolactone Formation in Photoirradiation of 5-Bromouracil-Containing Oligonucleotides by Direct C1′ Hydrogen Abstraction. Tetrahedron Lett. 2002, 43, 2243−2245. (17) Fujimoto, K.; Ikeda, Y.; Saito, I. Direct Strand Cleavage via Furanyladenine Formation in Anaerobic Photoirradiation of 5Bromouracil-Containing Oligonucleotides. Tetrahedron Lett. 2000, 41, 6455−6459. (18) Chen, T.; Cook, G. P.; Koppisch, A. T.; Greenberg, M. M. Investigation of the Origin of the Sequence Selectivity for the 5-Halo2-Deoxyuridine Sensitization of DNA to Damage by UV-Irradiation. J. Am. Chem. Soc. 2000, 122, 3861−3866.

damage since the efficiency of TEE in our system is probably very small.

4. CONCLUSIONS The 80 bp DNA fragment labeled with BrU via PCR mimics the genomic DNA that occurs after several cell divisions in a medium containing BrdU. Such DNA seems to be quite photoreactive, but only two types of damage are triggered by the UVB photons: strand breaks related to the long-range photoinduced electron transfer from a distant guanine to the photoexcited BrU, and intrastrand cross-links. The latter are formed in the cycloaddition reaction between the photoexcited BrU and the ground state pyrimidine nucleobase that adjoins the electronically excited molecule, mainly on its 3′-side. DHPLC turned out to be an exceptionally suitable tool for studies of DNA photodamage. Although DHPLC cannot decipher the identity of particular products in such a complicated multiply labeled system, it does enable damage yield to be accurately estimated since the signal corresponding to the original strands is easily quantifiable. With LC-MS, and especially the LC-MS/MS variant of mass spectrometry coupled to HPLC for the enzymatically digested photolytes, we were able to identify the end-products of the UVB photolysis as uridine as well as uridine−uridine and uridine-cytidine dimers. On the basis of these observations, together with literature data, we propound a mechanism of damage that explains rather well the value of the overall quantum yield measured for the system under study. The results of our studies suggest that the main degradation path of BrU labeled DNA is related to strand breaks, since QY for SSBs is much larger that for ICLs. However, the SSB type of damage could be harmful only if it leads to double strand breaks, since SSBs as such are easily repaired at the cellular level. On the other hand, sites susceptible to long-range electron transfer are significantly scarcer than those leading to ICLs. Therefore, despite the fact that there are more SSBs than ICLs in the irradiated cells, the ICLs are probably primarily responsible for lethal effects.



ASSOCIATED CONTENT

S Supporting Information *

Native PAGE electrophoresis, HPLC analysis of the studied DNA fragments, optimization of enzymatic digestion, and the degree of damage vs irradiation time. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*(J.R.) E-mail: [email protected]. Telephone: +4858 523 5118. Fax: +4858 523 5771. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the Polish National Science Center (NCN) under Grant No. N N204 156040.



REFERENCES

(1) Greer, S.; Zamenhof, S. Effect of 5-Bromouracil in Deoxyribonucleic Acid of E. coli on Sensitivity to Ultraviolet Irradiation. Abstracts of Papers, 131st National Meeting of the American Chemical 5015

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016

The Journal of Physical Chemistry B

Article

(19) Cook, G. P.; Greenberg, M. M. A Novel Mechanism for the Formation of Direct Strand Breaks upon Anaerobic Photolysis of Duplex DNA Containing 5-Bromodeoxyuridine. J. Am. Chem. Soc. 1996, 118, 10025−10030. (20) Cook, G. P.; Chen, T.; Koppisch, A. T.; Greenberg, M. M. The Effects of Secondary Structure and O2 on the Formation of Direct Strand Breaks upon UV Irradiation of 5-BromodeoxyuridineContaining Oligonucleotides. Chem. Biol. 1999, 6, 451. (21) Cecchini, S.; Masson, C.; La, M. C.; Huels, M. A.; Sanche, L.; Wagner, J. R.; Hunting, D. J. Interstrand Cross-Link Induction by UV Radiation in Bromodeoxyuridine-Substituted DNA: Dependence on DNA Conformation. Biochemistry 2005, 44, 16957−16966. (22) Watanabe, T.; Tashiro, R.; Sugiyama, H. Photoreaction at 5′(G/C)AABrUT-3′ Sequence in Duplex DNA: Efficient Generation of Uracil-5-yl Radical by Charge Transfer. J. Am. Chem. Soc. 2007, 129, 8163−8168. (23) Sugiyama, H.; Tsutsumi, Y.; Saito, I. Highly Sequence Selective Photoreaction of 5-Bromouracil-Containing Deoxyhexanucleotides. J. Am. Chem. Soc. 1990, 112, 6720−6721. (24) Wiczk, J.; Miloch, J.; Rak, J. DHPLC and MS Studies of a Photoinduced Intrastrand Cross-Link in DNA Labeled with 5-Bromo2′-Deoxyuridine. J. Photochem. Photobiol., B 2014, 130, 86−92. (25) Wiczk, J.; Miloch, J.; Rak, J. UV-Induced Strand Breaks in Double-Stranded DNA Labeled within 5-Bromouracil: Frank or Secondary? J. Phys. Chem. Lett. 2013, 4, 4014−4018. (26) Chomicz, L.; Rak, J.; Storoniak, P. Electron-Induced Elimination of the Bromide Anion from Brominated Nucleobases. A Computational Study. J. Phys. Chem. B 2012, 116, 5612−5619. (27) Polska, K.; Rak, J.; Bass, A. D.; Cloutier, P.; Sanche, L. Electron Stimulated Desorption of Anions from Native and Brominated Single Stranded Oligonucleotide Trimers. J. Chem. Phys. 2012, 136, 075101. (28) Lepczyńska, J.; Komodziński, K.; Milecki, J.; Kierzek, R.; Gdaniec, Z.; Franzen, S.; Skalski, B. Photoaddition of 5-Bromouracil to Uracil in Oligonucleotides Leading to 5,5′-Bipyrimidine Type Adducts: Mechanism of the Photoreaction. J. Org. Chem. 2012, 77, 11362−11367. (29) Skalski, B.; Rapp, M.; Suchowiak, M.; Golankiewicz, K. Photocycloaddition of 5-Bromouracil to Uracil in Dinucleotide Model Compound. Tetrahedron Lett. 2002, 43, 5127−5129. (30) Jiang, Y.; Hong, H.; Wang, Y. In-Vivo Formation and In-Vitro Replication of a Gua-Thy Intrastrand Cross-Link Lesion. Biochemistry 2007, 46, 12757−12763. (31) Bellon, S.; Gasparutto, D.; Saint-Pierre, C.; Cadet, J. GuanineThymine Intrastrand Cross-Linked Lesion Containing Oligonucleotides: From Chemical Synthesis to In Vitro Enzymatic Replication. Org. Biomol. Chem. 2006, 4, 3831−3837. (32) Gu, C.; Zhang, Q.; Yang, Z.; Wang, Y.; Zou, Y.; Wang, Y. Recognition and Incision of Oxidative Intrastrand Cross-Link Lesions by UvrABC Nuclease. Biochemistry 2006, 45, 10739−10746. (33) Colis, L. C.; Raychaudhury, P.; Basu, A. K. Mutational Specificity of γ-Radiation-Induced Guanine-Thymine and ThymineGuanine Intrastrand Cross-Links in Mammalian Cells and Translesion Synthesis Past the Guanine-Thymine Lesion by Human DNA Polymerase η. Biochemistry 2008, 47, 8070−8079. (34) Ausubel, F. M.; Brent, R.; Kingston, R. E.; Moore, D. D.; Seidman, J. G.; Smith, J. A.; Struhl, K. Short Protocols in Molecular Biology; John Wiley & Sons: New York, 2002. (35) Watanabe, T.; Bando, T.; Xu, Y.; Tashiro, R.; Sugiyama, H. Efficient Generation of 2′-Deoxyuridin-5-yl at 5′-(G/C)AAXUXU-3′ (X = Br, I) Sequences in Duplex DNA under UV Irradiation. J. Am. Chem. Soc. 2005, 127, 44−45. (36) Oefner, P. J.; Underhill, P. A. Comparative DNA Sequencing by Denaturing High-Performance Liquid Chromatography (DHPLC). Am. J. Hum. Genet. 1995, 57, A266. (37) Xiao, W. H.; Oefner, P. J. Denaturing High-Performance Liquid Chromatography: A Review. Hum. Mutat. 2001, 17, 439−474. (38) von Sontag, C. Free-Radicals-Induced DNA Damage and Repair. A Chemical Perspective; Springer-Verlag: Berlin and Heidelberg, Germany, 2006.

(39) Vayá, I.; Gustavsson, T.; Douki, T.; Berlin, Y.; Markovitsi, D. Electronic Excitation Energy Transfer between Nucleobases of Natural DNA. J. Am. Chem. Soc. 2012, 134, 11366−11368. (40) Markovitsi, D.; Thomas Gustavsson, T.; Akos Banyasz, A. Absorption of UV Radiation by DNA: Spatial and Temporal Features. Mutat. Res. 2010, 704, 21−28. (41) Storoniak, P.; Rak, J.; Polska, K.; Blancafort, L. Local Excitation of the 5-Bromouracil Chromophore in DNA. Computational and UV Spectroscopic Studies. J. Phys. Chem. B 2011, 115, 4532−4537.

5016

dx.doi.org/10.1021/jp500192z | J. Phys. Chem. B 2014, 118, 5009−5016