Full Papers
DOI: 10.1002/cbic.201600240
Coumarin-Induced DNA Ligation, Rearrangement to DNA Interstrand Crosslinks, and Photorelease of Coumarin Moiety Huabing Sun+, Heli Fan+, Hyeyoung Eom, and Xiaohua Peng*[a] DNA photoligation induced by the coumarin moiety was photoswitchable. Ligation products were formed between coumarin and dT or dC upon 350 nm irradiation but reverted to the original single-stranded oligodeoxyribonucleotides (ODNs) upon 254 nm irradiation. Rearrangement of ligated ODNs into ICL products occurred during the switchable (350 nm/254 nm) processes. Additionally, photoinduced cleavage of coumarin 3 occurred with dC-3 cycloadducts upon 254 nm irradiation, which was confirmed by mass spectrometry analysis.
Coumarin moieties react with thymine and cytosine in DNA by photoinduced [2+2] cycloaddition, which allows quantitative DNA interstrand crosslink (ICL) formation. Here, we report the application of coumarin analogues for DNA photoligation and the rearrangement of coumarin-induced ligation to ICL products. Both DNA sequences and the linker units at position 4 of the coumarin moieties affected coumarin-induced DNA photoligation. A flexible linker unit favored DNA ICL formation but led to inefficient photoligation, whereas coumarins without linker units greatly increased DNA photoligation efficiency.
Introduction DNA rearrangement, or switching, is defined as DNA structure or state change responding to external triggers, such as template, pH value, metal ions, and photonic or electrical stimuli. It has been widely applied in DNA hydrogels, DNA nanotechnology, controlled drug release, and antitumor drug migration.[1] Reversible Watson–Crick base pairing in DNA afforded opportunities for thermodynamics-based DNA strand rearrangement or displacement by toehold exchange.[2] During strand displacement, double-helical DNA duplex with partially complementary hybridization was exchanged with more efficiently pre-hybridized strands to form the thermally stable DNA duplex. This method has been widely used for nanomaterial self-assembly, molecular computation, and as the foundation of DNA nanotechnology.[3] However, non-covalent bonds, such as hydrogen bonds and base-stacking interactions occurring in DNA hybridization, cannot form stable products.[4] A variety of chemical reagents capable of forming covalent bonds with DNA have been employed for construction of stable DNA crosslinking products. Two major chemical reactions are usually involved in the DNA crosslinking process, including alkylation of nucleobases by an electrophile or by a [2+2] cycloaddition reaction. Alkylating reagents usually react with purines, such as quinone methide (QM) precursors, antitumor PtII complexes, nitrogen mustards, and mitomycin C.[5]
Many of these showed good migration ability under physiological conditions, such as reversible metal chelation and DNA alkylation by quinone methides (QMs), which allowed autonomous/spontaneous DNA rearrangement.[6] For example, Rokita and coworkers discovered that the reversible reactivity of QMs allowed for walking of QMs in DNA duplexes by multiple cycles of QM formation and trapping.[6a–c] Freccero et al. found that the reversible process of DNA alkylation by QMs reinforces the G-quadruplex structural rearrangement.[7] The antitumor PtII complexes can induce rearrangement of DNA intrastrand crosslinks or ligation products to interstrand crosslinking products under physiological conditions.[6d, e] Moreover, a number of photosensitive analogues efficiently form crosslink products with dT or dC by [2+2] cycloaddition reaction upon photo irradiation at wavelengths > 300 nm. Many of them showed photoreversibility, such as psoralen,[8] azobenzene,[9] p-stilbazole,[10] 3-cyanovinylcarbazole,[11] and others. The DNA crosslinking adducts formed with these compounds can be cleaved into single-stranded DNA upon photo irradiation at different wavelengths. Precise regulation of the light wavelength and intensity can efficiently control the reversibility.[12] For instance, 3-cyanovinylcarbazole can photo-crosslink the pyrimidine upon irradiation at 365 nm, and the yielded crosslinking adducts can revert to originally single-stranded DNA upon 312 nm irradiation.[11] Coumarin derivatives also exhibit potential photo reactivity towards DNA.[13] They have been widely used in the fields of biology, medicine, and cosmetics, and as fluorescent chemosensors for DNA, RNA, and protein detection.[14] Recently, our group reported that the coumarin derivatives allowed photoswitchable formation of DNA interstrand crosslinks (ICLs).[15] The coumarin analogues can react with dT or dC by photo-in-
[a] Dr. H. Sun,+ Dr. H. Fan,+ Dr. H. Eom, Prof. Dr. X. Peng Department of Chemistry and Biochemistry University of Wisconsin–Milwaukee 3210 N. Cramer St., Milwaukee, WI 53211 (USA) E-mail: [email protected] [+] These authors contributed equally to this work. Supporting information for this article can be found under http:// dx.doi.org/10.1002/cbic.201600240.
ChemBioChem 2016, 17, 1 – 9
These are not the final page numbers! ÞÞ
1
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers duced [2+2] cycloaddition, and the crosslinked products can be split into original forms upon photoirradiation at 254 nm. In this study, we applied coumarin moieties for DNA photoligation and we describe a photocontrollable rearrangement of DNA ligation to ICL adducts induced by coumarin moieties. The reversibility and stability variation of coumarin-induced ligation and ICL products can be applied in photoregulated DNA rearrangement, providing a novel opportunity for DNAbased nanotechnology.[16] Stable DNA devices can be achieved with coumarin-modified DNAs and can revert to original forms, if needed, after photoirradiation at 254 nm. Moreover, the reversibility is potentially useful for photorelease of drugs in biological systems. However, the compatibility should be considered in the process, as photoirradiation at 254 nm for a long time (e.g., > 0.5 h) can damage DNA or photosensitive molecules, and possible side photoreactions should be avoided.
ODNs containing 1–3 were synthesized by using commercially available b-cyanoethyl phosphoramidites with phenoxyacetyl protecting groups on the exocyclic amines of dA and dG (Scheme 2). All functionalized ODNs (7 b, 8 b, 11 b, 12 b, 13 c, 14 c, and 15 c) were deprotected, cleaved from solid support with NH3 (28 % aq.) at room temperature for 2 h, and purified by 20 % denaturing PAGE, with new ODNs characterized by ITTOF-MS (Figures S15–S18).
Results and Discussion Synthesis of the crosslinking precursors and their incorporation into ODNs To identify coumarin analogues capable of forming ligation and ICL products, we designed and synthesized three coumarin analogues (1–3). Compounds 1 and 2 were used to study linker effects on reactivity; 2 and 3 were used to investigate the effect of substituents at position 4. As previously reported, compounds 2 and 3 were synthesized by a Pechmann condensation between resorcinol and ethyl acetoacetate or malic acid. Compound 1 was prepared from 2 by Williamson ether synthesis.[15a] Compounds 1–3 were converted to the corresponding phosphoramidites (4–6) under standard conditions, which were used for solid-phase DNA synthesis of modified ODNs (Scheme 1). The compounds were confirmed by NMR and HRMS (Figure S14 A–D in the Supporting Information).
Scheme 2. Double-stranded DNAs (dsDNAs) used in this study.
Scheme 1. Coumarins used in the study and synthesis of phosphoramidites 4–6.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
2
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers Optimization of ODN sequences
Table 1. Interstrand crosslink (ICL) formation reaction rates for dsDNAs 7–12.
To identify DNA sequences that can form both ICL and ligation products for the study, the ODN sequences were first optimized. Previously, we showed that a cycloaddition reaction did not occur between coumarin and dG. In addition, the linker of the coumarin played an important role in increasing interstrand crosslinking efficiency.[15a] For example, a quantitative DNA ICL was generated by using coumarin moieties with a flexible linker of two or more carbons, whereas coumarin analogues without the linker led to less efficient ICL formation. Thus, three DNA duplexes containing 1 with a two-carbon linker (dsDNA 7, dsDNA 9, and dsDNA 11) and different numbers of dGs at the terminus were designed to study the effect of dT positions on ICL formation. Another three duplexes containing 2 without the linker (dsDNA 8, dsDNA 10, and dsDNA 12) were employed to investigate the linker effect in this process (Scheme 2). Upon photoirradiation at 350 nm, all DNA duplexes formed ICL products but with different efficiency and reaction rates. Almost quantitative DNA ICL formation was observed for 1containing dsDNA 7 (ICL yield: 98 %) and dsDNA 9 (ICL yield: 97 %). A doublet of ICL products was observed for dsDNA 7 (line 2, Figure 1), which is ascribed to the varied crosslinking
k [10 dsDNA 7 dsDNA 8 dsDNA 9
These are not the final page numbers! ÞÞ
42.0 4.0 4.8 0.4 6.1 0.5
t1/2 [min] 30 24 2 19 2
k [10 dsDNA 10 dsDNA 11 dsDNA 12
4
s 1]
0.63 0.03 1.6 0.07 0.84 0.04
t1/2 [min] 184 9 67 2 139 6
Coumarin-induced ICL and ligation reactions with dT or dC Having optimized ODN sequences, we designed dsDNAs 13– 15 for investigating DNA interstrand crosslinking and ligation reactions. These DNA duplexes contain dT 48 as a flanking nucleotide that can form ligation products with coumarin analogues and dT 16 as an opposing nucleotide capable of generating ICL products. Thus, the coumarin analogues had equal opportunity for generating ICL products or ligation products. The ICL and ligation reactions were studied separately by labeling different ODNs with g-32P. The ICL products were detected by using 32P-labeled ODN-13 a; the ligation products were detected by using 32P-labeled-13 b. As expected, dsDNA 13 containing 1 with a flexible linker showed poor reactivity and a low reaction rate in the ligation reaction (Figures 2, S5 and S6), due to competition with more favorable ICL reactions between dT and 1. Irradiation of dsDNA
sites. As coumarin 1 contains a flexible linker, it is highly likely to react with various dTs close to 3’ terminus in ODN 7 a, resulting in ICL products with different migration. Previous reports also showed that the oligodeoxynucleotides, with the same length but different structures, could migrate distinctly in denaturing PAGE.[15a, 17] Although dT is three nucleotides away from 1 in dsDNA 11, the ICL yield is still relatively high (78 %). These data showed that 1 might not be a good candidate for forming ligation products as dTs in a variety positions of the complementary strand would compete to form DNA ICL products. In contrast, the ICL formation induced by 2 is more dependent on the distance and positions of dT, though less efficient than that induced by 1. About 23 % decrease of ICL yield was observed with dsDNA 10 containing dT that is one nucleotide away from coumarin 1, and 39 % decrease for dsDNA 12 with dT two nucleotides away from 1 (Figure 1). Kinetic study showed that installation of dG flanking to dT slowed down the crosslinking reaction. Introduction of one dG next to dT led to an approximately sevenfold decrease in ICL www.chembiochem.org
s 1]
reaction rate (42.0 4.0·10 4 s 1 for dsDNA 7 vs. 6.1 0.5·10 4 s 1 for dsDNA 9; Table 1 and Figures S1 and S2). We propose that dsDNAs with highest reactivity for ICL reactions have poor reactivity for ligation reaction, whereas those with modest interstrand crosslinking efficiency are expected to be more favorable for ligation. In addition, the crosslinking efficiency of coumarin analogues without linkers can be better tuned by variation of dT positions and flanking sequences. Thus, we expect that compounds 2 and 3 are suitable for further study about competition between ICL and ligation reaction, as well as rearrangement study.
Figure 1. Phosphorimage autoradiogram of denaturing PAGE analysis of distance-dependent ICL formation. C: Control sample containing 100 nm dsDNA 7 without photoirradiation; 100 nm dsDNAs 7–12 was irradiated at 350 nm; ODNs without coumarins were 5’-[32P]-labeled.
ChemBioChem 2016, 17, 1 – 9
4
Figure 2. Phosphorimage autoradiogram of denaturing PAGE analysis of A) ICL and B) ligation reactions. C: Control sample containing 100 nm dsDNA 13 without photoirradiation; 100 nm dsDNAs 13–15 was irradiated at 350 nm for 4 h. For the ICL reactions, ODN-13 a was 5’-[32P]-labeled, whereas ODN-13 b was 5’-[32P]-labeled for ligation reactions.
3
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers Table 2. The rates of ICL/ligation product formation or cleavage reaction of dsDNA 13–18.[a] dsDNA
k (ICL formation) [10 4 s 1]
t1/2 [h]
kc (ICL cleavage) [10 3 s 1]
t1/2 [min]
k (ligation formation) [10 4 s 1]
t1/2 [h]
kc (ligation cleavage) [10 3 s 1]
t1/2 [min]
13 14 15 16 17 18
5.87 0.18 0.96 0.06 1.57 0.06 0.61 0.06 0.31 0.03 0.58 0.06
0.33 0.01 2.01 0.12 1.22 0.04 3.18 0.28 6.30 0.55 3.37 0.35
n.d. 5.12 0.27 4.75 0.34 n.d. 8.79 0.71 3.98 0.33
n.d. 2.26 0.12 2.44 0.17 n.d. 1.32 0.11 2.92 0.25
2.55 0.44 1.21 0.13 1.67 0.16 1.35 0.19 0.55 0.05 0.42 0.04
0.78 0.14 1.61 0.17 1.16 0.11 3.54 0.30 3.54 0.30 2.70 0.25
n.d. 5.94 0.45 9.78 0.43 n.d. 6.58 0.79 6.24 0.66
n.d. 1.96 0.15 1.18 0.05 n.d. 1.78 0.21 1.87 0.20
[a] n.d.: not determined.
13 with 350 nm light mainly generated ICL products in 77.8 % yield (Figure 2 A, lane 2), leading to a low ligation efficiency (12.3 %; Figure 2 B, lane 2). The rate of the interstrand crosslinking reaction of dsDNA 13 is two times that of the ligation reaction. In contrast, comparable ICL and ligation yields were observed with dsDNA 14 or 15 containing 2 or 3, which lack a flexible linker unit (Figure 2, 30.8 % vs. 26.5 % for dsDNA 14; 30.2 % vs. 34.1 % for dsDNA 15). Thus, dsDNAs 14 and 15 are preferable for further investigation on rearrangement between ICL products and ligation products. Kinetic studies indicated that the ligation reaction proceeded a little faster than the ICL reaction for dsDNAs 14 and 15 (Table 2 and Figures S3–S6). These data suggested that photo-induced ligation was more kinetically controlled for dsDNAs 14 and 15 than the ICL reactions. Previous studies showed that dC also formed [2+2] cycloaddition products with coumarin analogues in DNA. In order to evaluate the generality of the phenomena observed above, we investigated the competition between the dC-coumarin ligation and dT-coumarin ICL reaction. Thus, dsDNAs 16, 17, and 18 were designed to have a dC flanking two coumarin analogues and a dT in the opposing strand, which allowed the coumarin analogues to ligate with dC and crosslink with dT. Similar to dT-coumarin ligation, compound 1 in dsDNA 16 showed very low ligation efficiency (ligation yield = 3.0 %), and the ICL reaction was predominant (30.9 %). However, more efficient ligation reactions were observed for 2 (dsDNA 17: 8.8 %) and 3 (dsDNA 18: 12.5 %) than 1 (dsDNA 16: 3 %), whereas a less efficient ICL reaction was observed for 2 (9.9 %) and 3 (23.6 %) than 1 (30.9 %; Figure 3). The rate of ICL and ligation reactions for dsDNA 17 were (3.08 0.27) 10 5 s 1 and (5.48 0.46) 10 5, respectively, and those for dsDNA 18 were (5.76 0.60) 10 5 s 1 and (4.20 0.40) 10 5, respectively (Table 2 and Figures S3–S6). To provide further evidence that coumarins are essential for ICL and ligation product formation, we synthesized two dsDNAs 19 and 20 that did not contain coumarin moieties and investigated their photoreactivity under the same conditions used for dsDNAs 13–18. No ICL or ligation adducts were observed for dsDNAs 19 and 20, indicating that coumarins played a crucial role for photoinduced DNA ICL formation and ligation with dsDNAs 13–18 (Figure S11).
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Figure 3. Phosphorimage autoradiogram of denaturing PAGE analysis of A) ICL and B) ligation reactions. C: Control sample contained 100 nm dsDNA 16 without photoirradiation; 100 nm; dsDNAs 16–18 was irradiated at 350 nm for 20 h. For the ICL reactions, ODN-16 a was 5’-[32P]-labeled, whereas ODN-16 b was 5’-[32P]-labeled for ligation reactions.
Photoreversibility and rearrangement of ligation products to ICL products The photocycloaddition reactions between coumarin analogues and dT or dC were reversible, and rearrangement of ligation products to ICL products was observed during the process. The photocycloaddition products generated by 350 nm irradiation were split into single-stranded ODNs at a yield of 70 % upon 254 nm irradiation. The cleavage reactions followed first-order kinetics and were complete within 10 min (Table 2 and Figures S7–S10). The cleavage reaction rates for ligation products were similar to those for ICL products. For instance, the rate constant for cleaving ligation adducts formed with dsDNA 14 is (5.94 0.45) 10 3 s 1; that for ICL adducts is (5.12 0.27) 10 3 s 1. Similar results were observed for dsDNAs 15, 17, and 18 (Table 2). The reversible behavior was observed within three cycles of photoirradiation at 350 nm for 20/4 h and 254 nm for 10 min (Figure 4). However, after each cycle 254 nm/350 nm irradiation, the yields for interstrand crosslinking products increased, whereas those for ligation products decreased. For example, the ICL yield for dsDNA 14 increased from 41.5 % in the first circle to 47.8 % in the third cycle, while the ligation yield decreased from 32.5 to 9.9 % during this process (Figure 4 A). Similarly, three cycles of 254 nm/350 nm irradiation led to a 16 % increase in ICL yield and a 12 % decrease in photoligation yield 4
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers
Figure 4. Photoreversible process over three cycles for A) dsDNA 14, B) dsDNA 15, C) dsDNA 17, and D) dsDNA 18 upon photoirradiation at 350 nm for 4 h for dsDNAs 14/15 or 20 h for dsDNAs 17/18 (UV1) and 254 nm (UV2) for 10 min.
was observed, which migrated faster than the original singlestranded ODN-15 c after the reversible process (Figure 5). The mass analysis suggested that SS ODN* was caused by photorelease of coumarin 3 in ODN-15 c (Scheme 3 B and Figure S19). To gain more details on the photorelease reaction, such as which wavelength led to the cleavage of coumarin and which adduct was cleaved, we performed a more detailed study with dsDNA 18, with dsDNA 17 for comparison. The dsDNAs were separately irradiated with only 254 nm irradiation, only 350 nm irradiation, and both 350/254 nm irradiation (Figure 5). The results suggested that the photorelease of coumarin moiety only occurred with the dC-3 cycloaddition adducts upon 254 nm irradiation. First, cleaved product SS ODN* was not observed following 350 nm irradiation of the ICL/ligation products, indicating the coumarin-dC adduct was stable to 350 nm irradiation (Figure 5, lanes 2 and 6). Second, the SS ODN* was not observed with 254 nm irradiation of dsDNA 17 or 18, thus indicating that the coumarin was not cleaved in the single-stranded ODN (Figure 5, lanes 3 and 7). Third, the SS ODN* was not observed for dsDNA 17, thus indicating that dC-2 adducts were not cleaved (Figure 5, lanes 2–4). In addition, the photorelease products (SS ODN*) were not observed for dsDNAs 14 and 15, thus indicating that dT-coumarin adducts were not cleaved (Figure S13). Thus, we concluded that 254 nm irradiation of dC-3 cyclobutane adduct 21 formed with dsDNA 18 led to photorelease of coumarin 3 and SS ODN* (Scheme 3 B). The Schmidt group reported that (coumarin-4-yl)methyl esters
for dsDNA 17. The same trend was observed for dsDNA 15. These results indicated that rearrangement from kinetic-controlled ligation products to thermo-controlled ICL products occurred in the photoswitchable process (Figure 4 A–C). To provide direct evidence for the rearrangement of ligation products to ICL products, we isolated the ligation products formed with dsDNAs 14 and 15, which were used for photoreversibility and rearrangement studies. As expected, photoirradiation of the isolated ligation products at 254 nm led to photocleavage of ligation adducts, generating the single-stranded ODN (Figure S12). The rearrangement from ligation products to ICL products was also observed in the photoreversible process (254 nm/350 nm irradiation). After two cycles of 254 nm/ 350 nm irradiation, the ligation adducts obtained with dsDNA 14 were rearranged to ICL products in 26.7 % yield, whereas the ligation product yield dropped to 18.7 %. Similarly, two cycles of 254 nm/350 nm irradiation of the ligation adducts obtained with dsDNA 15 produced 28.1 % ICL products and 13.4 % ligation products (Figure S12). Photorelease of coumarin 3 in dsDNA 18 Surprisingly, both ICL and ligation yields decreased for dsDNA 18 during the photoreversible process (Figure 4 D). The ICL yield decreased from 40.4 % to 34.7 %, and the ligation yield decreased from 14.4 to 3.3 % after three cycles of 254 nm/ 350 nm irradiation. Meanwhile, a new ODN band (SS ODN*) ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
5
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers
Figure 5. Phosphorimage autoradiogram of denaturing PAGE analysis of photorelease reaction for dsDNAs 17/18 upon treatment with irradiation at different wavelengths.
Scheme 3. Photorelease of coumarin 3 from dsDNA 18 after photoirradiation.
Conclusion
(CM-A) are caged compounds that, upon excitation, release hemagglutinin and the (coumarin-4-yl)methyl alcohol via heterolytic ester cleavage (Scheme 3 A).[18] Similarly, we proposed that heterolytic ester cleavage occurred with the dC-3 cyclobutane adduct upon irradiation at 254 nm, generating free coumarin 3, ODN 16 b, and SS ODN*.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Three coumarin analogues (1–3), with or without the linker units at position 4, were employed for DNA photoligation. Both linker units in coumarin moieties and DNA sequences affected the ligation efficiency, due to the competition of DNA 6
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers temperature overnight (total volume of 50 mL). The 32P-labeled oligonucleotide duplex (0.1 mm, 2 mL) was mixed with NaCl (1 m, 2 mL), potassium phosphate (100 mm, pH 7.0, 2 mL), and autoclaved distilled water (14 mL) to give a final volume of 20 mL. The DNA duplexes were photoirradiated at 350 nm to form ICLs or ligation products (a control reaction was carried out without photoirradiation). The aliquots were taken at the desired time and immediately quenched with the equal volume of formamide (95 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis. For kinetics study, aliquots (final concentration: 50 nm 32P-labeled oligonucleotide duplex, NaCl (100 mm), potassium phosphate (10 mm)) of three independent reactions were taken at the prescribed times, immediately quenched by formamide (90 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis. Photoirradiation at 350 nm was employed for ICLs or ligation reactions, whereas photoirradiation at 254 nm was used for photocleavage reactions.
ICL formation. Coumarin analogues (2 and 3) without linker units are more favorable for DNA photoligation than 1, which has a two-carbon chain linker. Compounds 2 and 3 generated comparable photoligation products and DNA ICL adducts, but 1 induced almost quantitative DNA ICL formation with a very low yield of ligation products. DNA photoligation induced by coumarin moieties is reversible. Irradiation at 350 nm produced ligation products that could revert to the original singlestranded ODNs upon 254 nm irradiation. Several cycles of 350 nm/254 nm irradiation led to rearrangement of kineticcontrolled ligation products to thermocontrolled ICL adducts. Photorelease of the coumarin moiety was observed with 254 nm irradiation of the dC-3 products but not for dT-3 adducts. DNA photoligation induced by the coumarin moiety, rearrangement of DNA adducts, and photorelease of the coumarin moiety can serve as a novel tool for DNA nanomaterial research or drug-release studies.[16]
Isolation of the ligation products: The hybridized ODNs (10 mm), with or without a-32P-labeled strands, were photoirradiated at 350 nm with the desired time to generate the ICLs or ligation products. The products were isolated by using 20 % denaturing PAGE with a-32P-labeled ones as standards. No UV light at 254 nm can be used for product detection; otherwise photocleavage is induced. The ligation products were purified for further study via standard protocol. It should be mentioned that the hybridization of ligation products from 80 8C can result in a little DNA damage.
Experimental Section General methods: Unless otherwise specified, reagents were used directly as received without further purification. Columns and phosphoramidites for DNA synthesis were purchased from Glen Research. T4 polynucleotide kinase was obtained from New England Biolabs, and [g-32P]-ATP was from PerkinElmer Life Sciences. Water was purified via a Milli-Q purification system before use for preparation of ODN solutions.
Reversibility study: The hybridized ODNs were photoirradiated at 350 nm with the desired time to generate the ICLs or ligation products. Products formed by 350 nm irradiation were photocleaved upon photoirradiation at 254 nm with the required time estimated from kinetic studies. Aliquots were taken at the prescribed steps and immediately quenched with the equal volume of formamide (95 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis.
ODNs were synthesized with standard automated DNA synthesis techniques on a 1.0 mmol scale by using commercially available 1000 CPG-succinyl-nucleoside supports. The iPr-Pac-dA and iPrPac-dG phosphoramidites were used for the preparation of coumarin-containing ODNs. Deprotection of the nucleobases and phosphate moieties, as well as cleavage of the linker for normal ODNs, was carried out in a mixture of MeNH2 (40 % aq.) and NH3 (28 % aq.; 1:1) at room temperature for 2 h. Functionalized ODNs were deprotected and cleaved with NH3 (28 % aq.) at room temperature for 2 h. All crude ODNs were purified by 20 % denaturing polyacrylamide gel electrophoresis (PAGE) and then quantified in water. Radiolabeling was performed according to standard protocols.[19] The 32P-labeled ODN (100 nm) was annealed with 1.5 equiv of the complementary strand and related strands by heating to 80 8C for 3 min in a buffer of potassium phosphate (10 mm, pH 7.0) and NaCl (100 mm), followed by cooling slowly to room temperature overnight. Radiolabeled ODNs were quantified by using a Molecular Dynamics Phosphorimager.
7-Hydroxy-2 H-chromen-2-O-(2-cyanoethyl-N,N-diisopropyl)phosphoramidite (6): N,N-diisopropylethylamine (DIPEA; 156 mL, 0.9 mmol), and 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (167 mL, 0.75 mmol) were added to a solution of 7-hydroxy2H-chromen-2-one (3, 81 mg, 0.5 mmol) in CH2Cl2 (10 mL) under argon. The reaction mixture was stirred at room temperature for 3 h and diluted with CH2Cl2 (50 mL). The organic layer was washed with NaHCO3 (20 mL) and saturated aqueous NaCl (20 mL) and dried over anhydrous sodium sulfate. The solvent was removed under reduced pressure. Upon purification by column chromatography (EtOAc/hexane/Et3N 59:40:1), the product, 8, was isolated as a white solid (138 mg, 76 %). 31P NMR (122 MHz, CDCl3): d = 147.59. 1 H NMR (300 MHz, CDCl3): d = 7.56, 7.59 (d, J = 9.6 Hz, 1 H), 7.30– 7.33 (d, J = 8.4 Hz, 1 H), 6.90–6.94 (m, 2 H), 6.20, 6.24 (d, J = 9.6 Hz, 1 H), 3.87–3.90 (m, 2 H), 3.63–3.71 (m, 2 H), 2.61–2.65 (t, J = 6.3 Hz, 2 H), 1.10–1.19 (dd, J = 3.6 Hz, 12 H); 13C NMR (75 MHz, CDCl3): d = 160.0, 156.9, 156.8, 154.3, 142.3, 127.7, 116.3, 115.8, 115.7, 113.0, 106.4, 106.3, 58.3, 58.0, 43.0, 42.9, 23.6, 23.5, 23.4, 19.4, 19.3; HRMS-ESI (+): m/z calcd for C18H24N2O4P: 363.1468 [M+H] + ; found: 363.1470.
Photoirradiation of samples in colorless siliconized microcentrifuge tubes at 350 nm was carried out in a Rayonet Photochemical Chamber Reactor (Model RPR-100) equipped with 12 lamps, and photoirradiation at 254 nm was conducted with a Compact UV Lamp (UVGL-25) by using the short wavelength at room temperature. 1H NMR, 13C NMR, and 31P NMR analyses were performed on a 300 MHz spectrophotometer. Chemical shifts are reported in ppm relative to Me4Si (1H and 13C) or H3PO4 (31P). HRMS was performed at the Mass Spectrometry Laboratory of the University of Wisconsin–Milwaukee.
Acknowledgements
Photo-induced DNA ligation or ICL formation by coumarin moiety and kinetic studies: 5’-32P-labeled or 3’-32P-labeled ODNs (0.1 mm) were annealed with 1.5 equiv of the complementary strands in NaCl (100 mm) and potassium phosphate (10 mm, pH 7) by heating to 65 8C for 3 min, followed by cooling slowly to room
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
We are grateful for financial support from the University of Wisconsin-Milwaukee (UWM) Research Growth Initiative (RGI101234), the Greater Milwaukee Foundation (Shaw Scientist Award), and 7
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers a Wisconsin Applied Research Grant (ARG) Award. We thank Dr. Zhiqiang (Mark) Wang for mass analysis.
[11] a) K. Fujimoto, A. Yamada, Y. Yoshimura, T. Tsukaguchi, T. Sakamoto, J. Am. Chem. Soc. 2013, 135, 16161 – 16167; b) Y. Yoshimura, K. Fujimoto, Org. Lett. 2008, 10, 3227 – 3230; c) K. Fujimoto, K. Konishi-Hiratsuka, T. Sakamoto, Y. Yoshimura, ChemBioChem 2010, 11, 1661 – 1664. [12] W. Szyman´ski, J. M. Beierle, H. A. Kistemaker, W. A. Velema, B. L. Feringa, Chem. Rev. 2013, 113, 6114 – 6178. [13] a) J. Woo, P. B. Hopkins, J. Am. Chem. Soc. 1991, 113, 5457 – 5459; b) A. K. Thazhathveetil, S. T. Liu, F. E. Indig, M. M. Seidman, Bioconjugate Chem. 2007, 18, 431 – 437; c) M. Takasugi, A. Guendouz, M. Chassignol, J. L. Decout, J. Lhomme, N. T. Thuong, C. Helene, Proc. Natl. Acad. Sci. USA 1991, 88, 5602 – 5606; d) A. Majumdar, P. A. Muniandy, J. Liu, J. I. Liu, S. T. Liu, B. Cuenoud, M. M. Seidman, J. Biol. Chem. 2008, 283, 11244 – 11252; e) T. E. Haran, D. M. Crothers, Biochemistry 1988, 27, 6967 – 6971. [14] a) B. D. Wagner, Molecules 2009, 14, 210 – 237; b) H. Li, L. Cai, Z. Chen, Advances in Chemical Sensors (Ed.: W. Wang), InTech, Rijeka, 2012, pp. 121 – 150. [15] a) H. Sun, H. Fan, X. Peng, J. Org. Chem. 2014, 79, 11359 – 11369; b) M. M. Haque, H. Sun, S. Liu, Y. Wang, X. Peng, Angew. Chem. Int. Ed. 2014, 53, 7001 – 7005; Angew. Chem. 2014, 126, 7121 – 7125. [16] S. R. Trenor, A. R. Shultz, B. J. Love, T. E. Long, Chem. Rev. 2004, 104, 3059 – 3077. [17] Q. Zhai, P. Wang, Q. Cai, Y. Wang, Nucleic Acids Res. 2014, 42, 10529 – 10537. [18] R. Schmidt, D. Geissler, V. Hagen, J. Bendig, J. Phys. Chem. A 2007, 111, 5768 – 5774. [19] T. F. Maniatis, E. F. Sambrook, J. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, New York, 1982.
Keywords: coumarins · DNA rearrangement · interstrand crosslink · photoligation · photorelease [1] F. Wang, X. Liu, I. Willner, Angew. Chem. Int. Ed. 2015, 54, 1098 – 1129; Angew. Chem. 2015, 127, 1112 – 1144. [2] a) D. Y. Zhang, E. Winfree, J. Am. Chem. Soc. 2009, 131, 17303 – 17314; b) D. Y. Zhang, G. Seelig, Nat. Chem. 2011, 3, 103 – 113. [3] a) W. Shih, Nat. Mater. 2008, 7, 98 – 100; b) J. Bath, A. J. Turberfield, Nat. Nanotechnol. 2007, 2, 275 – 284. [4] S. Modi, D. Bhatia, F. C. Simmel, Y. Krishnan, J. Phys. Chem. Lett. 2010, 1, 1994 – 2005. [5] a) D. M. Noll, T. M. Mason, P. S. Miller, Chem. Rev. 2006, 106, 277 – 301; b) S. Cao, X. Peng, Curr. Org. Chem. 2014, 18, 70 – 85; c) M. Freccero, Mini-Rev. Org. Chem. 2004, 1, 403 – 415. [6] a) H. Wang, S. E. Rokita, Angew. Chem. Int. Ed. 2010, 49, 5957 – 5960; Angew. Chem. 2010, 122, 6093 – 6096; b) H. Wang, Curr. Org. Chem. 2014, 18, 44 – 60; c) W. B. Sherman, N. C. Seeman, Nano Lett. 2004, 4, 1203 – 1207; d) C. Perez, M. Leng, J. M. Malinge, Nucleic Acids Res. 1997, 25, 896 – 903; e) J. Malina, J. Kasparkova, N. P. Farrell, V. Brabec, Nucleic Acids Res. 2011, 39, 720 – 728; f) F. Fakhari, S. E. Rokita, Nat. Commun. 2014, 5, 5591. [7] M. Di Antonio, F. Doria, S. N. Richter, C. Bertipaglia, M. Mella, C. Sissi, M. Palumbo, M. Freccero, J. Am. Chem. Soc. 2009, 131, 13132 – 13141. [8] a) B. J. Parsons, Photochem. Photobiol. 1980, 32, 813 – 821; b) Y. H. Ping, T. M. Rana, Biochemistry 2005, 44, 2501 – 2509. [9] H. Asanuma, T. Ito, T. Yoshida, X. Liang, M. Komiyama, Angew. Chem. Int. Ed. 1999, 38, 2393 – 2395; Angew. Chem. 1999, 111, 2547 – 2549. [10] H. Kashida, T. Doi, T. Sakakibara, T. Hayashi, H. Asanuma, J. Am. Chem. Soc. 2013, 135, 7960 – 7966.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Manuscript received: April 21, 2016 Accepted article published: August 25, 2016 Final article published: && &&, 0000
8
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
FULL PAPERS UV control: Coumarin analogues were shown to photo-crosslink dT- or dCforming ligation products in addition to interstrand crosslink (ICL) adducts. The photoligation was photoswitchable: ligation products formed at 350 nm irradiation reverted to the original singlestranded ODNs upon 254 nm irradiation. Rearrangement of ligated ODNs into ICL products also occurred during the switchable processes.
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
H. Sun, H. Fan, H. Eom, X. Peng* && – && Coumarin-Induced DNA Ligation, Rearrangement to DNA Interstrand Crosslinks, and Photorelease of Coumarin Moiety
9
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
DOI: 10.1002/cbic.201600240
Coumarin-Induced DNA Ligation, Rearrangement to DNA Interstrand Crosslinks, and Photorelease of Coumarin Moiety Huabing Sun+, Heli Fan+, Hyeyoung Eom, and Xiaohua Peng*[a] DNA photoligation induced by the coumarin moiety was photoswitchable. Ligation products were formed between coumarin and dT or dC upon 350 nm irradiation but reverted to the original single-stranded oligodeoxyribonucleotides (ODNs) upon 254 nm irradiation. Rearrangement of ligated ODNs into ICL products occurred during the switchable (350 nm/254 nm) processes. Additionally, photoinduced cleavage of coumarin 3 occurred with dC-3 cycloadducts upon 254 nm irradiation, which was confirmed by mass spectrometry analysis.
Coumarin moieties react with thymine and cytosine in DNA by photoinduced [2+2] cycloaddition, which allows quantitative DNA interstrand crosslink (ICL) formation. Here, we report the application of coumarin analogues for DNA photoligation and the rearrangement of coumarin-induced ligation to ICL products. Both DNA sequences and the linker units at position 4 of the coumarin moieties affected coumarin-induced DNA photoligation. A flexible linker unit favored DNA ICL formation but led to inefficient photoligation, whereas coumarins without linker units greatly increased DNA photoligation efficiency.
Introduction DNA rearrangement, or switching, is defined as DNA structure or state change responding to external triggers, such as template, pH value, metal ions, and photonic or electrical stimuli. It has been widely applied in DNA hydrogels, DNA nanotechnology, controlled drug release, and antitumor drug migration.[1] Reversible Watson–Crick base pairing in DNA afforded opportunities for thermodynamics-based DNA strand rearrangement or displacement by toehold exchange.[2] During strand displacement, double-helical DNA duplex with partially complementary hybridization was exchanged with more efficiently pre-hybridized strands to form the thermally stable DNA duplex. This method has been widely used for nanomaterial self-assembly, molecular computation, and as the foundation of DNA nanotechnology.[3] However, non-covalent bonds, such as hydrogen bonds and base-stacking interactions occurring in DNA hybridization, cannot form stable products.[4] A variety of chemical reagents capable of forming covalent bonds with DNA have been employed for construction of stable DNA crosslinking products. Two major chemical reactions are usually involved in the DNA crosslinking process, including alkylation of nucleobases by an electrophile or by a [2+2] cycloaddition reaction. Alkylating reagents usually react with purines, such as quinone methide (QM) precursors, antitumor PtII complexes, nitrogen mustards, and mitomycin C.[5]
Many of these showed good migration ability under physiological conditions, such as reversible metal chelation and DNA alkylation by quinone methides (QMs), which allowed autonomous/spontaneous DNA rearrangement.[6] For example, Rokita and coworkers discovered that the reversible reactivity of QMs allowed for walking of QMs in DNA duplexes by multiple cycles of QM formation and trapping.[6a–c] Freccero et al. found that the reversible process of DNA alkylation by QMs reinforces the G-quadruplex structural rearrangement.[7] The antitumor PtII complexes can induce rearrangement of DNA intrastrand crosslinks or ligation products to interstrand crosslinking products under physiological conditions.[6d, e] Moreover, a number of photosensitive analogues efficiently form crosslink products with dT or dC by [2+2] cycloaddition reaction upon photo irradiation at wavelengths > 300 nm. Many of them showed photoreversibility, such as psoralen,[8] azobenzene,[9] p-stilbazole,[10] 3-cyanovinylcarbazole,[11] and others. The DNA crosslinking adducts formed with these compounds can be cleaved into single-stranded DNA upon photo irradiation at different wavelengths. Precise regulation of the light wavelength and intensity can efficiently control the reversibility.[12] For instance, 3-cyanovinylcarbazole can photo-crosslink the pyrimidine upon irradiation at 365 nm, and the yielded crosslinking adducts can revert to originally single-stranded DNA upon 312 nm irradiation.[11] Coumarin derivatives also exhibit potential photo reactivity towards DNA.[13] They have been widely used in the fields of biology, medicine, and cosmetics, and as fluorescent chemosensors for DNA, RNA, and protein detection.[14] Recently, our group reported that the coumarin derivatives allowed photoswitchable formation of DNA interstrand crosslinks (ICLs).[15] The coumarin analogues can react with dT or dC by photo-in-
[a] Dr. H. Sun,+ Dr. H. Fan,+ Dr. H. Eom, Prof. Dr. X. Peng Department of Chemistry and Biochemistry University of Wisconsin–Milwaukee 3210 N. Cramer St., Milwaukee, WI 53211 (USA) E-mail: [email protected] [+] These authors contributed equally to this work. Supporting information for this article can be found under http:// dx.doi.org/10.1002/cbic.201600240.
ChemBioChem 2016, 17, 1 – 9
These are not the final page numbers! ÞÞ
1
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers duced [2+2] cycloaddition, and the crosslinked products can be split into original forms upon photoirradiation at 254 nm. In this study, we applied coumarin moieties for DNA photoligation and we describe a photocontrollable rearrangement of DNA ligation to ICL adducts induced by coumarin moieties. The reversibility and stability variation of coumarin-induced ligation and ICL products can be applied in photoregulated DNA rearrangement, providing a novel opportunity for DNAbased nanotechnology.[16] Stable DNA devices can be achieved with coumarin-modified DNAs and can revert to original forms, if needed, after photoirradiation at 254 nm. Moreover, the reversibility is potentially useful for photorelease of drugs in biological systems. However, the compatibility should be considered in the process, as photoirradiation at 254 nm for a long time (e.g., > 0.5 h) can damage DNA or photosensitive molecules, and possible side photoreactions should be avoided.
ODNs containing 1–3 were synthesized by using commercially available b-cyanoethyl phosphoramidites with phenoxyacetyl protecting groups on the exocyclic amines of dA and dG (Scheme 2). All functionalized ODNs (7 b, 8 b, 11 b, 12 b, 13 c, 14 c, and 15 c) were deprotected, cleaved from solid support with NH3 (28 % aq.) at room temperature for 2 h, and purified by 20 % denaturing PAGE, with new ODNs characterized by ITTOF-MS (Figures S15–S18).
Results and Discussion Synthesis of the crosslinking precursors and their incorporation into ODNs To identify coumarin analogues capable of forming ligation and ICL products, we designed and synthesized three coumarin analogues (1–3). Compounds 1 and 2 were used to study linker effects on reactivity; 2 and 3 were used to investigate the effect of substituents at position 4. As previously reported, compounds 2 and 3 were synthesized by a Pechmann condensation between resorcinol and ethyl acetoacetate or malic acid. Compound 1 was prepared from 2 by Williamson ether synthesis.[15a] Compounds 1–3 were converted to the corresponding phosphoramidites (4–6) under standard conditions, which were used for solid-phase DNA synthesis of modified ODNs (Scheme 1). The compounds were confirmed by NMR and HRMS (Figure S14 A–D in the Supporting Information).
Scheme 2. Double-stranded DNAs (dsDNAs) used in this study.
Scheme 1. Coumarins used in the study and synthesis of phosphoramidites 4–6.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
2
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers Optimization of ODN sequences
Table 1. Interstrand crosslink (ICL) formation reaction rates for dsDNAs 7–12.
To identify DNA sequences that can form both ICL and ligation products for the study, the ODN sequences were first optimized. Previously, we showed that a cycloaddition reaction did not occur between coumarin and dG. In addition, the linker of the coumarin played an important role in increasing interstrand crosslinking efficiency.[15a] For example, a quantitative DNA ICL was generated by using coumarin moieties with a flexible linker of two or more carbons, whereas coumarin analogues without the linker led to less efficient ICL formation. Thus, three DNA duplexes containing 1 with a two-carbon linker (dsDNA 7, dsDNA 9, and dsDNA 11) and different numbers of dGs at the terminus were designed to study the effect of dT positions on ICL formation. Another three duplexes containing 2 without the linker (dsDNA 8, dsDNA 10, and dsDNA 12) were employed to investigate the linker effect in this process (Scheme 2). Upon photoirradiation at 350 nm, all DNA duplexes formed ICL products but with different efficiency and reaction rates. Almost quantitative DNA ICL formation was observed for 1containing dsDNA 7 (ICL yield: 98 %) and dsDNA 9 (ICL yield: 97 %). A doublet of ICL products was observed for dsDNA 7 (line 2, Figure 1), which is ascribed to the varied crosslinking
k [10 dsDNA 7 dsDNA 8 dsDNA 9
These are not the final page numbers! ÞÞ
42.0 4.0 4.8 0.4 6.1 0.5
t1/2 [min] 30 24 2 19 2
k [10 dsDNA 10 dsDNA 11 dsDNA 12
4
s 1]
0.63 0.03 1.6 0.07 0.84 0.04
t1/2 [min] 184 9 67 2 139 6
Coumarin-induced ICL and ligation reactions with dT or dC Having optimized ODN sequences, we designed dsDNAs 13– 15 for investigating DNA interstrand crosslinking and ligation reactions. These DNA duplexes contain dT 48 as a flanking nucleotide that can form ligation products with coumarin analogues and dT 16 as an opposing nucleotide capable of generating ICL products. Thus, the coumarin analogues had equal opportunity for generating ICL products or ligation products. The ICL and ligation reactions were studied separately by labeling different ODNs with g-32P. The ICL products were detected by using 32P-labeled ODN-13 a; the ligation products were detected by using 32P-labeled-13 b. As expected, dsDNA 13 containing 1 with a flexible linker showed poor reactivity and a low reaction rate in the ligation reaction (Figures 2, S5 and S6), due to competition with more favorable ICL reactions between dT and 1. Irradiation of dsDNA
sites. As coumarin 1 contains a flexible linker, it is highly likely to react with various dTs close to 3’ terminus in ODN 7 a, resulting in ICL products with different migration. Previous reports also showed that the oligodeoxynucleotides, with the same length but different structures, could migrate distinctly in denaturing PAGE.[15a, 17] Although dT is three nucleotides away from 1 in dsDNA 11, the ICL yield is still relatively high (78 %). These data showed that 1 might not be a good candidate for forming ligation products as dTs in a variety positions of the complementary strand would compete to form DNA ICL products. In contrast, the ICL formation induced by 2 is more dependent on the distance and positions of dT, though less efficient than that induced by 1. About 23 % decrease of ICL yield was observed with dsDNA 10 containing dT that is one nucleotide away from coumarin 1, and 39 % decrease for dsDNA 12 with dT two nucleotides away from 1 (Figure 1). Kinetic study showed that installation of dG flanking to dT slowed down the crosslinking reaction. Introduction of one dG next to dT led to an approximately sevenfold decrease in ICL www.chembiochem.org
s 1]
reaction rate (42.0 4.0·10 4 s 1 for dsDNA 7 vs. 6.1 0.5·10 4 s 1 for dsDNA 9; Table 1 and Figures S1 and S2). We propose that dsDNAs with highest reactivity for ICL reactions have poor reactivity for ligation reaction, whereas those with modest interstrand crosslinking efficiency are expected to be more favorable for ligation. In addition, the crosslinking efficiency of coumarin analogues without linkers can be better tuned by variation of dT positions and flanking sequences. Thus, we expect that compounds 2 and 3 are suitable for further study about competition between ICL and ligation reaction, as well as rearrangement study.
Figure 1. Phosphorimage autoradiogram of denaturing PAGE analysis of distance-dependent ICL formation. C: Control sample containing 100 nm dsDNA 7 without photoirradiation; 100 nm dsDNAs 7–12 was irradiated at 350 nm; ODNs without coumarins were 5’-[32P]-labeled.
ChemBioChem 2016, 17, 1 – 9
4
Figure 2. Phosphorimage autoradiogram of denaturing PAGE analysis of A) ICL and B) ligation reactions. C: Control sample containing 100 nm dsDNA 13 without photoirradiation; 100 nm dsDNAs 13–15 was irradiated at 350 nm for 4 h. For the ICL reactions, ODN-13 a was 5’-[32P]-labeled, whereas ODN-13 b was 5’-[32P]-labeled for ligation reactions.
3
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers Table 2. The rates of ICL/ligation product formation or cleavage reaction of dsDNA 13–18.[a] dsDNA
k (ICL formation) [10 4 s 1]
t1/2 [h]
kc (ICL cleavage) [10 3 s 1]
t1/2 [min]
k (ligation formation) [10 4 s 1]
t1/2 [h]
kc (ligation cleavage) [10 3 s 1]
t1/2 [min]
13 14 15 16 17 18
5.87 0.18 0.96 0.06 1.57 0.06 0.61 0.06 0.31 0.03 0.58 0.06
0.33 0.01 2.01 0.12 1.22 0.04 3.18 0.28 6.30 0.55 3.37 0.35
n.d. 5.12 0.27 4.75 0.34 n.d. 8.79 0.71 3.98 0.33
n.d. 2.26 0.12 2.44 0.17 n.d. 1.32 0.11 2.92 0.25
2.55 0.44 1.21 0.13 1.67 0.16 1.35 0.19 0.55 0.05 0.42 0.04
0.78 0.14 1.61 0.17 1.16 0.11 3.54 0.30 3.54 0.30 2.70 0.25
n.d. 5.94 0.45 9.78 0.43 n.d. 6.58 0.79 6.24 0.66
n.d. 1.96 0.15 1.18 0.05 n.d. 1.78 0.21 1.87 0.20
[a] n.d.: not determined.
13 with 350 nm light mainly generated ICL products in 77.8 % yield (Figure 2 A, lane 2), leading to a low ligation efficiency (12.3 %; Figure 2 B, lane 2). The rate of the interstrand crosslinking reaction of dsDNA 13 is two times that of the ligation reaction. In contrast, comparable ICL and ligation yields were observed with dsDNA 14 or 15 containing 2 or 3, which lack a flexible linker unit (Figure 2, 30.8 % vs. 26.5 % for dsDNA 14; 30.2 % vs. 34.1 % for dsDNA 15). Thus, dsDNAs 14 and 15 are preferable for further investigation on rearrangement between ICL products and ligation products. Kinetic studies indicated that the ligation reaction proceeded a little faster than the ICL reaction for dsDNAs 14 and 15 (Table 2 and Figures S3–S6). These data suggested that photo-induced ligation was more kinetically controlled for dsDNAs 14 and 15 than the ICL reactions. Previous studies showed that dC also formed [2+2] cycloaddition products with coumarin analogues in DNA. In order to evaluate the generality of the phenomena observed above, we investigated the competition between the dC-coumarin ligation and dT-coumarin ICL reaction. Thus, dsDNAs 16, 17, and 18 were designed to have a dC flanking two coumarin analogues and a dT in the opposing strand, which allowed the coumarin analogues to ligate with dC and crosslink with dT. Similar to dT-coumarin ligation, compound 1 in dsDNA 16 showed very low ligation efficiency (ligation yield = 3.0 %), and the ICL reaction was predominant (30.9 %). However, more efficient ligation reactions were observed for 2 (dsDNA 17: 8.8 %) and 3 (dsDNA 18: 12.5 %) than 1 (dsDNA 16: 3 %), whereas a less efficient ICL reaction was observed for 2 (9.9 %) and 3 (23.6 %) than 1 (30.9 %; Figure 3). The rate of ICL and ligation reactions for dsDNA 17 were (3.08 0.27) 10 5 s 1 and (5.48 0.46) 10 5, respectively, and those for dsDNA 18 were (5.76 0.60) 10 5 s 1 and (4.20 0.40) 10 5, respectively (Table 2 and Figures S3–S6). To provide further evidence that coumarins are essential for ICL and ligation product formation, we synthesized two dsDNAs 19 and 20 that did not contain coumarin moieties and investigated their photoreactivity under the same conditions used for dsDNAs 13–18. No ICL or ligation adducts were observed for dsDNAs 19 and 20, indicating that coumarins played a crucial role for photoinduced DNA ICL formation and ligation with dsDNAs 13–18 (Figure S11).
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Figure 3. Phosphorimage autoradiogram of denaturing PAGE analysis of A) ICL and B) ligation reactions. C: Control sample contained 100 nm dsDNA 16 without photoirradiation; 100 nm; dsDNAs 16–18 was irradiated at 350 nm for 20 h. For the ICL reactions, ODN-16 a was 5’-[32P]-labeled, whereas ODN-16 b was 5’-[32P]-labeled for ligation reactions.
Photoreversibility and rearrangement of ligation products to ICL products The photocycloaddition reactions between coumarin analogues and dT or dC were reversible, and rearrangement of ligation products to ICL products was observed during the process. The photocycloaddition products generated by 350 nm irradiation were split into single-stranded ODNs at a yield of 70 % upon 254 nm irradiation. The cleavage reactions followed first-order kinetics and were complete within 10 min (Table 2 and Figures S7–S10). The cleavage reaction rates for ligation products were similar to those for ICL products. For instance, the rate constant for cleaving ligation adducts formed with dsDNA 14 is (5.94 0.45) 10 3 s 1; that for ICL adducts is (5.12 0.27) 10 3 s 1. Similar results were observed for dsDNAs 15, 17, and 18 (Table 2). The reversible behavior was observed within three cycles of photoirradiation at 350 nm for 20/4 h and 254 nm for 10 min (Figure 4). However, after each cycle 254 nm/350 nm irradiation, the yields for interstrand crosslinking products increased, whereas those for ligation products decreased. For example, the ICL yield for dsDNA 14 increased from 41.5 % in the first circle to 47.8 % in the third cycle, while the ligation yield decreased from 32.5 to 9.9 % during this process (Figure 4 A). Similarly, three cycles of 254 nm/350 nm irradiation led to a 16 % increase in ICL yield and a 12 % decrease in photoligation yield 4
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers
Figure 4. Photoreversible process over three cycles for A) dsDNA 14, B) dsDNA 15, C) dsDNA 17, and D) dsDNA 18 upon photoirradiation at 350 nm for 4 h for dsDNAs 14/15 or 20 h for dsDNAs 17/18 (UV1) and 254 nm (UV2) for 10 min.
was observed, which migrated faster than the original singlestranded ODN-15 c after the reversible process (Figure 5). The mass analysis suggested that SS ODN* was caused by photorelease of coumarin 3 in ODN-15 c (Scheme 3 B and Figure S19). To gain more details on the photorelease reaction, such as which wavelength led to the cleavage of coumarin and which adduct was cleaved, we performed a more detailed study with dsDNA 18, with dsDNA 17 for comparison. The dsDNAs were separately irradiated with only 254 nm irradiation, only 350 nm irradiation, and both 350/254 nm irradiation (Figure 5). The results suggested that the photorelease of coumarin moiety only occurred with the dC-3 cycloaddition adducts upon 254 nm irradiation. First, cleaved product SS ODN* was not observed following 350 nm irradiation of the ICL/ligation products, indicating the coumarin-dC adduct was stable to 350 nm irradiation (Figure 5, lanes 2 and 6). Second, the SS ODN* was not observed with 254 nm irradiation of dsDNA 17 or 18, thus indicating that the coumarin was not cleaved in the single-stranded ODN (Figure 5, lanes 3 and 7). Third, the SS ODN* was not observed for dsDNA 17, thus indicating that dC-2 adducts were not cleaved (Figure 5, lanes 2–4). In addition, the photorelease products (SS ODN*) were not observed for dsDNAs 14 and 15, thus indicating that dT-coumarin adducts were not cleaved (Figure S13). Thus, we concluded that 254 nm irradiation of dC-3 cyclobutane adduct 21 formed with dsDNA 18 led to photorelease of coumarin 3 and SS ODN* (Scheme 3 B). The Schmidt group reported that (coumarin-4-yl)methyl esters
for dsDNA 17. The same trend was observed for dsDNA 15. These results indicated that rearrangement from kinetic-controlled ligation products to thermo-controlled ICL products occurred in the photoswitchable process (Figure 4 A–C). To provide direct evidence for the rearrangement of ligation products to ICL products, we isolated the ligation products formed with dsDNAs 14 and 15, which were used for photoreversibility and rearrangement studies. As expected, photoirradiation of the isolated ligation products at 254 nm led to photocleavage of ligation adducts, generating the single-stranded ODN (Figure S12). The rearrangement from ligation products to ICL products was also observed in the photoreversible process (254 nm/350 nm irradiation). After two cycles of 254 nm/ 350 nm irradiation, the ligation adducts obtained with dsDNA 14 were rearranged to ICL products in 26.7 % yield, whereas the ligation product yield dropped to 18.7 %. Similarly, two cycles of 254 nm/350 nm irradiation of the ligation adducts obtained with dsDNA 15 produced 28.1 % ICL products and 13.4 % ligation products (Figure S12). Photorelease of coumarin 3 in dsDNA 18 Surprisingly, both ICL and ligation yields decreased for dsDNA 18 during the photoreversible process (Figure 4 D). The ICL yield decreased from 40.4 % to 34.7 %, and the ligation yield decreased from 14.4 to 3.3 % after three cycles of 254 nm/ 350 nm irradiation. Meanwhile, a new ODN band (SS ODN*) ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
5
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers
Figure 5. Phosphorimage autoradiogram of denaturing PAGE analysis of photorelease reaction for dsDNAs 17/18 upon treatment with irradiation at different wavelengths.
Scheme 3. Photorelease of coumarin 3 from dsDNA 18 after photoirradiation.
Conclusion
(CM-A) are caged compounds that, upon excitation, release hemagglutinin and the (coumarin-4-yl)methyl alcohol via heterolytic ester cleavage (Scheme 3 A).[18] Similarly, we proposed that heterolytic ester cleavage occurred with the dC-3 cyclobutane adduct upon irradiation at 254 nm, generating free coumarin 3, ODN 16 b, and SS ODN*.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Three coumarin analogues (1–3), with or without the linker units at position 4, were employed for DNA photoligation. Both linker units in coumarin moieties and DNA sequences affected the ligation efficiency, due to the competition of DNA 6
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Full Papers temperature overnight (total volume of 50 mL). The 32P-labeled oligonucleotide duplex (0.1 mm, 2 mL) was mixed with NaCl (1 m, 2 mL), potassium phosphate (100 mm, pH 7.0, 2 mL), and autoclaved distilled water (14 mL) to give a final volume of 20 mL. The DNA duplexes were photoirradiated at 350 nm to form ICLs or ligation products (a control reaction was carried out without photoirradiation). The aliquots were taken at the desired time and immediately quenched with the equal volume of formamide (95 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis. For kinetics study, aliquots (final concentration: 50 nm 32P-labeled oligonucleotide duplex, NaCl (100 mm), potassium phosphate (10 mm)) of three independent reactions were taken at the prescribed times, immediately quenched by formamide (90 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis. Photoirradiation at 350 nm was employed for ICLs or ligation reactions, whereas photoirradiation at 254 nm was used for photocleavage reactions.
ICL formation. Coumarin analogues (2 and 3) without linker units are more favorable for DNA photoligation than 1, which has a two-carbon chain linker. Compounds 2 and 3 generated comparable photoligation products and DNA ICL adducts, but 1 induced almost quantitative DNA ICL formation with a very low yield of ligation products. DNA photoligation induced by coumarin moieties is reversible. Irradiation at 350 nm produced ligation products that could revert to the original singlestranded ODNs upon 254 nm irradiation. Several cycles of 350 nm/254 nm irradiation led to rearrangement of kineticcontrolled ligation products to thermocontrolled ICL adducts. Photorelease of the coumarin moiety was observed with 254 nm irradiation of the dC-3 products but not for dT-3 adducts. DNA photoligation induced by the coumarin moiety, rearrangement of DNA adducts, and photorelease of the coumarin moiety can serve as a novel tool for DNA nanomaterial research or drug-release studies.[16]
Isolation of the ligation products: The hybridized ODNs (10 mm), with or without a-32P-labeled strands, were photoirradiated at 350 nm with the desired time to generate the ICLs or ligation products. The products were isolated by using 20 % denaturing PAGE with a-32P-labeled ones as standards. No UV light at 254 nm can be used for product detection; otherwise photocleavage is induced. The ligation products were purified for further study via standard protocol. It should be mentioned that the hybridization of ligation products from 80 8C can result in a little DNA damage.
Experimental Section General methods: Unless otherwise specified, reagents were used directly as received without further purification. Columns and phosphoramidites for DNA synthesis were purchased from Glen Research. T4 polynucleotide kinase was obtained from New England Biolabs, and [g-32P]-ATP was from PerkinElmer Life Sciences. Water was purified via a Milli-Q purification system before use for preparation of ODN solutions.
Reversibility study: The hybridized ODNs were photoirradiated at 350 nm with the desired time to generate the ICLs or ligation products. Products formed by 350 nm irradiation were photocleaved upon photoirradiation at 254 nm with the required time estimated from kinetic studies. Aliquots were taken at the prescribed steps and immediately quenched with the equal volume of formamide (95 %) loading buffer, and stored at 20 8C until subjecting to 20 % denaturing PAGE analysis.
ODNs were synthesized with standard automated DNA synthesis techniques on a 1.0 mmol scale by using commercially available 1000 CPG-succinyl-nucleoside supports. The iPr-Pac-dA and iPrPac-dG phosphoramidites were used for the preparation of coumarin-containing ODNs. Deprotection of the nucleobases and phosphate moieties, as well as cleavage of the linker for normal ODNs, was carried out in a mixture of MeNH2 (40 % aq.) and NH3 (28 % aq.; 1:1) at room temperature for 2 h. Functionalized ODNs were deprotected and cleaved with NH3 (28 % aq.) at room temperature for 2 h. All crude ODNs were purified by 20 % denaturing polyacrylamide gel electrophoresis (PAGE) and then quantified in water. Radiolabeling was performed according to standard protocols.[19] The 32P-labeled ODN (100 nm) was annealed with 1.5 equiv of the complementary strand and related strands by heating to 80 8C for 3 min in a buffer of potassium phosphate (10 mm, pH 7.0) and NaCl (100 mm), followed by cooling slowly to room temperature overnight. Radiolabeled ODNs were quantified by using a Molecular Dynamics Phosphorimager.
7-Hydroxy-2 H-chromen-2-O-(2-cyanoethyl-N,N-diisopropyl)phosphoramidite (6): N,N-diisopropylethylamine (DIPEA; 156 mL, 0.9 mmol), and 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (167 mL, 0.75 mmol) were added to a solution of 7-hydroxy2H-chromen-2-one (3, 81 mg, 0.5 mmol) in CH2Cl2 (10 mL) under argon. The reaction mixture was stirred at room temperature for 3 h and diluted with CH2Cl2 (50 mL). The organic layer was washed with NaHCO3 (20 mL) and saturated aqueous NaCl (20 mL) and dried over anhydrous sodium sulfate. The solvent was removed under reduced pressure. Upon purification by column chromatography (EtOAc/hexane/Et3N 59:40:1), the product, 8, was isolated as a white solid (138 mg, 76 %). 31P NMR (122 MHz, CDCl3): d = 147.59. 1 H NMR (300 MHz, CDCl3): d = 7.56, 7.59 (d, J = 9.6 Hz, 1 H), 7.30– 7.33 (d, J = 8.4 Hz, 1 H), 6.90–6.94 (m, 2 H), 6.20, 6.24 (d, J = 9.6 Hz, 1 H), 3.87–3.90 (m, 2 H), 3.63–3.71 (m, 2 H), 2.61–2.65 (t, J = 6.3 Hz, 2 H), 1.10–1.19 (dd, J = 3.6 Hz, 12 H); 13C NMR (75 MHz, CDCl3): d = 160.0, 156.9, 156.8, 154.3, 142.3, 127.7, 116.3, 115.8, 115.7, 113.0, 106.4, 106.3, 58.3, 58.0, 43.0, 42.9, 23.6, 23.5, 23.4, 19.4, 19.3; HRMS-ESI (+): m/z calcd for C18H24N2O4P: 363.1468 [M+H] + ; found: 363.1470.
Photoirradiation of samples in colorless siliconized microcentrifuge tubes at 350 nm was carried out in a Rayonet Photochemical Chamber Reactor (Model RPR-100) equipped with 12 lamps, and photoirradiation at 254 nm was conducted with a Compact UV Lamp (UVGL-25) by using the short wavelength at room temperature. 1H NMR, 13C NMR, and 31P NMR analyses were performed on a 300 MHz spectrophotometer. Chemical shifts are reported in ppm relative to Me4Si (1H and 13C) or H3PO4 (31P). HRMS was performed at the Mass Spectrometry Laboratory of the University of Wisconsin–Milwaukee.
Acknowledgements
Photo-induced DNA ligation or ICL formation by coumarin moiety and kinetic studies: 5’-32P-labeled or 3’-32P-labeled ODNs (0.1 mm) were annealed with 1.5 equiv of the complementary strands in NaCl (100 mm) and potassium phosphate (10 mm, pH 7) by heating to 65 8C for 3 min, followed by cooling slowly to room
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
We are grateful for financial support from the University of Wisconsin-Milwaukee (UWM) Research Growth Initiative (RGI101234), the Greater Milwaukee Foundation (Shaw Scientist Award), and 7
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Full Papers a Wisconsin Applied Research Grant (ARG) Award. We thank Dr. Zhiqiang (Mark) Wang for mass analysis.
[11] a) K. Fujimoto, A. Yamada, Y. Yoshimura, T. Tsukaguchi, T. Sakamoto, J. Am. Chem. Soc. 2013, 135, 16161 – 16167; b) Y. Yoshimura, K. Fujimoto, Org. Lett. 2008, 10, 3227 – 3230; c) K. Fujimoto, K. Konishi-Hiratsuka, T. Sakamoto, Y. Yoshimura, ChemBioChem 2010, 11, 1661 – 1664. [12] W. Szyman´ski, J. M. Beierle, H. A. Kistemaker, W. A. Velema, B. L. Feringa, Chem. Rev. 2013, 113, 6114 – 6178. [13] a) J. Woo, P. B. Hopkins, J. Am. Chem. Soc. 1991, 113, 5457 – 5459; b) A. K. Thazhathveetil, S. T. Liu, F. E. Indig, M. M. Seidman, Bioconjugate Chem. 2007, 18, 431 – 437; c) M. Takasugi, A. Guendouz, M. Chassignol, J. L. Decout, J. Lhomme, N. T. Thuong, C. Helene, Proc. Natl. Acad. Sci. USA 1991, 88, 5602 – 5606; d) A. Majumdar, P. A. Muniandy, J. Liu, J. I. Liu, S. T. Liu, B. Cuenoud, M. M. Seidman, J. Biol. Chem. 2008, 283, 11244 – 11252; e) T. E. Haran, D. M. Crothers, Biochemistry 1988, 27, 6967 – 6971. [14] a) B. D. Wagner, Molecules 2009, 14, 210 – 237; b) H. Li, L. Cai, Z. Chen, Advances in Chemical Sensors (Ed.: W. Wang), InTech, Rijeka, 2012, pp. 121 – 150. [15] a) H. Sun, H. Fan, X. Peng, J. Org. Chem. 2014, 79, 11359 – 11369; b) M. M. Haque, H. Sun, S. Liu, Y. Wang, X. Peng, Angew. Chem. Int. Ed. 2014, 53, 7001 – 7005; Angew. Chem. 2014, 126, 7121 – 7125. [16] S. R. Trenor, A. R. Shultz, B. J. Love, T. E. Long, Chem. Rev. 2004, 104, 3059 – 3077. [17] Q. Zhai, P. Wang, Q. Cai, Y. Wang, Nucleic Acids Res. 2014, 42, 10529 – 10537. [18] R. Schmidt, D. Geissler, V. Hagen, J. Bendig, J. Phys. Chem. A 2007, 111, 5768 – 5774. [19] T. F. Maniatis, E. F. Sambrook, J. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, New York, 1982.
Keywords: coumarins · DNA rearrangement · interstrand crosslink · photoligation · photorelease [1] F. Wang, X. Liu, I. Willner, Angew. Chem. Int. Ed. 2015, 54, 1098 – 1129; Angew. Chem. 2015, 127, 1112 – 1144. [2] a) D. Y. Zhang, E. Winfree, J. Am. Chem. Soc. 2009, 131, 17303 – 17314; b) D. Y. Zhang, G. Seelig, Nat. Chem. 2011, 3, 103 – 113. [3] a) W. Shih, Nat. Mater. 2008, 7, 98 – 100; b) J. Bath, A. J. Turberfield, Nat. Nanotechnol. 2007, 2, 275 – 284. [4] S. Modi, D. Bhatia, F. C. Simmel, Y. Krishnan, J. Phys. Chem. Lett. 2010, 1, 1994 – 2005. [5] a) D. M. Noll, T. M. Mason, P. S. Miller, Chem. Rev. 2006, 106, 277 – 301; b) S. Cao, X. Peng, Curr. Org. Chem. 2014, 18, 70 – 85; c) M. Freccero, Mini-Rev. Org. Chem. 2004, 1, 403 – 415. [6] a) H. Wang, S. E. Rokita, Angew. Chem. Int. Ed. 2010, 49, 5957 – 5960; Angew. Chem. 2010, 122, 6093 – 6096; b) H. Wang, Curr. Org. Chem. 2014, 18, 44 – 60; c) W. B. Sherman, N. C. Seeman, Nano Lett. 2004, 4, 1203 – 1207; d) C. Perez, M. Leng, J. M. Malinge, Nucleic Acids Res. 1997, 25, 896 – 903; e) J. Malina, J. Kasparkova, N. P. Farrell, V. Brabec, Nucleic Acids Res. 2011, 39, 720 – 728; f) F. Fakhari, S. E. Rokita, Nat. Commun. 2014, 5, 5591. [7] M. Di Antonio, F. Doria, S. N. Richter, C. Bertipaglia, M. Mella, C. Sissi, M. Palumbo, M. Freccero, J. Am. Chem. Soc. 2009, 131, 13132 – 13141. [8] a) B. J. Parsons, Photochem. Photobiol. 1980, 32, 813 – 821; b) Y. H. Ping, T. M. Rana, Biochemistry 2005, 44, 2501 – 2509. [9] H. Asanuma, T. Ito, T. Yoshida, X. Liang, M. Komiyama, Angew. Chem. Int. Ed. 1999, 38, 2393 – 2395; Angew. Chem. 1999, 111, 2547 – 2549. [10] H. Kashida, T. Doi, T. Sakakibara, T. Hayashi, H. Asanuma, J. Am. Chem. Soc. 2013, 135, 7960 – 7966.
&
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
Manuscript received: April 21, 2016 Accepted article published: August 25, 2016 Final article published: && &&, 0000
8
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
FULL PAPERS UV control: Coumarin analogues were shown to photo-crosslink dT- or dCforming ligation products in addition to interstrand crosslink (ICL) adducts. The photoligation was photoswitchable: ligation products formed at 350 nm irradiation reverted to the original singlestranded ODNs upon 254 nm irradiation. Rearrangement of ligated ODNs into ICL products also occurred during the switchable processes.
ChemBioChem 2016, 17, 1 – 9
www.chembiochem.org
These are not the final page numbers! ÞÞ
H. Sun, H. Fan, H. Eom, X. Peng* && – && Coumarin-Induced DNA Ligation, Rearrangement to DNA Interstrand Crosslinks, and Photorelease of Coumarin Moiety
9
2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
