Protected 5-(hydroxymethyl)uracil nucleotides bearing visible-light photocleavable groups as building blocks for polymerase synthesis of photocaged DNA †

Nucleosides, nucleotides and 2 ’ -deoxyribonucleoside triphosphates (dNTPs) containing 5-(hydroxy-methyl)uracil protected with photocleavable groups (2-nitrobenzyl-, 6-nitropiperonyl or 9-anthrylmethyl) were prepared and tested as building blocks for the polymerase synthesis of photocaged oligonucleotides and DNA. Photodeprotection (photorelease) reactions were studied in detail on model nucleoside monophosphates and their photoreaction quantum yields were determined. Photocaged dNTPs were then tested and used as substrates for DNA polymerases in primer extension or PCR. DNA probes containing photocaged or free 5-hydroxymethylU in the recognition sequence of restriction endonucleases were prepared and used for the study of photorelease of caged DNA by UV or visible light at di ﬀ erent wavelengths. The nitropiperonyl-protected nucleotide was found to be a superior building block because the corresponding dNTP is a good substrate for DNA polymerases, and the protecting group is e ﬃ ciently cleavable by irradiation by UV or visible light (up to 425 nm).


Introduction
Photocleavable protecting groups 1 are gaining growing attention due to their applications in organic synthesis, chemical biology and many other areas. In biomolecules, photocaging 1,2 and cleavage of the photoremovable protective groups by light, which is both biocompatible and bioorthogonal, is often used for triggering or switching biological processes. Photocaging of nucleic acids 3 is a particularly active field. The photocaging on the nucleobase (mostly at N3 or exocyclic O 4 /N 4 atoms of pyrimidines) is used 4 to prevent the base-pairing of oligonucleotides which only after photodeprotection can hybridize with the complementary strand. On the other hand, photocaging of the phosphate backbone 5 increases the resistance to hydrolysis and influences the interactions with other biomolecules. Photochemical deprotection can therefore either trigger or block biological processes, i.e., primer extension, transcription, antisense effect, RNA interference, translation, etc. It is important to mention that all these types of caged oligonucleotides and nucleic acids can only be synthesized chemically because the corresponding caged (deoxy)nucleoside triphosphates (NTPs or dNTPs) are not suitable substrates for polymerases. [3][4][5] On the other hand, photocaging in the majorgroove site of nucleotides and nucleic acids has been studied less frequently. Photocaged 5-(aminopropargyl)-pyrimidine dNTPs were used for polymerase synthesis of DNA and photochemical release of the reactive amino functions in the major groove. 6 Nitrobenzyl-(NB), and -phenylethyl-caged 5-(hydroxymethyl)pyrimidine or 7-hydroxymethyl-7-deazapurine 2′-deoxyribonucleoside triphosphates (dNTPs) were used as reversible terminators of primer extension in sequencing. 7 On the other hand, 6-O-(2-nitrobenzyl)guanosine and 4-O-(2-nitrobenzyl)uridine triphosphates 8 were not substrates for RNA polymerases until photochemical decaging.
5-(Hydroxymethyl)uracil (5hmU) is a rare DNA base previously found in bacteriophages, 9 microorganisms, 10 and also eukarytotic genomes. 11 The Schultz group has even evolved 12 a bacterium which replaced thymines with 5hmU in genome showing that this base is biocompatible. Unlike the related 5-(hydroxymethyl)cytosine (5hmC), 13 which is a known epigenetic mark and intermediate in active demethylation of DNA, 14 the biological role of 5hmU is not known so far. 15 In order to create an artificial epigenetic system for the regulation or switching of gene expression, we have systematically studied 16 the influence of chemical modifications in the major groove of DNA on bacterial transcription and found that some small modifications are tolerated by the bacterial RNA polymerase and transcription factor and the modified DNA still serves as a template for transcription. Obviously, any bulkier modifications in the major groove inhibited the transcription. 16 Recently we found 17 that the presence of 5hmU or 5hmC in template DNA can even enhance (up to 3.5 times) or inhibit bacterial transcription (depending on the promoter) and thus even hmU could be an epigenetic mark.
Previously, we have also studied the influence of majorgroove modifications on recognition and cleavage of DNA by restriction endonucleases (REs) and found that some small modifications at T/U 18 or A 19 are tolerated (and the REs cleave the modified DNA), whereas all bulky modifications or any modifications at G 20 or C 18 bases inhibit the cleavage. We also developed 21 transient chemical protection of DNA from RE cleavage by polymerase incorporation of bulky 7-(triethylsilylethynyl)-7-deazaadenine which, after desilylation, released 7-ethynyl-7-deazaadenine-modified DNA being fully cleavable by several REs. In order to develop a more biocompatible and bioorthogonal switch, we recently reported 22 the use of 5-[(2nitrobenzyl)oxymethyl]uracil dNTP as a substrate for DNA polymerases leading to the enzymatic synthesis of DNA photocaged in the major-groove. The presence of the bulky NB group in the major groove blocked specific interactions and cleavage by restriction endonucleases (REs) but, after photodeprotection by irradiation using UV light (365 nm), the 5-hydroxymethyluracil-containing DNA was recognized and cut by REs. It was a good proof of the principle, but for any in cellulo or in vivo applications, the use of UV light is not desirable 1b because of its toxicity and limited penetration through tissues. Therefore, it is desirable to develop alternative photocaging groups for 5hmU which would be compatible with the synthesis of modified dNTPs, polymerase incorporation and would be cleavable by visible light. Major challenges in the design and development of new photocaging groups for nucleoside triphosphates are (i) limited chemical stability of some groups (e.g. carbonates) during the triphosphorylation steps and (ii) insufficient substrate activities of some very bulky dNTPs for polymerase incorporations to DNA (in particular by PCR). For example, the nitrophenethyl-caged dNTPs were incorporated by polymerases but terminated the primer extension. 7 Here we report on the synthesis of 5hmU dNTPs protected with 6-nitropiperonyl (NP) 23 and anthryl-9-methyl (An) 24 groups and comparison of substrate activities and photocleavage reactions with parent NB-caged nucleotides.

Synthesis
Although there were some reports on selective modification of hydroxymethyl groups in 5-hydroxymethyl-2′-deoxyuridine, 25 for the desired ether formation in the introduction of the photocaging groups, we chose an opposite approach based on nucleophilic substitution of a suitably protected 5-bromomethyl-2′deoxyuridine derivative (Scheme 1). In analogy to the previous related studies, 7,22,26 we started from 3′,5′-bis-O-(TBDMS)-3-Bocprotected 5-bromomethyl-2′-deoxyuridine 1 and performed substitution reactions with the corresponding benzyl-type alcohol in the presence of 2,2,6,6-tetramethylpiperidine in CH 3 CN at 50°C. The addition of AgOTf was important for the precipitation of AgBr and thus shifting the equilibrium towards products. The desired fully protected caged nucleosides 2a and 2b were isolated in moderate yields (10 and 21%). The An-caged intermediate 2c could not be isolated in pure form and was used in the next step in the crude form. Subsequently, the BOC group was removed using sodium bicarbonate 27 in MeOH at 50°C, whereas the common reagents such as CF 3 COOH or TBAF failed. The TBDMS group was then removed using Et 3 N·3HF 28 in THF. The desired photocaged nucleosides dU NB and dU NP were isolated in good yields of 64 and 57%, whereas the dU An was isolated in low 9% overall yield.
Phosphorylation reactions were carried out with POCl 3 in trimethyl phosphate at 0°C. Quenching of the reaction with Scheme 1 Synthesis of photocaged 2'-deoxyuridine monophosphates and triphosphates (i) ROH, 2,2,6,6-tetramethylpiperidine, silver triflate, CH 3 CN (ii) NaHCO 3 , MeOH (iii) Et 3 N.3HF, THF (iv) POCl 3 , PO(OMe) 3 , (NHBu 3 ) 2 H 2 P 2 O 7 (v) POCl 3 , PO(OMe) 3 , 2 M TEAB. *Prepared according to the literature. 22 2 M TEAB gave the desired monophosphates dU NB MP, dU NP MP and dU An MP in moderate yields after isolation by HPLC. To form dNTPs, phosphorylation was carried out in the same way followed by the addition of pyrophosphate and tributyl amine in DMF. The targets dU NB TP, dU NP TP and dU An TP were obtained in moderate yields after isolation by HPLC. The uncaged 5-hydroxymethyl-2′-deoxyuridine triphosphate dU hm TP was prepared in 27% yield by the photolysis of dU NP TP (with 59% recovery of the starting material) in water using irradiation by UV LEDs at either 375 or 385 nm followed by the separation on HPLC (Scheme 2).

Photophysical properties of modified monophosphates and quantum yields of uncaging reactions
Initially, we recorded the absorption spectra of the model caged monophosphates dU R MP in water and calculated the molar absorption coefficients for selected wavelengths in the range of 355-437 nm (Table 1 and Table S1 and Fig. S1-S7 in the ESI †). The parent dU NB MP exerted an absorption maximum at 264 nm and showed a very low absorption above 300 nm. At 365 nm (under which the NB group is cleaved 22 ), the molar absorption coefficient is only 238 M −1 cm −1 ( Table 1, Table S1 and Fig. S3 in the ESI †). On the other hand, the nitropiperonyl-caged nucleotide dU NP MP has a strong absorption maximum at 358 nm and still a relatively significant absorption at 400 nm (1940 M −1 cm −1 , Table 1 and Fig. S4 in the ESI †). dU An MP showed three absorption maxima at 349, 367 and 386 nm as expected for an anthracene moiety. The spectra of dU NP MP and dU An MP (Table S1, Fig. S4 and S6 in the ESI †) indicate that photochemical deprotection can be possible using light of wavelengths up to 425 nm.
Then we determined the quantum yields (QY) of uncaging. The measurements were carried out at 365 nm in water with a ferrioxalate actinometer 29 according to published procedures 30 giving an absolute value of QY. The QY of photocleavage of dU NP MP was moderate with the value of 4.4 ± 0.8%, whereas the QY value for dU An MP was negligible (in the range of an experimental error). Our attempts to measure QYs of previously published NB-group-protected dU NB MP failed because of two factors. (a) The molar absorption coefficient of dU NB MP at 365 nm was too low and the measurements had to be carried out in too concentrated solutions, and (b) the nitrosobenzene photoproduct formed during the reaction strongly absorbs at 365 nm. The fluorescence quantum yields of modified monophosphates in water were also determined at the excitation wavelength of 365 nm using quinine sulfate dissolved in 0.5 M H 2 SO 4 as a standard 31 giving relative values of QYs. U An MP is moderately fluorescent with the quantum yield of 4.48 ± 0.15%; U NB MP and U NP MP are not fluorescent.

Incorporation of caged nucleotides into DNA
Caged triphosphates dU NB TP, dU NP TP and dU An TP as well as the uncaged dU hm TP were tested as substrates for DNA polymerases. 32 At first, the primer extension (PEX) reaction was tested using a 19-mer template (for sequences of ONs see Table S4 in the ESI †) encoding for the incorporation of one modified dU R followed by three guanines. Fig. 1a shows the successful incorporation of all modified dU R nucleotides using KOD XL (for successful PEX using Vent (exo-) polymerase, see Fig. S12 in the ESI †). The extended modified oligonucleotide (ON) products were confirmed by the MALDI-TOF analysis showing masses corresponding to U hm -containing ONs (Fig. S13 in the ESI †) because the laser irradiation using in MALDI cleaves the photoremovable groups. In the case of An modification, both caged and uncaged products were detected (Fig. S14 in the ESI †). We also performed a simple kinetics of single nucleotide extension ( Fig. S15 and S16 in the ESI †) showing that the incorporation of modified dU R TPs was only slightly slower compared to that of TTP. Then, the PEX reaction was performed with a more demanding 31-mer template designed for the incorporation of 4 dU R modifications (Fig. 1b). This PEX proceeded successfully with dU hm TP, dU NB TP and dU NP TP giving a clean fully extended product. U AN TP appeared to be a slightly worse substrate for DNA polymerase giving the mixture of extended products also containing truncated ONs. The worse outcome was probably due to the bulkiness and hydrophobicity of the anthracene group.
Next, we tested dU hm TP, dU NB TP, dU NP TP and dU An TP in PCR using either a 98-mer or (a more demanding) 297-mer template. In agreement with our previous communications, 17,22 the PCR reactions with dU hm TP and dU NB TP, as well as with the new nitropiperonyl-caged dU NP TP, proceeded Scheme 2 Synthesis of dU hm TP. but the products of amplification were rather weak. The agarose gel of the 98-mer PCR (Fig. 1c) shows the bands of the modified PCR products only after the concentration of the sample, whereas the longer PCR ( Fig. 1d) showed only very weak bands with somewhat faster mobility even after the concentration of the sample. However, in all cases, the identity of the modified amplicons was confirmed by sequencing. The bulky and hydrophobic dU An TP did not give any amplification products in either of the PCR reaction.

Photochemical deprotection of caged DNA
In order to study the photocleavage of caged DNA, we used the previously described 22  . Uncaging experiments at 437 nm and higher wavelengths could not be tested because of the photobleaching of the 6-FAM used for labeling and visualization of ONs on gels (Fig. S17 †). Considering the negligible absorbances of dU R MPs at 437 nm (see Table S1 in the ESI † for molar absorption coefficients at 437 nm), we do not expect that liberation would proceed at 437 nm and higher wavelengths. Fig. 2 shows the results of photochemical deprotection reactions of modified DNA at 375 and 425 nm (b), whereas Table 1 (and Fig. S22-S27 in the ESI †) gives complete data at all tested wavelengths. ImageJ software was used to quantify the ratios of uncaged DNA (Table 1). Because of different optical power of the different LEDs, the reaction times differed; we tested at least 3 different irradiation times to reach (almost) complete conversion when possible. In all cases, the observed photochemical cleavage of the An group was very inefficient (even at 355 nm) and the conversions were very low (up to 15% even with prolonged time). The cleavage of NB and NP groups proceeded with comparable efficiencies at lower wavelengths (355-385 nm) reaching good conversions (92-95%) in reason- able reaction times (15-120 min). Only at higher wavelengths (400 and 425 nm), the nitropiperonyl group was cleaved significantly faster. At the highest wavelength of 425 nm, the cleavage of the nitrobenzyl group did not reach full conversion, whereas the cleavage of the NP group was still almost quantitative. Taking into account the almost negligible absorbance of the nitrobenzyl group at 400 and 425 nm, it was rather surprising that the NB group was still at least partly cleavable. However, as expected, uncaging of NPs was much faster and more efficient at these wavelengths, and we conclude that this group can be removed with visible light (Table 2).

Conclusions
We prepared three types of photocaged derivatives of 2′-deoxy-5-(hydroxymethyl)uridine nucleoside, mono-and triphosphates protected by 2-nitrobenzyl-, 6-nitropiperonyl-and anthryl-9-methyl groups, and tested their enzymatic incorporations to DNA and photochemical deprotections. The NB-and NP-caged dNTPs (dU NB TP and dU NP TP) were good substrates for DNA polymerases and worked in both PEX and PCR reactions. The An-linked dU An TP was a less efficient substrate but it was still possible to incorporate it to at least some sequences using PEX. The liberation of photocaged DNA was tested by monitoring the cleavage of DNA with REs. We show that both NB-and NP-caged DNA can be efficiently deprotected by UV irradiation, whereas the nitropiperonyl group is suitable for uncaging with visible light at 400 or 425 nm. Therefore, the NP-caging has a better potential for in cellulo or in vivo photocaging of 5-(hydroxymethyl)uracil in DNA. The An group does not apparently have much potential in DNA photocaging but still the An-substituted dNTP can find some use in polymerase synthesis of DNA bearing a fluorescent extended aromatic system for applications in imaging or selection of hydrophobic aptamers.
Reaction mixtures were separated using 12.5% denaturing PAGE.

Conflicts of interest
There are no conflicts to declare.