Nucleic Acid Conformation Influences Postsynthetic Suzuki–Miyaura Labeling of Oligonucleotides

Abstract

Chemoselective transformations that work under physiological conditions have emerged as powerful tools to label nucleic acids in cell-free and cellular environments. However, detailed studies investigating the influence of nucleic acid conformation on the performance of such chemoselective nucleic labeling methods are less explored. Given that nucleic acids adopt complex structures, it is highly important to study the scope of the chemical modification method in the context of nucleic acid conformations. Here we report a systematic study on the effect of local conformation on the postsynthetic Suzuki—Miyaura functionalization of human telomeric (H-Telo) DNA repeat oligonucleotide (ON) sequences, which form multiple G-quadruplex (GQ) structures. 5-Iodo-2’-deoxyuridine (IdU)-modified H-Telo ONs were synthesized by the solid-phase method, and when subjected to Suzuki—Miyaura cross-coupling reaction, its efficiency was found to depend on the type of conformation and the position of IdU label in different loops of the GQ structure. IdU-labeled GQs gave better yields as compared to single-stranded random coil structures. However, the IdU-labeled duplex under different ionic conditions did not undergo the coupling reaction. Further, using this method, we directly installed an environment-sensitive fluorescent probe, which photophysically reported the formation as well as distinguished different GQ topologies of telomeric repeat. Collectively, this systematic study underscores the influence of nucleic acid conformation, which has to be taken into account when establishing postsynthetic chemoselective functionalization strategies.

Introduction

Probing the structures of nucleic acids and their recognition properties is highly important, and in this context, oligonucleotides (ONs) labeled with biophysical probes greatly facilitate the basic understanding of structure—function relationships and also provide ways to develop diagnostic and therapeutic tools.^1–4 A technique, which has emerged in the past decade as a very power tool for labeling nucleic acids in vitro and in vivo is postsynthetic chemical labeling by using chemoselective reactions.^5–9 A reactive handle, which is compatible under chemical or enzymatic labeling methods, is introduced into ONs and further chemoselective reaction is carried out to functionalize the ONs with the desired tag. Few such reactions including azide—alkyne cycloaddition,^10–18 Staudinger^15,19–21 and inverse electron-demand Diels—Alder (IEDDA)^22–26 reactions have been demonstrated under biologically benign conditions and hence termed as bio-orthogonal reactions. These reactions have provided modern tools to (i) study nucleic acid localization, trafficking, and profiling, (ii) develop diagnostic and therapeutic agents, and (iii) develop functional nanomaterials.^27–29 While many of these benign chemoselective transformations have been established with different nucleic acid sequences and under different conditions including cell-free and cellular environments, the influence of nucleic acid conformation on the performance of such chemoselective reactions has not been well addressed. Such an effort will provide an understanding of the scope of a given postsynthetic reaction in modifying ONs, as different nucleic acid conformations are likely to affect the environment (e.g., steric, electronic, hydrophobic—hydrophilic balance, etc.) around the reactive handle and hence its reactivity with the cognate reaction partner.

The photoreactivity of 5-halouracil has been used to probe local DNA conformations such as A, B, and Z forms of the DNA duplex and different topologies of G-quadruplexes (GQ).³⁰ UV irradiation of 5-halouracil-containing ONs results in the formation of a uracil-5-yl radical, which in an atomspecific and conformation-dependent manner induces hydrogen abstraction and intrastand cross-links.³¹ Further, ligand-sensitized photochemical reactions with 5-bromouracil-modified GQ-forming ONs enabled the “photofootprinting” of GQ folding topologies.³² Per se this is not a labeling approach, and the end products including the deoxyribonolactone and side reactions such as strand cleavage limit the downstream utility of the ON products. GQ-forming sequences have also been used as scaffolds for carrying out template-assisted azide— alkyne cycloaddition reactions.³³ In a similar approach, GQ nanotemplates have been designed to facilitate regioselective formation of triazole products via click reaction between a series of alkyne and azide substrates.³⁴ Only the substrates that bind to a given GQ conformation undergo click reaction, thereby offering a target-guided synthesis approach to discover target-selective ligands. More recently, a chemoselective DNA labeling method was established by reacting 5-vinyl-2′-deoxyuridine-modified ONs with triazolinediones.³⁵ This reaction was found to be conformation-selective with the GQ structure giving higher yields compared to duplex DNA. However, the use of these reactions in attaching probes that can provide spectrum read-outs of a nucleic acid conformation was not explored and is not straightforward.

Chemoselective reactions mostly result in the formation of residual chemical groups (e.g., triazole), which however is not a major problem in many bioconjugation strategies but is not desirable when direct installation of biophysical probes is required. In this context, palladium-mediated C—C bond forming reactions have become popular in directly conjugating probes onto biomacromolecules including DNA ONs.^36–41 We recently established a chemoselective posttranscriptional functionalizing method to label RNA with biophysical probes by using Suzuki—Miyaura reaction.⁴² This technique allowed the direct synthesis of short RNA transcripts containing fluorogenic and affinity tags from 5-iodouridine-modified transcripts. Given the usefulness of Pd-mediated cross-coupling reactions in functionalizing ONs, we sought to explore the influence of nucleic acid conformational space on the efficiency of postsynthetic cross-coupling reactions.

Here we report the investigation on the effect of nucleic acid conformation on the efficiency of Suzuki—Miyaura coupling reaction using polymorphic GQ structures as the study model. 5-Iodo-2’-deoxyuridine (IdU) phosphoramidite was incorporated into human telomeric (H-Telo) DNA ONs, which upon annealing under different ionic conditions and in the presence of a synthetic molecular crowding agent (PEG) adopted different GQ topologies (Figure 1). Notably, the efficiency of coupling reaction was found to depend on the type of GQ conformation and the position of IdU in different loop regions. However, the Suzuki reaction did not work on the duplex structure of the telomeric DNA repeat. The environmentsensitive fluorescent probe-labeled telomeric ONs thus prepared faithfully reported the formation as well as distinguished different GQ topologies.

Figure 1.

(A) Synthesis of 5-iodo-2’-deoxyuridine (IdU) phosphoramidite 3 for solid-phase synthesis of H-Telo DNA ONs. DMTr = 4,4’-dimethoxytrityl, DMAP = 4-(dimethylamino)pyridine. (B) Flow diagram illustrating the synthesis of IdU (1)-labeled H-Telo DNA ONs, formation of respective GQstructures and their influence on postsynthetic Suzuki—Miyaura cross-coupling reaction. Position of IdU label in the loop region of each of the structures is shown in blue circle. The efficiency of Suzuki coupling to generate ONs labeled with an environment-sensitive probe depended on the GQ conformation. PDB ID is given in parentheses.

Results and Discussion

Synthesis and Incorporation of IdU Phosphoramidite into H-Telo DNA ONs

To evaluate the effect of nucleic acid conformation on the efficiency of Suzuki reaction, we chose H-Telo repeat sequence (TTAGGG)_n as the study system for the following reasons. H-Telo repeat sequence exhibits structural polymorphism, as it can adopt multiple structures depending on the ionic conditions, molecular crowding, and confined environments (Figure 1B).^43–46 In the presence of Na⁺ ions, it adopts an antiparallel GQ structure, whereas in the presence of K⁺ ions, it adopts mainly hybrid-type parallel-antiparallel structures (hybrids 1 and 2).⁴⁷ By slightly tweaking the 5’ and 3’ end nucleotides, one can also bias the repeat sequence to predominantly adopt either hybrid 1 or hybrid 2 structure.^48,49

In the presence of a synthetic crowding agent, polyethylene glycol (PEG), the sequence adopts an all-parallel stranded GQ structure, irrespective of the ionic conditions.⁵⁰ Further, in different GQ structures of H-Telo repeat, there are differences in the conformations of the loop nucleotides (TTA). The loop residues, especially T, replaced with an nucleoside analogue probe photophysically reported the formation of different GQstructures.^51–53 Similarly, a dual-app probe composed of a fluorophore and ¹⁹F label (fluorobenzofuran-modified nucleoside) placed in the loop position enabled two-channel detection of different GQ forms by fluorescence and NMR techniques.⁵⁴ Taking a cue from these observations, we decided to synthesize H-Telo DNA ONs labeled with IdU at different loop positions to study postsynthetic functionalization by Suzuki reaction.

5-Iodo-2′-deoxyuridine phosphoramidite 3 was synthesized in simple steps as shown in Figure 1A.⁵⁵5-Iodo-2′-deoxyur-idine 1 was protected with a DMTr group to obtain 5’-O-DMT-protected 5-iodo-2’-deoxyuridine 2, which was reacted with 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite to give phosphoramidite substrate 3 in good yields. A series of ONs 4—17 containing IdU in different loop regions were synthesized by using standard solid-phase synthesis protocol (Figure 2). ONs were deprotected using ammonia at room temperature for 48 h and then purified by polyacrylamide gel electrophoresis (PAGE) using urea as the denaturant. Deprotection at room temperature was necessary for minimizing the deiodination of IdU-modified ONs. The purity of ON products was analyzed by RP-HPLC, and their identity was confirmed by mass measurement (Figures S1-S4 and Table S1).

Figure 2.

Sequence of IdU-labeled H-Telo DNA ONs 4–17. The T residue in the first (4, 7, 10, 13), second (5, 8, 11, 14, 16), and third (6, 9, 12, 15’ 17) loop of H-Telo DNA ONs was replaced with IdU. 18–20 correspond to unmodified ONs. C-rich ON 21 is complementary to DNA ONs 4–20.

IdU Labeling and Coupling Reaction Conditions Did Not Affect the Formation and Stability of Respective GQ Structures

Before subjecting the labeled ONs to postsynthetic modification procedure, we first studied the effect of IdU modification on the formation of native GQ structures in coupling reaction conditions. As representative examples, IdU-labeled (12, 15, and 17) and control unmodified (18–20) H-Telo DNA ONs were annealed in 50 mM Tris-HCl buffer (pH 8.5) containing 100 mM KCl or 100 mM NaCl and 20% DMSO. The aqueous buffer used for the Suzuki coupling reaction contained 20% DMSO to solubilize the boronic acid/ester substrates. The CD spectra of both control and modified H-Telo ONs under K⁺ ionic conditions indicated the formation of hybrid-type GQ structures matching the literature reports for the same sequence (positive peak at ~290 nm with a shoulder at ~270 nm and a negative peak at 235—245 nm, Figure 3A).⁴⁷ Likewise, in the presence of Na⁺ ions, the ONs (except for 15 and 19) produced a CD profile with a positive peak at ~293 nm and a negative peak at ~260 nm, indicating the formation of an antiparallel GQ form (Figure 3B).⁴⁷ Further, UV-thermal melting analysis of control and modified H-Telo ONs under different ionic conditions (NaCl and KCl) gave similar T_m values, which were also in consensus with earlier reports (Figure 3C and 3D, Table 1).^56,57 Collectively, CD and T_m data prove that the replacement of a thymidine residue with IdU and Suzuki reaction conditions does not perturb the folding and stability of the GQ structures.

Figure 3.

(A and B) CD spectra of IdU-labeled ONs 12, 15, and 17 (solid lines) and control unmodified ONs 18–20 (dashed lines) in 50 mM Tris-HCl buffer (pH 8.5) containing 100 mM KCl or 100 mM NaCl and 20% DMSO. Consistent with a literature report,⁴⁷ ONs 15 and 19 in the presence of Na⁺ ions did not give a CD profile corresponding to typical GQstructures (Figure 3B). (C and D) UV-thermal melting profiles of IdU-labeled ONs 12, 15, and 17 (filled circles) and control unmodified ONs 18–20 (unfilled circles) under the above buffer conditions.

Table 1. T_m Values of IdU-Labeled and Control Unmodified GQs in the Presence of KCl and NaCl^a .

IdU-labeled ONs	Tm (°C)	control ONs	Tm (°C)
12 in KCl	70.5 ± 0.5	18 in KCl	69.2 ± 0.0
15 in KCl	59.3 ± 0.5	19 in KCl	61.8 ± 0.7
17 in KCl	66.8 ± 0.7	20 in KCl	65.3 ± 0.6
12 in NaCl	51.3 ± 0.6	18 in NaCl	50.0 ± 1.0
15 in NaCl	47.7 ± 0.9	19 in NaCl	48.7 ± 1.3*
17 in NaCl	56.6 ± 0.8	20 in NaCl	56.2 ± 0.8

^aIn NaCl, ONs 15 and 19 showed a CD pattern, which did not resemble a GQ topology (Figure 3B). However, the ONs in NaCl showed a UV-thermal melting profile possibly for a random structure.

Cross-Coupling Reaction Efficiency Depends on GQ Conformation

H-Telo DNA ONs containing an IdU label in the first and second T positions of different loops were annealed in Tris-HCl buffer containing LiCl/NaCl/KCl and reacted with benzofuran boronic acid (a) in the presence of Pd(OAc)₂L₂ (Figure 4). Although, phosphine-based Pd catalysts are useful in the postsynthetic modification of nucleic acids in vitro and in cellular setting,^36,37,58 they require an inert atmosphere, degassed buffers, or oxygen scavengers to perform the reaction. Therefore, we chose to use a catalytic system made of Pd(OAc)₂ and 2-aminopyrimidine-4,6-diol (L), which has been used for labeling protein and nucleic acids by Suzuki and Sonogashira reactions.^38,42,59,60 This catalytic system does not require an inert atmosphere or degassed buffer to perform coupling reactions. Benzofuran boronic acid (a) was preferred as a cross-coupling partner because the resultant ON products will be fluorescent and can sense GQ structures.51 The reaction mixtures were analyzed by RP-HPLC, and the fractions corresponding to the labeled ON products were isolated and quantified (Figures S5-S9, Table 2). The identity of the cross-coupled products were confirmed by MALDI-TOF or ESI-MS analysis (Figures S10-S12 and Table S2).

Figure 4. Postsynthetic Suzuki–Miyaura coupling reaction between IdU-labeled DNA ONs (4–17) and benzofuran boronic acid substrate a.

All the ONs used in this study form a random coil structure in the presence of LiCl and respective GQ structures in the presence of NaCl and KCl.⁴⁷‘61 Reactions performed with IdU-labeled G-quadruplex structures gave discernibly higher crosscoupled ON products as compared to a random coil structure formed by the same sequence (Figure 5, Table 2). GQ structures containing an IdU label at the first dT position of the first, second, and third loop (ONs 4–9) afforded the coupled products in reasonably good yields. For example, in NaCl, the antiparallel GQ topology gave similar yields when IdU was placed in first (4), second (5), or third loop (6) (Figure 5A). Similar results were observed in the case of hybrid 1 (ONs 4, 5, 6) and hybrid 2 (ONs 7, 8, 9) structures formed in the presence of KCl (Figure 5A and 5B). On the other hand, when the modification was placed at the second dT residue of different loops (ONs 10—15), the reaction efficiency was found to vary. A random structure of ONs 10—15 in the presence of Li⁺ ions showed lower coupling efficiency with the boronic acid substrate (Figure 5C and 5D). Interestingly, an antiparallel GQ topology formed by ON 12 in which IdU is present in the third loop gave 2- to 3-fold higher yields than ONs 10 and 11 in which the modification is present in the first and second loops, respectively (Figure 5C and Table 2). For the same set of sequences in the presence of K⁺ ions, the hybrid 1 form gave similar amounts of the coupled products irrespective of the modification on different loops. However, for reactions with hybrid 2-forming sequences 13–15, the trend was reversed. Under Na⁺ ionic conditions, the ONs gave similar amounts of the coupled products irrespective of the position of modification, whereas in the presence of K⁺ ions the ONs showed a difference in coupling efficiency (Figure 5D). There was a progressive increase in the product yield as the iodo position was moved from the first (13) to the second (14) and to the third loop (15).

Figure 5.

Bar diagram of isolated yields of cross-coupled products (4a–15a) obtained from reactions between IdU-labeled ONs (4–15) and benzofuran boronic acid substrate a. Reactions with ONs (A) 4, 5, and 6; (B) 7, 8, and 9; (C) 10, 11, and 12; (D) 13, 14, and 15 in the presence of differentions. See Figure S13 for a representative bar diagram showing the errors in the isolated yields.

Table 2. Isolated Yields of Cross-Coupled Products (4a-15a) Obtained by the Suzuki–Miyaura Reaction with ONs Containing IdU in Different Loops and under Different Ionic Conditions (Li⁺/Na⁺/K⁺)^a .

		isolated yields of cross-coupled product (%)
IdU (1)-labeled H-Telo DNA ONs in different salts	structures formed by ONs	first loop	second loop	third loop
Li+(1st T)b	random coil	36 (4a)	33 (5a)	28 (6a)
Na+(1st T)b	antiparallel	58 (4a)	52 (5a)	54 (6a)
K+(1st T)b	hybrid type 1	46 (4a)	62 (5a)	55 (6a)
Li+(1st T)b	random coil	17 (7a)	16 (8a)	31 (9a)
Na+(1st T)b	random coil	42 (7a)	32 (8a)	55 (9a)
K+(1st T)b	hybrid type 2	43 (7a)	46 (8a)	48 (9a)
Li+(2nd T)c	random coil	13 (10a)	11 (11a)	16 (12a)
Na+(2nd T)c	antiparallel	10 (10a)	18 (11a)	33 (12a)
K+(2nd T)c	hybrid type 1	21 (10a)	20 (11a)	22 (12a)
Li+(2nd T)c	random coil	5 (13a)	8 (14a)	7 (15a)
Na+(2nd T)c	random coil	15 (13a)	19 (14a)	15 (15a)
K+(2nd T)c	hybrid type 2	9 (13a)	15 (14a)	23 (15a)

^aAll reactions were performed on a 2.5 nmol scale of IdU-labeled DNA ONs (4–15).

^bIdU (1) is placed at the first dT position of different loops formed by TTA nucleotides.

^cIdU (1) is placed at the second dT position of different loops formed by TTA nucleotides.

Synthetic molecular crowding agents such as PEG can induce G-rich sequences including human telomeric repeat to adopt a parallel-stranded GQ topology.50 Labeled H-Telo ON 12 and control ON 18 were annealed in Tris-HCl buffer containing 40% PEG (200) and 100 mM KCl. Consistent with earlier reports, CD profiles of the ONs in PEG confirmed the formation of a parallel GQ structure (Figure S14A). Suzuki reaction between the parallel form of 12 and benzofuran boronic acid in the presence of Pd(OAc)₂L₂ catalyst did not produce any detectable coupled product (Figure S14B). Closer examination of the HPLC chromatogram revealed considerable reduction in the peak intensity of the Pd catalyst (retention time ~3.7 min), which could be due to the sequestering effect of PEG. To ascertain this, reactions were performed at different percents of PEG (5-30%, Figure S15). The formation of cross-coupled product 12a decreased with increase in the % of PEG in the reaction mixture. Higher amounts of PEG could decrease the diffusion rate and also could coordinate with palladium, thereby reducing the effective concentration of the catalyst.⁶²

Duplexes Made of IdU-Labeled H-Telo DNA ONs Did Not Undergo the Coupling Reaction

H-Telo DNA ONs were hybridized to complementary C-rich DNA ON 21 in Tris-HCl buffer (pH = 8.5) containing either 100 mM NaCl or 100 mM KCl. Both control (18•21, 19•21, 20•21) and modified DNA duplexes (12•21, 15•21, 17•21) showed similar CD profiles and were characteristic of a B form of duplex (Figure S16).⁶³ Interestingly, the duplexes formed using IdU-modified ONs were found to be completely unreactive under Suzuki reaction conditions even after 12 h of incubation (Figure S17). This observation further suggests the influence of conformation on the coupling reaction efficiency.

Change in Ionic Conditions Does Not Influence the Coupling Efficiency

It is reported that inorganic salts alter the efficiency of organic reactions such as Suzuki–Miyaura cross-coupling, Diels-Alder cycloaddition, Wittig reactions, to name a few.⁶⁴ To test the influence of added salts on the postsynthetic Suzuki–Miyaura cross-coupling, a short 5-iodouridine-modified ON 22, which does not form any secondary structure, was reacted with benzofuran boronic acid in the presence of LiCl/NaCl/KCl (Scheme S1). Different ionic conditions did not affect the reaction efficiency of the unstructured ON, indicating that the differences in reaction efficiency of GQs, random coil, and duplexes are due to the conformation adopted by ONs and not due to added salts (Figure S18 and Table S3).

Possible Reasons for the Differences in the Reactivity of the ONs

Iodo modification at the C5 position of the nucleoside incorporated into DNA duplexes should be present in the major groove, which is wide but is also deep. Due to this constraint in space, it is likely that the Pd(OAc)₂L₂ catalytic system and/or the intermediates in the coupling cycle may not attain the correct geometry for the coupling process. This restriction is somewhat released in the random coil structures, and hence we observed reasonable coupling efficiency.

The H-Telo ONs form GQs with three loops formed by TTA bases. IdU-modified GQs in general produced higher yields of the coupled ON products as compared to the random coil structure. When IdU was placed at the first dT in each of the three loops, the coupling efficiency was found to be the best, irrespective of the GQ topologies. On the basis of the 3D structures of antiparallel and hybrid GQs,^48,65,66 the first dT in all loops though show subtle differences in their interaction with neighboring bases, they are exposed, and hence the C5 iodo label is accessible for the coupling reaction (Figures S19 and S20). On the other hand, IdU placed at the second dT position in all the three loops reacted less efficiently with the boronic acid substrate. In the antiparallel form, the second dT in the first loop is almost buried and stacks on the G-tetrad core (Figure S19A). In the second loop the dT is stacked from one side with adjacent dA (Figure S19B), and in the third loop it is not stacked but is located in the groove wherein there is a void space around the C5 position (Figure S19C). Similarly, in the hybrid structure the second dT is either inaccessible or stacked over the G-tetrad (Figure S20). Hence, the differences in reactivity exhibited by the ONs is likely due to steric effects and differences in the interaction of the IdU label with adjacent nucleobases originating from the nucleic acid conformation.

Benzofuran Modification Introduced by Postsynthetic Suzuki Reaction Fluorescently Distinguishes Different GQ Topologies

To evaluate whether the modified ONs synthesized by Pd-catalyzed reactions are suitable for downstream biophysical analysis, benzofuran-modified H-Telo DNA ON 16a was subjected to fluorescence analysis. ON 16a containing the label at the first dT base of the second loop formed respective GQ topologies in different ionic conditions without hampering the native fold, which was confirmed by CD analysis (Figure S21). We then recorded the fluorescence spectra of the ON in the presence of K⁺ or Na⁺ ions by exciting the samples at 330 nm. The duplexes made of ON 16a in KCl and NaCl exhibited very low fluorescence (Figure 6). The mixed hybrid-type GQs (hybrids 1 and 2, blue line) formed by 16a in KCl exhibited ~6-fold enhancement in fluorescence intensity with a red shift in emission maximum (λ_em = 434 nm) as compared to the duplex form (λ_em for 16a•21 is 427 nm). Antiparallel topology displayed further enhancement in fluorescence intensity (~14-fold, black line) compared to its duplex form. The probe also distinguished between hybrid 1 and hybrid 2 forms (Figure 6). ON 5a, which predominantly forms the hybrid 1 structure in KCl, displayed an intense emission band (magenta line), nearly 5-fold higher than that of the hybrid 2 structure formed by ON 8a (red line). ON 16a, as before, showed intermediate fluorescence, as this sequence forms both hybrid 1 and 2 forms in KCl (blue line). The ability of the probe to fluorescently distinguish different GQ topologies is due to differences in the microenvironment of the probe in different conformations.^51,52 Collectively, these results clearly indicate the usefulness of Pd-catalyzed crosscoupling reactions in constructing ONs labeled with functional probes.

Figure 6.

Fluorescence spectra of GQs of H-Telo DNA ON 16a and corresponding duplexes under different ionic conditions. Spectra of GQs of ONs 5a (hybrid 1), 8a (hybrid 2), and 16a (mixed hybrids) in KCl. Samples (0.5 μM) were excited at 330 nm with an excitation and emission slit width of 5 and 6 nm, respectively.

Reaction yields of this conformation-selective modification and fluorescence properties of the ON products cannot be easily correlated, as factors that influence yields and fluorescence outcome could be different and also would depend on the boronic acid substrate. However, in our case we did see a correlation between the duplex form and the G-quadruplex structures. While the iodo-modified duplex did not yield any coupled product with benzofuran boronic acid, iodomodified GQ structures were found to be suitable for postsynthetic modification by Suzuki reaction. Further, benzofuran-modified GQs exhibited significantly higher fluorescence as compared to very weak fluorescence exhibited by the benzofuran-modified duplex under different ionic conditions (Figure 6).

Conclusions

We successfully performed postsynthetic Suzuki–Miyaura cross-coupling reactions on various DNA conformations such as random coil, GQ, and duplex structures. The differences in reactivity displayed by various DNA conformations of the human telomeric repeat ON sequences mainly depended on the position of the IdU label in the loop, and the electronic and steric environment around IdU. The results indicate that this Suzuki-based ON functionalization method is conformationdependent, as the efficiency of the reaction decreased in the following order: GQ topology > single-stranded random coil and no reaction with duplex DNA. Further, this approach is modular, as various functional tags can be directly installed by reacting boronic acid/ester substrates with easily accessible IdU/IU-labeled ONs. Here we synthesized telomeric repeats labeled with an environment-sensitive fluorescent probe, which faithfully reported the formation as well as distinguished different GQ structures via changes in its emission intensity and maximum. Taken together, this study highlights the importance of nucleic acid conformation, which needs to be taken into consideration when establishing postsynthetic functionalization strategies based on chemoselective reactions.

Experimental Section

Synthesis and characterization data of 5-iodo-2’-deoxyuridine phosphoramidite substrate 3 are given in Supporting Information. Detailed procedure for the incorporation of 3 into H-Telo DNA ON sequences by the solid-phase method, HPLC traces, and characterization by MALDI-TOF and ESI mass spectroscopy techniques are provided in Supporting Information. HPLC chromatograms of Suzuki–Miyaura crosscoupling reactions and mass spectra of cross-coupled products are also given in Supporting Information.

CD Measurements

Respective GQ structures were formed by annealing IdU-labeled DNA ONs 12/15/17 (8 μM) and control unmodified H-Telo DNA ONs 18-20 (8 μM) in Tris-HCl buffer (50 mM, pH 8.5) containing either 100 mM NaCl or 100 mM KCl and 20% DMSO. To obtain a parallel conformation, IdU-modified ON 12 and unmodified ON 18 were annealed in 50 mM Tris-HCl buffer (pH 8.5) containing 40% PEG 200, 100 mM KCl, and 20% DMSO. The corresponding IdU-modified (12•21, 15•21, 17•21) and unmodified DNA duplexes (18•21, 19•21, 20•21) were prepared by heating a 1:1.1 mixture of H-Telo DNA ONs 12/15/17/18-20 (8 μM) and complementary C-rich DNA ON 21 (8.8 μM) in Tris-HCl buffer (50 mM, pH 8.5) containing either 100 mM NaCl or 100 mM KCl and 20% DMSO. All the samples were heated at 90 °C for 3 min. The samples were slowly cooled to RT and kept on an ice bath for at least 1 h. The CD spectrum was recorded from 200 to 350 nm on a J-815 CD spectropolarimeter (Jasco, Portland, OR) using 1 nm bandwidth at 20 °C. Each CD profile is an average of three scans collected at a scan speed of 100 nm/min. All CD measurements were carried out in duplicate, and the spectra were corrected using an appropriate buffer solution without DNA ONs.

UV-Thermal Denaturation Experiments

The samples prepared for CD analysis were diluted with respective 1X buffer such that the final concentration of the ON samples was 1 μM. Thermal melting analysis was performed in duplicate by using a Cary 300 Bio UV-vis spectrophotometer. The temperature was increased from 20 °C to 90 °C at 1 °C/ min, and the absorbance was measured every 1 °C interval at 295 nm. Both forward and reverse cycles were used for the T_m determination.

Postsynthetic Suzuki–Miyaura Reactions

Reactions with IdU-Labeled DNA ONs 4–17

Samples of IdU-labeled DNA ONs 4–17 (2.5 nmol, 50 μM, 1 equiv) in Tris-HCl buffer (50 mM, pH 8.5) containing 100 mM salt (LiCl, NaCl, or KCl) and 20% DMSO were annealed to form the respective DNA structures by heating at 90 °C for 3 min. The samples were cooled to attain RT and placed in an ice bath for at least 1 h. Further, benzofuran boronic acid a (2.5 mM, 50 equiv) was added to the above solution, and the reaction was initiated by adding catalyst Pd(OAc)₂L₂(0.1 mM, 2 equiv). The final volume of the reaction was 50 μL, and the overall percentage of DMSO was kept at 20% v/v. All reactions were performed at 37 °C without degassing the buffer.

Reactions with IdU-Labeled DNA Duplexes

IdU-modified DNA ONs 12, 15 and 17 (2.5 nmol, 50 μM, 1 equiv) and complementary C-rich DNA ON 21 (2.75 nmol, 55 μM, 1.1 equiv) in Tris-HCl buffer (50 mM, pH 8.5) containing either NaCl or KCl (100 mM) and 20% DMSO were annealed to form duplexes as mentioned above. Boronic acid substrate a (2.5 mM, 50 equiv) and Pd(OAc)₂L₂(0.1 mM, 2 equiv) were added to a final volume of 50 μL. All reactions were performed at 37 °C without degassing the buffer.

Analysis of the Reaction

The reaction mixtures were filtered using spin filters (0.45 μm pore size), and the filters were washed with 50 μL of autoclaved water. The filtrates were analyzed by RP-HPLC (Phenomenex-Luna C18 column, 250 × 4.6 mm, 5 μm). Mobile phase A: 50 mM TEAA buffer (pH 7.0), mobile phase B: acetonitrile. Flow rate: 1 mL/min. Gradient: 0-30% B in 35 min, 30-100% B in 10 min and 100% B for 5 min. The run was monitored at 260 and 330 nm (for the benzofuran coupled ON products). The fractions corresponding to the products were collected and lyophilized at least three times to remove the volatile TEAA buffer. The ON products were quantified by UV absorption, and their identity was confirmed by mass analysis. See SI for HPLC profiles and mass data.

Detection of GQ Structures by Fluorescence

H-Telo ONs 5a/8a/16a (0.5 μM) were annealed at 90 °C for 3 min in 10 mM Tris-HCl buffer (pH 7.5) containing either 100 mM KCl or 100 mM NaCl. To obtain a DNA duplex (16a•21), ON 16a and complementary C-rich ON 21 were assembled by heating a 1:1.1 mixture in 10 mM Tris-HCl buffer (pH 7.5) containing either 100 mM KCl or 100 mM NaCl at 90 °C for 3 min. All the samples were then cooled slowly to attain RT and placed in an ice bath for 1 h. Each sample was excited at 330 nm with excitation and emission slit widths of 5 and 6 nm, respectively. All fluorescence experiments were performed in triplicate with appropriate blank corrections in a micro fluorescence cuvette (path length 1.0 cm, Hellma) on a Horiba Jobin Yvon, Fluoromax-4 at 20 °C.