An Emissive C Analog Distinguishes between G, 8-oxoG and T

Abstract

A minimally disruptive fluorescent dC analog provides a rapid and non-destructive method for in vitro detection of G, 8-oxoG and T, the downstream transverse mutation product.

One of the most thoroughly examined DNA base modifications is 7,8-dihydro-8-oxoguanine (8-oxoG), a mutagenic product of oxidative damage by reactive oxygen species.¹ The presence of 8-oxoG is frequently viewed as a marker for cellular oxidative stress, a condition that has been linked to carcinogenesis.² The significance of this seemingly minor base damage results from 8-oxoG’s ability to deceive DNA polymerases and form a stable (syn)8-oxoG•A base pair by presenting its Hoogsteen face thereby mimicking T (Figure 1).³ Unless repaired, this might cause G to T transversion mutations during DNA synthesis (Figure 1).⁴ Not surprisingly, base-excision repair mechanisms have evolved to correct such deleterious base modification products, and the bacterial and human enzymes have been thoroughly studied.⁵

Figure 1.

Base pairing along the DNA damage pathway from G•C to T•A, (its transverse mutation product) via 8-oxoG.

A non-destructive and real-time fluorescence-based detection of 8-oxoG, its repair and induced mutation processes in oligonucleotides, could significantly advance the in vitro biochemical evaluation of this important DNA lesion.⁶ It would complement existing methods that rely on chromatographic, electrophoretic and immunological methods.^7,8 Toward this end, we hypothesized that one could take advantage of the distinct redox properties of 8-oxoG.⁹ Electrochemical measurements show that 8-oxoG is more easily oxidized compared to G (E_1/2 ≈ 0.75 and 1.3 vs. NHE, respectively).^9a,10 Consequently, chemical approaches relying on 8-oxoG’s susceptibility to oxidation and covalent trapping of the oxidized products in duplex DNA have been reported.¹¹ Since fluorescence quenching frequently occurs via photoinduced electron transfer (PET) mechanisms,¹² we suspected that 8-oxoG is likely to be a more effective quencher of certain fluorophores compared to G, its precursor.¹³ Here we describe the design, synthesis, photophysical evaluation, incorporation and implementation of a simple isomorphic fluorescent dC analog 5 that, upon incorporation into an oligonucleotide, photophysically distinguishes between 8-oxoG and G on the complementary strand. Not only is the damaged 8-oxoG-containing duplex highly quenched and the “repaired” G-containing duplex more emissive, the transverse mutated duplex containing T instead of G displays the most intense emission. This furan-containing emissive nucleobase therefore provides signature emission profiles for all key nucleobases involved in this DNA damage pathway (Figure 1).¹⁴

We have been developing simple and minimally perturbing emissive nucleobases for the detection of nucleic acids lesions.^15,16 The primary design principle dictates maintaining the highest possible structural similarity to the natural nucleobases, while significantly improving their photophysical properties. Specifically, an isolated absorption band for selective excitation, enhanced quantum yield over the native nucleobases and sensitivity to changes in the microenvironment are desired. Useful uridine-based nucleosides, fulfilling these criteria, were obtained by conjugating five-membered aromatic heterocycles such as furan at the 5 position (e.g., 2).^15,16,17,18 Among the various heterocycles conjugated to dU (i.e., furan, thiophene, oxazole, thiazole) the furan moiety was found to yield the most favorable photophysical characteristics.^15,17a We therefore anticipated the analogous furan modified cytosine nucleobase to be emissive and responsive.

The 5-modified nucleosides are easily obtained using a coupling reaction between the 5-iodo substituted dU (1) and the corresponding stannylated heterocycles (Scheme 1). Conversion of the acetate protected furan-modified dU analog 3 to the desired dC analog 5 is accomplished by activation of the 4 position as an aryl sulfonate ester followed by a displacement reaction with ammonia,¹⁹ providing the fully deprotected furan-modified dC analog 5 (Scheme 1).^17a,20 Silyl protection of the furan-modified dU 4 facilitated the conversion to the dC analog 6 with retention of hydroxyl protection thus allowing for standard benzamide protection of the exocyclic amine to give 7. Desilylation and protection of the 5′-hydroxyl as the 4,4′-dimethoxytrityl derivative (9) followed by phosphitylation of the unprotected 3′-hydroxyl afforded 10, the building block necessary for automated DNA synthesis (Scheme 1).²⁰

Scheme 1.

Synthesis of furan dC analog 5 and its corresponding amidite.

The absorption spectrum of an aqueous solution of 5 shows, in addition to the typical high energy band seen in the parent nucleoside, a clear shoulder at ~310 nm (Figure 2). Excitation at this wavelength yields an emission profile with a maximum at 443 nm, which tails deeply into the visible range, with a relative quantum yield of 0.02 (Table 1).^20,21 Lowering solvent polarity results in a better defined long wavelength absorption band (λ_max = 309 nm) and a hypsochromic shift in emission maxima (λ_em = 421 nm), which is associated with a hypochromic effect (I_{Water/Dioxane} = 3) (Figure 2, Table 1).

Figure 2.

Absorption (dashed) and emission (solid) spectra of 5 in water (blue), methanol (green), dichloromethane (orange) and dioxane (red) at 2.4 × 10⁻⁵ M.

Table 1.

Photophysical properties of nucleoside 5 in various solvents.²⁰

Solvent	λabs (nm)	λem (nm)	Φ	I normalized
Water	310	443	0.020	1.00
Methanol	305	439	0.011	0.70
Dichloromethane	309	439	0.009	0.54
Dioxane	309	421	0.006	0.33

Emission spectra of 5 in dioxane-water mixtures provide a more detailed view of the hypsochromic shift that arises upon decreasing the polarity of the chromophore’s microenvironment (Figure S2.1).²⁰ Plotting the corrected emission energy maximum of 5 vs. the microenvironment polarity described by E_T(30) values,²² results in a linear correlation (Figure 3).²³ Stern-Volmer titrations were conducted to determine the differences between G and 8-oxoG’s quenching abilities (Figure 4). Rewardingly, while G minimally impacted the emission of 5 (K_sv = 0.004 mM⁻¹), 8-oxoG was found to be a very effective quencher, even at low concentrations (K_sv = 16.5 mM⁻¹).²⁰

Figure 3.

Correlating emission wavelengths with microscopic polarity E_T(30)²² for nucleoside 5 (filled circles and green line)²⁰ and interpolation of corrected emission maxima²³ of duplexes containing 5 (arrows).

Figure 4.

Steady-state Stern-Volmer plot for the titration of 5 with 8-oxo-2′-deoxyguanosine (stars, black dashed line), dGMP (X) and TMP (open circles). Error bars (dGMP and TMP) and data (dAMP and dCMP) have been omitted for clarity (Figure S3.1, Table S3.1).²⁰

To investigate the potential of the emissive nucleoside to photophysically discriminate between G, 8-oxoG and T, an oligonucleotide that contains 5 at a central position was synthesized using 10 and standard solid-phase synthesis protocols (Figure 5).²⁰ All modified oligonucleotides were characterized using MALDI TOF MS.²⁰ Since modified C residues can, in principle, deaminate to yield the corresponding U derivatives, special care was taken to unequivocally verify the presence of 5 and absence of 2 in the modified oligonucleotide 11. Enzymatic digestion reactions, followed by HPLC analysis against all authentic nucleosides verified the presence of intact 5 and absence of 2 in the modified oligonucleotide 11 (Figure S6.1, Table S6.1).²⁰

Figure 5.

Oligonucleotide sequences where Y = 8-oxoG.

Oligonucleotide 11, containing the furan functionalized dC 5 was hybridized to three complementary oligonucleotides that contain either G (oligo 12), 8-oxoG (oligo 13) or T (oligo 14) opposite the emissive nucleotide (Figure 5). Thermal denaturation (Table 2) and CD studies show no appreciable difference between the modified (11•12, 11•13 and 11•14) and unmodified duplexes (15•12, 15•13 and 15•14, respectively) suggesting the small furan modification does not disrupt duplex formation, stability or structure (Figure S8.1 and S9.1).²⁰

Table 2.

Thermal denaturation of control and modified oligonucleotides.

Duplexa	Tm(ΔTmb) (°C) 100 mM NaCl	Tm(ΔTmb) (°C) 500 mM NaCl
15•12	51.8	59.2
15•13	47.7	55.2
15•14	34.5	41.2
11•12	51.5 (−0.3)	59.0 (−0.2)
11•13	47.1 (−0.6)	54.5 (−0.7)
11•14	34.7 (+0.2)	42.0 (+0.8)

^a1.0 × 10⁻⁶ M duplex DNA aqueous buffer pH = 7.0.

^bΔT_m = modified = unmodified.

The emission spectra of the furan dC containing duplexes, in both low (100 mM) and elevated (500 mM) ionic strength are shown in Figure 6.²³ While the perfect duplex 11•12, where G is placed opposite 5, is significantly emissive, placing 5 opposite 8-oxoG (11•13) leads to considerable emission quenching (ca. two-fold), as hypothesized. Nucleoside 5 therefore clearly distinguishes between G and 8-oxoG, its oxidized product. In contrast, duplex 11•14, where 5 is placed opposite T, a substantial emission enhancement (ca. four-fold compared to 11•13) is observed (Figure 6, Table 3). Nucleoside 5 thus reports the presence of T, the ultimate transversion mutation product resulting from G oxidation to 8-oxoG, via enhanced emission. These observations, illustrating signature emission profiles for duplexes containing G, 8-oxoG and T, suggest that the furan-modified dC nucleoside 5 could be utilized to follow, in vitro, DNA damage and its repair via this pathway using such emissive oligonucleotides.

Figure 6.

Steady state emission spectra of oligonucleotides 11•12 (5•G - blue), 11•13 (5•8-oxoG - red) and 11•14 (5•T - orange) at 5.0 × 10⁻⁶ M in 1.0 × 10⁻² M phosphate aqueous buffer pH = 7.0 containing 1.0 × 10⁻¹ M NaCl (solid) and at 4.6 × 10⁻⁶ M in 1.0 × 10⁻² M phosphate aqueous buffer pH = 7.0 containing 5.0 × 10⁻¹ M NaCl (dashed).²⁰

Table 3.

Emission maxima of modified duplexes in phosphate buffers at different NaCl concentrations.

Duplex	λem/nm (cm−1) 100 mM NaCl	Intensity at λem/au	λem/nm (cm−1) 500 mM NaCl	Intensity at λem/au
11•12	439 (22,779)	1.00	440 (22,727)	1.00
11•13	435 (22,989)	0.52	437 (22,883)	0.55
11•14	446 (22,422)	2.05	447 (22,371)	2.15

The observed fluorescence quenching of 5 by 8-oxoG can be understood by the lower redox potential of 8-oxoG and concomitant higher potency as an excited state quencher compared to G (Figure 4). Several observations collectively suggest that the substantial fluorescence enhancement observed for the 5•T mismatch in 11•14 is due to exposure of the emissive nucleoside to a more polar environment, likely extrahelical. Thermal denaturation data illustrates that the dC/T containing duplex 15•14 (as well as the analogous modified duplex 11•14) are the least stable (Table 2). This is consistent with previous observations illustrating this particular dual pyrimidine mismatch to be rather unfavorable,²⁴ therefore suggesting a local structural perturbation. Fluorescence experiments carried out at elevated ionic strengths (500 mM) for all duplexes resulted in similar observations (Figure 6), rendering the potential higher abundance of the more emissive single strand due to poor hybridization unlikely.²⁰

Further support for the proposed extrahelical residency of the fluorescent nucleobase is obtained by interpolating the emission energy of modified duplexes containing 5 and environmental polarity using the linear correlation shown above (Figure 3, Table 3). The emission spectra of the furanyl modified duplexes 11•12 and 11•13, where furanyl dC 5 can pair in a Watson–Crick fashion, display emission maximum that correlate to relatively apolar environments (Figures 3 and 6, Table 3), while duplex 11•14, where furanyl dC 5 is unable to pair in a W-C fashion with T, shows a red shifted emission maximum that correlates to a polar environment, in agreement with an extrahelical disposition.

In summary, we have shown that an isomorphic fluorescent nucleoside 5, which upon incorporation results in no observable perturbation of the duplex shape and stability, is a valuable probe for the detection of G, 8-oxoG and its transverse mutation product T by eliciting markedly different emission intensities in conjunction with changes in emission maxima. The effective synthesis and incorporation of 5 could facilitate rapid and non-destructive real-time fluorescence-based methods for the in vitro monitoring of this DNA damage pathway.²⁵

Supplementary Material

1_si_001. Supporting Information Available.

Experimental procedures, spectra, oligonucleotide digestion, thermal denaturation and CD studies.