2.1. Synthesis and crystallization of nucleoside 1

7-Iodo-5-aza-7-de­aza­guanine (1) was synthesized as re­ported previously (Leonard et al., 2019). For crystallization, com­pound 1 was dissolved in a methanol/water (1:1 v/v) mixture (10 mg in 1 ml) and was obtained as colourless thin needles (m.p. 180–181 °C, decom­position) by slow evaporation of the solvent (room temperature, 24 h). A colourless needle-like specimen of 1 was used for the X-ray crystallographic analysis.

2.2. X-ray diffraction and refinement

Crystal data, data collection and structure refinement details are summarized in Table 1. The H atoms on N2, O2′, O3′ and O5′ were refined freely, but with N—H and O—H distance restraints and U_iso(H) values fixed. For atom C1′, the anisotropic displacement parameters were refined with damp­ing in order to hold this atom positive definite.

Table 1 Experimental details Crystal data Chemical formula C10H12IN5O5 Mr 409.15 Crystal system, space group Monoclinic, P21 Temperature (K) 100 a, b, c (Å) 9.1086 (2), 6.3149 (2), 11.0428 (3) β (°) 95.585 (1) V (Å3) 632.17 (3) Z 2 Radiation type Cu Kα μ (mm−1) 20.25 Crystal size (mm) 0.35 × 0.06 × 0.02   Data collection Diffractometer Bruker KappaCCD APEXII Absorption correction Multi-scan (SADABS; Bruker, 2014) Tmin, Tmax 0.39, 0.69 No. of measured, independent and observed [I > 2σ(I)] reflections 15087, 2216, 2144 Rint 0.072 (sin θ/λ)max (Å−1) 0.597   Refinement R[F2 > 2σ(F2)], wR(F2), S 0.036, 0.097, 1.07 No. of reflections 2216 No. of parameters 205 No. of restraints 12 H-atom treatment H atoms treated by a mixture of independent and constrained refinement Δρmax, Δρmin (e Å−3) 2.79, −1.19 Absolute structure Flack x determined using 919 quotients [(I+) − (I−)]/[(I+) + (I−)] (Parsons et al., 2013) Absolute structure parameter 0.040 (7) Com­puter programs: SAINT (Bruker, 2015), SHELXT2014 (Sheldrick, 2015a), SHELXL2014 (Sheldrick, 2015b), APEX3 (Bruker, 2016) and XP (Bruker, 1998).

2.3. Synthesis of benzo­furan conjugate 3

For the preparation of 2-amino-6-(benzo­furan-2-yl)-8-(β-D-ribo­furanos­yl)-8H-imidazo[1,2-a]-s-triazin-4-one (3), com­pound 1 (Leonard et al., 2019; 100 mg, 0.24 mmol), Pd(PPh₃)₄ (28 mg, 0.024 mmol), Na₂CO₃·10H₂O (352 mg, 1.23 mmol) and benzo­furan-2-boronic acid (158 mg, 0.97 mmol) in CH₃CN/H₂O (1:1 v/v, 20 ml) were refluxed for 30 min. After consumption of starting material 1 [monitored by thin-layer chromatography (TLC)], the solvent of the reaction mixture was evaporated under reduced pressure, and the residue was applied to flash chromatography (CH₂Cl₂/MeOH, 90:10→75:25 v/v). Com­pound 3 was obtained from the main zone as a white solid (yield 55 mg, 57%). TLC (silica gel, CH₂Cl₂/MeOH, 85:15 v/v) R_F 0.4. λ_max (MeOH)/nm 312 (∊/dm³ mol⁻¹ cm⁻¹, 22600), 327 (18800). ¹H NMR (600 MHz, DMSO-d₆): δ 8.00 (s, 1H, H-8), 7.71–7.65 (m, 2H, benzofuran Ar-H), 7.54 (dq, J = 8.4, 1.0 Hz, 1H, benzofuran Ar-H), 7.33 (ddd, J = 8.4, 7.2, 1.3 Hz, 1H, benzofuran Ar-H), 7.26 (td, J = 7.4, 1.0 Hz, 1H, benzofuran Ar-H), 7.12 (s, 2H, NH₂), 5.87 (d, J = 5.8 Hz, 1H, H-1′), 5.51 (d, J = 5.5 Hz, 1H, OH-2′), 5.20 (d, J = 4.6 Hz, 1H, OH-3′), 5.16 (t, J = 5.3 Hz, 1H, OH-5′), 4.37 (q, J = 5.4 Hz, 1H, H-2′), 4.11 (td, J = 4.7, 3.4 Hz, 1H, H-3′), 3.92 (q, J = 3.6 Hz, 1H, H-4′), 3.66 (ddd, J = 11.9, 5.2, 3.8 Hz, 1H, H-5′), 3.58 (ddd, J = 11.9, 5.3, 3.6 Hz, 1H, H-5′). ¹³C NMR (151 MHz, DMSO-d₆): δ 164.8, 153.9, 151.9, 150.1, 144.7, 128.4, 125.1, 123.2, 121.6, 115.7, 114.7, 110.7, 108.2, 86.5, 85.5, 85.4, 73.7, 70.2, 61.1, 61.0. HRMS (ESI–TOF) m/z: [M + Na⁺] calculated for C₁₈H₁₇N₅NaO₆, 422.1071; found, 422.1075.

3.1. Mol­ecular geometry and conformation of 7-iodo-5-aza-7-de­aza­guanosine (1)

The three-dimensional (3D) structure of 1 is shown in Fig. 3(a) and selected crystallographic data and structure refinement details are summarized in Table 1. According to the Flack parameter and the synthetic pathway, the anomeric centre at C1′ is in an R configuration confirming the β-D anomeric structure of 1. The structure of the related 2′-de­oxy­ribonucleoside has been reported previously (for selected geometric parameters of the single-crystal X-ray analysis of 2c, see Table S1 in the supporting information) (Leonard et al., 2019).

Figure 3 (a) Perspective view of 1, showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 50% probability level and H atoms are shown as small spheres of arbitrary size. (b) Overlay of ribonucleoside 1 and the related 2′-de­oxy­ribonucleoside 2c.

To evaluate the impact of the 2′-hy­droxy group of ribonucleoside 1 on the crystal structure, it was of inter­est to com­pare the geometric parameters of com­pound 1 with those of the 2′-de­oxy­ribonucleoside 2c (Leonard et al., 2019). An overlay of both mol­ecules visualizes the conformational differences mainly affecting the sugar moiety (Fig. 3b).

The nucleobase may adopt two principal orientations with respect to the sugar moiety (syn/anti) and it is defined by the torsion angle χ (O4′—C1′—N9—C4) (IUPAC–IUB Joint Commission on Biochemical Nomenclature, 1983). Purine nucleosides are found in a wide range of anti conformations (Blackburn et al., 2006). The 7-iodo-5-aza-7-de­aza­guanine moiety of both nucleosides also adopts an anti conformation, with χ = −120.6 (9)° for ribonucleoside 1 (Table 2) and χ = −139.9 (6)° for 2c (Leonard et al., 2019). The glycosylic bond lengths of 1 (Table 2) and 2c (Lin et al., 2004) are within the range of purine nucleosides (Saenger, 1984).

The ribose ring is twisted out of plane to minimize steric inter­actions of the substituents, referred to as sugar puckering (Altona & Sundaralingam, 1972). The more electronegative substituents of C2′ and C3′ prefer an axial orientation. Nucleosides can adopt two principal sugar puckering modes, namely C3′-endo (N) and C2′-endo (S) corresponding to the major displacement of C3′ or C2′ from the median C1′/O4′/C4′ plane. The C2′-endo conformation is the preferred puckering mode of 2′-de­oxy­ribonucleosides (Saenger, 1984), whereas ribonucleosides show a preference for the C3′-endo sugar pucker. Also, 7-iodo ribonucleoside 1 shows an N conformation, with a pseudorotational phase angle P = 77.87° and a maximum amplitude τ_m = 47.14°. However, the typical C3′-endo conformation is less pronounced and the sugar pucker is shifted towards O4′-endo. This is a sugar conformation where the nucleobase and the exocyclic 5′-OH group are arranged equatorially. The sugar moiety of 2′-de­oxy­ribonucleoside 2c also adopts an N-type sugar conformation (C3′-endo–C4′-exo, ³T₄, P = 28.55° and τ_m = 34.28°) (Leonard et al., 2019).

In many cases, the solid-state conformation of nucleosides differs from that in solution and is also different from that of nucleosides as constituents of a single- or double-stranded nucleic acid. Sugar puckering of 1 was determined in solution from the coupling constants of high-resolution ¹H NMR spectra (600 MHz) (for details, see Table S2 in the supporting information). The population of S versus N conformers was calculated on the basis of coupling constants using the program PSEUROT (Version 6.3; Van Wijk et al., 1999). Notably, the conformation found for ribonucleoside 1 in solution is S-type (69%) and is different from that in the solid state (N-type). The closely related 2′-de­oxy­ribonucleoside 2c also prefers an S conformation (62%) in solution (Lin et al., 2004).

The orientation of the exocyclic 5′-hy­droxy group relative to the sugar ring is described by the torsion angle γ (O5′—C5′—C4′—C3′). Surprisingly, γ is identical in both nucleosides despite the fact of their having different sugar moieties (ribose versus 2′-de­oxy­ribose). The 5′-hy­droxy group adopts, in both cases, an anti­periplanar (trans) conformation, with γ = −172.9 (8)° for 1 and γ = −172.7 (4)° for 2c (Leonard et al., 2019).

3.2. Hydrogen bonding and mol­ecular packing

In nucleoside solid-state structures, hydrogen bonds are formed between N and/or O atoms functioning as proton donors or acceptors. They can be provided by the nucleobases and/or the sugar moieties. Fig. 4(a) displays the hydrogen-bonding pattern of 7-iodo-5-aza-7-de­aza­guanosine (1) within one particular layer.

Figure 4 (a) Hydrogen bonding of 1, viewed along the b axis. (b) Multilayered mol­ecular packing of 1, shown along the slightly inclined b axis. (c) Detailed view of part (a), showing the nucleobase–sugar inter­action forming a central core.

Nucleobase-to-nucleobase inter­actions, such as self-pairing in a Watson–Crick-like fashion, are not observed for ribonucleoside 1. This is different from the solid-state structure of the 5-aza-7-de­aza­guanine nucleobase reported recently (Laos et al., 2019). The free nucleobase is able to adopt tautomeric forms with the H atom located either at N9 or at N3. The N9—H tautomer can base pair with the N3—H tautomer, forming a tridentate base pair (Laos et al., 2019). However, in the context of DNA or RNA structure, this is an artificial situation, as nucleosides glycosyl­ated at N9 lose the proton at the heterocycle and thus the ability to form these tautomers.

Com­pound 1 forms in the crystal a core of four mol­ecules displaying a `parallelogram'-like arrangement, with the O2′—H2′ and C2′—H2′1 groups as proton donors and atom O6 of the nucleobase as a bi-acceptor (Fig. 4a and Table 3). This structure is surrounded by `cyclo­hexa­ne'-like motifs formed by two mol­ecules connected via N2—H2A⋯O4′ⁱ and C1′—H1′⋯N2^v hydrogen bonds. Further stabilization of the solid-state structure is achieved by hydrogen bonds formed by the exocyclic 5′-OH group (N2—H2B⋯O5′ⁱⁱ and O5′—H5′⋯O3′^iv).

Table 3 Hydrogen-bond geometry (Å, °) D—H⋯A D—H H⋯A D⋯A D—H⋯A N2—H2A⋯O4′i 0.90 (3) 2.25 (10) 2.941 (11) 133 (11) N2—H2B⋯O5′ii 0.90 (3) 2.04 (5) 2.905 (12) 160 (12) O2′—H2′⋯O6iii 0.87 (3) 1.93 (8) 2.689 (9) 144 (12) O5′—H5′⋯O3′iv 0.87 (3) 1.87 (7) 2.673 (10) 151 (13) C1′—H1′⋯N2v 1.0 2.63 3.345 (15) 128 C2′—H2′1⋯O6vi 1.0 2.57 3.541 (12) 163 Symmetry codes: (i) x, y+1, z; (ii) ; (iii) ; (iv) ; (v) ; (vi) .

It is well documented that the iodo substituents of halogenated nucleosides can participate in mol­ecular inter­actions, sometimes even playing a predominant role. Strong iodo–iodo inter­actions have been reported for 5-iodo-α-2′-deoxycytidine (Müller et al., 2019), while the iodo substituent of 2′-de­oxy­ribonucleoside 2c shows a strong contact to atom O3′ [I7⋯O3′ = 2.794 (4) Å and C7—I7⋯O3′ = 169.2 (2)°] (Leonard et al., 2019). In the crystal structure of 7-iodo-5-aza-7-de­aza­guanosine (1), the iodo substituent I7 shows a contact to atom O2′ [2.936 (7) Å and C7—I7⋯O2′^vi = 170.4 (3)°; for symmetry code, see Table 3]. However, the I⋯O contact of ribonucleoside 1 is less strong than that in com­pound 2c.

The individual nucleoside layers are stacked upon each other in a very regular fashion, as demonstrated in Fig. 4(b). Fig. 5 shows that com­pound 1 forms chains which are arranged in an alternating reverse fashion (ABAB…). Two chains (A and B) are connected to each other by hydrogen bonds (C1′—H1′⋯N2^v, N2—H2B⋯O5′ⁱⁱ and O5′—H5′⋯O3′^iv; Table 3). However, as was discussed above, the ribonucleoside mol­ecules are not arranged in a Watson–Crick-like fashion. Instead, the nucleobase of one chain is placed opposite the sugar residue of the next chain. As a result, the iodo substituent points towoards the outside of each double chain (com­posed of A and B chains). The double chains are connected to each other by the C2′—H2′1⋯O6^vi hydrogen bond. This arrangement can also be visualized by the space-filling model shown in Fig. 5.

Figure 5 (Top) The crystal packing of 1, showing the hydrogen bonds along the a axis. (Bottom) Space-filling model of 1, viewed along the a axis.

3.3. Hirshfeld surface analysis of 1

Hirshfeld surface analysis is a valuable tool to obtain additional insight into the role of crystal packing forces and to visualize the inter­molecular inter­actions of crystalline com­pounds. The CrystalExplorer program (Version 17; Spackman & Jayatilaka, 2009; Turner et al., 2017) was used to conduct a Hirshfeld surface analysis mapped over d_norm (−0.5 to 1.5 Å), shape index (−1.0 to 1.0 Å) and curvedness (−4.0 to 0.4 Å) (Fig. S2 in the supporting information), as well as a two-dimensional (2D) fingerprint plot analysis. The Hirshfeld surface of com­pound 1 is shown in Figs. 6(a) and 6(b). Several red spots are visible on the Hirshfeld surface corresponding to the close O—H⋯O and N—H⋯O contacts of the nucleobase and sugar moieties. The shape index indicates π–π stacking inter­actions by the presence of adjacent red and blue triangles. Their absence clearly suggests that there are no π–π inter­actions in com­pound 1, as demonstrated by Figs. 6(c) and 6(d).

Figure 6 Hirshfeld surface of com­pound 1, mapped over dnorm (−0.5 to 1.5 Å), shown in (a) a front and (b) a back view. The shape index, shown in (c) a front and (d) a back view.

2D fingerprint plots can be resolved to highlight particular atom pair inter­actions. This enables the separation of contributions from different contact types that overlap in the full fingerprint plot. The overall 2D fingerprint plot and those resolved into H⋯C/C⋯H, H⋯H, I⋯O/O⋯I, N⋯H/H⋯N and O⋯H/H⋯O contacts are illustrated in Fig. 7, together with their relative contributions to the Hirshfeld surface. The proportions of O⋯H/H⋯O, H⋯H and N⋯H/H⋯N inter­actions com­prise 28.7, 19.1 and 15.9%, respectively, of the total Hirshfeld surfaces. Lower percentages are found for the H⋯C/C⋯H contacts (9.2%). Moreover, the inter­action between iodo substituent I7 and atom O2′ is confirmed and appears as two wings in the fingerprint plot, with a proportion of 4.1%.

Figure 7 2D fingerprint plots showing the percentages of the contribution of various interactions to the total Hirshfeld surface area of com­pound 1. Full inter­actions (upper row, left) and resolved contacts (upper row: middle, C⋯H/H⋯C; right, H⋯H; lower row: left, I⋯O/O⋯I; middle, N⋯H/H⋯N; right, O⋯H/H⋯O).

3.4. The stability of 5-aza-7-de­aza­guanine base pairs controlled by pK value differences of com­plementary bases

The Watson–Crick base-pair recognition of canonical DNA and RNA constituents depends on the hydrogen-bonding donor–acceptor pattern of purine and pyrimidine nucleosides, as well as on environmental influences caused by base-to-base stacking inter­actions, helix formation and various other factors. Among these, the pK values of nucleobases play an important role as they control the stability and lifetime of base pairs. Recently, the role of the pK values of nucleobases has been discussed with regard to the origin of chemical evolution. Strong base pairs are formed when the pK value difference (ΔpK) between the acceptor and donor sites of nucleobases is at least five units or more. A ΔpK < 5 results in weaker base pairs (Krishnamurthy, 2012).

Early studies on the physical properties of 5-aza-7-de­aza-2′-de­oxy­guanosine (2b) showed that the nucleobase is proton­ated on the s-triazine moiety having a pK_a value of 3.7 (Rosemeyer & Seela, 1987). Nitro­gen-1 was established as the protonation site using ¹³C NMR spectroscopy (Seela & Melenewski, 1999). The pK_a value and the protonation site is different from the closely related 2′-de­oxy­guanosine (pK_a value of protonation = 1.6 and deprotonation = 9.2) (Seela et al., 2003), as well as to 7-de­aza-2′-de­oxy­guanosine (1.1 and 10.3) (Ramzaeva & Seela, 1996), and is a direct consequence of the transposition of the nitro­gen-7 atom to bridgehead position-5.

The positional change of the N atom leads to the absence of a proton at N1 which now becomes a proton-acceptor site and not a donor site as in 2′-de­oxy­guanosine or 7-de­aza-2′-de­oxy­guanosine. Furthermore, position-7 cannot inter­act with protons or metal ions anymore, and similar to 7-de­aza-2′-de­oxy­guanosine, Hoogsteen base pairs cannot be formed. Consequently, 5-aza-7-de­aza-2′-de­oxy­guanosine exhibits a Watson–Crick recognition face similar to that of isocytidine. Thus, the com­plementary base has to provide a proton, as observed for base pairing with guanine or isoguanine, or with protonated cytidines or cytidine derivatives which carry a proton at nitro­gen-1 (Fig. 8a) (Hoshika et al., 2018). Silver ions can function as proton substitutes forming metal-mediated base pairs (Fig. 8b) (Guo et al., 2018).

Figure 8 (a) Protonated base pair of 5-aza-7-de­aza-2′-de­oxy­guanosine (2b) and 2′-de­oxy­cytidine formed in acidic medium. (b) Silver-mediated base pair of 2b and dC.

As ribonucleoside 1 is a promising candidate for the synthesis of `all-purine' RNA, it is of particular importance to identify the pK_a value of 1 and to com­pare this value to related purine and 7-de­aza­purine nucleosides. The pK_a value of com­pound 1 was determined by UV spectroscopic titration from absorbance changes at 280 nm at different pH values (Fig. S3 in the supporting information). According to Fig. 9, the pK_a value of 1 is 3.6. This pK_a value is almost identical to that of the corresponding 2′-de­oxy­ribonucleoside 2c (pK_a = 3.7) (Lin et al., 2004) and the purine nucleoside adenosine (pK_a = 3.6). Contrary to the related purine and 7-de­aza­purine nucleosides, the mol­ecule does not show a second pK value (for deprotonation), as the heterocyclic skeleton does not carry a proton which can be released in alkaline medium. Atom N1 is the protonation site, as was determined for 5-aza-7-de­aza­guanine 2′-de­oxy­ribonucleoside 2b (Seela & Melenewski, 1999).

Figure 9 pKa value of 1 determined by UV titration (1st derivative, red line). Spectroscopic changes determined at 280 nm at different pH values (black line) were measured in 0.1 M sodium phosphate buffer.

Table S3 (see supporting information) summarizes the pK values and protonation positions for 5-aza-7-de­aza­guanine, 7-de­aza­guanine, guanine and isoguanine nucleosides. It is obvious that the iodo substituent has no significant influence on the pK value of the nucleosides. However, com­pared to guanosine and 7-de­aza­guanosine, the pK for protonation of the unsubstituted and iodinated 5-aza-7-de­aza­guanosine base is strongly shifted and the mol­ecule is much easier to protonate. Also, the position of protonation (N1) is different from that in guanosine (N7) and 7-de­aza­guanosine (N3).

It has been shown that based on pK value differences of nucleosides participating in base-pair formation, predictions regarding base-pair stability can be made (Krishnamurthy, 2012). To this end, the ΔpK of base pairs of 7-iodo-5-aza-7-de­aza­guanosine with different nucleosides acting as donor or acceptor were calculated and com­pared (Fig. 10). Anti­parallel (aps) and parallel (ps) strand orientation of oligonucleotides within a duplex were considered.

Figure 10 Base-pairing motifs of 5-aza-7-de­aza­guanine nucleosides, pKa values of the base-pairing nucleosides and their pK value difference (ΔpK), as well as strand orientation of the respective duplex. Notes: R = β-D-ribo­furanosyl, C = cytidine, iG = isoguanosine, iC = isocytidine, G = guanosine, c7G = 7-de­aza­guanosine, aps = anti­parallel strands and ps = parallel strands.

The ΔpK of the 7-iodo-5-aza-7-de­aza­guanosine–isoguanosine base pair is 6.2 (motif I, Fig. 10) referring to an aps orientation of oligonucleotides. As a result, the stability of this purine–purine base pair is expected to be higher than for the Watson–Crick purine–pyrimidine G–C base pair (ΔpK = 4.8; motif VII, Fig. 10). A parallel arrangement of 1 with guanosine (ΔpK = 5.7; motif II) or 7-de­aza­guanosine (ΔpK = 6.6; motif III) also leads to stable purine–purine pairs. However, the formation of a purine–pyrimidine base pair with 1 results in a mismatch situation in an aps alignment with cytidine (ΔpK = 0.9; motif IV) or isocytidine (ΔpK = 0.7; motif VI). Consequently, the RNA constituent 1 is a promising candidate for the construction of stable purine–purine base pairs as building blocks for `all-purine' RNA with a conventional anti­parallel orientation of the strands. Moreover, even a parallel orientation of oligonucleotides might be applicable. The 7-iodo substituent allows the introduction of reporter groups to monitor structural characteristics of aps or ps `all-purine' RNA.

However, self-pairing of 1, as demonstrated by motif V (ps alignment; Fig. 10) or even in a shifted arrangement, leads to a mismatch situation. Moreover, none of the reported crystal structures of 5-aza-7-de­aza­guanine nucleosides show self-pairing of the nucleobases (Kojić-Prodić et al., 1982; Seela et al., 2002; Leonard et al., 2019). Only for the isolated 5-aza-7-de­aza­guanine nucleobase was self-pairing observed when one of the nucleobases forms a N3—H tautomer (Laos et al., 2019). In nucleosides, as well as in nucleic acids, this arrangement is not possible as the ribose moiety is attached to N9 and the N3 atom lacks a proton.

3.5. Functionalization of nucleoside 1 and fluorescence properties of the conjugate 3

The ability of 7-iodo­ribonucleoside 1 to act as a target for functionalization with functional reporter groups is demonstrated by use of the Suzuki–Miyaura cross-coupling reaction (Agrofoglio et al., 2003). As the benzo­furan residue is sensitive to mol­ecular crowding and the viscosity of solvents (Greco & Tor, 2007; Leonard et al., 2019), this residue was chosen for functionalization of ribonucleoside 1.

For this, com­pound 1 was used as the starting material (see Scheme 1). The synthesis of ribonucleoside 1 by nucleobase anion glycosyl­ation was reported recently, employing isobutyrylated iodo base 4 and bromo sugar 5 (Leonard et al., 2019). In order to introduce the benzo­furan side chain, Suzuki–Miyaura cross-coupling of ribonucleoside 1 and benzo­furan-2-boronic acid was performed. The reaction was carried out under reflux for 30 min in a CH₃CN/H₂O mixture (1:1 v/v) in the presence of Pd(PPh₃)₄ and Na₂CO₃·H₂O (see Scheme 1), yielding benzo­furan derivative 3 in 57%.

The new com­pound 3 was characterized by ¹H and ¹³C NMR spectroscopy, as well as by ESI–TOF mass spectroscopy. ¹H–¹³C correlated (HMBC and HSQC) NMR spectra were used to assign the ¹³C NMR signals (for spectra, see the supporting information, and for details, see the Experimental section).

The UV and fluorescence spectra of 3 were recorded in solvents of different polarity, and the corresponding Stokes shift and quantum yields were determined (Table S4 in the supporting information). The UV spectra of com­pound 3 show two maxima at around 312 and 327 nm in MeOH, MeCN and water (Fig. S4 in the supporting information). The maxima are bathochromally shifted (around 5 nm) in the less polar solvents dimethyl sulfoxide (DMSO) and di­methyl­formamide (DMF).

The fluorescence emission spectra for benzo­furan conjugate 3 in various solvents are displayed in Fig. 11(a). The highest quantum yields are obtained in DMF and DMSO (Table S4 in the supporting information), with emission maxima which are red-shifted com­pared to the other solvents. The fluorescence of 3 is very low in water com­pared to the other polar solvents.

Figure 11 Fluorescence emission spectra of com­pound 3 (1 µM) measured in (a) solvents of different polarity, (b) ethyl­ene glycol and (c) glycerol. Excitation wavelength of nucleoside 3 was 315 nm in ethyl­ene glycol and 345 nm in glycerol.

In the RNA constituent 3, the benzo­furan moiety and the 5-aza-7-de­aza­guanine base are separated by a freely rotatable ar­yl–aryl bond. The spatial positioning of both aromatic systems with respect to each other (π–π conjugation) has a direct impact on the fluorescence properties of conjugate 3. In order to study this effect for 3, fluorescence measurements were performed in solvents of similar polarity but different viscosity. Fig. 11(b) shows the fluorescence emission spectra of conjugate 3 determined in ethyl­ene glycol, water and their mixtures. The fluorescence of 3 is highest in ethyl­ene glycol and decreases with reduced viscosity of ethyl­ene glycol/water mixtures. However, the highest fluorescence is observed in a 50% glycerol–water mixture (Fig. 11c). Apparently, a rotational barrier exists for the benzo­furan residue which is sensitive to solvent viscosity, a phenomenon which has been reported for other nucleoside–benzo­furan conjugates (Tan­pure & Srivatsan, 2018; Tokugawa et al., 2016; Manna et al., 2018).