Pentacyclic adenine: a versatile and exceptionally bright fluorescent DNA base analogue† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c7sc05448c

A highly fluorescent, non-perturbing, pentacyclic adenine analog was designed, synthesized, incorporated into DNA and photophysical evaluated.


General Experimental Details
All reactions were performed in oven-dried glassware under a nitrogen atmosphere, unless otherwise stated. Reagents were purchased from various chemical vendors and used as received without further purification. Solvents were purchased as HPLC grade and used as received. Microwave reactions were performed in a Biotage Initiator Reactor using single mode irradiation with temperature and pressure control and with fixed hold time on. Reactions were monitored by TLC (Merck silica gel 60 F 254 ) analyzed under UV (254 nm), and by UPLC-MS (ESI/UV), using a Waters Acquity system equipped with either an Acquity UPLC HSS C 18 column (1.8 µm, length 50 mm, ID 2.1 mm) with a gradient of water/MeCN (95:5 to 5:95) , with the water eluent containing 1% formic acid (pH 3) or an Acquity UPLC BEH C 18 column (1.7μm, length 50 mm, ID 2.1 mm) with a gradient of water/MeCN (95:5 to 5:95), with the water eluent containing 1% ammonium hydroxide (pH 10). Flash chromatography was performed on a Grace Reveleris X2 instrument using pre-packed silica columns. HPLC purification was performed with formic acid (pH 3) or ammonia (pH 10) as modifier on a preparative HPLC system with an XBridge C 18 column (10 µm, 250 × 50 mm). 1 H and 13 C NMR spectra were recorded at 300 K on a Bruker 500 MHz system equipped with a CryoProbe, operating at 500 MHz and 126 MHz, respectively. The chemical shifts are recorded in ppm relative to the solvent residual peaks: CDCl 3 (7.26 ppm for 1 H and 77.15 ppm for 13 C) or DMSO-d 6 (2.50 ppm for 1 H and 39.52 ppm for 13 C). High resolution LC-MS was detected on a Waters LCTp ToF MS using electrospray ionization (ESI-MS). The MS inlet consisted of a Waters Acquity UPLC system, and the separation was performed on a Waters C18 XBridge at 45-50 °C. The separation was obtained with a 2-95% ACN gradient over 3 min at pH 10 (40 mM NH3 and 5 mM H2CO3). A measure of related impurities was assessed at 210 nM.

Oligonucleotide Synthesis
Oligonucleotide synthesis was performed on an Applied Biosystems 394 automated DNA/RNA synthesizer using standard 1.0 μmol phosphoramidite cycle of acid-catalyzed detritylation, coupling, capping and iodine oxidation. All native β-cyanoethyl phosphoramidite monomers were dissolved in anhydrous MeCN to a concentration of 0.1 M immediately prior to use under argon and coupled for 60 s. The pA phosphoramidite monomer 5, d(pA), was coupled for 10 min. Stepwise coupling efficiencies overall yields were determined by automated trityl cation conductivity monitoring and exceeded 98% for all oligonucleotides synthesized, including oligonucleotides containing d(pA). Cleavage of oligonucleotides from the solid support and deprotection were achieved by treating with concentrated aqueous ammonia for 1 h at RT followed by heating in a sealed tube for 4 h at 55 °C. Purification of oligonucleotides was carried out by reversed-phase HPLC on a Gilson system using a Brownlee Aquapore column (C 8 , 8 mm × 250 mm, 300Å pore diameter) with a gradient of MeCN in an aqueous solution of triethylammonium bicarbonate (TEAB, 0.1M) increasing from 0% to 50% buffer B over 30 min with a flow rate of 4 mL/min (buffer A: 0.1 M TEAB, pH 7.0, buffer B: 0.1 M TEAB, pH 7.0 with 50% MeCN). Elution of oligonucleotides was monitored by UV absorption at 295 or 300 nm. After HPLC purification, oligonucleotides were freeze-dried then dissolved in water without the need for desalting. All oligonucleotides were characterized by electrospray mass spectrometry using a Bruker micrOTOF II focus ESI-TOF MS instrument in ESI-mode. Data were processed using MaxEnt.

Oligonucleotide Analytical Data
The detailed analytical data for four representative modified oligonucleotides are shown.

Molar absorptivity and quantum yield of the pA monomer
To determine the molar absorptivity of the pA nucleoside in water, three samples with known concentration of pA (0.5, 0.75 and 1 μM) were prepared from a 1.0 mM stock in DMSO (the final samples contained 2% DMSO in water). Absorption was measured between 200 and 550 nm using a Cary 5000 (Varian Technologies) with a scan rate of 200 nm·min -1 . Using the Beer-Lambert law, the molar absorptivity at the maximum of the two lowest energy transitions and at 260 nm was determined. The molar absorptivity of the pA nucleoside monomer in other solvents (DMSO, EtOH, DCM, toluene, MeCN), were measured as above at 2 µM concentration.

Hybridization of DNA-strands
Sodium phosphate buffer (12.5 mM phosphate, 100 mM NaCl, 123 mM total Na + , pH 7.5) was used for all measurements unless otherwise stated. Before hybridization, absorption spectra between 200 and 550 nm were recorded on a Cary 4000 or Cary 5000 (Varian Technologies) for each of the single strands. The absorption at 260 nm was used for calculating the strand concentration, where the oligonucleotide molar absorptivity at 260 nm was taken as the linear combination of the molar absorptivity of the individual bases at this wavelength, multiplied by 0.9 to account for the effect of base stacking. The values used for the molar absorptivity of each base at 260 nm are: ε(T) = 9300 M -1 cm -1 , ε(C) = 7400 M -1 cm -1 , ε(G) = 11800 M -1 cm -1 , ε(A) = 15300 M -1 cm -1 , ε(pA) = 22300 M -1 cm -1 and ε(qA nitro ) = 12900 M -1 cm -1 . Hybridization was achieved by mixing each pA modified strand with 15% excess of its complementary strand (to assure full hybridization of the modified strands) at RT followed by heating to 95 °C and, after 10 minutes, cooling to 5 °C over a period of 12 hours. By measuring absorption on the single-stranded DNA and hybridized duplexes (assuming the concentration is given by the absorption at 260 nm using the molar absorptivities of the DNA bases as stated above) the molar absorptivities of pA and qA nitro in ssDNA and dsDNA at the maximum of the lowest energy transition were determined using the Beer-Lambert law.

DNA UV-melting and Circular Dichroism (CD)
DNA UV-melting curves were recorded on a Cary 4000 (Varian Technologies) with a programmable multi-cell temperature block, by heating from 10 or 15 °C (depending on T m ) to 85 °C with a rate of 0.5 °C/min and subsequent cooling to 10 or 15 °C at an identical rate. The absorption at 260 nm was recorded every 0.5 °C for two cycles. The duplex concentration was 2 μM in all measurements. The melting temperatures were calculated as the maximum of the first derivative of the UV-melting curves after FFT-filtered smoothing. Circular dichroism spectra were recorded on a Chirascan CD spectrometer (Applied Photophysics) scanning between 200-450 nm, using an integration time of 0.5 s and four repetitions. The duplex concentration was 4 μM and all spectra were corrected for background contribution.

Fluorescence measurements
Steady-state emission spectra were recorded on a SPEX Fluorolog 3 (JY Horiba) with an excitation wavelength of 353 nm. The concentration of the single-stranded/duplex samples were 4 µM in all steady-state and lifetime measurements. Emission was recorded between 365 and 690 nm at a scan rate of 600 nm min -1 , with the excitation and emission monochromator slit widths set to 1.6 nm. All quantum yields were determined using quinine sulfate ( F = 54.6%) in 0.5 M H 2 SO 4 as reference. The quantum yield of the pA nucleoside monomer was measured as above with an excitation wavelength of 368 nm, recording the emission between 378 and 690 nm.
Fluorescence lifetimes were determined using time-correlated single-photon counting (TCSPC). The samples were excited by a PicoQuant pulsed (10 MHz) laser diode emitting at 377 nm and the emission monochromator was set to 408 nm. The counts were collected by a R3809U-50 microchannel-plate photomultiplier tube (Hamamatsu) and fed into a Lifespec multi-channel analyzer (Edinburgh Analytical Instruments) with 2048 channels. 10000 counts were recorded in the top channel. Reconvolutional fitting to mono-, bi-or triexponential functions was performed using Fluofit Pro v.4 software (PicoQuant GmbH). The average lifetimes were amplitude weighted according to equation S1: where <> is the average lifetime,  i is the i th lifetime and α i is the amplitude of the i th lifetime. Measurements were duplicated. The radiative decay rate (k r ) is calculated using equation S2: The nonradiative decay rate (k nr ) is calculated using equation S3:

FRET measurements
Using the eight sequences shown in Table S5, fifteen duplexes were hybridized: Three with only the donor pA present (at three different positions) and twelve with the donor pA in one strand and the acceptor qA nitro in the opposite strand, resulting in duplexes with 2-13 bases separating the donor and acceptor. Hybridization was performed as described above, using 15% excess of the unmodified or acceptor strands compared to those containing the donor pA. Steady state emission spectra were measured for all samples as described above, but with an excitation wavelength of 370 nm (at which the ratio of absorption between pA and qA nitro is greatest). Emission was recorded between 374 and 690 nm at a scan rate of 600 nm·s -1 , with the excitation and emission monochromator slit widths set to 1.4 nm. Time-resolved decays were measured for all samples using the same settings as described above. The duplex concentration was 4 μM in all measurements. Quantum yields were measured on the three duplexes with only pA present (i.e. no qA nitro ) using an excitation wavelength of 353 nm. Emission was recorded between 360 and 650 nm at a scan rate of 600 nm·min -1 , with the excitation and emission monochromator slit widths set to 1.5 and 1.8 nm respectively.

Data evaluation/curve fitting
The average FRET efficiency (E) of the sample was calculated from the measured data using equation S4.
Where I is the integrated steady-state emission and <> is the average lifetime (calculated using eq S1 above). DA refers to sample containing both donor and acceptor, D refers to samples containing only donor. The FRET efficiency is theoretically given by equation S5.
where R 0 is the Förster radius, the distance between donor and acceptor that gives 50% FRET efficiency, and R DA is the actual distance between donor and acceptor. R 0 is calculated using equation S6.
where J DA is the overlap integral,  2 is an orientation factor that takes the orientation between the transition dipole moments of the donor and acceptor into account, n is the refractive index (set to 1.4) [1] and  D is the quantum yield of the donor.  2 was approximated by equation S7, assuming straight, B-form DNA and planar, static probes positioned inside DNA. [2] where n DA is the number of base pairs between the donor and acceptor,  is the helical rise angle (34.3˚/base pair),  is the phase angle (i.e. the angle between the donor and acceptor transition dipole moments projected onto the base-pair plane) and a is the distance between the center of the chromophore to the center of the DNA helix (3.9 Å). To calculate the distance between donor and acceptor, we use that the average distance between donor and acceptor in the base-pair plane can be expressed as , which allows R DA to be

Vesicle preparation
Lipid vesicles were prepared by the lipid film hydration and extrusion method. [3] Briefly stream to form a lipid film and further dried in vacuum (minimum 2 hrs). This thin dried lipid film was then rehydrated in 1 mL phosphate buffered saline (PBS: 10 mM phosphate buffer, 2.7 mM KCl, 137 mM NaCl, pH 7.4) by vortexing, for a total lipid concentration of 1 mg/mL. The resulting suspension was extruded 11 times through a 30 nm polycarbonate filter at a pressure of 1 bar and stored at 4 °C until usage.

Size distribution and concentration of vesicles with Nanoparticle Tracking Analysis
The size distribution of the extruded vesicles was determined with a Nanoparticle Tracking Analysis (NTA) LM10 device equipped with a 488 nm laser (Malvern, UK) and analyzed using the NTA software version 3.1.
The camera and analysis settings were optimized to enable size distribution and concentration according to the manufacturer recommendations: Camera shutter: 696, camera gain: 73, detection threshold: 2, blur and minimum track length: automatic, minimum expected particle size: 30 nm. The lipid vesicle suspension, measured at a lipid concentration of 0.2 µg/mL, had a vesicle concentration of 3.18×10 9 ±0.04×10 9 particles/mL and a size distribution with a mean diameter of 91.9 ±0.7 nm and a full width at half maximum of 46 nm.

Vesicle functionalization with pA-cholesterol-DNA complex
Prior to observation in the microscope, the vesicles were modified with pA-cholesterol-DNA complex. First, the pA-cholesterol-DNA complex was formed by hybridizing the pA-modified sequence AA to a pair of cholesterolterminated DNA strands (CholC and CholD, the latter containing 5 binding sites for the AA sequence) in a molar ratio of 5:1:1, in a 1:4 volume ratio of Milli-Q water and PBS, for 2 hours. The lipids vesicles were then incubated with the pA-cholesterol-DNA-complex at a 1:1 volume ratio and a molar ratio of 500 pA-cholesterol-DNA complex per vesicle for 30 minutes (final vesicle concentration: 5 nM), leading to an irreversible selfincorporation of the pA-cholesterol-DNA complex into the vesicle lipid membrane. [4]

Total internal reflection fluorescence (TIRF) microscopy and image analysis
All fluorescence micrographs were acquired with an inverted microscope (Nikon Ti Eclipse, Japan) equipped with TIRF system, a 60× magnification (Numerical aperture (NA) = 1.49) oil immersion objective (Nikon Corporation, Tokyo, Japan), TRITC (Rhodmaine-DHPE) and DAPI (pA probe) filter cubes (Nikon Corporation), perfect focus S6 system and an Andor DU-897 X-3530 EMCCD camera (Andor Technology, Belfast, Northern Ireland). A mercury lamp (Intensilight C-HGFIE; Nikon Corporation) connected to the microscope via an optical fiber served as illumination source. Micrographs containing 512 × 512 pixels (0.41 μm/pixel) were acquired with an exposure time of 500 ms. In TIRF mode, only vesicles close to the surface (~100 nm) are excited by the evanescent field. This enables detection of individual vesicles bound to the surface even in the presence of a bulk concentration of suspended, unbound vesicles.

Sample holder and binding of vesicles to surface
Vesicles decorated with pA-cholesterol-DNA complex were unspecifically bound to borosilica coverslips forming the bottom of wells made of polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning, Midland, MI). Briefly, borosilicate coverslips (no. 1; Brand, Germany) were cleaned in boiling 1% Liquinox (Alconox) in water for 30 min, rinsed thoroughly with Milli-Q water then stored in Milli-Q for a maximum of 2 weeks before usage. PDMS wells were formed by first casting PDMS polymer in a 1-3 mm thick sheet which was then cut to fit the geometry of the coverslip and punctured with a hole-punch to create wells. The perforated PDMS sheet was let to adhere to the cleaned borosilica surface to form wells. The wells were then filled with 20 µL PBS buffer and the coverslip was mounted on the sample holder of the microscope. Fluorescently labeled vesicles decorated with pA-cholesterol-DNA complex were then injected in the well (10 µL) to reach a final vesicle concentration of 50 fM. The binding of vesicles was monitored in fluorescence mode in the TRITC filter channel, corresponding to the excitation and collection wavelength of the rhodamine fluorophore. The vesicles were unspecifically adsorbed to the glass surface for approximately 30 seconds to reach adequate surface coverage (~100 particles in field of view) followed by thorough rinsing with PBS buffer to stop the binding of unbound vesicles. Fluorescence signal of the pA probe was observed with DAPI filter. Chart S1. Fluorescent nucleobase analogs listed in Table 5.       Figure S5. Fluorescence quantum yields with standard errors of pA in dsDNA compared with qAN1, one of the brightest adenine FBAs in DNA. [5] Letters on the x-axis denote the bases surrounding pA. [a] Values are reported with their 95% confidence intervals. Duplexes were formed as described in the experimental section, and were measured at RT in phosphate buffer, pH 7.5, 123 mM Na + .

Two-Photon Excitation
A mode-locked Ti:sapphire laser (Coherent Mira pumped by Coherent Verdi), producing pulses of duration ~ 150 fs at a repetition rate of 76 MHz, was used as excitation source. A variable reflective neutral density filter was used to attenuate the excitation beam, which then passed through a dichroic mirror (Semrock Brightline FF735-Di02) and was focused by a 10 objective (Olympus) into the sample solution, contained in a 1 cm pathlength cuvette. Fluorescence emission was collected by the same objective, reflected from the dichroic mirror, passed through a shortpass filter (Semrock Brightline FF01-720/SP-25) and detected by a fibre-coupled spectrometer (Ocean Optics USB2000+), with an acquisition time of 2 s. The incident power was measured using a Coherent FieldMaster power meter. The pA base was dissolved in EtOH (99.9%, Fisher) and the pAmodified oligonucleotides in sodium phosphate buffer, as described above.
The two-photon absorption cross section ( 2 ) of the sample was calculated relative to a reference standard, using equation S10. (S10) where  is the fluorescence quantum yield, η is a term that accounts for the wavelength-dependent collection efficiency of the fluorescence, n is the refractive index of the solvent, C is the concentration, F is the integrated fluorescence signal from the recorded spectrum, P is the excitation power, and superscripts S and R refer to sample and reference, respectively.
The measurements were made at an excitation wavelength of 780 nm, against three different standards, Rhodamine 6G in MeOH ( 2 = 70 GM;  = 0.93) 1 , coumarin 153 in DMSO ( 2 = 11 GM;  = 0.76) 1 and coumarin 153 in toluene ( 2 = 14 GM;  = 0.87). 1 The emission intensities (F) of sample and reference standard were measured as a function of incident laser power (P), for at least 10 different laser powers, and the relative cross section was calculated using the ratio of the gradients of the respective plots of F versus P 2 . The two-photon cross sections have an estimated accuracy of ±10%, due to uncertainty in the cross-sections of the standards and errors in the measurement of the spectral throughput, absorption spectra and emission spectra.
6.1. Table S5   Table S5. Two-photon cross section,  2 , and two-photon brightness,  2 , for pA base measured relative to three reference standards, at excitation wavelength 780 nm.
[b] Using a quantum yield of 80%. Table S6. Two-photon cross section,  2 , and two-photon brightness,  2 , for pA-modified single-strand oligonucleotide GA, measured relative to two reference standards, at excitation wavelength 780 nm.
[b] Using a quantum yield of 42%.
6.3. Table S7   Table S7. Two-photon cross section,  2 , and two-photon brightness,  2 , for pA-modified double-strand oligonucleotide GA, measured relative to two reference standards, at excitation wavelength 780 nm. [a] From reference [5]. The error in the standard values of  2 is reported as 8% [b] Using a quantum yield of 14%.