Synthesis, Characterization and Application of a Highly Hydrophilic Triarylmethyl Radical for Biomedical EPR

Abstract

Stable tetrathiatriarylmethyl radicals have significantly contributed to the recent progress in biomedical EPR due to their unmatched stability in biological media and long relaxation times. However, the lipophilic core of the most commonly used structure (Finland trityl) is responsible for its interaction with plasma biomacromolecules such as albumin, and self-aggregation at high concentrations and/or low pH. While Finland trityl is generally considered inert towards many reactive radical species, we report that sulfite anion radical efficiently substitutes the three carboxyl moieties of Finland trityl with a high rate constant of 3.53 ×10⁸ M⁻¹s⁻¹ leading to a tri-sulfonated Finland trityl radical. This newly synthesized highly hydrophilic trityl radical shows an ultranarrow linewidth (ΔB_pp=24 mG), a lower affinity for albumin than Finland trityl, and a high aqueous solubility even at acidic pH. Therefore this new tetrathiatriarylmethyl radical can be considered as a superior spin probe in comparison to the widely used Finland trityl. One of its potential applications was demonstrated by in vivo mapping oxygen in a mouse model of breast cancer. Moreover, we showed that one of the three sulfo groups can be easily substituted with S-, N-, P- nucleophiles opening access to various mono-functionalized sulfonated trityl radicals.

Graphical Abstract

1. Introduction

Triarylmethyl radicals (TAM, trityl) of type tetrathiatriarylmethyl are unique spin probes and spin labels used for biomedical EPR applications. These radicals exhibit unmatched properties, such as ultranarrow linewidth (e.g. <40mG for deuterated Finland trityl (dFT), Fig. 1) due to their long relaxation times and exceptional stability within biological media.¹ These features make tetrathiatriarylmethyl radicals superior spin probes and labels by comparison to the commonly used nitroxide radicals. Indeed, the latter being generally hampered by their fast reduction in biological media and their broad linewidths.² The most representative structures of tetrathiatriarylmethyl radicals are the Finland trityl (FT), OX063 and their deuterated analogs (dFT and OX071)(Fig. 1).

Figure 1.

Structure of FT, OX063 radicals, and their deuterated analogs.

Because of their unprecedented properties, tetrathiatriarylmethyl radicals have been widely used as spin probes for in vivo low-field EPR to measure physiologically relevant parameters such as pO₂, pH, and inorganic phosphate concentration (Pi).^3–4 Besides, as spin labels of biomacromolecules, they have enabled distance measurement in DNA and proteins for structural biology studies using dipolar EPR spectroscopy.^5–7 Tetrathiatriarylmethyl radicals are efficient hyperpolarizing agents for dissolution dynamic nuclear polarization (d-DNP)^8–9 and Overhauser-enhanced magnetic resonance imaging (OMRI).^10–11 Very recently, we demonstrated that a tetrathiatriarylmethyl radical labeled ¹³C at the central carbon enables the measurement of the probe rotational correlation time with applications to measure microviscosity and molecular dynamics by EPR.¹²

The developments of tetrathiatriarylmethyl radicals relied exclusively on modifications of the Finland trityl but not of the OX063/71 derivatives.^13–18 This is mostly due to the availability of large scale chromatographic column-free syntheses of (d)FT^19–21, while the syntheses of hydroxyethyl derivatives OX063/71 are twice longer and have only been reported very recently.²² Unfortunately, for in vivo applications, FT-based derivatives show unwanted lipophilic interactions with plasma biomacromolecules, such as albumin, which in turn result in a broadening of the EPR line.²³ Therefore when injected in vivo only the unbound fraction of FT-based TAMs is detected. Also, Finland trityl aggregates at high saline concentration and at low pH.²⁰ Therefore, a more hydrophilic structure of FT would be highly desirable to address these limitations. Hereby, we report on the synthesis, characterization and application of a highly hydrophilic sulfonated triarylmethyl radical for biomedical EPR applications.

2. Experimental Section

2.1. Generals.

HPLC analysis was carried out using a Waters Alliance e2695 separation module, equipped with a 2998 PDA detector and an SQD2 Mass Detector. A XBridge BEH C18 4.6 mm × 50 mm, 2.5 μm column was used for separation with the following conditions: flow rate, 1.5 mL/min; column temperature, 40°C, gradient, t_0min: H₂O – 80%, ACN – 10%, H₂O with 1% TFA – 10% ; t_5min: H₂O – 0%, ACN – 90%, H₂O with 1% TFA – 10%; t_6min: H₂O – 0%, ACN – 100%, H₂O with 1%TFA –0%. HRMS spectra were recorded using a Thermofisher Scientific Q Exactive Mass Spectrometer with an Electron Spray Ionization (ESI) source. CW X-band EPR spectra were recorded using an ELEXSYS E580 EPR spectrometer (Bruker, Germany). For EPR spectra recorded under controlled gas composition, the solutions were filled into a gas-permeable Teflon tube (diameter, 1.14 mm, wall thickness, 60 μm, Zeus, Inc., USA) and a gas flow of a controlled composition was applied using a temperature and gas controller (Noxygen, Germany). The g-factors were determined using dFT as an internal standard (g=2.00307).²⁴ UV-Vis measurements were performed using an Agilent Carry 60 spectrophotometer. Preparative purification of dFT-(SO₃)₃ was carried out using a Teledyne CombiFlash Rf+ purificator and a C18 column. dFT radical was synthesized according to reported procedures.^20–21 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was purchased from Dojindo Molecular Technologies. All other reagents and solvents were purchased from Sigma Aldrich and Fisher Scientific and used without further purification.

2.2. Chemistry.

2.2.1. dFT in aerated Na₂SO₃ solution.

dFT was stirred for 2.5 hours in aerated aqueous sodium sulfite solution, the final concentrations were 200 μM for dFT and 200 mM for sodium sulfite. Then, an aliquot was transferred into a gas-permeable EPR capillary and EPR spectra were recorded under constant nitrogen flow and also analyzed by HPLC-MS. The EPR acquisition parameters were: ModAmp 0.020 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 2.5 G, 2048 pts.

2.2.2. dFT in sulfite anion radical (SO₃^•-) generating system.

Potassium dichromate in water was added to a solution of dFT (final concentration = 200 μM) in aqueous sodium sulfite (final concentration = 200 mM) under nitrogen. The final concentration of K₂Cr₂O₇ were 1, 3, 5, 10, 20 mM. Shortly after mixing, an aliquot was transferred into a gas-permeable EPR capillary, and EPR spectra were recorded at 37°C under constant nitrogen flow. In addition, HPLC-MS analysis were performed for [K₂Cr₂O₇] = 10 mM after 5 and 90 minutes. EPR spectrum was recorded using the following acquisition parameters: ModAmp 0.030 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 2.5 G, 1024 pts. The confirmation of the formation of sulfite anion radical in the system was verified by spin trapping of SO₃^•- using DMPO. K₂Cr₂O₇ (final concentration = 3 mM) was added to a solution containing DMPO (final concentration = 20 mM), dFT (final concentration = 200 μM) and sodium sulfite (final concentration = 200 mM) and under nitrogen at room temperature. EPR spectrum was recorded using the following acquisition parameters: ModAmp 0.7 G, ModFreq 30.00 kHz, ConvTime 20.00 ms, Power 15 mW, SweepWidth 70 G, 2048 pts.

2.3. Determination of rate constant of the mono-sulfonation reaction.

DMPO was used as a competitor for sulfite anion radical. In the presence of a competitor, the generated sulfite anion radical reacts with DMPO and dFT with a second-order rate constants k₁ and k_2, correspondingly.

$$S{O}_3^{·-}+\mathit{\text{DMPO}} \to \mathit{\text{DMPO}}-S{O}_3{k}_1$$

$$S{O}_3^{·-}+dFT \to dFT-{\left(S{O}_3\right)}_1{k}_2$$

With DMPO: in the steady-state regime, the rate of generation of sulfite anion radical, V_gen, is equal to its consumption by dFT and DMPO.

$$0=V_{gen}-k_1\left[\mathit{\text{DMPO}}\right]\left[S{O}_3^{·-}\right]-k_2\left[dFT\right]\left[S{O}_3^{·-}\right]$$ (1)

The steady-state concentration of sulfite anion radical can be derived from equation 1.

$$\left[S{O}_3^{·-}\right]=\frac{V_{gen}}{k_1\left[\mathit{\text{DMPO}}\right]+k_2\left[dFT\right]}$$ (2)

Using equation 2, the rate of dFT-(SO₃)₁ formation in the presence of the competitor, V_DMPO, equals:

$$\frac{d}{dt}[dFT-\left(S{O}_3{)}_1\right]=V_{\mathit{\text{DMPO}}}=k_2\left[dFT\right]\left[S{O}_3^{·-}\right]=k_2\left[dFT\right]\frac{V_{gen}}{k_1\left[\mathit{\text{DMPO}}\right]+k_2\left[dFT\right]}$$ (3)

Without DMPO: in the steady-state regime, the rate of sulfite anion radical generation equals the rate of dFT-(SO₃)₁ formation, V_{no DMPO}:

$$0=V_{gen}-k_2\left[dFT\right]\left[S{O}_3^{·-}\right]$$ (4)

$$\frac{d}{dt}[dFT-\left(S{O}_3{)}_1\right]=V_{\mathit{\text{no DMPO}}}=k_2\left[dFT\right]\left[S{O}_3^{·-}\right]$$ (5)

$$V_{gen}=k_2\left[dFT\right]\left[S{O}_3^{·-}\right]=V_{\mathit{\text{no DMPO}}}$$ (6)

By combining equations 3 and 6, we can derive the rate of dFT-(SO₃)₁ formation in the presence of the DMPO competitor:

$$V_{\mathit{\text{DMPO}}}=k_2\left[dFT\right]\frac{V_{\mathit{\text{no DMPO}}}}{k_1\left[\mathit{\text{DMPO}}\right]+k_2\left[dFT\right]}$$ (7)

The ratio between the rate of dFT-(SO₃)₁ formation in the absence and presence of the competitor can be derived from equation 7:

$$\frac{V_{\mathit{\text{no DMPO}}}}{V_{\mathit{\text{DMPO}}}}-1=\frac{k_1\left[\mathit{\text{DMPO}}\right]}{k_2\left[dFT\right]}$$ (8)

K₂Cr₂O₇ (final concentration = 3 mM) was added to a solution containing dFT (final concentration = 200 μM), Na₂SO₃ (final concentration = 200 mM) and DMPO (final concentrations = 0, 5, 10, 15, 20 mM) in water under nitrogen at room temperature. The total volume was 100 μL. Immediately after mixing, the solution was filled into a gas-permeable Teflon EPR capillary and placed inside the X-band EPR cavity. A flow of nitrogen was maintained throughout the experiment. The decay of dFT concentration was followed for 25 minutes, corresponding to a maximum of 15% conversion. The dFT decay allows to determine V_{no DMPO} (0 mM DMPO) and V_DMPO (5, 10, 15, 20 mM DMPO). The plot $\frac{V_{\mathit{\text{no DMPO}}}}{V_{\mathit{\text{DMPO}}}}-1$ vs. [DMPO] allows to determine k₂=3.53 ×10⁸ M⁻¹s⁻¹ using k₁=1.2×10⁷ M⁻¹s⁻¹.²⁵ EPR spectrum was recorded using the following acquisition parameters: ModAmp 0.020 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 2.5 G, 2048 pts.

2.4. EPR characterization.

The anoxic peak-to-peak (ΔB_pp) linewidth of dFT-(SO₃)₃ was measured from a spectrum of 50 μM in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) using the following parameters: ModAmp 0.010 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 0.5 G, 1024 pts. For the measurement of the ¹³C and ³³S satellites, an aliquot from the C18 purification was analyzed directly without evaporation of the solvent. The solvent composition was 13% ACN in water containing 0.3% TFA. EPR acquisition parameters were: ModAmp 0.010 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 18.0 G, 8192 pts.

2.4.1. Computational chemistry.

The geometry of dFT-(SO₃)₃ was optimized at the UB3LYP/6–31G* level of theory using ORCA 4.2.0 computational package.²⁶ The isotropic hyperfine splittings with ¹³C and ³³S were calculated using the “eprnmr” ORCA keyword for a single point calculation using the IGLO-III basis sets for the optimized geometry. The UB3LYP functional have previously given good predictions of experimental results for this family of trityl radicals.^27–28 The sulfo groups were treated as protonated rather than sodium salts.

2.5. Interaction with albumin, solubility at low pH, and stability under reducing environment and in cell lysate.

To investigate the interaction with albumin, dFT, dFT-(SO₃)₃ and OX063 (final concentration = 50 μM) were incubated with various concentrations of bovine serum albumin (BSA) (final concentration from 0 to 1 mM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) under nitrogen. In order to compare TAMs with different intrinsic linewidths, the EPR signal intensities were measured for each BSA concentration and normalized to EPR intensity without BSA (I_BSA/I_{No BSA}). The EPR acquisition parameters were: ModAmp 0.020 G, ModFreq 10.00 kHz, ConvTime 20.00 ms, Power 0.02993 mW, SweepWidth 0.5 G, 1024 pts. To investigate the solubility of dFT-(SO₃)₃ at low pH, the EPR spectra of dFT-(SO₃)₃ (100 μM) were recorded at pH 1.1 and 7.4 and the EPR signal intensities were compared. The EPR acquisition parameters were: ModAmp 0.100 G, ModFreq 30.00 kHz, ConvTime 40.00 ms, Power 0.09464 mW, SweepWidth 5.0 G, 1024 pts. The stability of dFT-(SO₃)₃ under reducing environment was investigated by mixing dFT-(SO₃)₃ (final concentration = 50 μM) with ascorbate (final concentration = 500 μM) in mixture with reduced glutathione (final concentration = 650 μM).²⁹ The EPR time-course was recorded under nitrogen flow for half an hour. The EPR acquisition parameters were: ModAmp 0.025 G, ModFreq 30.00 kHz, ConvTime 40.00 ms, Power 0.09464 mW, SweepWidth 5.0 G, 1024 pts. The stability in cell lysate was investigated by incubating dFT-(SO₃)₃, dFT and OX063 (200 μM) with MDA-MB-231 cell lysate at 37°C under nitrogen for 30 min. Briefly, 2 × 10⁷ MDA-MB-231 cells grown in DMEM supplemented with 10% FBS were mechanically collected by scraping in PBS and subsequently centrifuged at 450 rcf for 5 min at 4°C to pellet the cells. The packed cell volume of the cells was about 75 μL. The cells were resuspended in 500 μL ice-cold lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40 (IGEPAL CA630), 1 mM EDTA, 1.5 mM MgCl₂, 5% glycerol and Halt protease inhibitor cocktail (Pierce) and incubated on ice for 30 min. The lysate was clarified by centrifugation at 20.000 rcf for 10 min at 4°C and the supernatant was collected. 1 μL of a 10 mM stock solution of dFT-(SO₃)₃, dFT or OX063 was added to 49 μL of cell lysate. The mixture was filled into a gas-permeable Teflon EPR capillary and placed inside the X-band EPR cavity and EPR time-courses were recorded under nitrogen flow for half an hour at 37°C. The EPR acquisition parameters were: ModAmp 0.030 G, ModFreq 30.00 kHz, ConvTime 30.00 ms, Power 0.09464 mW, SweepWidth 5.0 G, 1024 pts.

2.6. Oxygen sensitivity and in vivo oxygen imaging.

To determine the oxygen sensitivity, dFT-(SO₃)₃ (50 μM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) was filled into a gas-permeable Teflon tube and placed in the EPR cavity. EPR spectra were recorded under various oxygen-nitrogen ratio. The peak-to-peak linewidths were measured for pO₂ = 0, 34.2, 76, 114, 152 mmHg. The EPR acquisition parameters were: ModAmp 0.015 G, ModFreq 30.00 kHz, ConvTime 40.00 ms, Power 0.09464 mW, SweepWidth 3.0 G, 1024 pts.

2.6.1. In vivo oxygen mapping.

Animal studies were performed according to the approved WVU Institutional Animal Care and Use Committee (IACUC) protocol.

5×10⁵ C57Bl/6J PyMT mouse breast tumor cells (ATCC® CRL-3278™) were orthotopically implanted into the number 4 abdominal mammary glands of female C57Bl/6J mice. Tumor volumes were determined using calipers measuring the shortest and perpendicular longest dimensions. The tumor oxygen was imaged when the tumor reached a volume of approximately 257 mm³ using the equation V=0.5 × (shortest dimension² × longest dimension). Briefly, 50 μL of a 2 mM solution of sulfonated trityl radical dFT-(SO₃)₃ in 10 mM phosphate buffer saline was injected intratumorally in an anesthetized mouse. Immediately after, the animal was placed in a custom-built Rapid-Scan EPR imager operating at 800 MHz.³⁰ Acquisition parameters were as follows: rapid scan frequency, 9.4 kHz; number of projections, 2546; maximum gradient, 3G/cm; acquisition time, 8.25 min. An integral intensity threshold of 30% was implemented to remove low signal-to-noise data.

2.7. Generation of tri-sulfonated TAM cation and Ipso-substitution by S-, N-, P- nucleophiles.

The stability of dFT-(SO₃)₃ cation was investigated by UV-Visible spectroscopy. A quartz cuvette was filled with dFT-(SO₃)₃ in phosphate buffer (100 mM, pH=7.4) at room temperature, then one equivalent of K₂IrCl₆ was added. Immediately after mixing the spectra were recorded every 1 minute for 10 minutes. The total volume was 3 mL and the final concentrations of dFT-(SO₃)₃ and K₂IrCl₆ were 50 μM.

3. Results and Discussion

3.1. Synthesis

In order to overcome the issue arising from the lipophilicity of Finland trityl, we sought to substitute the three carboxyl groups by a more hydrophilic moiety, which does not give rise to an additional hyperfine splitting in the EPR spectrum. The sulfo functional group seemed to be the ideal candidate to meet these requirements. Our first attempt to introduce the sulfo group through a classical addition of the aryllithium 2, generated from the deprotonation of 1 with tert-Butyl lithium, to sulfur trioxide pyridine complex (Scheme 1) failed to provide the desired compound 3 as a complex mixture was obtained.

Scheme 1.

The unsuccessful attempt of the synthesis of sulfonated trityl 3.

Serendipitously, we noticed that a solution of deuterated Finland trityl (dFT) (200 μM) in aerated sodium sulfite (200 mM) stirred for 2.5 hours yielded to the formation of a new EPR line observed at X-Band (Fig. 2A). The HPLC-MS analysis of the mixture revealed the presence of 17% of a new peak with a retention time (RT) of 3.4 min and a molecular weight (m/z=1071.4) before the peak of dFT (RT=4.1 min, m/z=1035.1), consistent with the substitution of one carboxyl group by a sulfo group (Fig. 2B and Fig. S1). The same reaction carried out under strictly deoxygenated conditions did not result in the formation of this new compound (data not shown).

Figure 2.

A) EPR spectrum of dFT radical (200 μM) in aerated sodium sulfite (200 mM) after stirring for 2.5 hours B) Reverse phase HPLC-MS chromatogram of dFT radical under the same conditions.

By analogy with the sequential substitution of the carboxyl groups of dFT by a nitro group with NO₂^• described previously^{17, 31}, we hypothesized the mono-sulfonated TAM resulted from the reaction of dFT with sulfite anion radical (SO₃^•-) generated in aerated sodium sulfite solutions. In order to drive further the reaction, we incubated dFT (200 μM) in a sulfite anion radical generating system composed of sodium sulfite (200 mM) and potassium dichromate (from 1 to 20 mM) under nitrogen. The reactions were monitored by X-band EPR and HPLC-MS. The EPR spectra (Fig. 3A) show the conversion of the single EPR line of dFT to three new lines at higher magnetic fields with different g-factors (Table S1). At a high concentration of potassium dichromate (20 mM), we observed a fast conversion of dFT EPR line to the line at the highest magnetic field. The HPLC-MS chromatograms show similar results (Fig. 3B and Fig. S2), the dFT peak (RT=4.1 min, m/z=1035.1) was progressively converted to three more polar compounds with peaks (RT₁=3.4 min, m/z=1071.4; RT₂=2.6 min, m/z=1107.3, RT₃=1.8 min, m/z=1144.3) corresponding to the mono-, di-, and tri-sulfonated TAM respectively (Fig. 3C) as determined by their mass spectra.

Figure 3.

A) X-band EPR of the dFT (200 μM) in sodium sulfite (200 mM) in the presence of various concentration of potassium dichromate (1, 3, 5, 10, 20 mM) at 37°C under nitrogen showing the sequential sulfonation of dFT radical into tri-sulfonated TAM radical B) Reverse phase HPLC-MS of a solution containing dFT (200 μM) in sodium sulfite (200 mM) in potassium dichromate (10 mM) under nitrogen after 5 and 90 minutes C) Chemical scheme of the sequential sulfonation of dFT radical into tri-sulfonated dFT radical.

In order to verify that sulfite anion radical was generated in this system, a solution of dFT (200 μM) was incubated with Na₂SO₃ (200 mM) and K₂Cr₂O₇ (3 mM) in the presence of the DMPO (20 mM) spin-trap at room temperature. The EPR spectrum showed the formation of the characteristic DMPO-SO₃ nitroxide adduct³² (Fig. S3), confirming the formation of sulfite anion radical. Based on this result, our proposed mechanism involves the addition of sulfite radical at the para-position of the aryl ring, which triggers an oxidative decarboxylation as reported previously for the substitution of carboxyl moieties of Finland trityl by nitro groups with NO₂^• (Scheme 2).^{17, 31}

Scheme 2.

Proposed mechanism for the substitution of the carboxyl moieties of dFT by sulfo groups.

The reaction was easily scaled up to 1 g without any difficulty allowing to isolate the tri-sulfonated TAM in 95% yield after purification using a C18 column (see S.I. for detailed procedure).

3.2. Determination of the rate constant

In order to determine the rate constant of the reaction between dFT and sulfite anion radical, competition kinetics method was used with DMPO as a competitor (Fig. 4A). To isolate the first step of the reaction (mono-sulfonation), the reactions were monitored by EPR to a maximum of 15 percent conversion of dFT which does not result in the formation of the di-sulfonated trityl radical (dFT-(SO₃)₂). The reactions were performed with various DMPO concentrations (0, 5, 10, 15, 20 mM) and the initial rates were determined for each kinetics (Fig. 4B). The rate ratio without and with the competitor allows to determine the second-order rate constant of the reaction which equals 3.53 ×10⁸ M⁻¹s⁻¹ (Fig. 4C). It is worth noting that such a high rate constant will allow further applications of dFT to trap sulfite anion radicals with EPR or HPLC detection.^32–33

Figure 4.

A) Scheme of the reaction of dFT with SO₃^•- in the presence of DMPO as a competitor B) Initial rates of the reaction with various concentrations of DMPO (0, 5, 10, 15, 20 mM) C) Rates ratio without and with various concentrations of DMPO allowing to determine the rate constant of 3.53 ×10⁸ M⁻¹s⁻¹ for the mono-sulfonation reaction.

3.3. EPR spectrum

The EPR spectrum of dFT-(SO₃)₃ (50 μM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) recorded at X-band under nitrogen shows a single line with an ultranarrow peak-to-peak linewidth of ΔB_pp=24 mG (Fig. 5A). To our knowledge, this is the narrowest linewidth of all TAMs reported to date. A narrow linewidth leads to a high-intensity signal which is of first importance for in vivo applications where the signal-to-noise ratio (SNR) is always a critical issue. Also, the spectrum shows additional low-intensity peaks corresponding to the ¹³C satellites (the ¹³C natural abundance is 1.1 %, I=1/2) of the aryl rings (Fig. 5B). The values of all ¹³C hyperfine splitting constants (hfs) show high similarity with those reported for dFT (Table 1). Interestingly, a quartet: 1:1:1:1 resulting from the interaction between free electron with the three ³³S (the ³³S abundance is 0.75%, I=3/2) of the sulfo groups is also visible (Fig. 5B, red dots). In order to confirm the attribution of this quartet to the ³³S of the sulfo groups, we computed the hfs using density functional theory calculations. First, the geometry of dFT-(SO₃)₃ was optimized at the UB3LYP/6–31G* level of theory (Fig. S4). Then, we performed a single-point calculation using IGLO-III basis set to compute the isotropic hfs. The measured hfs=0.65 G is in good agreement with the value calculated: −0.70 G. Note that the measurements were performed using CW EPR which does not allow to determine the sign of the hfs. In addition, all calculated ¹³C satellites hfs are in good agreement with the measured hfs. Notably, similar small ³³S coupling was not observed for a sulfonated perchlorinated TAM because of it ~ 30 times larger linewidth compared to dFT-(SO₃)₃.³⁴

Figure 5.

A) X-band EPR spectrum of dFT-(SO₃)₃ (50 μM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) under nitrogen B) Spectrum with 18 G sweep width zoomed on the ¹³C and ³³S satellites.

Table 1.

Experimental and calculated hfs values for ¹³C and ³³S satellites.

Nucleus (spin)	Hyperfine splitting, G
	Experimentala		Calculated for dFT-(SO3)3b	Degeneracy	Assignment
	Finland tritylc	dFT-(SO3)3
33S (3/2)	N.A.	0.65	−0.70	3	● Sulfonate
13C (1/2)	0.18	0.18	−0.27	6	○ Thio-ketal
13C (1/2)	2.35	2.18	−2.28	6	♦ 3,5-Phenyl
13C (1/2)	3.36	3.58	3.47	3	▲ 4-Phenyl
13C (1/2)	9.03	9.00	9.77	6	■ 2,6-Phenyl
13C (1/2)	11.46	11.10	−11.41	3	◊ 1-Phenyl

^a-The measurements were performed using CW EPR which does not allow to determine the sign of the hfs.

^b-Calculated at UB3LYP/IGLO-III//UB3LYP/6–31G* level of theory.

^c-Converted from the values in MHz measured at 250 MHz.²⁷

3.4. Interaction with albumin, solubility at low pH, and stability under reducing environment and in cell lysate

To investigate whether the substitution of the three carboxyl groups of dFT by sulfo groups decreases the interaction with albumin, EPR spectra of dFT, and dFT-(SO₃)₃ in the presence of various concentrations of bovine serum albumin (BSA) were recorded (Fig. 6). While dFT loses 80% of its EPR signal intensity upon incubation with one equivalent (BSA) due to a line broadening, indicating a strong interaction²³, dFT-(SO₃)₃ loses less than 25% showing a lower affinity to albumin. This result can be explained by the higher hydrophilicity of sulfo group by comparison to the carboxyl moiety. Conversely, the EPR intensity of OX063 recorded under the same conditions did not change upon incubation with BSA, showing the total absence of interaction with albumin. Spin labels based on FT have shown unwanted non-specific bindings to hydrophobic sites of proteins and membranes.^{23, 35} OX063/71 are more hydrophilic but have a slightly larger molecular volumes (50% larger molecular volumes, considering radius of 7 Å and 8 Å for dFT-(SO₃)₃ and OX063/71³⁶, respectively) which could interfere with the structure of macromolecule under investigation. dFT-(SO₃)₃ has the advantage of high hydrophilicity without increasing the size of the label.

Figure 6.

The normalized intensity of the EPR signal of dFT, dFT-(SO₃)₃ and OX063 (50 μM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) upon incubation with various concentrations of BSA (0 – 1mM).

The protonation of the carboxyl groups of dFT is responsible for its aggregation at pH lower than 5.²⁰ The substitution of the carboxyl functions by sulfo groups, with a pKa<0 keep the molecule charged at any physiological pH, prevents this aggregation. Indeed, no change in the EPR signal intensity was observed between the spectra of dFT-(SO₃)₃ (100 μM) recorded at pH 7.4 and 1.1 (data not shown). Importantly, the introduction of three sulfo withdrawing groups, do not result in an instability toward reducing agents. Indeed, dFT-(SO₃)₃ (50 μM) incubated in the presence of an excess of ascorbic acid (500 μM) in mixture with glutathione (GSH) (650 μM)²⁹ for more than 30 minutes under nitrogen did not result in a noticeable decrease of the EPR signal intensity, demonstrating its stability toward those biological reducing agents (Fig. S5A). Note that this result contrast with the fast reduction by ascorbic acid observed for a sulfonated perchlorinated TAM.³⁴ Also, dFT-(SO₃)₃ (200 μM) incubated with lysate of MDA-MB-231 triple-negative breast cancer cell line for 30 minutes under nitrogen at 37°C resulted in less than 3% decrease of the EPR signal intensity (Fig. S5B). Similar results were obtained with OX063 with 1% decay while dFT decayed by 9% under the same conditions (Fig. S5B).

3.5. Oxygen sensitivity and EPR imaging of tumor oxygenation in vivo

While molecular oxygen is a paramagnetic substance, it cannot be measured directly in solution by EPR due to its fast relaxation times. However, the Heisenberg spin exchange between dissolved molecular oxygen and a free radical such as a TAM or a nitroxide results in a broadening of the EPR linewidth of the radicals.^{2, 37} This effect can be used to measure oxygen concentration, including in vivo.² Comparing to nitroxides, tetrathiatriarylmethyl radicals show a higher sensitivity to oxygen due to one order of magnitude smaller anoxic linewidth (typically ~100 mG vs. ~1 G for nitroxides) resulting in a relative larger linewidth variation.^{2, 37} The EPR spectra of dFT-(SO₃)₃ (50 μM) in PBS were recorded with different oxygen partial pressures (pO₂) allowing to determine an oxygen peak-to-peak linewidth broadening of 0.52 mG/mmHg pO₂ (Fig. 7). This sensitivity is similar to the 0.53 mG/mmHg pO₂ previously reported for dFT³⁸ demonstrating that dFT-(SO₃)₃ can be used as an oximetric spin probe.

Figure 7.

X-band calibration curve of the peak-to-peak linewidth of dFT-(SO₃)₃ (50 μM) in phosphate buffer saline (10 mM, pH=7.4, NaCl 140 mM) at room temperature with respect to pO₂ leading to a sensitivity of 0.52 mG/mmHg.

To demonstrate the application of dFT-(SO₃)₃ for EPR oximetry in vivo, we performed 4D spectral-spatial imaging (3D-spatial, 1D-spectral) in a breast cancer tumor of an MMTV-PyMT mouse model using a home-build 800 MHz EPR rapid-scan imager, the linewidth in each voxel allows for pO₂ mapping. Figure 8A shows the oxygen distribution in a slice of the tumor in the x,z-plane. Figure 8B shows the histogram of pO₂ for the entire image. The pO₂ values extracted are consistent with those previously reported in the literature for the same tumor model.⁴

Figure 8.

A) In vivo 4D spectral-spatial imaging (3D-spatial, 1D-spectral) in an MMTV-PyMT mouse model of breast cancer. A slice of oxygen distribution in the x,z-plane is represented. B) Histogram of oxygen pO₂ distribution for the whole image.

3.6. Generation of tri-sulfonated TAM cation and Ipso-substitution by S-, N-, P- nucleophiles

The ability to replace selectively one sulfo group by a different functional group would expand the applications of the sulfonated tetrathiatriarylmethyl radicals for biomedical EPR applications. Previously, an Ipso-nucleophilic substitution of one carboxyl or hydrogen of Finland trityl derivatives by various nucleophiles was reported.^{17, 39–40} This strategy is based on the reaction of the trityl cation, generated in situ, with S-, N-, P- nucleophiles. We investigated whether a similar strategy could be applied to dFT-(SO₃)₃. First, dFT-(SO₃)₃ (50 μM) was oxidized to its cation using one equivalent of the Ir(IV) water-soluble oxidizing agent K₂IrCl₆ (50 μM) in 100 mM phosphate buffer (Fig. 9A), this transformation is accompanied by decrease of the radical UV-Vis absorption peak at λ=465 nm and an increase of a peak at λ=790 nm, corresponding to the trityl cation⁴¹ (data not shown). If no nucleophile is added, a water molecule can add at the para-position of the aryl group (Fig. 9A, step 2), leading to 4, which undergoes an oxidative desulfonation yielding the phenol 5, which in turns is further oxidized to the quinone methide QM with a UV-Vis absorption peak at λ=520 nm (Fig. 9B).^{17, 41} The half-life of the trityl cation was estimated to be 2 minutes in phosphate buffer pH=7.4 at room temperature (Fig. 9B). Monitored by UV-Vis, the decay of the trityl cation peak at λ=790 nm is accompanied by an increase of two peaks, one at λ=465 nm and the other one at λ=520 nm corresponding to the trityl radical and QM respectively.⁴¹ The formation of QM involves 2 oxidation steps, the HPLC-MS analysis at the end of the reaction shows a ~ 2:1 ratio between dFT-(SO₃)₃ radical and QM (m/z=1079.1) (data not shown). The presence of dFT-(SO₃)₃ radical that was generated back from the cation indicates that the trityl cation is the oxidant for the steps 3 and 4 in figure 9.

Figure 9.

A) Oxidation of dFT-(SO₃)₃ (50 μM) in phosphate buffer (100 mM, pH=7.4) at room temperature with K₂IrCl₆ (50 μM) and conversion to QM. B) 10 minutes time course of UV-Vis spectra of the conversion of trityl cation (50 μM) to QM and dFT-(SO₃)₃.

In order to investigate whether a strong nucleophile can outcompete the addition of water on the trityl cation, various nucleophiles (20 equivalents) were added immediately after the oxidation of the radical. Gratifyingly, we found that S-, N-, P- nucleophiles such as glutathione (GSH), dimethylamine (NH(CH₃)₂), tri-n-butylphosphine (P(n-Bu)₃), or trimethylphosphite (P(OMe)₃) were able to yield the mono-functionalized sulfonated trityl radicals (Fig. 10A). Note that the trimethylphosphite leads to the mono-methylphosphonic ester derivative. The EPR spectra and the HRMS of the purified compounds were in agreement with their molecular structure (Fig. 10B and Fig. S7–S10), the hyperfine splitting constants were obtained from the simulations (Fig. S11). Notably, the mono-GSH derivative 6 exists in a 50:50 mixture of two diastereoisomers arising from the chirality of trityl core (M, P propeller conformations)^{28, 42} and the L-glutathione. Note that, the EPR spectrum of 6 reveals a conjugation through the thiol, as the amine would give additional hyperfine splitting to the nitrogen nucleus (I_N=1). The mono-substitution of dFT-(SO₃)₃ will allow for the selective synthesis of unsymmetrical sulfonated trityl radicals with numerous functionalities.

Figure 10.

A) Ipso-nucleophilic substitution of dFT-(SO₃)₃ with glutathione (GSH), dimethylamine (NH(CH₃)₂), tri-n-butylphosphine (P(n-Bu)₃) or trimethylphosphite (P(OMe)₃) nucleophiles. B) X-band EPR spectra of the mono-functionalized sulfonated trityl. Hfs obtained from simulations: for 6 (GSH) diastereoisomer 1: a_H1=0.03 G, a_H2=0.10 G; diastereoisomer 2: a_H1=0.06 G, a_H2=0.10 G; for 7 a_N=0.18 G, a_H(6)=0.40 G; for 8 a_P=3.38 G, a_H(6)=0.05 G; for 9 a_P=3.77 G.

4. Conclusions

In conclusion, we have shown that sulfite anion radical substitutes the carboxyl groups of Finland trityl with a high rate constant leading to the highly hydrophilic tri-sulfonated TAM probe, dFT-(SO₃)₃, showing an ultra-narrow linewidth, and a lower affinity toward albumin than Finland trityl. We demonstrated the application of dFT-(SO₃)₃ for mapping oxygen concentration in vivo in a mouse model of breast cancer. Moreover, we showed the dFT-(SO₃)₃ can be selectively mono-functionalized allowing to expand its applications for biomedical EPR. Our work will enable the synthesis of new sulfonated TAM spin probes, such as a monophosphonated derivative for concurrent pO₂, pH, and Pi measurements or methanethiosulfonate (MTS), maleimide spin labels for distance measurement in biomacromolecules using dipolar EPR spectroscopy.

ABBREVIATIONS

EPR

electron paramagnetic resonance

FT

Finland trityl

TAM

triarylmethyl

OMRI

Overhauser-enhanced magnetic resonance imaging

DMPO

5,5-dimethyl-1-pyrroline N-oxide

hfs

hyperfine splitting

BSA

bovine serum albumin

QM

quinone methide

GSH

glutathione

TFA

trifluoroacetic acid

ACN

acetonitrile

FBS

Fetal Bovine Serum‎