Direct Comparison of Triggering Motifs on Chemiluminescent Probes for Hydrogen Sulfide Detection in Water

Abstract

Hydrogen sulfide (H₂S) is an important biomolecule and significant efforts have focused on developing chemical tools to aid different biological investigations. Of such tools, there are relatively few chemiluminescent or bioluminescent methods for H₂S detection. Here we report two dioxetane-based chemiluminescent probes for H₂S detection. With these probes, we directly compare the probe response to H₂S-mediated azide reduction and nucleophilic displacement of 2,4-dinitrophenyl motifs and demonstrate that the S_NAr cleavage of the DNP group results in a larger response and greater stability in water.

Graphical Abstract

1. Introduction

Hydrogen sulfide (H₂S) is the third recognized gasotransmitter and an important endogenously-produced biomolecule,^1–2 with roles in cardioprotection,³ neuroprotection,⁴ wound healing,^5–6 and mitigating oxidative stress and associated damage.^7–8 For example, H₂S-producing enzymes are overexpressed in diabetic rats models and subsequent inhibition of enzymatic H₂S production reduces hyperglycemia.⁹ Conversely, in rat models of Parkinson’s disease (PD), brain H₂S levels are significantly lower than in healthy control animals, and treatment with exogenous NaSH reverses the progression of PD symptoms.^10–12 Motivated by the growing and diverse roles of H₂S in different biological systems, the last decade has witnessed significant development of chemical tools for detection and delivery of H₂S and related reactive sulfur species. Of such species, particular attention has focused on the development and expansion of biocompatible methods for H₂S detection and quantification, with activity-based fluorescent probes providing useful platforms for visualizing H₂S accumulation in cell culture and more complex experiments.^13–17

In parallel to fluorescent probe development, chemiluminescent probes for H₂S have also been developed, which provide an alternative approach for biological imaging that often results in lower background signal and greater tissue penetration due to the lack of excitation requirements.^18–19 Because chemiluminescent and bioluminescent platforms utilize a chemical reaction to generate an excited state intermediate, which emits photons upon relaxation back to ground state, such systems typically result in reduced autofluorescence, photobleaching and phototoxicity, and background interference.²⁰ These favorable properties have led to the wide adoption of such probes as important tools for bioimaging and biochemical studies.^21–24 Demonstrating the diversity of this approach, prior probes have been developed for a wide array of small biomolecules, including hydrogen peroxide,²⁵ singlet oxygen,²⁶ formaldehyde,²⁷ nitroxyl,²⁸ as well as other analytes. For H₂S detection, there are relatively few chemiluminescent or bioluminescent methods when compared to the number of available fluorescent probes. The first report of chemiluminescent H₂S detection was described by our group in 2013, in which H₂S-mediated azide reduction was used to convert azidoluminol to luminol.^29–30 Bioluminescent H₂S probes based on caged-luciferin were first reported by Lu and Li in 2015 and also leveraged H₂S-mediated azide reduction.^31–32 In 2014, the first chemiluminescent H₂S probe based on Schaap’s dioxetane³³ was reported by Sozmen and coworkers, which utilized the H₂S-responsive 2,4-dinitrophenol (DNP) based S_NAr trigger.³⁴ Schaap’s dioxetane employs spiro-adamantyl dioxetanes, and the rigidity of this structure imparts enhanced thermal stability to the dioxetane.³⁵ This probe, however, required harsh basic conditions (pH 12) to generate a chemiluminescence response, which limits the potential utility for biological applications.³⁴ The Lippert group reported a dioxetane-based chemiluminescent H₂S probe in 2015, which utilized H₂S-mediated azide reduction to initiate a 1,6-self-immolative elimination to reveal the luminogenic phenoxide.³⁶ These probes displayed moderate chemiluminescence under biologically relevant conditions, however the response could be enhanced and red-shifted by the addition of 20% of the surfactant-dye adduct Emerald II Enhancer (Figure 1a).³⁷ Motivated by the scarcity of chemiluminescent probes for H₂S detection and lack of evaluation of H₂S-mediated modes of activation, we report here the preparation and direct comparison of two bright chemiluminescent probes for H₂S and evaluate their efficacy in aqueous solution with the goal of streamlining access to these useful chemical tools for H₂S-related investigations.

Figure 1.

a) Selected examples of prior chemiluminescent or bioluminescent H₂S probes. b) The two dioxetane-based chemiluminescent probes for H₂S with different triggering mechanisms used for direct comparison in this work.

2. Results and Discussion

To compare the efficacy of different H₂S sensing strategies, we chose to use the bright chemiluminescent dioxetane scaffold recently reported by the Shabat lab.³⁸ This luminophore, which is functionalized with a slightly deactivating chloride and an extended electron-withdrawing group para to the dioxetane, is ~10³-fold brighter than previously prepared similar cores. The additional electron-withdrawing groups and extended conjugation on the aromatic luminophore core red-shift the emission wavelength of this dioxetane probe by 55 nm when compared to the unfunctionalized phenol core.^{37, 39–40} Although H₂S-mediated azide reduction is the most common approach to H₂S detection, the rate of this reaction is significantly slower than S_NAr-based methods.⁴¹ Based on these considerations, our goal was to compare the rate and efficacy of the H₂S-responsive azide system with the 2,4-dinitrophenol (DNP) electrophilic system, which undergo H₂S-mediated reduction to the parent amine and nucleophilic aromatic substitution, respectively (Figure 2a). An additional key distinction between these systems is that the S_NAr electrophiles only require one equivalent of H₂S to initiate luminescence,⁴² whereas azide reduction requires two equivalents of H₂S. Moreover, azide reduction results in H₂S oxidation to generate polysulfides, which are biologically-relevant reactive sulfur species.⁴³ The mechanisms of activation of both the DNP cleavage and azide reduction by H₂S have been reported previously, with examples provided in the references above. In addition, the 4-azido benzyl carbonate used to trigger self-immolation results in release of a para imino-quinone methide, which could potentially react with nucleophiles, although is likely scavenged by water to generate the corresponding benzyl alcohol.

Figure 2.

a) Mechanism and byproducts of H₂S-mediated turn-on of chemiluminescent probes CL-N3 and CL-DNP. b) Synthesis of the two probes CL-N3 and CL-DNP from a known phenol intermediate, EE-OH.

To prepare these chemiluminescent probes, we reacted H₂S-responsive motifs with the previously-reported enol-ether phenol core EE-OH,³⁸ followed by singlet oxygen oxidation (Figure 2b). We initially used the Acid Red/Rose Bengal photosensitizer system for the generation of singlet oxidation, which has been used previously in the presence of azides,³⁶ but in our hands were unable to access clean alkene oxidation without azide photoreduction. We found, however, that using tetraphenyl porphyrin (TPP)⁴⁴ as the photosensitizer allowed for the reaction to be run in CH₂Cl₂, which slows the decay of the generated singlet oxygen, reduces the reaction time, and decreases the amount of azide photodegredation.⁴⁵ These same photooxidation conditions were also used to prepare CL-DNP. We note that the final DNP product is significantly less photosensitive than CL-N3, which enabled greater scalability and a 76% yield over two steps compared to the 16% yield for the azide system.

With the two target probes in hand, we next measured the chemiluminescent response from the reaction of CL-N3 and CL-DNP with H₂S. To simplify our initial investigations, we first measured the response in organic solution to eliminate potential complications with solubility, aggregation, or quenching in water. Our expectation was that the rate difference between CL-N3 and CL-DNP would be smaller in organic solvents due to prior reports demonstrating the enhanced rate of H₂S-mediated azide reduction in organic solution, when compared to aqueous systems.⁴³ Upon treatment of a 25 μM solution each probe with 100 equivalents of NaSH in THF we observed a significant increase in luminescence. Over the course of 30 minutes, the CL-N3 probe resulted in a luminescence turn on of over 2600-fold, which was much larger than the approximately 1450-fold luminescence turn on for CL-DNP (Figure 3a, and SI). We attribute the difference in observed emission to the absorbance of 2,4-dinitrothiophenol (λ_max = 450 nm)⁴⁶ generated from CL-DNP, which overlaps with the probe emission spectrum.

Figure 3.

a) Luminescent response of 25 μM solutions of CL-N3 and CL-DNP in THF to 100 equiv. of NaSH over 30 minutes at 37 °C. b) Luminescent response of 25 μM solutions of CL-N3 and CL-DNP in degassed 10 mM PBS 7.4 buffer with 5% DMSO to 100 equiv. of NaSH over 30 minutes at 37 °C. The inset shows the first 1.5 minutes of the experiment. The probe was added after about 20 s, and the NaSH was added after about 40 s, as denoted in the plot.

Having demonstrated that both the azide- and DNP-based probes function in THF, we next investigated the chemiluminescent responses in buffered aqueous conditions. Upon addition of CL-N3 to degassed PBS buffer, we were surprised to observe a moderate, but rapid, increase in luminescence prior to the addition of NaSH. This reproducible result suggests that the triggering group on CL-N3 may be unstable under aqueous conditions, possibly due to hydrolysis at the electrophilic carbonate. This observation limits the suitability of CL-N3 probe for use in more complex biological imaging studies (Figure 3b, inset). Due to the background turn on of CL-N3 observed in buffer, the normalized turn-on response of CL-N3 upon addition of 100 equiv. of NaSH over that of the background signal was only a 7-fold enhancement. By contrast, the CL-DNP probe was found to be stable in PBS pH 7.4 buffer, and resulted in a 100-fold increase in normalized luminescence after addition of 100 equiv. of NaSH (Figure 3b). Importantly, this large change in luminescence does not require the addition of enhancers, unlike prior diooxetane-based H₂S probes.

Based on the greater stability of CL-DNP and significant luminescent response, we next carried out selectivity studies to confirm that the primary response is observed for H₂S. In these experiments, we treated CL-DNP with 100 equiv. of different analytes at 37 °C and measured the integrated luminescence response over 30 minutes. We chose to limit our selectivity investigations to these biological nucleophile analytes due to the significant body of prior work in the literature that has demonstrated that primary cross-reactivity can occur with other biological nucleophiles, whereas little or no reactivity of the DNP group is observed with other potential biological reactants, including metal ions, reactive oxygen species, and reducing agents.^47–48 As expected, we observed a high selectivity for H₂S over GSH, Cys, and Lys (Figure 4a). For comparison, we have also included the integrated response in THF under identical conditions, which shows significant increase in overall luminescent signal. Notably, this probe does not require an enzyme activator nor exogenous luminescence enhancer to function in aqueous environments. Moreover, the luminescent turn on of CL-DNP is significantly larger than previously-reported chemiluminescent or bioluminescent H₂S probes, as shown in Table 1.

Figure 4.

a) Selectivity studies for CL-DNP. The THF data were acquired in air-free unstabilized THF and all other data were acquired in 10 mM degassed PBS 7.4 buffer with 5% DMSO. Each experiment was performed at 37 °C with 100 equiv. of NaSH and integrated over 30 minutes. Each bar represents the average normalized turn-on response of three independent trials relative to 25 μM CL-DNP in PBS 7.4 buffer with 5% DMSO with no analyte added. b) Normalized chemiluminescent response of 25 μM CL-DNP with increasing [NaSH]. NaSH experiments were performed in triplicate at 37 °C in 10 mM degassed PBS 7.4 buffer with 5% DMSO and integrated over 30 minutes.

Table 1.

Emission wavelengths and turn-on responses of CL-DNP and the previously reported chemiluminescent and bioluminescent H₂S probes.

Probe	λem (nm)	Turn-on (fold)	Notes	Ref
CLSS-1	425	45	50 μM probe, 33 equiv. H2S, 5 min. integration, 37 °C, pH 7.4 Requires H2O2 and horseradish peroxidase	Pluth 201329
Probe 6	475	0	100 μM probe, 10 equiv. H2S. 25 °C, pH 12. No significant difference in signal with or without H2S.	Sozmen 201434
CHS-3	545	7	40 μM probe, 5 equiv. H2S, 20 min. integration, 25 °C, pH 7.4 with 20% Emerald II Enhancer.	Lippert 201536
Probe 1	588	8	40 μM probe, 250 equiv. H2S, 60 min. integration, 37 °C, pH 7.4. Requires ATP and luciferase.	Li 201631
CL-DNP	525	100	25 μM probe, 100 equiv. H2S, 30 min. integration, 37 °C, pH 7.4 with 5% DMSO.	this work
		1600	25 μM probe, 100 equiv. H2S, 30 min. integration, 37 °C, THF.

In addition, we also measured the integrated chemiluminescent response to varying concentrations of H₂S in buffer over the range of 0 – 167 equiv. of NaSH over 30 minutes. At lower concentrations of NaSH, it is likely that the triggering S_NAr reaction may not by fully complete, resulting in a lower integrated emission than expected. Nonetheless, a significant increase in luminescence is observed even when only 10 equiv. of NaSH are added to CL-DNP. Although the curvature of the chemiluminescence response precludes a definitive limit of detection, the chemiluminescence signal is sufficient to readily detect sub-micromolar levels of H₂S.

3. Conclusions

In summary, we prepared two bright chemiluminescent probes for H₂S detection that function without the need of brightness enhancers, surfactants, or enzyme activation. We demonstrated that in these systems, the more commonly used azide-trigger displayed moderate auto-activation in water, whereas the DNP-triggered probe is more stable. Moreover, the DNP-based system is significantly more synthetically accessible and shows good selectivity for H₂S over common biological nucleophiles. In a broader context, aryl azides can be reduced to the parent amine by the by cytochrome P450 enzymes,⁴⁹ whereas activation DNP groups by P450s has not been demonstrated to the best of our knowledge. These factors suggest that alternative approaches to the commonly used azide-reduction strategy may be more fruitful in biological settings.

4. Experimental Section

Materials and Methods

Reagents were purchased from Sigma-Aldrich, Tokyo Chemical Industry (TCI), Fisher Scientific, or VWR and used directly as received. Silica gel (SiliaFlash F60, Silicycle, 230−400 mesh) was used for column chromatography. Deuterated solvents were purchased from Cambridge Isotope Laboratories (Tewksbury, Massachusetts, USA). ¹H and ¹³C{¹H} NMR spectra were recorded on Bruker 500 MHz or Bruker 600 MHz NMR instruments at the indicated frequencies. Chemical shifts are reported in parts per million (δ) and are referenced to residual protic solvent resonances. The following abbreviations are used in describing NMR couplings: (s) singlet, (d) doublet, (b) broad, and (m) multiplet. IR spectra were measured on a Thermo Scientific Nicolet 6700 RT-IR using an ATR attachment. Mass spectrometric measurements were performed by the University of Illinois, Urbana Champaign MS facility, or on a Xevo Waters ESI LC/MS instrument. Phosphate buffered saline (PBS) tablets (1X, CalBioChem) were used to prepare buffered solutions (140 mM NaCl, 3.0 mM KCl, 10 mM phosphate, pH 7.4) in deionized water. Buffer solutions were sparged with nitrogen to remove dissolved oxygen and stored in an Innovative Atmosphere nitrogen-filled glovebox. All stock solutions were freshly prepared using degassed solvents immediately before use. Anhydrous sodium hydrogen sulfide (NaSH) was purchased from Strem Chemicals and handled under nitrogen. L-Cysteine and L-Lysine were purchased from TCI. Reduced glutathione was purchased from Aldrich. Stock solutions of the analytes were prepared in 10 mM PBS 7.4 buffer or DMSO under nitrogen immediately prior to use and were introduced into buffered solutions with an air-tight Hamilton syringe. Note: CL-N3 and CL-DNP are not air-sensitive, but protection of reaction solution from O₂ was to prevent H₂S oxidation. To ensure accurate measurements and to prevent decomposition of potentially reactive species, all experiments were performed under an inert atmosphere unless otherwise indicated. Chemiluminescence measurements were measured using a Quanta Master 40 spectrofluorometer (Photon Technology International) equipped with a Quantum Northwest TLC-50 temperature controller at 37.0 ± 0.05 °C. All chemiluminescent measurements were made under an inert atmosphere in septum-sealed cuvettes obtained from Starna Scientific and were repeated at in triplicate.

Chemiluminescence Studies

For all chemiluminescence experiments, excitation slits were closed, and the excitation wavelength set to 800 nm. An excitation wavelength input was required for the instrument to run, however this should in no way interfere with the measurement of chemiluminescent output. Emission slits were set to 4.0 mm, and the wavelength measured at was 525 nm. Scans were taken every second for at least 30 minutes. All experiments performed in triplicate.

General Procedure in THF:

In a septum sealed cuvette, 3.0 mL of degassed air-free THF was incubated for 5 minutes at 37 °C, after which data collection was started. To the cuvette 15 μL of a 5 mM THF stock of probe was added using an airtight Hamilton syringe to make a 25 μM solution, then approximately 20 seconds later analyte was added with an airtight Hamilton syringe. Data was collected for at least 30 minutes after analyte was added.

General Procedure in PBS:

In a septum sealed cuvette, 3.0 mL of a solution of degassed 10 mM PBS 7.4 buffer with 5% DMSO was incubated for 5 minutes at 37 °C, after which data collection was started. To the cuvette 15 μL of a 5 mM THF stock of probe was added using an airtight Hamilton syringe to make a 25 μM solution, then approximately 20 seconds later analyte was added with an airtight Hamilton syringe. Data was collected for at least 30 minutes after analyte was added.

Preparation of Stock Solutions:

In small HPLC vials, 500 mM NaSH, L-Cys, and L-Lys stocks were prepared in degassed Millipore water, and 15 μL were added for 100 equiv. experiments. Due to poor solubility, 250 mM GSH stocks were prepared in degassed Millipore water, and 30 μL GSH stock was added to the cuvette to reach 100 equiv. For the variable concentration NaSH experiments, the aliquots added were 0, 1, 2.5, 5, 10, 15, and 25 μL, to reach 0, 7, 17, 33, 67, 100, and 167 equivalents, respectively.

Normalized Turn-On Response

Data for four blank baseline response trials were collected with 25 μM of either CL-DNP or CL-N3 in either THF or PBS 7.4 with 5% DMSO at 37 °C for 30 minutes, with no analyte added. Identical fluorimeter parameters were used for each experiment: the excitation slits were closed, and the excitation wavelength set to 800 nm. Emission slits were set to 4.0 mm, and the wavelength measured at was 525 nm. Scans were taken every second for at least 30 minutes. Background luminescence measurements were recorded and subtracted from all experiments when calculating the normalized luminescence response. The tabulated baseline responses are listed in Table S1.

Syntheses

CL-DNP.

Core phenol EE-OH was prepared according to the literature procedures. Spectral data agrees with those reported in the literature.^{38, 50} EE-OH (100 mg, 0.281 mmol, 1.0 equiv.), 2,4-dinitrobromobenzene (76 mg, 0.31 mmol, 1.1 equiv.), and K₂CO₃ (78 mg, 0.56 mmol, 2.0 equiv.) were dissolved in dry DMF (3 mL) and stirred overnight under N₂. The reaction mixture was quenched with brine and extracted with EtOAc (3 × 10 mL). The combined organic layers were washed with 5% aqueous LiCl (4 × 10 mL), dried over anhydrous MgSO₄, concentrated under vacuum, and purified by silica column chromatography to yield 113 mg of a white solid. The crude DNP-enol ether product (110 mg, 0.211 mmol, 1.0 equiv.) was dissolved in CH₂Cl₂ (40 mL) and transferred to a large test tube. TPP (10 mg, 0.016 mmol, 0.05 equiv.) was added and mixed. The reaction tube was clamped with the bottom 2 inches in an ice water bath and a steady stream of O₂ was bubbled through a 9-inch Pasteur pipette into the solution while the reaction mixture was illuminated with a flood lamp. The reaction was run for three hours and with more CH₂Cl₂ was added every 20 minutes to maintain an approximately constant volume. The reaction was monitored by removing aliquots and measuring the ¹H NMR spectrum. After completion of the reaction, the crude reaction mixture was concentrated and purified by preparatory TLC (1:1 Hex:EtOAc, 1000 μm thick silica) to yield CL-DNP as a white solid (110 mg, 76% yield over two steps). ¹H NMR (600 MHz, CDCl₃) δ (ppm): 8.93 (d, J=2.73 Hz, 1H), 8.28 (bs, 1H), 8.23 (d, J=8.49 Hz, 1H), 7.73 (d, J=8.49 Hz, 1H), 7.47 (d, J=16.72 Hz, 1H), 6.48 (bs, 1H), 6.19 (d, J=16.72 Hz, 1H), 3.19 (s, 3H), 2.99 (s, 1H), 2.09 (d, J=13.31 Hz, 1H), 1.93 (bs, 1H), 1.83 (d, J=13.31 Hz, 2H), 1.76−1.56 (m, 7H), 1.48 (dd, J=13.05, 2.97 Hz, 1H), 1.40 (d, J=12.98 Hz, 1H), 1.17 (d, J=13.05 Hz, 1H). ¹³C{¹H} NMR (151 MHz, CDCl₃) δ (ppm): 153.97, 147.20, 142.23, 141.84, 138.63, 137.68, 131.99, 130.20, 129.02, 127.08, 125.76, 122.83, 116.81, 115.43, 111.15, 102.94, 96.32, 49.86, 36.30, 33.85, 33.67, 32.65, 32.02, 31.55, 31.49, 26.00, 25.67. IR (cm⁻¹) 2916.7, 2859.4, 2222.0, 1736.4, 1610.0, 1537.6, 1392.5, 1346.2, 1266.3, 1227.4, 1217.3, 1068.9. HRMS m/z [M + Na]⁺ calcd. for [C₂₇H₂₄N₃O₈ClNa]⁺ 576.1150, found 576.1151.

CL-DNP.

EE-OH (100 mg, 0.281 mmol, 1.0 equiv.), 2,4-dinitrobromobenzene (76 mg, 0.31 mmol, 1.1 equiv.), and K₂CO₃ (78 mg, 0.56 mmol, 2.0 equiv.) were dissolved in dry DMF (3 mL) and stirred overnight under N₂. The reaction mixture was quenched with brine and extracted with EtOAc (3 × 10 mL). The combined organic layers were washed with 5% aqueous LiCl (4 × 10 mL), dried over anhydrous MgSO₄, concentrated under vacuum, and purified by silica column chromatography to yield 113 mg of a white solid. The crude DNP-enol ether product (110 mg, 0.211 mmol, 1.0 equiv.) was dissolved in CH₂Cl₂ (40 mL) and transferred to a large test tube. TPP (10 mg, 0.016 mmol, 0.05 equiv.) was added and mixed. The reaction tube was clamped with the bottom 2 inches in an ice water bath and a steady stream of O₂ was bubbled through a 9-inch Pasteur pipette into the solution while the reaction mixture was illuminated with a flood lamp. The reaction was run for three hours and with more CH₂Cl₂ was added every 20 minutes to maintain an approximately constant volume. The reaction was monitored by removing aliquots and measuring the ¹H NMR spectrum. After completion of the reaction, the crude reaction mixture was concentrated and purified by preparatory TLC (1:1 Hex:EtOAc, 1000 μm thick silica) to yield CL-DNP as a white solid (110 mg, 76% yield over two steps). ¹H NMR (600 MHz, CDCl₃) δ (ppm): 8.93 (d, J=2.73 Hz, 1H), 8.28 (bs, 1H), 8.23 (d, J=8.49 Hz, 1H), 7.73 (d, J=8.49 Hz, 1H), 7.47 (d, J=16.72 Hz, 1H), 6.48 (bs, 1H), 6.19 (d, J=16.72 Hz, 1H), 3.19 (s, 3H), 2.99 (s, 1H), 2.09 (d, J=13.31 Hz, 1H), 1.93 (bs, 1H), 1.83 (d, J=13.31 Hz, 2H), 1.76−1.56 (m, 7H), 1.48 (dd, J=13.05, 2.97 Hz, 1H), 1.40 (d, J=12.98 Hz, 1H), 1.17 (d, J=13.05 Hz, 1H). ¹³C{¹H} NMR (151 MHz, CDCl₃) δ (ppm): 153.97, 147.20, 142.23, 141.84, 138.63, 137.68, 131.99, 130.20, 129.02, 127.08, 125.76, 122.83, 116.81, 115.43, 111.15, 102.94, 96.32, 49.86, 36.30, 33.85, 33.67, 32.65, 32.02, 31.55, 31.49, 26.00, 25.67. IR (cm⁻¹) 2916.7, 2859.4, 2222.0, 1736.4, 1610.0, 1537.6, 1392.5, 1346.2, 1266.3, 1227.4, 1217.3, 1068.9. HRMS m/z [M + Na]⁺ calcd. for [C₂₇H₂₄N₃O₈ClNa]⁺ 576.1150, CL-N3. EE-OH (515 mg, 1.45 mmol, 1.2 equiv.) and the azide carbonate coupling partner (350 mg, 1.21 mmol, 1.0 equiv.)³⁶ were dissolved in 20 mL of 4:1 THF:CH₂Cl₂ under N₂. DMAP (221 mg, 1.81 mmol, 1.5 equiv.) and Et₃N (0.67 mL, 4.83 mmol, 4.0 equiv.) were added, and the resultant reaction mixture was stirred overnight protected from light. The reaction mixture was quenched with brine and extracted with EtOAc (3 × 20 mL). The combined organic layers were dried over anhydrous MgSO₄, concentrated under vacuum, and purified by preparatory TLC (1:1 Hex:EtOAc, 1000 μm silica thickness). The resulting N3-enol ether (120 mg, 0.226 mmol, 1.0 equiv.) was then dissolved in CH₂Cl₂ (40 mL) and transferred to a large test tube. TPP (7 mg, 0.01 mmol, 0.05 equiv.) was added and mixed. The reaction tube was clamped with the bottom 2 inches in an ice water bath, and a steady stream of O₂ was bubbled through a 9-inch Pasteur pipette into the solution while the reaction mixture was illuminated with a flood lamp. The reaction was run for three hours and with more CH₂Cl₂ was added every 20 minutes to maintain an approximately constant volume. The reaction was monitored by removing aliquots and measuring the ¹H NMR spectrum. The crude reaction mixture was concentrated and purified by preparatory TLC (2:1 Hex:EtOAc, 1000 μm silica thickness) to yield CL-N3 as a white solid (87 mg 16% yield as a white solid) ¹H NMR (500 MHz, CDCl₃) δ (ppm): 8.08 (d, J=8.44 Hz, 1H), 7.56 (d, J=8.44 Hz, 1H), 7.42 (m, 3H), 7.07 (d, J=8.48 Hz, 2H), 6.01 (d, J=16.76 Hz, 1H), 5.29 (s, 2H), 3.19 (s, 3H), 3.01 (bs, 1H), 2.21 (d, J=13.28 Hz, 1H), 1.92 (bs, 1H), 1.83 (m, 2H), 1.73 (m, 4H), 1.65 (m, 1H), 1.59 (m, 2H), 1.46 (dd, J=12.92, 3.00 Hz, 1H), 1.33 (dd, J=13.48, 3.00 Hz, 1H). ¹³C{¹H} NMR (126 MHz, CDCl₃) δ (ppm): 151.68, 146.16, 142.70, 141.10, 136.39, 131.06, 130.69, 130.34, 129.53, 127.42, 124.67, 119.40, 117.14, 111.35, 101.63, 96.39, 70.93, 49.81, 36.50, 33.90, 33.61, 32.41, 32.14, 31.56, 31.52, 26.11, 25.76. IR (cm⁻¹) 2910.9, 2858,4, 2221.32, 2110.0, 1767.4, 1508.4, 1453.1, 1397.9, 1376.5, 1210.7, 1180.2, 1128.8, 1104.6, 1068.6. HRMS m/z [M + Na]⁺ calcd. for [C₂₉H₂₇N₄O₆ClNa]⁺ 585.1517, found 585.1529. found 576.1151.