A chemiluminescent probe for cellular peroxynitrite using a self-immolative oxidative decarbonylation reaction

Peroxynitrite is a damaging agent of oxidative stress that has been difficult to monitor in living cells. Here, an isatin-based chemiluminescent probe for peroxynitrite is reported.


Introduction
Peroxynitrite (ONOO À ) is a highly reactive oxygen species that can be formed biologically from a diffusion-controlled reaction between superoxide (O 2 c À ) and nitric oxide (NOc). 1 In biological systems, ONOO À has long been known as a deleterious species due to its oxidative damage to lipids, proteins, and nucleic acids. 2 Abnormal regulation of ONOO À in living systems is associated with diseases such as ischemia-reperfusion, 3 diabetes, 4 cardiac dysfunction, 5 inammatory conditions, 6 autoimmune, and neurodegenerative diseases. 7,8 However, ONOO À has been proposed to act as a physiological mediator in certain contexts, for example in immune system function 9 and ischemic preconditioning. 10 Macrophages produce nitric oxide and superoxide to form ONOO À to attack invading pathogens. 11 Macrophage cells recognize lipopolysaccharide (LPS), a component of the bacterial cell wall, which induces high expression of inducible nitric oxide synthase (iNOS). 12 The effect of LPS is amplied by release of IFN-b, providing a mechanism for robust local production of NOc and ONOO À . Although important biological functions of ONOO À are still emerging, there remains a lack of efficient and selective methods to monitor its production in real time to study its effects in living systems. Therefore, the development of precise approaches for detecting ONOO À is crucial to provide a more detailed understanding of its complex biological effects.
Much of the difficulty of studying the biological roles of ONOO À stems from its complex chemistry. ONOO À coexists with its protonated form ONOOH (pK a 6.8) under physiological pH, and can directly undergo two-electron reactions with thiols, 13 CO 2 , 14 and carbonyls. 15 The H + and CO 2 adducts, ONOOH and ONOOCO 2 À , rapidly decompose via homolytic cleavage reactions to form secondary radicals such as hydroxyl radical (HOc), carbonate radical (CO 3 c À ), and nitrogen dioxide (NO 2 c), which mediate potent one-electron chemistry, including tyrosine nitration. 16 ONOO À can also form from reaction between O 2 and NO À produced by photolysis of Angeli's salt under alkaline conditions. 17 Under physiological conditions, Angeli's salt has been widely used as a nitroxyl (HNO) donor for biological studies. Reaction of HNO with O 2 also proceeds at neutral pH to form a potent oxidative species and represents an important decomposition pathway of HNO. 18, 19 While there are similarities and differences between ONOO À and the autooxidation product of HNO formed at neutral pH, the details of this chemistry and the exact species eliciting the oxidative effect of Angeli's salt is still under debate.
The classical approach for studying biological ONOO À relies on immunostaining of 3-nitrotyrosine residues, 20 which lacks temporal precision and is an indirect method tracking the "footprint" of ONOO À . A large number of reaction-based uorescent probes have been developed in an attempt to solve this problem. These small molecule tools can react with chemical species as they are produced, enabling detection within living intact systems. Several families of reaction-based uorescent probes for ONOO À detection have been reported and applied in biological studies, such as triuorocarbonyl-based, 21 N-dearylation-based, 22 and boronic ester-based 23 probes. Organoselenium-based, 24 and organotellurium-based 25 probes are advantageous because the uorescence signal is reversible, but depend upon reaction with cellular reductants to reverse the signal. 26 This limited reversibility, as well as other intrinsic drawbacks of uorescent detection methods such as photobleaching and autouorescence highlights the need for new methods of ONOO À detection. Chemiluminescence provides an attractive solution for studying biological molecules and processes due to the low background, high signal-to-noise ratio, and reversible signal of this technique. 27 Nanoparticle-based methods have recently been reported for chemiluminescent ONOO À detection. 28,29 These probes rely on an interesting radical-pair mechanism to generate signal from the ONOO À decomposition products HOc and O 2 c À . Although these systems do display good sensitivity, complementary approaches that operate via direct reaction with ONOO À are desirable. Chemiluminescent probes based on sterically hindered 1,2-dioxetanes offer opportunities to incorporate specic reaction-based triggers and have already been applied for detection of analytes in cellular systems and for live animal imaging. [30][31][32][33][34] One fruitful design strategy uses Schaap's adamantylidene-dioxetane attached to an analyte-specic reactive handle. 35 A striking substituent effect for Schaap's 1,2-dioxetanes has been reported, where incorporation of an electron withdrawing group such as an acrylonitrile at the ortho position of the phenol results in a signicant increase in the overall chemiluminescent quantum yield of 1,2-dioxetanes. 36 As a result, chemiluminescent emission can be observed in aqueous conditions without the addition of any polymeric enhancer solutions. We herein describe the development and application of a selective chemiluminescent probe for ONOO À detection, PNCL, with both in vitro and live cell applications.

Results and discussion
Design and synthesis of PNCL As part of a program to develop chemical probes for imaging reactive sulphur, 29,37 oxygen, 30,38,39 and nitrogen species, our laboratory has previously reported that isatin reacts with ONOO À via an oxidative decarbonylation reaction to generate anthranilic acid. 40 19 F NMR probes, 5-uoroisatin and 6-uorisatin, have been developed for selective detection of ONOO À that utilize this newly discovered transformation. In order to expand this reaction-based detection strategy for luminogenic detection, we adopted a self-immolative 1,6-elimination strategy for chemiluminescent ONOO À detection that relies on tethering an isatin moiety to a 1,2-dioxetane chemiluminescent scaffold via an ether linkage (Scheme 1). The isatin serves as the reaction handle, and 1,6-elimination can be triggered aer reacting with ONOO À to form the anthranilic acid with the amino group positioned para to the benzylic ether. Luminescent emission is then initiated upon the decomposition of the 1,2-dioxetane through a chemically initiated electron exchange luminescence (CIEEL) mechanism. Synthesis of a peroxynitrite chemiluminescent probe (PNCL) proceeded from commercially available 5-iodoisatin 1 by protection of the ketone group as the dimethylacetal 2 (Scheme 2). The protected 5-iodoisatin was subjected to palladium catalyzed reductive formylation with Nformyl saccharin as the CO source to form the aldehyde 3, 41 followed by reduction of the aldehyde with NaBH 4 and deprotection to afford benzyl alcohol 4. The alcohol 4 was linked to the phenol 5 (ref. 36) through a Mitsunobu reaction to deliver the enol ether precursor 6. This precursor underwent [2 + 2] cycloaddition with photogenerated singlet oxygen using Rose bengal as a sensitizer to provide the nal 1,2-dioxetane PNCL.
Spectral response of the probes to ONOO À Aer obtaining PNCL, its photophysical properties and reactivity towards ONOO À were characterized. Luminescence emission spectra were collected using an F-7000 Hitachi spectrophotometer by treating 20 mM PNCL with 0-200 mM ONOO À (Fig. 1). An emission peak centered at 525 nm was observed in the presence of ONOO À with $28-fold increase in peak luminescence intensity compared to a blank control (Fig. 1A). The reaction between PNCL and ONOO À was analyzed by GC-MS, conrming the generation of 2-adamantanone and 2-chloro-4acrylonitrile-3-(methoxycarbonyl)phenol (Fig. S1 †). The luminescence emission slowly decayed over a course of 20 min (Fig. 1B), consistent with fast reaction kinetics of ONOO À with the isatin moiety of PNCL, 40 followed by a slower 1,6-elimination. Interestingly, ONOO À can be added to a solution of PNCL over several cycles with good recovery of chemiluminescent signal (Fig. S2 †). This shows that the reversible chemiluminescent emission has good reproducibility and can be used monitor ONOO À uxes over time. The dose dependence of the peak chemiluminescent emission at 525 nm measured 10 Scheme 1 Spiroadamantane 1,2-dioxetane-based probe for chemiluminescent ONOO À detection. seconds aer ONOO À addition showed excellent linearity, demonstrating that ONOO À can be quantied using PNCL (Fig. 1C). Calibration of the response of PNCL with lower concentrations of ONOO À revealed a detection limit of 6 nM (3s) bolus ONOO À (Fig. S3 †), comparing favorably to previous uorescent and chemiluminescent methods (Table S1 †). The reactivity of PNCL towards SIN-1 was also evaluated (Fig. 1D), showing that PNCL can monitor the slow release of ONOO À over time, as would be observed in biological systems.
The selectivity of PNCL for ONOO À was tested against a panel of biologically relevant reactive sulphur, oxygen, and nitrogen species and metal cations. The response of PNCL to 200 mM ONOO À was compared against other analytes in HEPES buffer (20 mM, pH 7.4), by adding 5 mM reduced glutathione (GSH), 1 mM L-cysteine, or 200 mM other reactive species. Most of the species tested displayed no signicant increase in luminescence intensity over the blank control within 20 min (Fig. 2). Given the destructive reactivity of HOc, we examined the interference of this species by treating PNCL with ONOO À in the presence of the HOc-generating Fenton system (Fig. S4 †), where only a minimal reduction in signal was observed. A small luminescence increase was observed when treating PNCL with Angeli's salt, commonly used as a nitroxyl (HNO) donor. This was consistent with previous observations using HPLC to track the reactivity of 5-uoroisatin with Angeli's salt. 37 In light of studies exploring the generation of ONOO À and other nitrogen oxide species from the autooxidation of NO À and HNO, 17-19 we performed a series of competition experiments with HCO 3 À and GSH to better understand the luminescence increase resulting from reaction of PNCL with Angeli's salt. ONOO À reacts rapidly with CO 2 , but slowly with GSH. HNO, on the other hand, reacts rapidly with GSH. 42 We treated PNCL with ONOO À (Fig. 3A and B) or Angeli's salt ( Fig. 3C and D) in the presence of various amount of HCO 3 À (1 mM, 2 mM, 3 mM, 4 mM, and 5 mM). HCO 3 À decreased the chemiluminescence emission intensity generated by both ONOO À and Angeli's salt, providing evidence that Angeli's salt can generate ONOO À . A similar experiment was performed by reacting PNCL (20 mM) with ONOO À (Fig. 4A and B) or Angeli's salt ( Fig. 4C and D) and various amounts of GSH (50 mM, 100 mM, 200 mM, and 1 mM). However, the GSH only quenched the luminescence intensity from Angeli's salt (Fig. 4B) and no signicant luminescence decrease was observed from ONOO À (Fig. 4A). Interestingly, substoichometric quantities of GSH would not only reduce the total amount of luminescence, but displayed a lag time before producing a chemiluminescent signal (Fig. 4C). Angeli's salt generates HNO with a half-life of 2.3 minutes at 37 C. The observed lag time of the chemiluminescent response is Scheme 2 Synthesis of PNCL.  consistent with GSH scavenging the newly formed HNO more rapidly than the autooxidation reaction of HNO. Once the GSH is consumed, HNO can be oxidized to ONOO À or another nitrogen oxide species capable of two-electron chemistry and reaction with PNCL proceeds to produce a luminescent signal. In order to conrm this selectivity and that HCO 3 À /CO 2 does not scavenge HNO, we prepared a uorescent nitroxyl (HNO) probe, herein called XF1 (Fig. 5A). HNO reacts with the diphenylphosphinobenzoyl group to form an aza-ylide that will nucleophilically attack the ester to trigger the release of the uorophore. 43,44 A time-dependent uorescence turn-on was observed when treating 10 mM XF1 with 200 mM Angeli's salt ( Fig. S5A and B †), and selectivity tests demonstrated that XF1 can distinguish HNO from other reactive sulfur, oxygen, and nitrogen species (Fig. S5C †). Inhibition experiments were conducted by treating XF1 with various amounts of HCO 3 À or GSH together with Angeli's salt. While no obvious inhibition of the uorescence response was observed with the addition of HCO 3 À , the uorescence was completely quenched with 200 mM GSH ( Fig. 5B and C). These results conrm that GSH scavenges HNO generated from Angeli's salt but HCO 3 À /CO 2 does not scavenge HNO. Taken together, the results from Fig. 3-5 are consistent with the luminescence increase from Angeli's salt being due to the generation of ONOO À . We do note that alternative oxidative species formed from HNO have been proposed. 18 Although we cannot completely rule this out, the results here are consistent with the formation of ONOO À from HNO and support the high specicity of PNCL for detecting ONOO À . As a nal mechanistic investigation, we examined the selectivity of ONOO À versus NO 2 c, HOc, and CO 3 c À (the decomposition products of ONOO À ) by treating isatin with ONOO À in the presence of HCO 3 À /CO 2 or TEMPOL, a free radical scavenger (Fig. S6 †). While HCO 3 À /CO 2 inhibits the reaction, TEMPOL has no effect. This suggests that isatin reacts directly with ONOO À   and not with free radical species like NO 2 c, HOc, or CO 3 c À that are formed via spontaneous ONOO À decomposition.

Evaluation of PNCL for ONOO À detection in living cells
Aer demonstrating that PNCL reacts with ONOO À with high sensitivity and selectivity, we then proceeded to investigate the ability of PNCL for ONOO À detection in living cells using a BioTek Cytation 5 plate reader. Toxicity was determined by the MTT assay, which indicated low toxicity aer 24 hour treatment with PNCL at concentrations less than 100 mM and more than 50% viability aer treating with high concentrations of 1 mM PNCL for 24 hours (Fig. S7 †). Cellular internalization was veri-ed by imaging the intrinsic uorescence emission of PNCL (Fig. S8 †). In order to evaluate the ability of PNCL to detect cellular delivery of ONOO À from donor compounds, RAW 264.7 macrophages were incubated with PNCL for 30 minutes, washed, and then treated with 0, 200 mM, 400 mM, 800 mM, 1 mM, or 2 mM SIN-1, a commonly used ONOO À donor that operates by simultaneously generating NOc and O 2 c À . A time course of the chemiluminescent emission intensity increased to reach a maximum at 40 minutes (Fig. 6A). Both the peak and integrated luminescence intensities increased with increasing concentrations of SIN-1 ( Fig. 6A and B), and a 5-fold luminescence increase was obtained with 2 mM SIN-1 compared with a blank control. We also performed the same experiment in A549 cells (Fig. S9 †). Similar results were observed and excellent linearity was achieved in living cells between 0 and 400 mM SIN-1 (Fig. S9C †). This increase in chemiluminescent emission from SIN-1 and PNCL was attenuated by the ONOO À scavenger Mn(III) TMPyP (Fig. S10 †), conrming that the response is due to ONOO À . Finally, we investigated the ability of PNCL to detect ONOO À generated by macrophages stimulated to mount an immune response. Lipopolysaccharide (LPS) binds to the toll-like receptors of macrophages, activating the transcription factor NF-kb ultimately leading to the expression of iNOS and generation of reactive oxygen species that can produce ONOO À . 12 Macrophages stimulated with LPS produced increased chemiluminescent signal from PNCL (Fig. 6C), showing a 36% increase in integrated luminescence intensity with 1000 ng mL À1 LPS. Treatment with the ONOO À scavenger Mn(III)TMPyP reduces this signal back to control levels and co-incubation with the selective iNOS inhibitor 1400 W also reduces the signal from LPS and PNCL (Fig. S11 †). These control experiments validate the ability of PNCL to detect cellular production of ONOO À . Bright eld images of the cells aer LPS stimulation show the formation of phagosomes (Fig. 6D), a phenotypic response to LPS exposure.

Conclusions
We have designed and synthesized a reaction-based chemiluminescent probe for ONOO À , PNCL. PNCL reacts with ONOO À through an oxidative decarbonylation reaction to initiate light emission that can be observed instantly with high sensitivity and selectivity against other reactive sulphur, oxygen, and nitrogen species. The observed reactivity towards Angeli's salt is attributed to the generation of ONOO À , which is supported by inhibitor experiments with HCO 3 À and GSH, providing experimental evidence of an autooxidative product of HNO. Several studies have purported that the autooxidation of HNO produces a product unique from ONOO À , as evidenced by data showing no effect of HCO 3 À on the reaction between Angeli's salt and reactive oxygen species probe dihydrorhodamine 123. 18 In our studies, however, we observe a drastic HCO 3 À dependent reduction in chemiluminescent signal from PNCL and Angeli's salt, consistent with ONOO À formation from HNO under physiological conditions. 19 PNCL was used to detect ONOO À generated from donor compounds in multiple cell lines, as well as cellular ONOO À generated by macrophages stimulated with LPS. Conrmation of the chemical specicity towards ONOO À over downstream free radical products, coupled with operational compatibility in living cells establishes PNCL a useful and easy-to-use tool for studying ONOO À in chemical and biological systems.

Conflicts of interest
A. R. L. discloses a nancial stake in BioLum Science, LLC, a company developing asthma monitoring technology. Error bars are S.D. from n ¼ 3-7 wells and 3 biological replicates (n ¼ 10 wells from 6 biological replicates for the control experiment). Statistical significance was assessed using a two-tailed student's t-test. ***p < 0.001, **p < 0.01.