Highly specific C–C bond cleavage induced FRET fluorescence for in vivo biological nitric oxide imaging† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c6sc04071c Click here for additional data file.

A novel FRET fluorescence “off–on” system based on the highly specific, sensitive and effective C–C bond cleavage of certain dihydropyridine derivatives was reported for real-time quantitative imaging of nitric oxide (NO).


Introduction
Nitric oxide (NO), a very active radical signaling molecule, has been found to exist and play a signicant role in living body pathophysiology. 1 It has been gradually realized that NO is involved in different physiological processes including vasodilation, wound healing, immune responses and nerve cell communication. 2 Therefore, NO detection in vitro or, more ideally, 3D imaging the distribution and concentration of NO in vivo would be of a great help in understanding the metabolism of NO.
Various methods, such as electrochemistry, 3 optical imaging (uorescence, chemiluminescence and bioluminescence imaging), 4 electron paramagnetic resonance (EPR), 5 and magnetic resonance (MR), have been developed to image NO. 6 Among these methods, optical probe-based techniques are the most available for NO imaging. With intrinsic molecular sensitivity, high resolution, repeatability, a high level of safety, and a relatively low instrumentation cost, optical probes have become important tools in the study of cells and small animals. Both uorescence and chemiluminescence imaging have been widely used for elucidating the function of NO in different biological systems. 7 Among the existing ratiometric probes, most of them are designed according to a photo-induced electron transfer (PET) mechanism, during which donor-p-acceptor structures are formed aer the reaction between NO and the probe. Numerous efforts have been made on o-diamino aromatic compounds which could be oxidized by NO into triazole derivatives in the presence of oxygen. 8 Although uorescein based probes have been effective in zebrash, 9 there are no results regarding mammalian models yet. The speculated reason is the deciency of oxygen at site for NO detection. Furthermore, o-diamino aromatics are easily disturbed by some oxidative species secreted in cells, such as H 2 O 2 and OONO À . 10 A series of copper complexes were tested in mouse models, which could obviously distinguish a normal liver from a liver with acute severe hepatic injury (ASHI). 11 Yet, the uorescence images were obtained from excised liver slices harvested from a living animal using an ex vivo method. There are other limitations of PET probes: 12 (1) the measured excitation and emission wavelengths are oen different from the expected values; (2) relatively broad uorescence spectra are oen detected for PET probes; (3) although remarkable advances have been made in computational chemistry, the quantum yield of uorophores cannot be accurately predicted.
On the other hand, FRET mechanism based uorescent probes have been used in the elds of chemistry and biology for decades. The commercially available FRET uorophores have been proven to have certain excitation and emission wavelengths and an exact quantum yield. With different uorophores, the available detection range could be extended to the near-infrared region, which makes FRET probes more promising for in vivo imaging than PET probes. A typical FRET quencher-uorophore pair is FITC and DABCYL. This pair can directly indicate if the bonds between the two moieties are broken. A probe based on coumarin decorated 1,4-dihydropyridine was synthesized and evaluated in vitro, which was designed based on stoichiometry. 13 The probe transforms into a conjugated uorophore aer NO stimulation, but the in vivo application was largely restricted by the short emission wavelength. Therefore, we designed and synthesized a FRET based probe to conquer the previous obstacles.
Herein, we present a uorophore-quencher FRET "off-on" system for detecting and imaging NO in vivo in a non-invasive and real-time way. The 1,4-dihydropyridine derivatives can react with NO specically. Furthermore, as the C4-position is substituted with a benzyl group, the C-C bond between 1,4dihydropyridine and the benzyl group can be cleaved through NO stimulation with a high yield (Scheme 1). 14 The reaction is very specic and sensitive.

Results and discussion
In our probes, FITC and DABCYL were linked via 1,4-dihydropyridine within a distance of 10 nm. Because of the FRET mechanism, DABCYL quenches the uorescence emitted by FITC in the absence of NO. The cleavage of the C-C bond leads to the release of DABCYL and the uorescence can be detected again (Fig. S4 †). To achieve an image in vivo, the emission wavelength is tuned to increase the signal-noise ratio through using different uorophores (Scheme 2).
To obtain the NO target probe, an asymmetric synthesis strategy was used in the Hanztzsch reaction. An equivalent amount of aldehyde, b-keto ester and methyl 3-aminobut-2enoate ester were combined into an asymmetrical 1,4-dihydropyridine (compound 4) which had only one azido moiety. 15 Compound 6 was designed as a linker (Scheme 2). The amino group on compound 6 is the binding site for the uorophore, because many commercial uorophores are designed for labeling the amino groups on peptides and proteins. The azido group would react with an yne group which was linked to a quencher. DABCYL-yne and FITC were selected to react with the linker to obtain the target probe. The reaction between DHPFQ (dihydropyridine-uorophore-quencher) and NO was evaluated in phosphate buffer using a uorospectrometer, and the reaction proceeded in a quantitative way. The C-C bond cleavage was conrmed from the increase in the intensity of the uorescence.
As expected, an enhancement of the uorescence intensity at certain excitation (490 nm) and emission (525 nm) wavelengths was detected in 2 min aer DHPFQ reacted with NO (Fig. 1a). The uorescence intensity increases with an increasing amount of NO (diluted NO saturated solution). 16 Most small molecule inammatory factors in the complicated internal environment are reactive oxygen/nitrogen species (ROS/ RNS). Therefore, specic recognition and reaction between the probes and NO is essential for in vivo applications. To evaluate the reaction selectivity, compound S (Scheme S2 †) as a DHPFQ analog was treated with different reactive species rstly (Fig. S1 †). About 98% of compound S was found to be decomposed when treated with NO, which was measured using HPLC. Furthermore, DHPFQ was treated with NO and many other reactive species (Fig. 1b). The decomposition was measured using a uorescence spectrometer. DHPFQ showed a signicant increase in uorescence intensity with NO. No signicant decomposition was found when DHPFQ and compound S were treated with other reactive species. The results reveal the exclusive response of 1,4-dihydropyridine derivatives to no ROS or RNS but NO. The UV spectra of DHPFQ before and aer NO treatment were also investigated,  which showed that activated and inactivated DHPFQ share similar proles (Fig. S3 †).
With excellent responsiveness to NO and exceptional selectivity in vitro conrmed, the ability of DHPFQ to visualize NO in live cells was further tested. Sodium nitroprussiate (SNP) as an exogenous NO donor was used to validate the high contrast. Compared to the control group (Fig. S7a †), the SNP treated group (Fig. S7b †) showed strong uorescence, which implies that DHPFQ could respond to exogenous NO in a physiological environment.
Since NO is an inammatory factor generated by macrophages when stimulated with an antigen, in order to test the feasibility of application in vitro, raw 264.7 cells were stimulated with lipopolysaccharide (LPS) to generate endogenous NO. 17 A signicant enhancement of the uorescence intensity was observed when the macrophages were incubated with DHPFQ and LPS (Fig. 2, middle). There was a high correspondence between the green uorescence and the Lytracker Red signal which suggests this probe is metabolized in lysosomes. In contrast, when the macrophages were pretreated with N G -monomethyl-L-arginine (L-NMA), an inhibitor of nitric oxide synthase (NOS), 18 only a weak uorescence signal above the background was detected (Fig. 2, bottom). This result demonstrated that the uorescence is indeed from the switch-on reaction between DHPFQ and NO. DHPFQ has the ability to detect endogenous NO in vitro, suppress interference from other inammatory factors, and has the potential to detect NO in vivo.
NO, as a signicant inammatory factor, is concentrated in inamed areas. 19 An inamed mouse model was then examined to test whether DHPFQ is effective in vivo. Freund's adjuvant was subcutaneously injected into the le rear paws of mice to trigger inammation. 20 Aer two weeks, another portion of Freund's adjuvant was subcutaneously injected. At 3 days aer the second injection, DHPFQ was intravenously injected (injected dose: 0.5 mg kg À1 , 10% ethanol, 90% physiological saline, 100 mL) and the mice were imaged aer 10 min to 60 min using an in vivo imaging system (Fig. 3). To reduce the background absorption, depilatory paste was used to partially remove hair from the mice. 470 nm and 600 nm were chosen as   the excitation and emission wavelengths to improve the signal to noise ratio. The uorescence intensity observed for the inamed region was about 8-times higher than that of the normal area in the rst 10 min aer injection.
Inammation region kinetics and normal region curves were obtained using semiquantitative analysis of the uorescence image, which was analyzed by selecting the region of interest (ROI) (Fig. 4a). Aer iv injection, DHPFQ was switched on immediately in the inamed region. The uorescence intensity was about 8-times higher in the inamed region over the normal region within the rst 10 min post-injection. Aer 60 min, about 2-fold uorescence intensity was observed in the inamed region compared to the normal region.
The immunohistology staining (Fig. 4b) also indicated that NO was generated within inammatory tissues where macrophages were crowded (which is indicated from F4/80 staining) and the NO concentration was signicantly higher than in the normal region. The concentration of DHPFQ decreased with time, which diminished the uorescence intensity of the inamed tissue. Because of the metabolism, switch-on DHPFQ can be transported to normal tissue through blood circulation, leading to a uorescence signal at normal organs as well. Some studies suggested that the concentration of NO in physiological environments ranges from 100 pM up to 5 nM, 21 which indicates that DHPFQ has a lower limit of detection at the nM level.
The comparison of existing imaging methods and the probe developed in this study is summarized in Table 1. DHPFQ possesses the advantages of optical imaging, and low specicity can be overcome with the application of the 1,4-dihydropyridine structure. In this study, FITC and DABCYL were used to demonstrate the feasibility of the uorescence imaging of NO in vivo. NIR uorophores and quenchers could substitute FITC and DABCYL to obtain a FRET probe with good penetration.

Conclusions
A novel C-C bond cleavage induced FRET uorescent probe DHPFQ based on 1,4-dihydro-pyridine was synthesized. In vitro assessments have shown a ratiometric and specic result in responding to NO over other RNS/ROS. The uorescence intensity increased linearly with the increasing amount of NO. Cell assay also supports the specicity. With improved speci-city, DHPFQ is the rst uorescent probe applied in mammals. Furthermore, semiquantitative results were obtained through ROI analysis. The penetration property of NIR probes is tunable simply through the substitution of the uorophores and quenchers, which will make the probe more practical for clinical applications.