Hemicyanine-Based Near-Infrared Fluorescence Off–On Probes for Imaging Intracellular and In Vivo Nitroreductase Activity

Nitroreductase (NTR) has the ability to activate nitro group-containing prodrugs and decompose explosives; thus, the evaluation of NTR activity is specifically important in pharmaceutical and environmental areas. Numerous studies have verified effective fluorescent methods to detect and image NTR activity; however, near-infrared (NIR) fluorescence probes for biological applications are lacking. Thus, in this study, we synthesized novel NIR probes (NIR-HCy-NO2 1–3) by introducing a nitro group to the hemicyanine skeleton to obtain fluorescence images of NTR activity. Additionally, this study was also designed to propose a different water solubility and investigate the catalytic efficiency of NTR. NIR-HCy-NO2 inherently exhibited a low fluorescence background due to the interference of intramolecular charge transfer (ICT) by the nitro group. The conversion from the nitro to amine group by NTR induced a change in the absorbance spectra and lead to the intense enhancement of the fluorescence spectra. When assessing the catalytic efficiency and the limit of detection (LOD), including NTR activity imaging, it was demonstrated that NIR-HCy-NO2 1 was superior to the other two probes. Moreover, we found that NIR-HCy-NO2 1 reacted with type I mitochondrial NTR in live cell imaging. Conclusively, NIR-HCy-NO2 demonstrated a great potential for application in various NTR-related fields, including NTR activity for cell imaging in vivo.


Introduction
Reductase is a kind of enzyme that chemically reduces a substrate. Nitroreductase (NTR) is a type of reductase and a type of flavoenzyme, which involves a nicotinamide adenine dinucleotide (phosphate) hydrate (NAD(P)H)-dependent reduction for nitro groupcontaining compounds, such as nitroaromatic and nitroheterocyclic molecules [1]. The determination of NTR activity is especially significant in vivo because most nitro groupcontaining compounds exhibit high cytotoxicity [2]. NTR has attracted great attention since it has started to be used as an activator of nitro group-containing prodrugs and a decomposer of explosives, such as trinitrotoluene (TNT) [3,4]. Therefore, the detection of 2 of 12 NTR is highly significant in pharmaceutical and environmental areas. NTR is expressed in some bacteria and eukaryotic species, and it has been particularly utilized in therapeutic technologies for tumor-targeted delivery, such as cancer chemotherapy, also known as gene-directed enzyme prodrug therapy (GDEPT) [5][6][7][8][9].
NTR is classified into two different types: oxygen-insensitive type I NTR and oxygensensitive type II NTR. Type I NTR is mainly used as a nitroaromatic prodrug activator in GDEPT, and type II NTR is used for the selective imaging of hypoxic tumors because of its overexpression and the little interference that oxygen has under hypoxic conditions [10][11][12][13][14][15][16][17][18][19]. Further, previous studies on NTR activity fluorescence imaging have mostly focused on type II NTR [7][8][9] because fluorescence imaging for the activity of type I NTR is relatively insufficient. Previous reports have suggested the possibility of the mitochondrial existence of type I NTR based on the bacterial origin of mitochondria in fluorescence imaging [20][21][22]. As a prodrug activator, type I NTR can potentially present activity for nitroaromatic prodrugs when employing fluorescence imaging.
Functional fluorescent probes have the ability to identify and distinguish species of interest (SOI) in complex systems, such as intracellular or in vivo systems [23]. To utilize the advantages of functional fluorescent probes, various SOI have been applied to sensors and bioimaging, including for the detection and imaging of NTR. The probes for NTR detection are enzymatically activated by NTR, reducing the nitro group to the amine group, which induces a conversion from the electron-withdrawing group (EWG) to the electrondonating group (EDG) [2,24,25]. The enzymatic reduction induced by the activity of NTR results in a remarkable enhancement in the fluorescence spectra due to the intramolecular charge transfer (ICT) of the amine group, known as a strong EDG, utilizing the change in the functional group induced by NTR. However, most of the probes still suffer from the limitations of a high fluorescence background and a slow response, making the majority of probes useful only for the detection and imaging of NTR in cytoplasm. Moreover, the fluorescence probes proposed in previous studies were mostly viable in the visible range in order to assess the detection and imaging of NTR activities; however, fluorescence in the visible region is unfavorable for in vivo imaging due to the transmission and absorbance of biomolecules [26]. Therefore, near-infrared (NIR) probes are much more suitable for in vivo imaging due to their relatively low absorbance and good transmission in the NIR region.
In this study, we describe the novel NIR fluorescence probe NIR-HCy-NO 2 for the intracellular and in vivo imaging of NTR activity through the highly selective enzymatic reaction of NTR ( Figure 1). We used a hemicyanine skeleton as a fluorogenic backbone, introducing a nitro group as a selective NTR-responsive moiety and fluorescence quencher. NIR-HCy-NO 2 derivatives (NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3) are also designed to increase water solubility, which was achieved by introducing a sulfonate (-SO 3 − ) and quaternary ammonium group to the indolium part of NIR-HCy-NO 2 [27,28]. NIR-HCy-NO 2 showed a low fluorescence because of the interference of ICT by the nitro group and the reduction induced by NTR, which induced an enhancement in the fluorescence spectra due to the effect of ICT on the amine group. According to previous studies, indolium cation as a mitochondria tracker enables the verification of mitochondrial NTR [29]. This approach, which utilizes indolium cation in NIR-HCy-NO 2 , can be also applied to the mitochondrial imaging of NTR activity. Previously reported NTR probes have shown that the 4-nitrobenzene group, as an enzyme response moiety, should be introduced to the fluorophore backbone, and an additional elimination reaction is essential for the enhancement of the fluorescence [30,31]. In our newly synthesized NIR-HCy-NO 2 derivatives, the nitro group, the enzymatically reactive part, was directly conjugated to the NIR fluorophore backbone in order to achieve a fast response. The hemicyanine skeleton-based NIR probe has been previously applied to live zebrafish larvae for in vivo imaging. However, larval zebrafish have limitations in the in vivo imaging model due to their nonmammalian status and relatively thin skin, which is highly penetrable by light. Thus, fast responsive NIRemitted NTR sensors should be used in NTR activity-related practical applications, such as in live cell and animal imaging. NIR probe has been previously applied to live zebrafish larvae for in vivo imaging. However, larval zebrafish have limitations in the in vivo imaging model due to their nonmammalian status and relatively thin skin, which is highly penetrable by light. Thus, fast responsive NIR-emitted NTR sensors should be used in NTR activity-related practical applications, such as in live cell and animal imaging. Herein, the three derivatives of NIR-HCy-NO2 are applied to intracellular and in vivo mouse imaging, respectively. It was hypothesized that NIR-HCy-NO2 has high sensitivity and selectivity for NTR in the presence of NADH and that it would provide intracellular and in vivo NTR activity imaging.

Design and Synthesis of NIR-HCy-NO2 1-3
To design and synthesize NIR fluorophore for NTR detection, the hemicyanine skeleton was chosen as a fluorescence unit because of its long wavelength (longer than λ = 650 nm) and ability to minimize autofluorescence and biological damage [32,33]. First, the bottom-up approach was reported to synthesize the NIR-emitted hemicyanine structure [34], and the structural flexibility was one of the attractive points for the use of the bottomup approach. Previously, we developed novel alkaline phosphatase (ALP)-targeted NIR fluorescent probes (NIR-Phos-1 and NIR-Phos-2) using the same synthesis approach, and the hydroxyl group, which introduced the hemicyanine skeleton, was used as an NIR fluorophore for the ALP activity bioimaging [35]. In this study, we designed NIR-HCy-NO2 derivatives, which are direct nitro group-modified hemicyanines. To achieve their synthesis, the nitro group, containing a tricyclic compound as a core intermediate, was combined with indolium derivatives which each have different functional groups (Scheme S1). NIR-HCy-NO2 2 and NIR-Hcy-NO2 3 had a better water solubility than NIR-Hcy-NO2 1 due to the introduction of sulfonate and quaternary ammonium to the indolium part in NIR-Hcy-NO2. As expected, the introduction of polar functional groups increased the water solubility of NIR-Hcy-NO2. NIR-Hcy-NO2 3 only reacted with NTR in PBS, and the NIR-Hcy-NO2 1 and NIR-Hcy-NO2 2 reactions were conducted in a cosolvent of PBS and ACN. However, NIR-HCy-NO2 1, the most nonpolar probe, responded with the highest enhancement of fluorescence by NTR in the developed probes ( Figure 2). Herein, the three derivatives of NIR-HCy-NO 2 are applied to intracellular and in vivo mouse imaging, respectively. It was hypothesized that NIR-HCy-NO 2 has high sensitivity and selectivity for NTR in the presence of NADH and that it would provide intracellular and in vivo NTR activity imaging.

Design and Synthesis of NIR-HCy-NO 2 1-3
To design and synthesize NIR fluorophore for NTR detection, the hemicyanine skeleton was chosen as a fluorescence unit because of its long wavelength (longer than λ = 650 nm) and ability to minimize autofluorescence and biological damage [32,33]. First, the bottomup approach was reported to synthesize the NIR-emitted hemicyanine structure [34], and the structural flexibility was one of the attractive points for the use of the bottom-up approach. Previously, we developed novel alkaline phosphatase (ALP)-targeted NIR fluorescent probes (NIR-Phos-1 and NIR-Phos-2) using the same synthesis approach, and the hydroxyl group, which introduced the hemicyanine skeleton, was used as an NIR fluorophore for the ALP activity bioimaging [35]. In this study, we designed NIR-HCy-NO 2 derivatives, which are direct nitro group-modified hemicyanines. To achieve their synthesis, the nitro group, containing a tricyclic compound as a core intermediate, was combined with indolium derivatives which each have different functional groups (Scheme S1). NIR-HCy-NO 2 2 and NIR-Hcy-NO 2 3 had a better water solubility than NIR-Hcy-NO 2 1 due to the introduction of sulfonate and quaternary ammonium to the indolium part in NIR-Hcy-NO 2 . As expected, the introduction of polar functional groups increased the water solubility of NIR-Hcy-NO 2 . NIR-Hcy-NO 2 3 only reacted with NTR in PBS, and the NIR-Hcy-NO 2 1 and NIR-Hcy-NO 2 2 reactions were conducted in a cosolvent of PBS and ACN. However, NIR-HCy-NO 2 1, the most nonpolar probe, responded with the highest enhancement of fluorescence by NTR in the developed probes ( Figure 2).

Optical Properties of NIR-HCy-NO 2 1-3
To confirm a change in the optical properties following the enzymatic reduction induced by NTR, we examined the differences in the absorbance spectra in the absence or presence of NTR. In the absorbance spectra, NIR-HCy-NO 2 1-3 absorbed at λ abs = 552, 594, and 600 nm in the absence of NTR, and the absorbance peak increased at λ abs = 604, 662, and 658 nm in the presence of NTR and NADH ( Figure S30) [36]. To explain the relationship between the change in the absorbance spectra and the enhancement in the fluorescence spectra, the reduction of NIR-HCy-NO 2 1-3 by NTR was carried out under the same conditions, and the fluorescence signals of all NIR-HCy-NO 2 1-3 were enhanced 15-, 7-, and 9-fold, respectively, compared to the absence of NTR ( Figure 2B-D). Additionally, following the reduction induced by NTR, the color of the NIR-HCy-NO 2 1 solution changed from violet to blue, NIR-HCy-NO 2 2 changed from navy blue to blue, and NIR-HCy-NO 2 3 changed from navy blue to emerald. These color changes in NIR-HCy-NO 2 1-3 were caused by the bathochromic effect (i.e., redshift), as confirmed in the absorbance spectra. Additionally, the red light-absorbed fluorophores were generally observed as a blue-green color in the solution; a similar trend was observed in NIR-HCy-NO 2 1-3. High-resolution mass spectrometry (HR-MS) was used to show that the bathochromic effect was caused by the change from the nitro group to the amine group; NIR-HCy-NO 2  NIR-HCy-NO 2 1-3 were reduced, which was in line with the expected process ( Figure 1) and confirmed in the abovementioned results ( Figures S31-S33). In previous studies, NIR-HCy-NHOH 1-3 were reduced via four-electron transfer, and NIR-HCy-NH 2 1-3 were reduced through two-electron transfer from hydroxylamine intermediates. Specifically, the formation of azoxy compounds was induced by the reaction between hydroxylamine and nitroso intermediates to produce more stable azoxy compounds than the intermediate forms. This can be explained indirectly through the azoxy formation of NIR-HCy-NO 2 1, in which nitroso intermediates were produced during the NTR reduction process. Specifically, the bathochromic effect and fluorescence enhancement occurred when the nitro group (a strong EWG) was reduced to an amine group (a strong EDG) by NTR. As a result, the fluorescence emissions of NIR-HCy-NO 2 1-3 themselves were very weak, which is consistent with their nonemissive characteristic; this was due to the quenching effect of the six-nitro substitution on the hemicyanine skeleton. However, reductive NIR-HCy-NO 2 1-3 showed fluorescent properties due to the removal of ICT interference by the nitro group ( Figure 2B-D) [37,38].

Optical Properties of NIR-HCy-NO2 1-3
To confirm a change in the optical properties following the enzymatic reduction induced by NTR, we examined the differences in the absorbance spectra in the absence or presence of NTR. In the absorbance spectra, NIR-HCy-NO2 1-3 absorbed at λabs = 552, 594, To confirm the exact excitation and emission wavelengths of reduced NIR-HCy-NO 2 1-3 by NTR, the emission spectra were recorded using the same excitation wavelength (λ ex = 672 nm) ( Figure S34). All three probes emitted at longer wavelengths, in order from longest to shortest, of 3, 2, and 1 in the NIR region. Additionally, NIR-HCy-NO 2 1-3 exhibited a Stokes shift of more than 20 nm, similar to the Stokes shift of other NIR dyes. Additionally, to cross-check the fluorescence-emitted form, fully reduced forms (HCy-NH 2 1-3) were obtained using the chemical reduction method (Figures S35-S37). The spectra of HCy-NH 2 1-3 were almost the same as the NTR reduction result ( Figures S34 and S38), and the conversion from a strong EWG to a strong EDG was quite important to emit fluorescence due to the ICT effect. Moreover, kinetically, the reduction reaction of NIR-HCy-NO 2 1-3 samples by NTR was dependent on time and was saturated in less than 20 min ( Figure S39). Specifically, the reductive reaction of NIR-HCy-NO 2 1 and NIR-HCy-NO 2 3 was saturated within 10 min ( Figure S39A,C).

Selectivity Study
After confirming the reaction condition and optical properties of NIR-HCy-NO 2 1-3, including the reaction time, excitation, and emission wavelength, we evaluated the selectivity of NIR-HCy-NO 2 1-3 for NTR over other various types of biological and chemical species. The fluorometric change in NIR-HCy-NO 2 1-3 by NTR and other analytes was measured. Metal cations (Na + , K + , Mg 2+ , Ca 2+ , and Hg 2+ ), halogen anions (Br − and I − ), amino acids (L-cysteine, DL-homocysteine, L-phenylalanine, and glycine), and proteins (BSA, ALP, GOx, thrombin, AchE, lysozyme, and trypsin) were used as the analytes. The fluorometric response of NIR-HCy-NO 2 1-3 by NTR was much higher than the other analytes ( Figure 3). Thus, NIR-HCy-NO 2 1-3 reacted selectively with NTR, and, specifically, NIR-HCy-NO 2 1 showed the highest fluorescence response compared to the others. when the nitro group (a strong EWG) was reduced to an amine group (a strong EDG) by NTR. As a result, the fluorescence emissions of NIR-HCy-NO2 1-3 themselves were very weak, which is consistent with their nonemissive characteristic; this was due to the quenching effect of the six-nitro substitution on the hemicyanine skeleton. However, reductive NIR-HCy-NO2 1-3 showed fluorescent properties due to the removal of ICT interference by the nitro group ( Figure 2B-D) [37,38].
To confirm the exact excitation and emission wavelengths of reduced NIR-HCy-NO2 1-3 by NTR, the emission spectra were recorded using the same excitation wavelength (λex = 672 nm) ( Figure S34). All three probes emitted at longer wavelengths, in order from longest to shortest, of 3, 2, and 1 in the NIR region. Additionally, NIR-HCy-NO2 1-3 exhibited a Stokes shift of more than 20 nm, similar to the Stokes shift of other NIR dyes. Additionally, to cross-check the fluorescence-emitted form, fully reduced forms (HCy-NH2 1-3) were obtained using the chemical reduction method (Figures S35-S37). The spectra of HCy-NH2 1-3 were almost the same as the NTR reduction result ( Figures S34  and S38), and the conversion from a strong EWG to a strong EDG was quite important to emit fluorescence due to the ICT effect. Moreover, kinetically, the reduction reaction of NIR-HCy-NO2 1-3 samples by NTR was dependent on time and was saturated in less than 20 min ( Figure S39). Specifically, the reductive reaction of NIR-HCy-NO2 1 and NIR-HCy-NO2 3 was saturated within 10 min ( Figure S39A,C).

Quantitative Analysis
The quantitative analysis of NIR-HCy-NO 2 1-3 was performed with various concentrations of NTR under the physiological condition (PBS (pH 7.4) in the presence of 50 µM of NADH at 37 • C). The fluorescence signals of NIR-HCy-NO 2 1-3 increased the NTR concentration dependently. The trend line R-square (R 2 ) values of NIR-HCy-NO 2 1-3 were 0.9888, 0.9969, and 0.9964, respectively, and all three probes showed good linearity over the concentration range of 0.125-5 µg/mL of NTR ( Figure S40). The detection limits of NIR-HCy-NO 2 1-3 were 8, 114, and 181 ng/mL NTR, respectively. The sensitivity was the opposite trend to the probe polarity. Thus, the probe polarity was quite an important factor in deciding the detection limit.

Michaelis-Menten Kinetics
To elucidate the enzyme-substrate interaction between NTR and NIR-HCy-NO 2 1-3, Michaelis-Menten kinetics were conducted. In Figure 4, the initial reaction rates of the enzymatic reduction were dependent on the concentration of NIR-HCy-NO 2 1-3 (0-40 µM), and the reduction induced by NTR was saturated above 40 µM of NIR-HCy-NO 2 1-3. Consequently, the kinetic curves ( Figure 4) followed the Michaelis-Menten equation. In the kinetics results, the Michaelis constant (K m ) and the catalytic rate constant (k cat ) of NTR for NIR-HCy-NO 2 1-3 were obtained.
of NADH at 37 °C). The fluorescence signals of NIR-HCy-NO2 1-3 increased the NTR concentration dependently. The trend line R-square (R 2 ) values of NIR-HCy-NO2 1-3 were 0.9888, 0.9969, and 0.9964, respectively, and all three probes showed good linearity over the concentration range of 0.125-5 µg/mL of NTR ( Figure S40). The detection limits of NIR-HCy-NO2 1-3 were 8, 114, and 181 ng/mL NTR, respectively. The sensitivity was the opposite trend to the probe polarity. Thus, the probe polarity was quite an important factor in deciding the detection limit.

Michaelis-Menten Kinetics
To elucidate the enzyme-substrate interaction between NTR and NIR-HCy-NO2 1-3, Michaelis-Menten kinetics were conducted. In Figure 4, the initial reaction rates of the enzymatic reduction were dependent on the concentration of NIR-HCy-NO2 1-3 (0-40 µM), and the reduction induced by NTR was saturated above 40 µM of NIR-HCy-NO2 1-3. Consequently, the kinetic curves ( Figure 4) followed the Michaelis-Menten equation. In the kinetics results, the Michaelis constant (Km) and the catalytic rate constant (kcat) of NTR for NIR-HCy-NO2 1-3 were obtained.  Table S1, the catalytic efficiency (kcat/Km) increased in the order of NIR-HCy-NO2 1-3, and it was indicated that NIR-HCy-NO2 1 was reduced more effectively by NTR compared to NIR-HCy-NO2 2 and NIR-HCy-NO2 3. Additionally, the Km value of NTR for all three probes was lower than the kinetic values obtained in previous NTR probe studies under similar conditions. Thus, NIR-HCy-NO2 1-3 had a better affinity to NTR compared to the other probes reported in previous studies (Table S2). While NIR-HCy-NO2 1 was superior to nitrofurazone in terms of its catalytic efficiency, NIR-HCy-NO2 2 and NIR-HCy-NO2 3 were lower. The above results suggest that NIR-HCy-NO2 1  Table S1, the catalytic efficiency (k cat /K m ) increased in the order of NIR-HCy-NO 2 1-3, and it was indicated that NIR-HCy-NO 2 1 was reduced more effectively by NTR compared to NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3. Additionally, the K m value of NTR for all three probes was lower than the kinetic values obtained in previous NTR probe studies under similar conditions. Thus, NIR-HCy-NO 2 1-3 had a better affinity to NTR compared to the other probes reported in previous studies (Table S2). While NIR-HCy-NO 2 1 was superior to nitrofurazone in terms of its catalytic efficiency, NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 were lower. The above results suggest that NIR-HCy-NO 2 1 is a more suitable substrate for NTR kinetics analysis than the other two probes. It was also suggested that the charged function groups (sulfonate and quaternary ammonium) in indolium interrupt the reduction induced by NTR.

NTR Activity Imaging in Live Cells
We expanded the biological application of NIR-HCy-NO 2 1-3 to NTR activity imaging in live cells. The A549 cell line was selected for confocal fluorescence images. First, the cytotoxicity tests of NIR-HCy-NO 2 1-3 were performed to determine the probes' nontoxic concentration level. All the probes were nontoxic at high concentrations (≥20 µM) for cell imaging after 1 h, suggesting that NIR-HCy-NO 2 1-3 were biocompatible ( Figure S41) for live cell NTR activity imaging.
At a low concentration (<5 µM), NIR-HCy-NO 2 1 induced sufficient fluorescence signals for imaging after 10 min ( Figure 5A). However, the fluorescence signals of NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 were not detected at 40 µM after 30 min of treatment ( Figure S42A). The arithmetic mean intensity of NIR-HCy-NO 2 1 increased in a concentration-dependent manner; the arithmetic mean intensity of each concentration of NIR-HCy-NO 2 1 increased 10.1-, 19.8-, and 28.8-fold, respectively, compared to the untreated group ( Figure 5B). However, the arithmetic mean intensities of NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 increased up to about 6.2-and 2.7-fold, respectively ( Figure S42B), which was too small a fold change considering their treatment concentrations. In the Michaelis-Menten kinetics results, the catalytic efficiencies of NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 were relatively lower than NIR-HCy-NO 2 1, and the trend was the same as live cell imaging ( Figures 4 and 5 and Figure S42 and Table S2). Additionally, NIR-HCy-NO 2 1 was localized in mitochondria, which was overlapped with a MitoTracker ( Figure 5C). The scatter plot of the two channels (MitoTracker (MTGFM) and NIR-HCy-NO 2 1 (Cy5.5)) showed a linear form and tendencies to synchronize with a Mender's colocalization coefficient of 0.9748 ( Figure 5D).

In Vivo Imaging
We next applied NIR-HCy-NO2 to in vivo fluorescence imaging in a xenograft model using an IVIS Spectrum system. Prior to imaging, we first determined the adequate concentration of each probe for injection. As shown in Figure S45, the fluorescent background signals from prereacted NIR-HCy-NO2 1-3 were not observed from 5-50 µM in PBS. Based on their cytotoxicity results, 20 µM was decided on for NIR-HCy-NO2 1-3, considering a sufficiently strong fluorescence signal for in vivo imaging. Then, nude mice bearing A549 xenograft tumors were intratumorally injected with each probe and monitored in a treatment time-course manner. In Figure 6D, the fluorescence signal of all three NIR-HCy-NO2 initially enhanced after the injection, and NIR-HCy-NO2 1 showed the strongest fluorescence signal among them. The fluorescence of NIR-HCy-NO2 1 was not saturated until 20 min post-injection, while that of NIR-HCy-NO2 2 and NIR-HCy-NO2 3 were saturated at 10 and 5 min, respectively. Thus, NIR-HCy-NO2 1 showed the best performance for in vivo tumor imaging among all NIR-HCy-NO2 probes due to the strong signal and the continuous reaction with the reductase (Figure 6D). Interestingly, the NIR-HCy-NO 2 1 signal was oxygen concentration-independent in the live cell ( Figure S43). In a previous study, a Cy7-based fluorescence probe detected mitochondrial NTR in an A549 cell line under normoxia, and it was identified as type I NTR, which is oxygen-independent [20]. In this study, NIR-HCy-NO 2 1 was also able to detect type I mitochondrial NTR under the same conditions and cell line. Moreover, the fluorescence signals of NIR-HCy-NO 2 1 were similar under normoxia and hypoxia, and this suggested that NIR-HCy-NO 2 1 was mainly reduced by type I mitochondrial NTR ( Figure S43B). To verify the relationship between the signal enhancement and reductase, dicoumarol was pretreated as the reductase inhibitor, and the signal in the treated group decreased by up to 40% compared to the untreated group ( Figure S44). As with previous results, NIR-HCy-NO 2 1 is more suitable for NTR activity imaging than NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3.

In Vivo Imaging
We next applied NIR-HCy-NO 2 to in vivo fluorescence imaging in a xenograft model using an IVIS Spectrum system. Prior to imaging, we first determined the adequate concentration of each probe for injection. As shown in Figure S45, the fluorescent background signals from prereacted NIR-HCy-NO 2 1-3 were not observed from 5-50 µM in PBS. Based on their cytotoxicity results, 20 µM was decided on for NIR-HCy-NO 2 1-3, considering a sufficiently strong fluorescence signal for in vivo imaging. Then, nude mice bearing A549 xenograft tumors were intratumorally injected with each probe and monitored in a treatment time-course manner. In Figure 6D, the fluorescence signal of all three NIR-HCy-NO 2 initially enhanced after the injection, and NIR-HCy-NO 2 1 showed the strongest fluorescence signal among them. The fluorescence of NIR-HCy-NO 2 1 was not saturated until 20 min post-injection, while that of NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 were saturated at 10 and 5 min, respectively. Thus, NIR-HCy-NO 2 1 showed the best performance for in vivo tumor imaging among all NIR-HCy-NO 2 probes due to the strong signal and the continuous reaction with the reductase ( Figure 6D).

Discussion
In summary, we succeeded in the design and synthesis of novel NIR off-on probes NIR-HCy-NO2 1-3 for NTR activity imaging. The water solubility of NIR-HCy-NO2 derivatives was different depending on the functional group introduced to NIR-HCy-NO2, and all NIR-HCy-NO2 1-3 reacted selectively with NTR. The fluorescence intensity and absorbance spectra of NIR-HCy-NO2 1-3 changed due to the reduction reaction induced by NTR, and the LODs of NIR-HCy-NO2 1-3 were under 200 ng/mL of NTR. Among them, the LOD of NIR-HCy-NO2 1 was the lowest with 8 ng/mL of NTR, and the catalytic efficiency was the highest at 0.22 ± 0.03 µM −1 ·s −1 . In intracellular NTR activity imaging, the performance of NIR-HCy-NO2 1 was overwhelmingly good, and the signal was reduced with a dicoumarol treatment known as the reductase inhibitor. Additionally, the NIR-HCy-NO2 1 response was oxygen-independent, and it was considered that NIR-HCy-NO2 1 should be used to detect and image type I mitochondrial NTR as the potential target in live cells. NIR-HCy-NO2 1 showed a strong fluorescence intensity and sustained reactivity in vivo. In conclusion, NIR-HCy-NO2 1 showed a good performance among the three derivatives, and similar trends were observed in the Michaelis-Menten kinetics and intracellular and in vivo NTR activity imaging. In further studies, NIR-HCy-NO2 has the potential to be applied to various fields related to NTR, and it is able to be used for intracellular and in vivo NTR activity imaging.

Discussion
In summary, we succeeded in the design and synthesis of novel NIR off-on probes NIR-HCy-NO 2 1-3 for NTR activity imaging. The water solubility of NIR-HCy-NO 2 derivatives was different depending on the functional group introduced to NIR-HCy-NO 2 , and all NIR-HCy-NO 2 1-3 reacted selectively with NTR. The fluorescence intensity and absorbance spectra of NIR-HCy-NO 2 1-3 changed due to the reduction reaction induced by NTR, and the LODs of NIR-HCy-NO 2 1-3 were under 200 ng/mL of NTR. Among them, the LOD of NIR-HCy-NO 2 1 was the lowest with 8 ng/mL of NTR, and the catalytic efficiency was the highest at 0.22 ± 0.03 µM −1 ·s −1 . In intracellular NTR activity imaging, the performance of NIR-HCy-NO 2 1 was overwhelmingly good, and the signal was reduced with a dicoumarol treatment known as the reductase inhibitor. Additionally, the NIR-HCy-NO 2 1 response was oxygen-independent, and it was considered that NIR-HCy-NO 2 1 should be used to detect and image type I mitochondrial NTR as the potential target in live cells. NIR-HCy-NO 2 1 showed a strong fluorescence intensity and sustained reactivity in vivo. In conclusion, NIR-HCy-NO 2 1 showed a good performance among the three derivatives, and similar trends were observed in the Michaelis-Menten kinetics and intracellular and in vivo NTR activity imaging. In further studies, NIR-HCy-NO 2 has the potential to be applied to various fields related to NTR, and it is able to be used for intracellular and in vivo NTR activity imaging.

Mass Analysis of Reduced NIR-HCy-NO 2 1-3
To prepare the reaction mixture for mass analysis, 500 µM of NIR-HCy-NO 2 1-3 were reduced by 1 µM of NTR with 500 µM of NADH in PBS (10 mM, pH 7.4) at room temperature for 3 min. To quench the enzymatic reaction, β-mercaptoethanol was added to the reaction mixture until the concentration was 2%. The product mass was measured using a high-resolution mass spectrometer (micrOTOF-QII, Bruker Daltonik, Bremen, Germany) in the electrospray ionization (ESI) mode.

Cell Culture
Non-small cell lung cancer adenocarcinoma A549 cell lines were obtained from the Bioevaluation Center at the Korea Research Institute of Bioscience and Biotechnology (KRIBB). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin/streptomycin (P/S). The cells were cultured at 37 • C under 5% CO 2 .

Cytotoxicity Assay
The cytotoxicity assay was carried out using the methylene blue staining method (Methylene blue, Sigma-Aldrich, St. Louis, MO, USA). A549 cells were seeded into 96-well cell culture plates at 2 × 10 4 /well and preincubated in DMEM (10% (v/v) FBS and 1% (v/v) P/S). After the medium in the wells was removed, NIR-HCy-NO 2 1-3 (100 µL/well) at concentrations of 0-40 µM were added to the wells of the treatment group, respectively. The cells were incubated for 1 h at 37 • C under 5% CO 2 and then fixed for over 1 h by adding 10% formalin solution (50 µL/well). After each well attached to fixed cells was washed with PBS (pH 7.4), the fixed cells were stained with 2% methylene blue working solution (50% (v/v) methanol) for 1 h. The stained cells were washed strongly with distilled water and dried sufficiently at room temperature. The cells were lysed using 0.5% hydrogen chloride, and an imaging reader (CYTATION5, BioTek, Winooski, VT, USA) was used to measure the OD600 (absorbance value) of each well. Cell viability was calculated using the following formula: cell viability (%T) = A t /A c × 100 (%), where A t denotes the absorbance value of the treated group, and A c denotes the absorbance value of the untreated group.

Confocal Fluorescence Imaging in Living Cells
A549 cells (4 × 10 4 /well) were plated on µ-Slide 4 Well (ibidi, Gräfelfing, Germany) and were allowed to adhere for 24 h. All staining procedures were carried out under the normoxia condition. The cells were incubated in serum-free DMEM at 37 • C with NIR-HCy-NO 2 1 (0.5, 1, and 2 µM) for 10 min, or NIR-HCy-NO 2 2 and NIR-HCy-NO 2 3 (40 µM) for 30 min. Then, the cells were washed with DPBS (0.4 mL × 2 times) and were further incubated with 5 µg/mL of Hoechst 33,342 in a serum-free DMEM at 37 • C for 5 min. After washing with DPBS (0.4 mL × 3 times), fluorescence imaging was performed with an LSM 800 confocal fluorescence microscope (ZEISS, Jena, Germany) at a 40× water immersion objective lens. The fluorescence signal of the cells incubated with Hoechst 33,342 and NIR-HCy-NO 2 1-3 was collected at 400-600 nm using a semiconductor laser at 405 nm as an excitation resource of Hoechst 33,342 and at 645-700 nm using a semiconductor laser at 640 nm as an excitation resource of NIR-HCy-NO 2 1-3, respectively.
For the comparison of the fluorescence images between normoxia and hypoxia, the hypoxic culture was carried out for 12 h at 37 • C under 5% CO 2 and 2% O 2 . The staining procedure of NIR-HCy-NO 2 1 and Hoechst 33,352 and the condition of confocal fluorescence imaging were performed in the same way as previously outlined.
A549 cells were used for the colocalization imaging, and the confocal imaging for colocalization was carried out in the same culture and using the NIR-HCy-NO 2 1 staining condition with fluorescence confocal imaging. Then, the cells were treated with 0.5 µM of MitoTacker ® Green FM (MTGFM) for 30 min and were washed with DPBS (0.4 mL × 3 times). Fluorescence imaging was carried out in the same conditions outlined for the previous experiment; the fluorescence signal was collected at 400-650 nm using a semiconductor laser at 490 nm as an excitation resource of MTGFM.

Reductase Inhibition Test in Live Cells
Dicoumarol (Sigma-Aldrich) was used for reductase inhibition. A549 cells were pretreated for 4 h with dicoumarol (500 µM), and then NIR-HCy-NO 2 1 (10 µM) was added for 20 min. The fluorescence intensity in A549 cells was measured using a Synergy H1 multimode plate reader (BioTek instruments, Winooski, VT, USA).

Fluorescence Imaging in Xenograft Mice
in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. Institutional Review Board Statement: All animals were cared for in accordance with the guidelines provided by the Korea Research Institute of Bioscience and Biotechnology (KRIBB), and all experiments using mice were approved by KRIBB-IACUC (approval number: KRIBB-AEC-19084).

Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.