Engineering a nanolab for the determination of lysosomal nitric oxide by the rational design of a pH-activatable fluorescent probe† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c5sc04415d

A pH-activatable fluorescent probe, Rhod-H-NO, was designed and synthesized for the determination of lysosomal NO in living cells and in vivo.


Introduction
Nitric oxide (NO), a small uncharged free radical, which is one of the mediators of cellular responses, produced by nitric oxide synthases, plays a very important role as a signaling molecule in a variety of physiological and pathological processes that take place in the cardiovascular, nervous, and immune systems. 1 In a recent study, it was discovered that lysosomal functions are subtly regulated by NO. 2 These functions include the degradation of a cells own components to supply energy and nutrients for its growth through the lysosomal machinery, 3 and can cause various diseases (such as Fabry, Gaucher, or Danon diseases) due to lysosomal disorders. 4 Recently, numerous uorescent probes for intracellular NO have been developed and applied to intracellular sensing and imaging, 5,6 which has signicantly enriched our knowledge about NO homeostasis and the crucial roles of NO in many biological processes. It is worth noting that obtaining convincing evidence about the interrelation between the variations of NO levels in lysosomes and different physiological processes remains a challenge, because available probes for the selective tracking of NO in the lysosomes of living cells have rarely been reported hitherto. One of the bottlenecks is the lack of organelle-specicity, which results in a high background signal from cytosol and other organelles, meaning that the probes cannot provide high spatial resolution in lysosomes. To tackle this challenge, Xiao et al. conjugated a lysosomal-specic morpholine moiety as the guiding unit with a NO probe to target the subcellular organelle recently, 7 but this strategy suffers from a serious problem: the morpholine moiety exhibits an alkalizing effect on the lysosomes so that longer incubation times with these probes can induce an increase in lysosomal pH and result in cell death. 8 In addition, existing research shows that organic probes are susceptible to acidic pH and interferences from other intracellular species, and are even prone to being degraded by substantial hydrolase in lysosomes, 9 which makes the detection unsatisfactory for application to living cells. Therefore, the development of an ideal uorescent probe with lysosomal targeting is highly desirable for quantifying the variations of NO levels in real time and in a dynamic range of concentrations.
Mesoporous silica nanoparticles (MSNs), popular smallmolecule drug reservoir systems which show excellent biocompatibility, ease of functionalization and are nontoxic to cells, have attracted widespread interest for drug delivery purposes within the past decade. 10 In particular, MSNs can undergo cellular uptake into acidic lysosomes by endocytosis when the nanoparticles are below 200 nm in diameter, 11 which could provide a potential way to study lysosomal tracking and explore its role in invasion with high-resolution spatial images.
Recently, we have developed two uorescent probes for the detection of lysosomal H 2 S and Cu 2+ based on simultaneous target and location activation. 12,13 One of the targeting strategies is via entrapping the uorescent probe into the nanopores of MSNs. The results were inspiring to us, showing that the intact probe molecule can be stored in the nanopores without interference and degradation and then automatically accumulated in lysosomes, which successfully realized the detection of lysosomal Cu 2+ and also provided an efficient method to address these challenges in the determination of lysosomal NO.
As a part of our ongoing interest in uorescent probes for application in lysosomes, we herein fabricate a nanolab for the determination of lysosomal NO by engineering a pH-activatable uorescent probe into the nanopores of MSNs (Scheme 1). Firstly, a pH-activatable uorescent probe, namely Rhod-H-NO, was designed and synthesized for the determination of NO. To avoid the unpredictable uorescence signals from other organelles and to make full use of the acidic conditions of lysosomes, we adopted a strategy to lock the widely applied NO recognition site moiety of o-phenylenediamine by a pH-sensitive imine bond. 14 The presence of the imine bond shows a silent uorescence response to NO in neutral and alkaline conditions, while the lysosomal pH (4.0-6.0) of which mediates hydrolysis of the imine group affords a large, rapid, NO-induced uorescence response. Then, in order to realize the accumulation of probes in lysosomes, MSNs were applied as protective nanocoats to entrap Rhod-H-NO in the nanopores, which prevented the probes from interference and degradation, and thus provided a reaction lab for the determination of NO. Finally, b-cyclodextrin (b-CD), a widely reported "gatekeeper" for closing the gates of the pores of MSNs, 15 was conjugated on the surface of the MSNs to stop external species. This strategy holds the promise of real application in biological research and medical diagnosis.

Results and discussion
Rhod-H-NO was synthesized according to Scheme 1A. The reaction between rhodamine B and o-phenylenediamine led to the spiro compound Rhod-NO, and further condensation reaction between Rhod-NO and propionic aldehyde afforded the probe Rhod-H-NO in 65.2% yield. The puried product was fully characterized by NMR spectroscopy and ESI-MS to conrm the structure (Fig. S1-S3 †). With the probe Rhod-H-NO in hand, we rstly veried the feasibility of the above-mentioned design that NO couldn't "switch-on" the uorescence of Rhod-H-NO unless being activated by acidic pH. As shown in Fig. S4, † Rhod-H-NO exhibits no absorption features in the visible region and is essentially non-uorescent in neutral aqueous solution (pH ¼ 7.0) owing to the spirolactam structure. Upon the addition of NO, there is no signicant absorption or uorescence change indicating that the spirolactam structure remained intact. When Rhod-H-NO was treated with an acidic buffer (pH ¼ 5.0), a strong absorption band centered at 552 nm (3 ¼ 6.50 Â 10 4 M À1 cm À1 ) and a bright uorescence emission with peak maximum at 590 nm (f F ¼ 0.63) were observed obviously, indicating rhodamine ring-opening via a NO-induced reaction in acidic medium. HPLC analysis further demonstrated that the reaction of Rhod-H-NO and NO in acidic medium was identical with the proposed mechanism ( Fig. S5 †). Based on this, we anticipate that the activation response model is benecial towards fabricating an effective lysosome-targetable molecular tool for exploring NO biology.
To ensure that Rhod-H-NO can accumulate effectively and exist stably in lysosomes without interference and degradation, a novel nanolab was fabricated by engineering the probe into the nanopores of MSNs and using b-CD as the gatekeeper of the nanopores, which allows NO to pass through the cavities of b-CD and enter the nanopores to react with Rhod-H-NO. The fabrication process is shown in Scheme 2. MSNs with typical MCM-41 hexagonal arrangements were rst prepared according to the literature protocol. 16 Then the surface of the MSNs was functionalized with chloride groups by treatment with (3-chloropropyl)triethoxysilane to afford MSN-Cl. The azide terminated MSN-N 3 was obtained from the reaction of as-prepared MSN-Cl and sodium azide. Aer removing the surfactant template (n-cetyltrimethylammonium bromide, CTAB) from MSN-N 3 , Rhod-H-NO was then entrapped in MSN-N 3 by a diffusion experiment. Subsequently, the CD-alkyne, which was synthesized according to a previously reported method ( Fig. S6 †), 17 was attached to the MSN surface using a click chemistry approach, giving the nanolab Rhod-H-NO@MSN-CD.
Transmission electron microscopy (TEM) images showed an obvious border aer b-CD was conjugated to the surface of the MSNs, but the average diameter did not show a signicant difference ($80 nm) (Fig. 1A). In addition, MSN-Cl only showed the silica framework vibrations, whereas MSN-N 3 exhibited the characteristic azide stretch signal at 2110 cm À1 , but this band was strongly reduced in intensity upon modication with b-CD (Fig. 1B). The zeta potential values of MSN-Cl, MSN-N 3 and MSN-CD were +18.5 mV, À20.6 mV, and +30.2 mV respectively (Fig. S7 †). These results conrmed that the surface of MSNs was well modied with b-CD. When Rhod-H-NO was loaded into the nanopores of MSN-CD, XRD patterns showed that the low-angle reections indexed as (110) and (200) had disappeared ( Fig. S8 †), and nitrogen adsorption-desorption isotherms and pore size distributions showed that the average pore size decreased with an increase in surface density and inlling ( Fig. S9 †). Fig. 1C further demonstrates that Rhod-H-NO is effectively entrapped in the nanopores without being released for more than 48 h, which is attributed to the size of Rhod-H-NO (1.84 nm Â 1.29 nm Â 0.96 nm, calculated by Gaussian 09 programs) which is larger than the nanopore gates aer modi-cation with b-CD. The loading efficiency of Rhod-H-NO in MSN-CD was estimated to be $15.6 wt%. Owing to the protective effect of the MSNs and the enhanced water-solubility of b-CD, Rhod-H-NO@MSN-CD exhibited an improved long-term stability without degradation in cell lysate compared with free Rhod-H-NO (Fig. 1D). Most importantly, NO molecules could freely diffuse into the nanolab through the cavities of the b-CD rings to react with Rhod-H-NO (Fig. S10 †).
The real-time kinetics of Rhod-H-NO@MSN-CD and its response toward NO in a buffer solution with different pH environments further conrmed that Rhod-H-NO@MSN-CD is pH-activatable for NO determination ( Fig. 2A). Aer changing the neutral solution of Rhod-H-NO@MSN-CD to acidic conditions, there was slight, quick, increase in the uorescence intensity which then remained steady over tens of minutes. However, the uorescence intensity dramatically increased and reached a plateau in $30 min aer being subsequently incubated with NO. When NO was rst added into the neutral solution of Rhod-H-NO@MSN-CD and the solution was then adjusted to acidic pH, the uorescence change further affirmed the pH activation characteristics (Fig. S11 †). Then, we got the optimal pH range by measuring the uorescence intensity enhancements of Rhod-H-NO@MSN-CD towards NO. As seen in Fig. S12, † a signicant enhancement of F/F 0 was observed over the pH range from 6.0 to 4.0, which is in good accord with lysosomal pH conditions, while Rhod-NO@MSN-CD was pH insensitive for NO response, suggesting that Rhod-H-NO@MSN-CD is a good candidate for the detection of lysosomal NO.
Next, we performed uorescence titration studies of Rhod-H-NO@MSN-CD for NO in a buffer solution (pH ¼ 5.0). As shown in Fig. 2B, the addition of NO with increasing concentrations from 0 to 20.0 mM NO, elicited a gradual enhancement of the emission band at 590 nm. The uorescence enhancement (F/F 0 ) reached a maximal value (26-fold) in the presence of 20.0 mM NO. Moreover, there was an excellent linear correlation between F/F 0 and NO concentration in the range 0.2 to 6.0 mM with a detection limit (3s/slope) of 100 nM, indicating that Rhod-H-NO@MSN-CD would be a potential tool to monitor endogenous NO in lysosomes. Moreover, Rhod-H-NO@MSN-CD exhibited high selectivity for NO with various biologically relevant species in an aqueous buffer solution at pH ¼ 5.0 (Fig. S13 †). The  presence of reactive oxygen (H 2 O 2 , $OH, O 2 À , 1 O 2 , ClO À ), reactive nitrogen (NO 2 À , NO 3 À , ONOO À ), reactive sulfur (SO 3   2À , Cys, Hcy, GSH), and some common metal cations (K + , Na + , Ca 2+ , Mg 2+ , Zn 2+ , Fe 2+ , Cu 2+ ), didn't cause an observable uorescence enhancement, which was due to the specicity of the o-phenylenediamine receptor for NO.
Rhod-H-NO@MSN-CD was then investigated for the ability to both target the lysosomes and respond to NO in living cells. The cytotoxicities of Rhod-H-NO@MSN-CD on living cells were rst evaluated by employing standard cell viability protocols (MTT assay) (Fig. S14 †). Aer being cultured for 24 h, the cellular viability of HeLa cells was over 95% and no signicant difference in the morphology was observed even when the concentration of Rhod-H-NO@MSN-CD was increased up to 100 mg mL À1 , showing a very low cytotoxicity. Next, co-localization experiments were performed by co-staining HeLa cells with LysoTracker Green, MitoTracker Green and Rhod-H-NO@MSN-CD. From Fig. 3 and S15, † it can be seen that HeLa cells stained with Rhod-H-NO@MSN-CD for 2 h at 37 C in the presence of 100 mM exogenous NO displayed signicant red uorescence, which merged well with the image from staining with LysoTracker Green but not with MitoTracker Green. Moreover, the intensity proles of the linear regions of interest across the HeLa cells stained with Rhod-H-NO@MSN-CD and LysoTracker Green vary in close synchrony (Fig. S16 †). The Pearson's colocalization coefficient and overlap coefficient were 0.865 and 1.528 respectively, as calculated using Autoquant X2 soware, indicating that Rhod-H-NO@MSN-CD existed predominantly in the lysosomes and was activated by the coexistence of acidic pH and NO. TEM provides further additional physical evidence for the nanoparticles residence in the lysosomes (Fig. 3d). In view of the Rhod-H-NO probe selectively responding to the coexistence of H + and NO in vitro, HeLa cells were pretreated with 1,8-bis(dimethylamino) naphthalene (DMAN), a common proton sponge, to promote lysosomal damage due to the proton-sponge effect. 18 As expected, no obvious uorescence was observed aer incubation with Rhod-H-NO@MSN-CD in the presence of exogenous NO (Fig. S17 †) because the probe escaped into the cytoplasm when the lysosomes were disrupted, further proving that the intracellular response behaviour of Rhod-H-NO@MSN-CD happened in acidic lysosomes rather than in cytoplasm and other organelles with neutral environment.
Next, further experiments were performed to verify the viability of Rhod-H-NO@MSN-CD to detect variations of NO level in living RAW 264.7 macrophages, which are well-known as NO producing cells. As shown in Fig. 4, the image of the probe-loaded macrophages gave barely detectable uorescence ( Fig. 4A(a)). Nevertheless, a signicant increase in the intracellular uorescence was visualized aer further treatment with the widely used NO donor diethylamine NONOate (Fig. 4A(b)), and the intensity presented a tendency toward increased uorescence over time and reached a maximum aer 20 min (Fig. S18 †), indicating that NO released from NONOate was able to diffuse into the lysosomes and then enter into the nanolab to react with Rhod-H-NO. According to the spontaneous NO releaser which in principle releases two equivalents of NO, 19 the linearity between the relative uorescence intensity and NO concentration was established, as shown in Fig. 4B and C, by pretreating the cells with different concentrations of NONOate,  which demonstrated that the probe could quantitatively evaluate the intracellular NO content. Thus, to initiate physiological NO production, the well-known external stimuli, bacterial endotoxin lipopolysaccharide (LPS), L-arginine (L-Arg) and pro-inammatory cytokine interferon-gamma (IFN-g), were pretreated with cells for 12 h to promote NO production by inducible nitric oxide synthase (iNOS), 20 and Rhod-H-NO@MSN-CD was subsequently introduced to the cells, showing a strong red uorescence (Fig. 4A(c)). According to the linearity established in Fig. 4C, the average basal NO levels were estimated as ca. 8.36 mM. It should be conrmed that the uorescence increase shown in Fig. 4A(c) was caused by NO generation and not by environmental changes. A NO scavenger, 2-(4-carboxyphenyl)-4,4,5,5-tetramethyl-imidazoline-1-oxyl-3oxide (PTIO), 21 was further introduced to the LPS and IFN-g treated cells, which lead to the uorescence being weakened obviously (Fig. 4A(d)). Moreover, we further quantitatively evaluated the endogenous NO in individual cells on a large scale using ow cytometry analysis. Fig. 4D demonstrates the intracellular accumulation of NO where an increase in intracellular uorescence is indicated by a shi in the uorescence signal measured in the FITC channel, which shows the same trend as the microscopy imaging of Fig. 4A. These results indicate that uorescence imaging using Rhod-H-NO@MSN-CD as a nanolab is an effective tool for measuring different NO levels produced in lysosomes.
With an excellent lysosomal-targeting nanoprobe in hand, we used Rhod-H-NO@MSN-CD to investigate the produced NO level in lysosomes in an inamed mouse model. 200 mL of LPS (1 mg mL À1 ) was subcutaneously injected into the le rear leg of a mouse to cause inammation for 12 hours, while the right rear leg was treated with saline simply as the control. From in vivo imaging (Fig. 5A), one can see that strong uorescence was observed from the le leg but not from the right leg aer Rhod-H-NO@MSN-CD was intravenously injected, suggesting that the NO level increased signicantly during inammation. To further examine whether the signal came from the lysosomes of macrophage cells, the le leg skin was then sectioned to get the inammation tissue. As shown in Fig. 5B, the uorescence signal from Rhod-H-NO@MSN-CD was consistent with the results of the immune-staining histological sections with macrophage marker CD11b. From the results we came to the conclusion that the lysosomal NO level of macrophage cells during inammation could be detected by Rhod-H-NO@MSN-CD.

Conclusions
In conclusion, we have presented a novel strategy to fabricate a nanolab for the determination of intracellular bioactive molecules by engineering the highly selective and sensitive small molecular probe into mesoporous nanomaterials. The nanolab was capable of detecting lysosomal NO changes in the presence of exogenous or endogenous NO in living cells and in vivo. Signicantly, this paper successfully addresses some challenges in intracellular imaging analysis. On the one hand, the location-activatable approach could effectively eliminate false signals and improve spatial resolution. On the other hand, the hermetical nanolab can protect the stability of the functional probe and avoid interference from a complex biological environment. In addition, the modication with b-CD on the surface of MSNs can improve the stability and biocompatibility of the nanomaterials in a physiological environment. In view of these merits, we anticipate that this design strategy would open up a new train of thought for the development of efficient molecular tools for bioanalytical and biomedical applications.