Biocompatible ZnS:Mn quantum dots for reactive oxygen generation and detection in aqueous media

The formation and morphology of Mn-doped ZnS QDs synthesized via a one-pot wet chemical method are depicted in Fig. 1. The XRD and selected area electron diffraction (SAED) patterns show three major diffraction peaks and rings, which are indexed to the diffraction planes of zinc blende ZnS [(111), (220), and (311)] according to the JCPDS card, File No. 65-0309 and space group (see Fig. 1a). Both techniques confirmed that ZnS:Mn crystallized preferentially in the more stable cubic ZnS phase. No detectable lattice distortion (shift) was observed in the host ZnS indicative of the successful random substitution of Mn²⁺ into Zn²⁺ sites at 5 wt% Mn doping. The well-defined peaks and rings found and the absence of parasitic phases and signal of ZnS wurtzite in the XRD and SAED patterns evidence that ZnS:Mn QDs are of high crystalline quality and purity. The calculated lattice constant of ZnS:Mn was a = 0.5384 ± 0.0007 nm consistent with its standard value (Beltran-Huarac et al. 2013a, b). The prominent broadening of the XRD diffraction peaks is ascribed to the nanocrystalline nature of the QDs. The average crystallite size was calculated to be ~4 nm by means of Scherrer’s formula.

The microstructure quality of ZnS:Mn QDs was studied by Raman scattering spectroscopy (Fig. 1b). The well-pronounced low-frequency Raman bands observed were simulated using the damped harmonic oscillator phonon model (DHOPM) (Beltran-Huarac et al. 2014). From the simulation, the Raman modes are determined to be at 261 and 342 cm⁻¹, which can be ascribed to the characteristic transverse and longitudinal optical modes of cubic ZnS, respectively (Nilsen 1969). The corresponding fit shows that those bands are red-shifted by ~10 cm⁻¹ and narrow (full width at half maximum, FWHM ~23 cm⁻¹) when compared to bulk ZnS (Nilsen 1969), which clearly indicates that the zinc blende ZnS is under tensile stress as a result of the growth process. The morphology and size distribution of ZnS:Mn QDs were studied by electron microscopy. The HRTEM images depicted in Fig. 1c show well-dispersed QDs with a high degree of crystallinity whose sizes range from 2 to 7 nm. A closer look of representative QDs shows that they are almost spherical with lattice fringes readily observable and a diameter of ~4 nm. The surfaces of QDs were clean, smooth, and atomically resolved with d-spacing of 0.31 nm corresponding to the major plane (111) of ZnS, consistent with the XRD and SAED analyses. The statistical size distribution (see lower inset of Fig. 1c) obtained by image analysis further confirmed that the average size of QDs was ~4 nm in accordance with the XRD results. Taken altogether, the XRD, Raman, and HRTEM analyses indicate the formation of narrowly distributed high-quality ZnS:Mn QDs with a zinc blende structure.

The assembly of the amperometric biosensor based on Tyr enzyme immobilized onto ZnS:Mn QDs to detect H₂O₂ is depicted in Fig. 2. Prior to the covalent immobilization, the surface of the as-cleaned polycrystalline Pt electrodes was modified using 4-ATP that enables the formation of a NH₂-terminated AM, followed by covalent linkage of QDs. The carboxylic acid groups on the surface of QDs were crosslinked to the primary amines of enzyme through an EDC-mediated coupling, allowing for a strong covalent immobilization of Tyr onto QDs (see Fig. 2a). Each immobilization step was monitored by CV via direct detection of the electron transfer of the active ferrocyanide species present in the solution, as shown in Fig. 2b. The voltammogram of the bare Pt electrode shows an anodic peak at 280 mV and a cathodic peak at 220 mV that arises from [Fe(CN)₆]⁴⁻ and [Fe(CN)₆]³⁻, respectively, and is characteristic of the quasi-reversible one-electron transfer redox behavior of the electroactive species. It was observed that such redox peaks disappear when the 4-ATP electrode modification occurs, which is attributable to the insulating effect produced by the densely packed SAM formation on the electrode surface that hinders the interfacial charge transfer between the ferricyanide ion and the electrode. SAM formation provides the chemical environment necessary to covalently attach MPA-capped QDs onto the electrode. A relative increase in current was detected in Pt/SAM/QDs, which was associated to the electron injection into sub-bandgap states of the metal (Zn,Mn) sulfide (Topoglidis et al. 2001; Boschloo and Fitzmaurice 1999). A decrease in current was observed when Tyr (IEP of ~4.7–5.0) is covalently linked to the QDs. This is due to the enzymatic blockage that is produced on the surface of QDs, thus corroborating both the enzyme deposition and the proposed biosensor assembly (Hernandez-cancel et al. 2015) (see Fig. 2a).

Figure 3a shows the electrocatalytic response of the biosensor to H₂O₂ at concentrations ranging from 0 to 75 μM and in the presence of 0.1 M PBS. An enzyme redox peak centered at ~63 mV in the absence of H₂O₂ was observed, which signifies that the electroactive prosthetic group of the enzyme is present in close proximity to the modified electrode. An enhancement of the anodic peak current (~4.61 μA) was obtained when a 12 μM H₂O₂ solution was gradually added, indicating that the current response of the biosensor represents the characteristic electrochemical behavior of H₂O₂ (Kafi et al. 2008), as expected. This is due mainly to the catalytic effect of Tyr in the absence of a one-electron donor substrate, which remains active and exhibits an enhanced stability as a result of the effective immobilization of ZnS:Mn QDs. At higher concentrations (as high as 75 μM), a net anodic current of ~13.01 μA was detected. This demonstrates that ZnS:Mn-immobilized Tyr exhibits improved catalytic properties in the reduction of H₂O₂, and is partially ascribed to the increase of binding sites of the QDs for enzyme loading. It is noteworthy that although the catalytic cycle of Tyr involves H₂O₂ consumption and water generation (see Fig. 3b), no inactivation of Tyr was observed up to 100 μM of H₂O₂ (a current of ~18.18 μA). This tendency clearly indicates that H₂O₂ is acting as both oxidizing and reducing substrate in solution and is not irreversibly adsorbed on the biosensor surface, i.e., the reaction between H₂O₂ and Tyr is a diffusion-controlled process at this scan rate. The catalase activity found in mushroom Tyr/QDs for H₂O₂ can be understood in terms of a catalase-cycled kinetic mechanism, as follows. It is known that under anaerobic conditions Tyr or polyphenol oxidase exhibits two enzymatic species, met-Tyr and deoxy-Tyr, which can initiate the reaction indistinctly. Thus, a H₂O₂ molecule of the solution at pH 7.0 can reduce met-Tyr to deoxy-Tyr through the formation of oxy-Tyr, which releases oxygen into the medium and is subsequently transformed into deoxy-Tyr. Then, this is oxidized to met-Tyr upon binding another H₂O₂ molecule (see Fig. 3b). In this context, H₂O₂ can be bound to the axial position of one of the active copper atoms of the oxy-Tyr site during the catalase cycle to produce deoxy-Tyr, which correlates well with the recent results obtained for biosensors based on Tyr/ZnO (Rather et al. 2014; Garcia-Molina et al. 2005). This is also consistent with the preparation of binuclear copper active sites of tyrosinase, being able to act on both monophenols and o-diphenols in the oxygenated form, oxy-Tyr, whereas the met and deoxy forms need the assistance of oxygen and H₂O₂ (Garcia-Molina et al. 2005).

In the previous discussion, we furnished evidence on the possibility of detecting ROS, such as H₂O₂, based on Tyr and ZnS:Mn QDs. We evaluate here the ability of QDs to generate ROS when used as PS, employing both a chemical scavenger (DPBF, a sensor highly sensitive to ¹O₂) and a standard PS (RB, a dye with a known ¹O₂ quantum yield) in organic and aqueous media and through a photooxidation approach. To do this, we specifically chose the RB/DPBF pair system because its photo-oxygenation pathway can solely produce ¹O₂ via a Type-II mechanism (Beltran-Huarac et al. 2010; Ding et al. 2011). Thus, using ZnS:Mn as a PS, excited singlet-state oxygen can be produced by energy transfer, which in turn oxidizes the ground-state substrate. Figure 4 depicts the absorption spectra of the DPBF oxidation photosensitized by RB and ZnS:Mn QDs in water and buffer solution, when exposed to 532 nm laser irradiation for every 2 s under vigorous stirring. In order to confirm that the reaction is induced by ¹O₂ instead of a direct interaction between QDs and DPBF, air-saturated solutions were prepared. Oxygen-depleted solutions (via inert gas purge) showed no detectable changes in the absorption spectra. Figure 4a depicts the absorption profile of DPBF (400 μL of a 6.0 × 10⁻⁵ M solution) in 2 mL of water in the presence of RB (1.0 × 10⁻⁵ M). This spectrum shows the typical absorption band edges of DPBF centered at 264, 317, and 413 nm, and of RB peaking at 555 nm (Beltran-Huarac et al. 2010, 2013a, b; Tuncel et al. 2011). No photooxidation of RB was observed when the samples were exposed to light (10 mW) for 16 s, as expected. However, the absorption bands of DPBF dropped significantly indicating that the scavenger is being oxidized by singlet oxygen. A more conspicuous photooxidation was seen in the broader absorption band centered at 413 nm, which is used as a reference for the ¹O₂ QY calculation. A similar behavior but with faster reaction rates was observed when a buffer solution is used instead of water (see Fig. 4b). Figure 4c depicts the absorption profile of DPBF (400 μL of a 6.0 × 10⁻⁵ M solution) in 2 mL of water in the presence of ZnS:Mn QDs (10 mg/10 mL). This spectrum shows the typical absorption band edges of DPBF centered at 264 and 413 nm, and of ZnS:Mn peaking at 323 nm (Beltran-Huarac et al. 2010, 2013a, b). It was difficult to differentiate the absorption band of DPBF at 317 nm given that the spectral linewidth of the ZnS:Mn absorption band was further increased. A more regular DPBF oxidation with larger changes of absorption was also observed, when compared to RB spectra, suggesting that the photoluminescent QDs are multiplet-state photoactive and can be used as a stable, more efficient PS. A similar behavior but with faster reaction rates was observed when the buffer solution is used instead of water (see Fig. 4d). Note that the dependence of the reaction on dissolved O₂ indicates that ¹O₂ is generated by the QDs during the photosensitization. No oxidation of DPBF was detected in the absence of light.

In order to calculate the ¹O₂ QY (Φ _Δ) of the photoexcited ZnS:Mn QDs in organic and aqueous media, we have expressed the oxidation of DPBF in terms of the decrease of absorbance (A) at 413 nm as a function of time (Δ(ln[A _o /A _t ]/Δ(t)). The profiles of the time-dependent oxidation of DPBF induced by singlet oxygen, obtained by this procedure, are depicted in Fig. 5a, b. It was observed that ZnS:Mn shows high reaction rates and a marked linear correlation, i.e., a first-order kinetic reaction, which is a clear indicator that ZnS:Mn QDs can yield high ¹O₂ values in both buffer and water. To quantify this observation, we employ a widely validated chemical trapping method (Beltran-Huarac et al. 2010; Nardello et al. 1997; Xiao et al. 2011; Nadhman et al. 2014) that consists in the comparative use of a standard PS of a known Φ _Δ (Φ _Δstandard), such as RB, and the slopes determined from the decay curves of a chemical scavenger of ¹O₂, such as DPBF, which irreversibly undergoes a 1,4-cycloaddition that is detected as an intensity drop of its absorption band at 413 nm (Xiao et al. 2011). Thus, the Φ _Δ (number of ¹O₂ molecules formed per absorbed photon) of ZnS:Mn QDs (Φ _ΔZnS:Mn) is calculated, as follows (Nardello et al. 1997):where k _standard and k _ZnS:Mn represent the slopes determined from the decay curves for RB and ZnS:Mn QDs, respectively. By introducing the value of Φ _Δstandard (RB = 0.76) (Xiao et al. 2011) and the determined slopes, one obtains Φ _ΔZnS:Mn = 0.62 ± 0.02 in buffer and Φ _ΔZnS:Mn = 0.54 ± 0.03 in water. These values are higher than those reported for some semiconductors and metals employed as PSs (see Table 1). Thus, based on our observations, we suspect that the ¹O₂ generation by ZnS:Mn QDs in water and buffer solution involves the following (Beltran-Huarac et al. 2013; Beltran-Huarac et al. 2010; Xiao et al. 2011; Nadhman et al. 2014; DeRosa and Crutchley 2002): (i) the promotion of electrons toward the conduction band of ZnS:Mn by light excitation; (ii) nonradiative decay of the photoexcited QDs into a multiplet energy state, ⁴ T ₁ (first excited state with spin 3/2), due to the internal Mn²⁺ ion transition; (iii) intersystem crossing of QDs to a long-lived multiplet state, ⁶ A ₁ (ground state with spin 5/2); (iv) energy transfer from the multiplet state of QDs to the triplet ground state (T _o) of oxygen (³O₂); and (v) subsequent yield of singlet oxygen (¹O₂), which irreversibly reacts with DPBF via a 1,4-cycloaddition yielding an endoperoxide that decomposes in turn to produce o-dibenzoylbenzene. A schematic of these processes is depicted in Fig. 5c. Given that the reaction occurs in solution, we assume that the excited state (¹O₂) eventually relaxes back to the ground state (³O₂) via nonradiative deactivation (Hurst and Schuster 1983). In this proposed mechanism, no response of hydroxyl radical and superoxide species was considered since the DPBF is highly selective for ¹O₂. We believe that the powerful visible light supplied by the laser which is used to avoid the self-oxidation process of DPBF (through absorption) and assisted by the 150 mW continuous ozone-free Xe lamp light (200–800 nm) can induce energy transfer from the particles to the molecular oxygen and DPBF. Taken together, this method provides an efficient pathway to generate ¹O₂, which rapidly reacts with unsaturated carbon–carbon bonds present in DPBF via the energy transfer from the ⁴ T ₁ state to molecular oxygen: ⁴ T ₁ (PS) + ³O₂ → ⁶ A ₁ (PS) + ¹O₂, wherein ZnS:Mn QDs act as an efficient PS.

In the previous discussion, we have accounted for the high ¹O₂ QY of ZnS:Mn QDs in terms of a chemical trapping approach, which in turn involves a high PL QY that promotes the ⁴ T ₁–⁶ A ₁ transition. The PL QY of ZnS:Mn QDs falls in the range of 13.2–70 % at room temperature with a decay time of a few milliseconds (Shen et al. 2012; Beltran-Huarac et al. 2013a, b; Geszke-Moritz et al. 2011, 2013; Bhargava and Gallaguer 1994). In this section, we further exploit additional PL characteristics of ZnS:Mn and evaluate their cytotoxicity in order to employ them as luminescent nano-probes for imaging P. aeruginosa cells. The PL spectrum of ZnS:Mn QDs dispersed in water is depicted in Fig. 6a, which shows two emission bands in the visible region. The blue band at ~416 nm (labeled as defects) is ascribed to ZnS host, corresponding to the self-activated recombination centers arising from shallow donors and zinc vacancies (Beltran-Huarac et al. 2013a, b), consistent with the photooxidation analysis. The orange emission band at ~598 nm is due to the internal Mn²⁺ ion transition, which causes energy transfer from the s-p electron–hole pair band states (ZnS host) to the Mn²⁺ ion d-electron states (Beltran-Huarac et al. 2013a, b). Both bands in the PL spectrum confirm the homogeneous incorporation of Mn²⁺ ions into the semiconductor. The optical images of ZnS QDs (blue) and ZnS:Mn QDs (orange) in aqueous solution when exposed to UV light are shown in the inset of Fig. 6a. Prior to evaluating ZnS:Mn QDs as luminescent nano-probes for bioimaging, we first study their cytotoxic effect on human cervical adenocarcinoma (HeLa) cells. Cell viability was examined through the measurement of cell metabolic activity. The mitochondrial function was measured via MTS viability assay after incubating HeLa cells with different concentrations (5.2–1000 µg/mL) of QDs for 24 h. Viable cells convert MTS tetrazolium into formazan dyes, which can be spectrophotometrically detected (Cory et al. 1991). The cell viability results are depicted in Fig. 6b. It was observed an average of 95 % of viable cells at 5.2 µg/mL of QDs, and 88 % at 300 µg/mL. The cell viability slightly decreased until a value of ~68 % at an excessively high concentration (1000 µg/mL). These findings reflect that ZnS:Mn QDs are not toxic for human cells and cause a negligible cytotoxicity (~32 % of reduction in cell viability) at elevated doses, which are not expected to be administered in clinical trials. The analysis of variance (ANOVA) test indicated that there are no statistically significant differences attributed to QDs’ concentrations (p = 0.1457). To evaluate the ability of ZnS:Mn QDs as luminescent nano-probes, we incubated P. aeruginosa cells with 5 mL of QD solution at 40 mM for 3 h, and monitored both the corresponding suspensions by measuring the optical absorbance and the internalization of the QDs by the cells via confocal microscopy. To do this, we made the QDs positively charged by protonating the amine groups of chitosan, which bind to the phosphoryl and carboxyl groups present in the cellular walls. Figure 6c shows the suspensions of cells (control) and chitosan-capped QDs–cells monitored by absorption at 640 nm. The decrease in the absorbance of the post incubation process indicates that the QDs were able to either attach to the cells or move inside the cells from the suspension. After washing, the supernatant showed a slight absorbance, as expected, whereas the remaining suspension showed an absorbance of ~55 %, indicating that QDs were able to enter the cells. The internalization of the 4 nm ZnS:Mn QDs by the cells was confirmed by confocal microscopy, as shown in Fig. 6d, e. It was clearly observed that the QD-labeled cells emit a stronger fluorescent signal (reflected as high-contrast images) when compared to the control (unlabeled cells), revealing the dimension and rod-like morphology characteristic of P. aeruginosa strains, which assume a wide range of morphologies and lengths dependent on the growth parameters. The well-defined geometry of the cells suggests that the fluorescence signature was homogeneously distributed throughout the bacterial cell during incubation. No photobleaching effect was observed after a few hours of continuous irradiation exposure, evidencing that the QDs are an ideal alternative to conventional organic dyes that are highly susceptible to light-induced oxidation. Even though the amine groups present in the QDs serve as effective sites for bioconjugation, further studies are needed to fully understand the nature of the molecular associations of the chitosan-capped ZnS:Mn QDs with the P. aeruginosa cells.