Intracellular Zn²⁺ detection with quantum dot-based FLIM nanosensors

Abstract

Fluorescence Lifetime Imaging Microscopy (FLIM) has been employed for the detection of intracellular Zn²⁺ levels, implicated in various signalling pathways, using a family of quantum dot (QD) nanosensors. The sensing mechanism was based on photoinduced electron transfer (PET) between an azacycle receptor group and the QD nanoparticles.

---

The search for new nanosensors for specific species with minimal local perturbation is of special interest in fields such as chemistry, biology or biomedicine. In particular, the development of appropriate intracellular ion sensors is singularly important because some ions play significant roles in cellular biology.¹ Multimodal fluorescence imaging, based on intensity, lifetime and polarization detection, combined with high spatial resolution enables quantitative readouts that allow for real time monitoring of biological processes through biosensing and bioimaging. One of the most interesting materials for these sensing applications are semiconductor nanocrystals or quantum dots (QDs).² QDs have been proposed as an alternative to conventional molecular probes because of their unique optical properties, which are ideally suited for the long-term monitoring of intracellular processes.³ Among these features, QDs show long photoluminescence (PL) decay times, typically from five to hundreds of nanoseconds, which make them especially interesting for fluorescence lifetime imaging microscopy (FLIM).⁴ Most of the published intracellular QD nanosensors are based on changes in the fluorescence intensity when a target ion or molecule interacts with a nanoparticle.^5–7 However, fluorescence intensity-based measurements suffer from uncertainties because they can be altered by fluctuations in the excitation light, the probe concentration and heterogeneities in the optical properties of the medium. As an alternative, the use of time-resolved fluorescence spectroscopy and FLIM imaging can overcome many of these limitations.^8,9 The average PL lifetime of QDs is significantly longer than the lifetime of the cell autofluorescence, which makes QDs easily discernible from the background signal. Although some QD-based luminescence lifetime sensors have been reported for the determination of pH¹⁰ and detection of ions, such as Cl⁻ or Cu²⁺,^11,12 the advantages of the FLIM technique and its use with QDs as intracellular probes have not yet been extensively exploited.⁸ A few examples where QD nanoparticles were employed with FLIM have been described but only for intracellular detection and not for the quantification of the target molecules.^13,14 To date, FLIM has been used as an effective technique only for the quantitative real-time sensing of intracellular pH.⁴ Long-decay, near-infrared, emitting QDs have also been applied for in vivo pH sensing through FLIM imaging into a nude mouse.¹⁵ These works demonstrated the high sensitivity of these QD-based nanosensors and the great potential of FLIM for intracellular applications.

In this communication, we have shown that this methodology can be extended to the sensing of other ions inside cells. We have focused our attention on Zn²⁺ homeostasis as a crucial intracellular catalytic and signalling mechanism that controls many important cellular events.¹⁶ Deregulation of Zn²⁺ levels can cause several health problems, such as neuronal dysfunction and immunodeficiency. Relevant concentrations of Zn²⁺ range from fM to almost 0.5 mM in some mammalian brain cells. Hence, many efforts are currently in place to develop Zn²⁺ sensors with a wide quantitation range.¹⁷ In this communication, the surface of CdSe/ZnS QDs was properly modified to create a family of intracellular-sensitive Zn²⁺ probes. Two azamacrocycles molecules, cyclen (1,4,7,10-tetraazacyclododecane, 1) and cyclam (1,4,8,11-tetraazacyclotetradecane, 2), have been attached by EDC/NHS covalent coupling to the surface of mercaptopropionic acid (MPA) capped-QDs (Fig. 1a, and ESI† for more details), for the preparation of QD–1 and QD–2 conjugates. We have employed QDs of two different sizes that emit at 520 (QD₅₂₀) and 600 nm (QD₆₀₀). Upon attachment, a proximal nitrogen lone pair acts as an electron donor to the excited QD, producing a quenching of the PL emission via photoinduced electron transfer (PET).^18–20 These azamacrocycle groups act as receptor units for the Zn²⁺ ion, and after coordination, a dramatic increase in the photoluminescence emission occurs as PET quenching is deactivated. However, in addition to the changes in the steady-state emission, the PL decay traces of QD have also been modified after assembly. In agreement with previous reports,^9,21 QD nanoparticles showed long decay times and multi-exponential behaviour. The best fits of the PL decay traces of QD–MPA required a sum of three exponential functions (Tables S1 and S2 for QD₅₂₀–MPA and QD₆₀₀–MPA, respectively, ESI†) to reach low χ² values, as well as random distributions of the weighted residuals and auto-correlation function, which are indicators of the goodness of the fits. Prior to the immobilization of the azacycle, the intensity-weighted average PL lifetimes, τ_ave (see Methods of analysis in the ESI†), of QD₅₂₀–MPA and QD₆₀₀–MPA were 15.8 ± 0.1 and 18.4 ± 0.1 ns, respectively. The three individual decay time components, as well as the average PL lifetime of both the QDs, gradually decreased by increasing the concentration of the azacycle attached on the surface as a result of the PET mechanism (Tables S1 and S2, ESI†). After attachment of the optimized amount of azacycle 1 or 2 on the surface of QD₅₂₀–MPA to obtain the proposed conjugates (see Experimental section in ESI†), the calculated τ_ave decreased to 8.6 ± 0.1 ns (QD₅₂₀–1) or 12.0 ± 0.1 (QD₅₂₀–2) (Fig. S1, ESI†). For QD₆₀₀–MPA, the corresponding QD–azacycle showed a decrease in τ_ave down to 9.2 ± 0.1 ns or 13.2 ± 0.2 ns for QD₆₀₀–1 or QD₆₀₀–2, respectively (Fig. S1, ESI†).

Fig. 1 (a) Scheme of covalent coupling reaction for the formation of the QD–azacycle conjugates. (b) PL decay traces of QD₆₀₀–1 conjugates in the presence of 0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, and 2 mM of Zn²⁺ (λ_ex = 440 nm and λ_em = 598 nm, Tris buffer pH 7.2). Average lifetime calculated from the decays for (c) QD₆₀₀–1 and (d) QD₅₂₀–2 conjugates.

Taking into account that the azacycles can coordinate some metals, the proposed QD–azacycle conjugates show a response in the presence of metals with d¹⁰ electronic configuration, such as zinc.¹⁸ Indeed, an enhancement of the emission intensity²⁰ and the average lifetime of QDs were detected upon the addition of Zn²⁺ as a result of the interruption of the PET mechanism between the QD and azacycle (Fig. 1b). The three individual decay time components of the QD–azacycle conjugates were gradually augmented when the concentration of Zn²⁺ of the medium increased (Tables S1 and S2, ESI†). The τ_ave of the QD–azacycle conjugates showed a dependency on the concentration of Zn²⁺ with a linear response versus the logarithm of the concentration in a range covering approximately three orders of magnitude, between 2 μM and 1 mM for QD₆₀₀–1 and between 10 μM and 1 mM for QD₅₂₀–2 conjugates (Fig. 1c and d). These ranges suggest the potential application of zinc detection in some Zn²⁺-rich intracellular media, such as brain tissue.²² The range of values (10–16 ns) and variation of the PL average lifetime make these nanosensors much more sensitive than previously published fluorescence lifetime probes. For instance, changes of 0.4 ns are usually reported for fluorescence protein-²³ or organic fluorophores-based FLIM sensors.²⁴ In some occasions, FRET-based FLIM sensors have been reported to work with a total change in the sensor's lifetime as low as 0.06 ns.²⁵ Our family of QD–azacycle conjugates exhibit a sensitivity that is between 10 and 100 times those figures.

The response of the QD–azacycle conjugates toward Zn²⁺ was in the μM concentration range, with a higher sensitivity of the QD–1 conjugates than that of the QD–2 conjugates. This is in agreement with the fact that the binding constant of Zn²⁺ with cyclen 1 is higher than the corresponding binding constant for cyclam 2.²⁶ Nevertheless, the response towards Zn²⁺ shown by the QDs of different wavelengths coupled to the same azacycle was very similar, suggesting that the PET mechanism does not depend on the QD size and allowing for the use of different QD conjugates depending on the spectral needs. When compared to other Zn²⁺ sensors, the response of our QD–azacycle conjugates are comparable to other sensors in which the dissociation constant, K_d, of the sensor with the metal ion is in the μM range. The commercially available FluoZin-1 and Newport Green exhibit K_d values in this range,¹⁷ and hence, response to μM concentrations of Zn²⁺. These commercial sensors are fluorescent indicators with a single Zn²⁺ binding site (1 : 1). They can be responsive to a Zn²⁺ variation along two orders of magnitude (around pK_d ± 1). Hence, increasing the linear range for Zn²⁺ quantitation can only be performed with multiple binding sites. In our case, as each QD contains several azacycle units, the quantitation range reaches almost three orders of magnitude. Newly designed PET sensors have decreased the interaction K_d of Zn²⁺ into the nanomolar range, allowing for an enhanced sensitivity and quantitation of sub-micromolar Zn²⁺ levels.^27,28 Commercially available FluoZin-3 exhibits a K_d value around 15 nM.¹⁷ However, in such cases, saturation is reached at lower Zn²⁺ concentrations, leading to unsuitable sensors for high Zn²⁺ levels. This situation establishes the need for a careful selection of the suitable fluorescent sensor, depending on the aimed concentration levels of analyte.

The selectivity of the Zn²⁺ nanosensors was also considered in the presence of other metals and molecules that can be found in biological samples. QD–1 and QD–2 conjugates were exposed to several potential interfering agents at different concentrations, and the PL decay traces were collected. Fig. S2 (ESI†) shows a summary of the potential interfering species that were tested with no effect in the average PL lifetime of QDs. The τ_ave of the proposed nanosensors showed negligible response toward the major intra- and extracellular cations (Na⁺, K⁺, Ca²⁺ and Mg²⁺). Metals such as Fe²⁺, Fe³⁺, Co²⁺ or Cu²⁺ produce some quenching of the photoluminescence of QD–MPA due to some adsorption and inner filter effects;²⁹ thus, this effect was also detected with the QD–azacycle conjugates. Fig. S2 (ESI†) show the maximum concentration of such interferents that has no effect on the nanosensor τ_ave. All these interferents are commonly found in other Zn²⁺ sensors, such as commercial FluoZin-3 or Newport Green.³⁰ Other transition metals, such as Ni²⁺ or Mn²⁺, could also coordinate into the azacycle,³¹ slightly affecting the τ_ave of the QDs. However, these free metal cations are not present, to a large extent, in the majority of biological systems; thus, not representing a major issue for the use of these nanosensors in physiological samples. Moreover, the effect of changes in the pH value of the medium in the response of the Zn²⁺ nanosensors was also evaluated. The pH value may have an effect because it alters the coordination efficiency of the azacycles toward Zn²⁺, and because of the presence of remaining free carboxylic groups on the surface from nonreacted MPA. Nevertheless, at the different pH tested, the τ_ave of the QD–azacycle conjugates after reaction with 0.1 mM Zn²⁺ increased a similar amount (55% at pH 5.47, 65% at pH 6.49, and 47% at pH 7.20, Fig. S3, ESI†), indicating that the Zn²⁺ response was not altered. Therefore, the pH of the medium must be known and controlled while determining Zn²⁺ concentrations with the proposed conjugates.

We tested the usefulness of the QD–azacycle nanosensors for the detection of Zn²⁺ concentrations by employing FLIM imaging, especially in intracellular applications. First, QD₆₀₀–1 nanosensors dissolved in 10 mM TRIS buffer at pH 7.2 with different Zn²⁺ concentrations and deposited on a glass slide were imaged using a confocal FLIM microscope. The increase in the τ_ave with increasing concentrations of Zn²⁺ was visible in the arbitrary colour scale image, as well as in the lifetime distributions obtained upon analysis of the images (Fig. 2). These results suggested that the proposed nanosensors could be promising as FLIM-based intracellular Zn²⁺ probes.

Fig. 2 FLIM images and the corresponding lifetime distributions of QD₆₀₀–1 conjugates suspended in 10 mM Tris buffer pH 7.2 deposited on glass slides in the presence of different concentrations of Zn²⁺: (i) 0 mM Zn²⁺; (ii) 0.1 mM Zn²⁺; and (iii) 1 mM Zn²⁺. White scale bars represent 5 μm.

The QD–1 nanosensors were introduced into live HepG2 cells for intracellular Zn²⁺ sensing. After the incubation, the cells internalized the nanoparticles mainly by cellular endocytosis (Fig. S4, ESI†). After the 2 hour of incubation time, the QDs exhibited negligible toxicity on the cells (see ESI† and Fig. S5 for survival rates). Fig. 3 shows FLIM images of QD₆₀₀–1 in the absence and presence of 1 mM Zn²⁺. The images show an enhancement of the average PL lifetime of QDs in cells treated with Zn²⁺. The changes in the PL lifetime distributions of the regions of interest indicate the excellent response of the proposed QD lifetime-based nanosensors inside the cells. It is important to note that the long-lived PL of the QD nanosensors can be easily distinguished from the interfering cellular autofluorescence. In the absence of QDs, the cells displayed minimal emission with lifetimes of approximately 2.6 ns (Fig. 3). In contrast, the τ_ave associated with the QDs was significantly longer than that of cell autofluorescence, even in the absence of Zn²⁺. The commercial probe, Newport Green, has been previously used to measure intracellular Zn²⁺ levels using phase-modulation FLIM,³² with the fluorescence lifetime increasing from 0.88 to 2.93 ns; values which lie in the same range than cellular autofluorescence, presumably causing interferences. The use of the long lifetime of the QD sensors is a much more powerful approach, as the cell autofluorescence is discarded by applying suitable time windows.

Fig. 3 FLIM images and the corresponding lifetime distributions of HepG2 cells in PBS pH 8.0: (i) autofluorescence; (ii) cells with QD₆₀₀–1 conjugates; (iii) cells with QD₆₀₀–1 conjugates incubated with 1 mM Zn²⁺. White scale bars represent 5 μm.

In conclusion, we have demonstrated the advantages of the FLIM methodology, particularly in combination with QD nanoparticles, whose very long PL lifetime greatly enhances the sensitivity and selectivity of the nanosensors. These long decay times facilitate the discrimination between the signal from the sensor and the intrinsic fluorescence of the cells and produce an enhanced signal-to-background ratio. Moreover, the FLIM technique also eliminates the need for a near-infrared probe because it can use fluorophores with emission within the range of the green cellular autofluorescence. However, all of these excellent advantages have not yet been exploited because recently, only a couple of studies using the combination of QDs and FLIM microscopy have been published.^4,15 Nevertheless, those works are aimed at pH sensing. To the best of our knowledge, this is the first time that QD-based FLIM sensors are specifically designed for intracellular detection of an important metal ion. This methodology can be extended to the development of other nanosensors for the detection of a wide range of molecules of interest. Other photophysical processes, such as charge transfer or energy transfer (FRET), that may take place at the QD surface with other ligands have an effect on the PL decay time.⁹ Any of these processes can be optimized for developing specific FLIM nanosensors. Nevertheless, for intracellular application, special attention to excessive aggregation or cell intake must be taken into account.

This work was supported by Fundación Ramon Areces and grant CTQ2014-56370-R from Ministerio de Economia y Competitividad of Spain. We also thank Dr M. D. Giron and Dr R. Salto for kindly supplying the cell cultures.

Notes and references

1. M. J. Ruedas-Rama, J. D. Walters, A. Orte and E. A. H. Hall, Anal. Chim. Acta, 2012, 751, 1–23.
2. A. P. Alivisatos, W. Gu and C. Larabell, Annu. Rev. Biomed. Eng., 2005, 7, 55–76.
3. U. Resch-Genger, M. Grabolle, S. Cavaliere-Jaricot, R. Nitschke and T. Nann, Nat. Methods, 2008, 5, 763–775.
4. A. Orte, J. M. Alvarez-Pez and M. J. Ruedas-Rama, ACS Nano, 2013, 7, 6387–6395.
5. R. Freeman, R. Gill, I. Shweky, M. Kotler, U. Banin and I. Willner, Angew. Chem., Int. Ed., 2009, 48, 309–313.
6. Y. Wang, H. Mao and L. B. Wong, Talanta, 2011, 85, 694–700.
7. Y.-S. Liu, Y. Sun, P. T. Vernier, C.-H. Liang, S. Y. C. Chong and M. A. Gundersen, J. Phys. Chem. C, 2007, 111, 2872–2878.
8. M. Ruedas-Rama, J. Alvarez-Pez, L. Crovetto, J. Paredes and A. Orte, in Advanced Photon Counting, ed. P. Kapusta, M. Wahl and R. Erdmann, Springer International Publishing, 2015, pp. 191–223.
9. M. J. Ruedas-Rama, A. Orte, E. A. H. Hall, J. M. Alvarez-Pez and E. M. Talavera, ChemPhysChem, 2011, 12, 919–929.
10. M. J. Ruedas-Rama, A. Orte, E. A. H. Hall, J. M. Alvarez-Pez and E. M. Talavera, Chem. Commun., 2011, 47, 2898–2900.
11. M. J. Ruedas-Rama, A. Orte, E. A. H. Hall, J. M. Alvarez-Pez and E. M. Talavera, Analyst, 2012, 137, 1500–1508.
12. J. U. Sutter, D. J. S. Birch and O. J. Rolinski, Meas. Sci. Technol., 2012, 23, 055103.
13. R. K. Pai and M. Cotlet, J. Phys. Chem. C, 2011, 115, 1674–1681.
14. L. Carlini and J. L. Nadeau, Chem. Commun., 2013, 49, 1714–1716.
15. C. Chen, P. Zhang, L. Zhang, D. Gao, G. Gao, Y. Yang, W. Li, P. Gong and L. Cai, Chem. Commun., 2015, 51, 11162–11165.
16. T. Fukada, S. Yamasaki, K. Nishida, M. Murakami and T. Hirano, J. Biol. Inorg. Chem., 2011, 16, 1123–1134.
17. L. Zhu, Z. Yuan, J. T. Simmons and K. Sreenath, RSC Adv., 2014, 4, 20398–20440.
18. E. G. Moore, P. V. Bernhardt, A. Fürstenberg, M. J. Riley, T. A. Smith and E. Vauthey, J. Phys. Chem. A, 2005, 109, 3788–3796.
19. P. V. Bernhardt, E. G. Moore and M. J. Riley, Inorg. Chem., 2002, 41, 3025–3031.
20. M. J. Ruedas-Rama and E. A. H. Hall, Anal. Chem., 2008, 80, 8260–8268.
21. M. Dahan, T. Laurence, F. Pinaud, D. S. Chemla, A. P. Alivisatos, M. Sauer and S. Weiss, Opt. Lett., 2001, 26, 825–827.
22. C. J. Frederickson, J.-Y. Koh and A. I. Bush, Nat. Rev. Neurosci., 2005, 6, 449–462.
23. M. Tantama, Y. P. Hung and G. Yellen, J. Am. Chem. Soc., 2011, 133, 10034–10037.
24. J. M. Paredes, M. D. Giron, M. J. Ruedas-Rama, A. Orte, L. Crovetto, E. M. Talavera, R. Salto and J. M. Alvarez-Pez, J. Phys. Chem. B, 2013, 117, 8143–8149.
25. C. D. Harvey, A. G. Ehrhardt, C. Cellurale, H. Zhong, R. Yasuda, R. J. Davis and K. Svoboda, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 19264–19269.
26. M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., 1977, 2269–2276.
27. J. Fan, X. Peng, Y. Wu, E. Lu, J. Hou, H. Zhang, R. Zhang and X. Fu, J. Lumin., 2005, 114, 125–130.
28. M. Taki, J. L. Wolford and T. V. O'Halloran, J. Am. Chem. Soc., 2004, 126, 712–713.
29. Y. Chen and Z. Rosenzweig, Anal. Chem., 2002, 74, 5132–5138.
30. J. Zhao, B. A. Bertoglio, M. J. Devinney Jr, K. E. Dineley and A. R. Kay, Anal. Biochem., 2009, 384, 34–41.
31. S. Mossin, H. O. Sørensen, H. Weihe, J. Glerup and I. Søtofte, Inorg. Chim. Acta, 2005, 358, 1096–1106.
32. R. B. Thompson, D. Peterson, W. Mahoney, M. Cramer, B. P. Maliwal, S. W. Suh, C. Frederickson, C. Fierke and P. Herman, J. Neurosci. Methods, 2002, 118, 63–75.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, methods of analysis, Tables S1 and S2, and Fig. S1–S5. See DOI: 10.1039/c5cc06676j

---