Ion-Selective Optical Nanosensors Based on ... - ACS Publications

Article pubs.acs.org/acssensors

Ion-Selective Optical Nanosensors Based on Solvatochromic Dyes of Different Lipophilicity: From Bulk Partitioning to Interfacial Accumulation Xiaojiang Xie,* Istvan Szilagyi, Jingying Zhai, Lu Wang, and Eric Bakker* Department of Inorganic and Analytical Chemistry, University of Geneva, CH1211 Geneva, Switzerland S Supporting Information *

ABSTRACT: The sensing mechanism of fluorescent ionselective nanosensors incorporating solvatochromic dyes (SDs), with K+ as model ion, is shown to change as a function of dye lipophilicity. Water-soluble SDs obey bulk partitioning principles where the sensor response directly depends on the lipophilicity of the SD and exhibits an influence on the phase volume ratio of nanosensors to aqueous solution (dilution effect). A lipophilization of the SDs is shown to overcome these limitations. An interfacial accumulation mechanism is proposed and confirmed with Fö rster resonance energy transfer (FRET) with a ratiometric near-infrared fluorescence FRET pair. This work lays the foundation for operationally more robust ion-selective nanosensors incorporating SDs. KEYWORDS: nanosensor, solvatochromic, ionophore, mechanism, ion-selective

A

dye leakage. In this work, we provide a mechanistic insight into ion-selective optical nanosensors based on SDs. Lipophilized SDs were found to lead to a distinct sensing mechanism and overcome the dye-leakage disadvantage. The proposed theoretical framework was confirmed by a Förster resonance energy transfer (FRET) pair, which also results in a ratiometric fluorescence readout in the near-infrared (NIR) region.

long with the expanding demands in environmental, biological, clinical, and food chemistry for chemical analysis, the development of ion-selective optical sensors has received enormous attention. A number of film-based ionselective sensors and miniaturized micro/nanosensors have been reported taking advantages of the highly selective ionophores.1−5 Nanosensors require minimized sample amount compared with sensing films. Compared with synthetic probes for bioimaging, ion-selective nanosensors can be brighter, less cytotoxic and do not suffer from nonspecific binding and reactions with biomacromolecules or unwanted sequestration.6,7 The potential of solvatochromic dyes (SDs) as signal transducers in ionophore-based ion-selective optical sensors has been assessed by some researchers.8−12 The fluorescence (or absorbance) of SDs is sensitive to solvent polarity.13,14 The introduction of SDs into ionophore-based ion-selective sensors have allowed one to put forward novel pH-insensitive optical ion sensors that overcome the well-known pH cross-response of conventional chromoionophore-based optical ion sensors.4 Promising hydrogel sensing films have been reported for ions including K+, NO3−, and Cl−.10−12 However, sensor development in this direction was limited, partly due to an unestablished theoretical framework and insufficient detection limits, response times and reversibility. Recently, the use of SDs in ionophore-based ion sensors was revisited at the nanoscale with encouraging improvements and promising in vitro and in vivo applications.15,16 Meanwhile, a thorough understanding of the sensing mechanism has not yet been achieved and the sensors still suffer from the drawback of © 2016 American Chemical Society

■

EXPERIMENTAL SECTION

Reagents. Solvatochromic dyes from SD 1 to SD 4 were synthesized in the lab (see the Supporting Information (SI) for synthetic routes, procedures, and structural confirmation). Acetic anhydride, 4-(dimethylamino)cinnamaldehyde, valinomycin, Pluronic F-127 (F127), bis(2-ethylhexyl) sebacate (DOS), tetrahydrofuran (THF), ethanol, methanol, sodium tetrakis-[3,5-bis(trifluoromethyl)phenyl]borate (NaTFPB), 4-dimethylaminopyridine (DMAP), dichloromethane triethylamine, 2-methylbenzothiazole, acetonitrile, stearoyl chloride, 4-[N,N-bis(2-hydroxyethyl)amino]benzaldehyde, 3ethyl-2-methylbenzothiazolium iodide, N-methyl-N-(2-hydroxyethyl)4-aminobenzaldehyde, 1-iodooctadecane, 4-(dimethylamino)cinnamaldehyde, and cation exchange resin (Dowex MAC-3 hydrogen form) were obtained from Sigma-Aldrich. Lumogen Red was obtained from BASF. All solutions were prepared by dissolving appropriate salts into deionized water (Mili-Q). All salts used were analytical grade or higher. Instrumentation and Measurements. Fluorescence responses of the nanosensors were measured with a fluorescence spectrometer Received: January 5, 2016 Accepted: March 3, 2016 Published: March 3, 2016 516

DOI: 10.1021/acssensors.6b00006 ACS Sens. 2016, 1, 516−520

Article

ACS Sensors Scheme 1. Schematic Illustration for the Sensing Mechanisms of Ionophore-Based Ion-Selective Optical Nanosensors Containing Water-Soluble SDs (a) and Lipophilized SDs (b)a

a

For (a), SD 1 is expelled into and distribute evenly in the aqueous phase as K+ being extracted by valinomycin. For (b), only the charged chromophore is pushed into the aqueous phase, resulting in an accumulation of SDs on the surface of the nanospheres. The nanospheres (gray area) are composed of bis(2-ethylhexyl) sebacate and Pluronic F-127 (structures omitted for simplicity).

■

(Fluorolog3, Horiba Jobin Yvon) using disposable poly(methyl methacrylate) cuvettes or quartz cuvettes (for organic solvent) with path length of 1 cm as sample container. The absorbance was measured with a UV−vis spectrometer (SPECORD 250 plus, Analytic Jena, AG, Germany). For the calibration of the nanosensors, stock solutions were added stepwise to reach the desired concentration, causing negligible volume variation. Dynamic light scattering (DLS) measurements were performed at a scattering angle of 173° with the Zetasizer Nano ZS (Malvern Instrument). This instrument uses a He/ Ne laser operating at 633 nm as a light source and an avalanche photodiode as a detector. The cumulant method was used to fit the correlation function and subsequently determine the hydrodynamic diameter, which could be typically measured with an error of 1 nm. Standard plastic cuvettes were used for all light scattering experiments. Relative lipophilicity of the SDs were modeled in ChemBioDraw Ultra 12.0. The concentration of the nanosensors was calculated based on the amount of anionic sites (TFPB) with respect to the total volume. To compare the dye leakage, nanosensors incorporating SD 1 and 2 were dialyzed with cellulose tubing (MWCO = 14 000 Da) in 0.1 M KCl solution. After 1 day, the color of the SD 1 based nanosensors became very weak while there was no apparent change for nanosensors incorporating SD 2. In addition, nanosensors were mixed with cation exchange resin and the absorbance spectra of the nanosensor suspensions were measured after storing in the dark overnight. The influence of pH on the nanosensors was evaluated by calibrating the K+ responses in phosphate buffer at pH 5.0, 7.0, and 9.0, respectively. Nanosensor Preparation. The K+-selective nanosensors were prepared by following previously reported solvent displacement steps.17 Typically, 0.6 mg of NaTFPB, 0.2 mg of SD, 8.0 mg of DOS, 5 mg of F127, and 1.2 mg of valinomycin were dissolved in 3.0 mL of methanol to form a homogeneous solution. Then 0.4 mL of this solution was pipetted and injected into 4.0 mL of deionized water on a vortex with a spinning speed of 1000 rpm. The resulting clear mixture containing self-assembled nanospheres (nanoemulsions) was blown with compressed air on the surface for at least 30 min to remove methanol, giving a clear particle suspension. The average hydrodynamic radius of the nanospheres was determined to be ca. 43 nm by DLS. For FRET based nanoensor preparation, the cocktail contained 0.28 mg of Lumogen Red, 0.13 mg of SD 2 (or SD 1), 0.84 mg of NaTFPB, 1.48 mg of valinomycin, 7 mg of DOS, and 5 mg of F127.

RESULTS AND DISCUSSION We first consider nanosensors containing SD 1 that can readily dissolve in water (Scheme 1a). The solvatochromic effect of the dyes containing the same core structures have been characterized in our previous work.9 In general, the fluorescence emission intensity becomes much higher when the dyes are dissolved in nonpolar solvents than in water. Here, the nanosensors contains valinomycin as K+ ionophore and tetrakis-[3,5-bis(trifluoromethyl)phenyl]borate as anionic sites (ion exchanger). Ion transfer equilibrium between the organic nanospheric phase and the aqueous phase controls the partition of SD 1 (see Appendix 1 in the SI). An increase in K+ concentration in the aqueous phase results in more SD 1 entering the aqueous phase and thus in a decrease in fluorescence intensity. From this model, the response range was found to be dependent on nanosensor concentration (volume of the aqueous phase over the volume of the organic nanospheres). As shown in Figure 1, this dilution effect is indeed experimentally observed. The detection limit shifted by one logarithmic unit upon 10-fold nanosensor dilution. Like a double-edged sword, the dilution effect can be used to tune the detection range but also makes calibration impossible for unknown nanosensor concentrations. Moreover, the water

Figure 1. Fluorescence response to various K+ concentrations for the K+ selective nanosensors containing SD 1 at two different nanosensor concentrations (10 and 1 μM). I0 is the maximum emission intensity in the absence of K+. 517


Article

ACS Sensors solubility of the SDs makes dye leakage problematic for in vivo studies.15 Dye leakage occurs by the ion exchange between a lipophilic cation in the aqueous solution and the SD cation in the nanospheres. Increasing the lipophilicity of the SDs appears to be the obvious solution to overcome undesired loss of SDs.18 However, according to the above-mentioned model, it will become much more difficult to displace, by ion-exchange, highly lipophilic SDs with ions such as K+, and thus the nanosensors may show no response to the desired ion concentration changes. To test this assumption, SDs with enhanced lipophilicity were synthesized (SD 2, 3, and 4 in Scheme 1; see the SI for synthesis). Compared with SD 1 and 2, the core structure of SD 3 and 4 contained one less carbon−carbon double bond, which resulted in a blue-shift in the absorption and emission spectra. Nevertheless, these compounds remained solvatochromic. Figure 2a shows a comparison in K+ responses between

Figure 3. Fluorescence responses to various K+ concentrations for the K+ selective nanosensors containing SD 3 and 4. The dashed line is the theoretical response modeled from Appendix 1 in the SI. I0 is the maximum emission intensity in the absence of K+.

existence of a different sensing mechanism for the lipophilized SDs. Upon close inspection, SD 2, 3, and 4 all contain a hydrophobic tail and a relatively hydrophilic head containing the chromophore. A possible mechanism is shown in Scheme 1b. As K+ is extracted into the nanospheres, these SDs, unable to entirely enter the aqueous phase owing to the overall high lipophilicity, are forced to transfer just the hydrophilic head (the hemicyanine chromophore) into the aqueous environment so that the overall charge balance in the nanospheres is maintained. This process should result in an accumulation of the SDs at the nanosphere−water interface, with the chromophores in the more polar aqueous environment, causing a signal change in absorbance (Figure 2b) and fluorescence mode (Figure 2c). According to this model (see Appendix 2 in the SI), where the SDs are confined at the nanosphere surface, it is expected that nanosphere concentration no longer matters because the sensor response no longer relies on the volume ratio of aqueous and nanosphere phase. Indeed, as shown in Figure S1, the K+ response for nanosensors containing SD 2 remains the same even after the nanosensors were diluted 200 times. Dialysis of the nanosensors against 0.1 M KCl showed no leakage for SD 2 but a severe leakage for the water-soluble SD 1 (Figure S2). To mimic an environment where the sequestration of the positively charged solvatochromic dyes could occur, the nanosensors containing various SDs were exposed to a cation exchange resin. Since the resin is lipophilic, positively charged SDs tend to adsorb on the resin surface in time. As shown in Figure S3, the nanosensors containing hydrophilic SD 1 showed a dramatic decrease in absorption spectra, indicating severe dye leakage. However, nanosensors containing the lipophilized SDs did not suffer from this problem. Lipophilic SDs bearing hydrophilic chromophores were also used in previously reported hydrogel sensing films.11 A similar interfacial process was proposed as well but not experimentally confirmed. Here, the interfacial mechanism is confirmed for the first time by Förster resonance energy transfer (FRET). FRET has been widely used in various areas to study distance related phenomena since the FRET efficiency is known to be extremely sensitive to the distance between the FRET donor and acceptor.19−21 The FRET process is illustrated in Scheme 2. An inert reference dye acting as the FRET donnor, Lumogen Red, was incorporated together with SD 2 (FRET acceptor) into the K+ selective nanospheres. As shown in Figure 4a, the emission spectrum of Lumogen Red in the nanospheres and the

Figure 2. (a) Comparison between K+ responses for nanosensors containing SD 1 and 2 with different lipophilicity (log P). I0 is the maximum emission intensity in the absence of K+. Absorption (b) and fluorescence emission (c) spectra of the K+ selective nanosensors containing SD 2 measured at various K+ concentrations indicated on the x-axis of (a).

nanosensors incorporating SD 1 and 2. According to the model developed for water-soluble SDs, the distance along the x-axis between the two curves should be proportional to the lipophilicity difference between SD 1 and 2. While the relative partition coefficient of the SDs between octanol and water (log P) can be easily calculated, the coefficient for nanospheres and water (log K) is only approximately inferred using the analogy of a water−bis(2-ethylhexyl) sebacate plasticized PVC system (see eq 10 in the SI). The response window predicted from the model for nanosensors incorporating SD 2 was found to be different from the experimental results (by at least 0.7 logarithimic units), indicating a different sensing mechanism. Moreover, convincingly, as shown in Figure 3, nanosensors containing SD 3 and 4 (log P = 10.2 and 18.8, respectively) exhibited nearly the same K+ response. From the theoretical model for water-soluble SDs (see Appendix 1 in the SI), they should respond only at much higher concentrations (dashed line in Figure 3). These observations further confirm the 518


Article

ACS Sensors Scheme 2. Schematic Illustration of the FRET Process in K+ Selective Nanospheres Containing SD 2 as FRET Acceptor and Lumogen Red as FRET Donora

a

As SD 2 accumulates at the interfacial region, the average distance between donor and acceptor increases, resulting in a dramatic decrease in FRET efficiency.

Figure 5. Fluorescence emission spectra at various K+ concentrations for the K+ selective nanosensors containing SD 1 (a) and SD 2 (b). Intensities are normalized according the maximum emission intensities from the SDs. Arrows indicate the change in fluorescence with increasing K+ concentration. (c) Calibration curves for (a) and (b) using the ratio of the emission intensity at 686 nm over 600 nm.

Additionally, the FRET pairs represent a new family of ratiometric signal reporters in the NIR region, which is advantageous over conventional fluorescent K+ indicators with respect to bioapplications.22−24 The nanosensors exhibited excellent selectivity for K+ (Figure 6a) and very fast response Figure 4. (a) Overlay of the absorbance of nanospheres containing SD 2 at 0 M K+ concentration (blue solid trace), at 1 M K+ concentration (pink solid trace) and the normalized fluorescence emission spectrum of Lumogen Red (black dashed trace). (b) Fluorescence emission spectra for the nanospheres containing the donor Lumogen Red only (orange trace) and containing Lumogen Red and SD 2 at 0 M K+ (black trace).

absorption spectra of SD 2 in the absence of K+ exhibited a large spectral overlap integral (J) of ca. 3.2 × 1017, J was reduced to 1.6 × 1017 in the presence of 1 M of K+. The Förster distance (R0) was calculated as 12.5 and 11.5 nm. The average distance between adjacent donor and acceptor (RDA) in the absence of K+ was estimated to be ca. 0.2 nm (see the SI). Comparing the R0 values with RDA gives almost 100% FRET efficiency, which accounts for the extremely weak emission from the donor Lumogen Red observed in the absence of K+ in the aqueous sample (Figure 4b). If the SDs still stay in the interior of the nanospheres in the presence of 1 M of K+, the FRET efficiency should still be ca. 100% because R0 (11.5 nm) is so much larger than RDA. However, a gradual recovery in the emission peak of Lumogen Red was observed with increasing K+ concentration (Figure 5b). This means that a relatively large separation between Lumogen Red and SD 2 occurred (RDA = ca. 8.2 nm). While Lumogen Red is electrically neutral, this change evidences the accumulation of SD 2 in the interfacial region. It is further confirmed with a similar FRET system using SD 1. Since SD 1 is water-soluble, the separation between Lumogen Red and SD 1 is expected to be much more significant. Indeed, as shown in Figure 5, SD 1-based nanosensors exhibited much higher recovered emission intensity from the donor Lumogen Red.

Figure 6. (a) Fluorescence response of the K+ selective nanosensors containing SD 2 and Lumogen Red to various ions (Mn+) as indicated. (b) Response time (from 0 to 0.01 M of K+) of the K+ selective nanosensors containing SD 2 and Lumogen Red determined in a vigorously stirred nanosphere suspension using an optical fiber to follow the emission intensity at 686 nm.

(t95% = ca. 400 ms, Figure 6b), suggesting promising applications in complex sample media and for real time monitoring. As shown in Figure S4, changing the sample pH from 5.0 to 7.0 and to 9.0 had little influence on the sensor response to K+, confirming the pH independence of this type of nanosensors. The nanosensors without the reference dye showed a decrease in the fluorescence intensity as the K+ concentration became higher and a moderate dynamic range (from 8- to 13fold), which is not yet optimal for the purpose of bioimaging. However, the problem is circumvented with the nanosensors containing Lumogen red, which showed a ratiometric response.

■

CONCLUSIONS To summarize, the fundamental sensing mechanisms of SDbased fluorescent ion-selective nanosensors were established 519


Article

ACS Sensors using K+ as a model ion. With an increasing lipophilicity of the SDs, the sensor mechanism switches from a bulk partitioning of the dyes to an interfacial dye accumulation. The surface accumulation of SDs was confirmed by FRET, resulting in a ratiometric K+ selective nanosensor in the NIR region. This work lays the foundation for future investigations of ion selective optical nanosensors based on lipophilized SDs.

■

(10) Huber, C.; Werner, T.; Krause, C.; Wolfbeis, O. S. Novel chloride-selective optode based on polymer-stabilized emulsions doped with a lipophilic fluorescent polarity-sensitive dye. Analyst 1999, 124, 1617−1622. (11) Krause, C.; Werner, T.; Huber, C.; Wolfbeis, O. S. Emulsionbased fluorosensors for potassium featuring improved stability and signal change. Anal. Chem. 1999, 71, 5304−5308. (12) Huber, C.; Klimant, I.; Krause, C.; Werner, T.; Wolfbeis, O. S. Nitrate-selective optical sensor applying a lipophilic fluorescent potential-sensitive dye. Anal. Chim. Acta 2001, 449, 81−93. (13) Reichardt, C. Solvatochromic dyes as solvent polarity indicators. Chem. Rev. 1994, 94, 2319−2358. (14) Reichardt, C.; Welton, T. Solvents and solvent effects in organic chemistry, 4th ed.; Wiley-VCH: Weinheim, Germany, 2010. (15) Xie, X.; Gutiérrez, A.; Trofimov, V.; Szilagyi, I.; Soldati, T.; Bakker, E. Charged solvatochromic dyes as signal transducers in pH independent fluorescent and colorimetric ion selective nanosensors. Anal. Chem. 2015, 87, 9954−9959. (16) Xie, X.; Zhai, J.; Bakker, E. Potentiometric response from ionselective nanospheres with voltage-sensitive dyes. J. Am. Chem. Soc. 2014, 136, 16465−16468. (17) Xie, X.; Mistlberger, G.; Bakker, E. Ultrasmall fluorescent ionexchanging nanospheres containing selective ionophores. Anal. Chem. 2013, 85, 9932−9938. (18) Klymchenko, A. S.; Roger, E.; Anton, N.; Anton, H.; Shulov, I.; Vermot, J.; Mely, Y.; Vandamme, T. F. Highly lipophilic fluorescent dyes in nano-emulsions: towards bright nonleaking nano-droplets. RSC Adv. 2012, 2, 11876−11886. (19) Roy, R.; Hohng, S.; Ha, T. A practical guide to single-molecule FRET. Nat. Methods 2008, 5, 507−516. (20) Sarabipour, S.; Del Piccolo, N.; Hristova, K. Characterization of membrane protein interactions in plasma membrane derived vesicles with quantitative imaging Förster resonance energy transfer. Acc. Chem. Res. 2015, 48, 2262−2269. (21) Zhou, X.; Herbst-Robinson, K. J.; Zhang, J. Visualizing dynamic activities of signaling enzymes using genetically encodable FRET-based biosensors: from designs to applications. In Methods in Enzymology; Conn, P. M., Ed.; Elsevier Inc.: Amsterdam, 2012; Vol. 504, pp 317− 340. (22) Xia, W.-S.; Schmehl, R. H.; Li, C.-J. A highly selective fluorescent chemosensor for K+ from a bis-15-crown-5 derivative. J. Am. Chem. Soc. 1999, 121, 5599−5600. (23) Baruah, M.; Qin, W.; Vallee, R. A. L.; Beljonne, D.; Rohand, T.; Dehaen, W.; Boens, N. A Highly potassium-selective ratiometric fluorescent indicator based on BODIPY azacrown ether excitable with visible light. Org. Lett. 2005, 7, 4377−4380. (24) Zhou, X.; Su, F.; Tian, Y.; Youngbull, C.; Johnson, R. H.; Meldrum, D. R. A new highly selective fluorescent K+ sensor. J. Am. Chem. Soc. 2011, 133, 18530−18533.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssensors.6b00006. Synthetic information on the SDs: including synthetic routes, NMR, and mass spectra; supplementary figures as noted in the main text; derivation of the theoretical sensor response: models for hydrophilic SDs and for highly hydrophobic SDs (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank the Swiss National Science Foundation (SNF) and the University of Geneva for financial support of this study. J.Z. gratefully acknowledges the Chinese Scholarship Council.

■

REFERENCES

(1) Zamaleeva, A. I.; Despras, G.; Luccardini, C.; Collot, M.; Waard, M. d.; Oheim, M.; Mallet, J.-M.; Feltz, A. FRET-based nanobiosensors for imaging intracellular Ca2+ and H+ microdomains. Sensors 2015, 15, 24662−24680. (2) Xie, X.; Li, X.; Ge, Y.; Qin, Y.; Chen, H.-Y. Rhodamine-based ratiometric fluorescent ion-selective bulk optodes. Sens. Actuators, B 2010, 151, 71−76. (3) Brasuel, M.; Kopelman, R.; Miller, T. J.; Tjalkens, R.; Philbert, M. A. Fluorescent nanosensors for intracellular chemical analysis: decyl methacrylate liquid polymer matrix and ion-exchange-based potassium PEBBLE sensors with real-time application to viable rat C6 glioma cells. Anal. Chem. 2001, 73, 2221−2228. (4) Xie, X.; Bakker, E. Ion selective optodes: from the bulk to the nanoscale. Anal. Bioanal. Chem. 2015, 407, 3899−3910. (5) Xie, X.; Crespo, G. A.; Zhai, J.; Szilágyi, I.; Bakker, E. Potassiumselective optical microsensors based on surface modified polystyrene microspheres. Chem. Commun. 2014, 50, 4592−4595. (6) Wolfbeis, O. S. An overview of nanoparticles commonly used in fluorescent bioimaging. Chem. Soc. Rev. 2015, 44, 4743−4768. (7) Kantner, K.; Ashraf, S.; Carregal-Romero, S.; Carrillo-Carrion, C.; Collot, M.; Pino, P. d.; Heimbrodt, W.; Aberasturi, D. J. D.; Kaiser, U.; Kazakova, L. I.; Lelle, M.; Baroja, N. M. d.; Montenegro, J. M.; Nazarenus, M.; Pelaz, B.; Peneva, K.; Gil, P. R.; Sabir, N.; Schneider, L. M.; Shabarchina, L. I.; Sukhorukov, G. B.; Vazquez, M.; Yang, F.; Parak, W. J. Particle-based optical sensing of intracellular ions at the example of calcium - what are the experimental pitfalls? Small 2015, 11, 896−904. (8) Wolfbeis, O. S. Fluorescence-based ion sensing using potentialsensitive dyes. Sens. Actuators, B 1995, 29, 140−147. (9) Murkovic, I.; Lobnik, A.; Mohr, G. J.; Wolfbeis, O. S. Fluorescent potential-sensitive dyes for use in solid state sensors for potassium ion. Anal. Chim. Acta 1996, 334, 125−132. 520
