COMMUNICATION

www.rsc.org/softmatter | Soft Matter

Biocompatible uorescent nanoparticles for pH-sensoring

fe,c Jana Wotschadlo,a Tim Liebert,a Stephanie Hornig,a Christoph Biskup,b Anja Gra c a Gerhard J. Mohr* and Thomas Heinze*
Received 8th January 2008, Accepted 28th March 2008 First published as an Advance Article on the web 14th April 2008 DOI: 10.1039/b800276b

Dialysis of a mixture of uorescein and sulforhodamine B marked dextran derivatives yields biocompatible and tuneable nanosensors that can be used for ratiometric pH measurements. The exploration of chemical microenvironments in organisms as well as on the cellular level is of great interest in medical and biological research. Every organism carefully controls a seemingly endless list of parameters including the pH and the concentration of ions in the body uids.1 Homeostasis is also achieved at the level of the single cell, which regulates many of the same parameters the body regulates. Especially the concentration of protons inside a cell and in the body as a whole is tightly regulated. This is vitally important because the activity of most enzymes is highly sensitive to changes in pH. Aberrations from the pH homeostasis caused by mutations in protontranslocating enzymes, disease, or by pharmacological means can lead to severe functional changes.1,2 To understand physiological and pathological mechanisms on the molecular level, tools to measure the pH in the extra- and intracellular space are an essential prerequisite. In particular uorescent pH indicators, among them carboxyuorescein derivatives (BCECF),3,4 seminaphthorhodauors (SNARF dyes) or seminaphthouoresceins (SNAFL dyes) are widely used to measure proton concentrations.5 Cells can be easily loaded with the membrane-permeant esters of these dyes.6 However, this approach has several shortcomings: after loading, the dyes redistribute and cannot be targeted to individual organelles. Moreover, the dyes can aggregate, which complicates or even invalidates pH measurements.7 These drawbacks can be overcome by embedding the indicator dyes in a polymer matrix.8 During recent years, optical analyte detection via uorescent sensor particles has become a very promising method, because it is precise, reproducible, and enables real-time observation of a large number of different cells simultaneously.9 In addition, the incorporation of the dyes in the particles minimizes organelle sequestration and protects the cells from the mostly cytotoxic organic uorophores. By combining an indicator dye with an inert reference dye, the ratiometric detection of analytes is possible. Changes in the uorescence emission of the indicator dye that are not caused by the analyte
a Center of Excellence for Polysaccharide Research, Member of the European Polysaccharide Network of Excellence, Friedrich Schiller University of Jena, Humboldtstrasse 10, 07743 Jena, Germany. E-mail: [email protected] b Institute of Physiology II, Universita tsklinikum Jena, Teichgraben 8, 07740 Jena, Germany c Institute of Physical Chemistry, Friedrich Schiller University of Jena, Lessingstrasse 10, 07743 Jena, Germany. E-mail: gerhard.mohr@ uni-jena.de Electronic supplementary information (ESI) available: Synthesis, materials and methods. See DOI: 10.1039/b800276b

but by uctuations in the sensor concentration or uctuations in the light source intensity can be corrected for by taking advantage of the uorescence signal of the reference dye. From the great variety of uorescent dyes and indicators various indicatorreference dye combinations can be tailored to specic applications. Among this concept are microparticle sensors like lipobeads (phospholipid coated polystyrene particles)10 and polyelectrolyte microcapsules11 which allow the measurement of intracellular pH values. One further approach is the encapsulation of uorescent probes in a polyacrylamide matrix during emulsion polymerization.8 Drawbacks may arise from toxicity because of monomer- and surfactant residues, and leaching of the non-covalently bound dye molecules from the polymer matrix. Recently, novel core/shell silica nanoparticles were developed incorporating the reference (tetramethylrhodamine) and sensor dye (uorescein) by covalent bonding in the coreshell architecture.12 However, the silica particle uptake for pH sensing in intracellular compartments is only feasible after addition of stimulating agents to the cells. The use of a biocompatible natural polymer as a basis for functionalized nanoparticle systems may overcome these problems. Our investigation employs novel polysaccharide based uorescent nanoparticles for sensor applications in living cells. The polyglucan dextran is a water soluble, well-established biopolymer in medical elds, e.g. in combination with bioactive molecules as a prodrug in pharmaceutical devices13 and as a coating material for the encapsulation of drugs.14 Dextran consists of a-1,6glycosidically linked D-glucose units with varying branches depending on the dextran producing bacterial strain. The availability of narrow distributed molecular weight fractions and the high amount of hydroxyl groups throughout the molecule qualies dextran as an excellent starting material for polymer-analogous reactions leading to highly functionalized dextran derivatives. Surprisingly, it was found that such highly hydrophobic polymers, e.g. propionylated dextrans, form well dened nanospheres with a narrow size distribution.15 Because it is known that ferrouids coated with dextran derivatives are easily incorporated into cells, the question arose whether the hydrophobic full-polymer particles show comparable behaviour, which could open up new paths for intracellular analysis and release devices for pharmaceutical agents.16 Here, we document results towards a nanoscale, intracellular uorescence sensor system containing both a pH-sensitive and a reference dye in one particle. Therefore, esterication of dextran leading to nanoparticle forming derivatives was conducted by reaction of the polymer with propionic anhydride in pyridine as the base. This procedure does not lead to degradation according to our experiences in the eld of polysaccharide chemistry.17 The covalent attachment of small amounts of uorescent dye can be carried out either before or after propionylation to prevent undesired acid catalysed hydrolysis of the dyes. Because of their high uorescence quantum yield, only low
Soft Matter, 2008, 4, 11691172 | 1169

This journal is The Royal Society of Chemistry 2008

Scheme 1 Synthesis of sulforhodamine B and uorescein labelled dextran propionate and the preparation of nanospheres; one anhydroglucose unit is schematically represented.

degrees of substitution with dyes are necessary, which guarantees that intra- and intermolecular interactions between the dyes are unlikely. The functionalization of dextran with sulforhodamine B acid chloride and subsequent perpropionylation leads to a completely substituted sulforhodamine B labeled dextran propionate (SRB-dextran propionate), which is used as the reference system (Scheme 1, for details see Supporting Information). The pH indicator dye uorescein isothiocyanate was allowed to react with a dextran propionate possessing a degree of substitution of 1.81 (FITC-dextran propionate). The desired ratio of dyes and hence the desired uorescence absorption can be simply adjusted during dialysis as mentioned below. It was possible to shape both uorescence labelled dextran propionates into spherical nanoparticles via dialysis of an N,N-dimethylacetamide (DMAc) solution against water.15,18 It is a general result of our studies that the exchange of water-miscible solvents like DMAc and acetone of a dilute polymer solution by water leads to nanoprecipitation of functionalized polysaccharide particles.19 The resulting SRB-dextran propionate particles are 542 nm in size and exhibit a polydispersity index (PdI) of 0.225 as examined by dynamic light scattering (DLS). FITC-Dextran propionate particles are smaller (374 nm) and more uniformly distributed as indicated by the lower PdI of 0.133. Furthermore, the particles were characterized by SEM showing regular nanospheres with a size in the same range as determined by DLS (Fig. 1a,b). The nanoparticle suspensions are even stable during a standard sterilising procedure in an autoclave at 120  C and 2 bar for removing germs. Generally, the adjustment of uorescence intensities in particle systems for ratiometric measurements is not easy, because only small amounts of uorescent dyes are needed to label the particles.58 Furthermore, the reaction of dyes with the polymer backbone is often not quantitative and can hardly be predicted exactly. In our concept,
1170 | Soft Matter, 2008, 4, 11691172

Fig. 1 SEM images of a) FITC-dextran propionate particles (d 374 nm), b) SRB-dextran propionate particles (d 524 nm), and c) dextran propionate particles (d 501 nm) bearing both uorescein and sulforhodamine B as sensor and reference dye, respectively (ratio FITCdextran propionate : SRB-dextran propionate 1 : 7). Scale bar 1 mm.

however, the incorporation of both reference and sensor dyes in one particle can be achieved by mixing two dextran derivatives in the desired ratio. It can be shown that this is simply realized by dissolving the mixture in DMAc with subsequent dialysis leading to nanoparticle suspensions. By variation of the weight percentages of the particular dextran propionate, the optimal peak intensities can be adjusted. The regulation of the uorescence intensities of the mentioned dextran nanoparticles was conducted by mixing SRB-dextran propionate and FITC-dextran propionate in the ratios 7 : 1, 3 : 1, 1 : 1, and 1 : 3, with subsequent dialysis of the DMAc solution against water. Fig. 2 shows the uorescence emission spectra of the resulting nanoparticle suspensions at pH 8.1. The spectra have been obtained at an excitation wavelength of 488 nm, which is commonly available in confocal laser scanning microscopes equipped with an Ar+ laser or other setups used to image biological samples. Reasonable spectra for ratiometric pH measurements were obtained in the ratios 7 : 1 and 3 : 1. Since SRB absorption is maximal at 560 nm and low at 488 nm, the
This journal is The Royal Society of Chemistry 2008

Fig. 2 Normalized uorescence emission spectra (lexc 488 nm) of nanoparticle suspensions at pH 8.1 with varying ratios (w/w) of SRBdextran propionate (lmax 581 nm) to FITC-dextran propionate (lmax 516 nm); a) 7 : 1, b) 3 : 1, c) 1 : 1, and d) 1 : 3 prior to the dialysis process.

uorescence intensity of SRB is still low with respect to the uorescence intensity of uorescein despite the higher fraction of SRBdextran propionate. In the following experiments, the 7 : 1 (SRB- : FITC-dextran propionate) nanoparticle suspension with particles having a size of 501 nm was used (Fig. 1c). To assess the potential of the FITC-/SRB-dextran propionate nanoparticles as a ratiometric pH-sensor, the particle suspension was calibrated against sodium phosphate buffer solutions of pH 4.98.2 as depicted in Fig. 3a. The uorescence intensity of uorescein decreases with decreasing pH, whereas the sulforhodamine B intensity remains stable. The small decrease of the sulforhodamine B emission peak results from the overlap with the uorescein emission. A calibration curve (Fig. 3b) was obtained by calculating the ratio of the emission peaks of uorescein at 516 nm (lexc 488 nm) and sulforhodamine B at 581 nm (lexc 488 nm). The effective pKa value of the ratiometric dextran propionate particles at both SRB excitation wavelengths is 6.4, which matches exactly the literature values for the transition of the mono- and dianionic state of unbound uorescein dissolved in aqueous solution.20 As demonstrated by Fig. 3a the emission spectra of uorescein and SRB overlap. In experiments on the stage of a confocal laser-scanning microscope (LSM) this cross talk between the sensor and the reference channel can complicate the measurement. To overcome this problem, we separated FITC- and SRB-uorescence by exploiting the different excitation properties of the dyes. With the 543 nm line of He/Ne-lasers, which are available in most LSMs, only SRB but not uorescein is excited, so that the SRB emission is not overlapped by uorescein uorescence. All SRB emission spectra can be perfectly superimposed at all pH values without the need to correct for a spectral overlap (Fig. 3a). Because of this advantage, the dextran particles were excited in all confocal measurements subsequently with the Ar-laser at 488 nm and with the He/Ne-laser at 543 nm. Like in the experiments in the cuvette, the ratio of both signals was used to calculate the pH-calibration curve (Fig. 3b). A crucial point is the homogeneous distribution of both dyes in the particles. Fig. 4 shows a confocal cross section of a single FITC-/ SRB-dextran propionate nanoparticle. Fluorescein uorescence was excited with the Ar 488 nm line and was collected with a 505530 nm bandpass lter (Fig. 4a). SRB uorescence was excited with the He/ Ne 543 nm line and collected with a 560 nm longpass lter (Fig. 4b).
This journal is The Royal Society of Chemistry 2008

Fig. 3 Calibration of the dextran particles (FITC-dextran propionate : SRB-dextran propionate 1 : 7) via uorescence spectroscopy: a) uorescence emission spectra excited at 488 nm and at 543 nm of the nanoparticle suspension at pH 4.9 (bottom), 5.4, 6.0, 6.5, 7.0, 7.5 and 8.2 (top), b) calibration curve showing the ratio of uorescence intensities of the indicator dye (determined at 516 nm, lexc 488 nm) to the reference dye (determined at 581 nm, lexc 488 nm and 543 nm, respectively) versus pH.

Fig. 4 Confocal laser scanning micrographs of dextran propionate nanoparticles (FITC-dextran propionate : SRB-dextran propionate 1 : 7); a) uorescein (sensor) channel, b) SRB (reference) channel, c) overlay of both channels, d) uorescence prole of FITC (green curve) and SRB (red curve) along the indicated line in panel c. Scale bar 200 nm.

Soft Matter, 2008, 4, 11691172 | 1171

Moreover, the particles might also be used for drug release studies in cells as they contain hydrophobic microdomains for the inclusion of hydrophobic substances.21 Changes in the chemical microenvironments can be detected in situ. In conclusion, dextran nanoparticles provide a promising biocompatible and easily functionalizable alternative towards conventional nanoparticle systems.

Acknowledgements
Fig. 5 Confocal micrographs of human broblasts incubated with FITC-/SRB-dextran propionate nanoparticles; a) uorescein (sensor) channel, b) SRB (reference) channel, c) overlay of both channels. Scale bar 5 mm.

The overlay of both images (Fig. 4c) shows that both dyes are uniformly distributed throughout the particle, which is conrmed by the uorescence prole (Fig. 4d) along the indicated line in Fig. 4c. To assess if the particles can be incorporated into cells, human foreskin broblasts were incubated with FITC/SRB-dextran propionate nanoparticles. After one day, non-incorporated particles were removed by extensive washing, and the cells were allowed to grow further in cell culture medium under standard cell culture conditions. Fig. 5 shows a confocal micrograph of broblasts, which was obtained 7 days after incubation. The micrograph shows that the cells are highly loaded with particles. Confocal z-stacks proved that the particles were located inside the cells and did not stick to the cell surface. In the overlay of the sensor and reference channel (Fig. 5c) particles located in the more acidic cytosol have an orange color, whereas particles in the more alkaline surrounding solution exhibit a more greenish color. Long-term experiments over a period of 22 days show still healthy cells, which are able to divide, and evidence the absence of cytotoxic effects. The cells are still loaded with the particles. In all cells incubated for several days or longer, we noticed, however, that SRB uorescence decreases with respect to uorescein uorescence. This might be due to the fact that the thiocarbamate bond of the uorescein derivative is more stable than the sulfonic acid ester bond of the SRB labelled dextran. No substantial leaching of the uorescence dyes, which would have caused a diffuse uorescence pattern in the cell and the surrounding medium, was observed. These experiments demonstrate the potential of uorescence labelled polysaccharide nanoparticles in biological experiments. They are non-toxic and easily incorporated by cells without any additional agents. In this study, we employed the particles as ratiometric pHsensors. But, by using other indicator dyes analogue nanosensors for other analytes of interest can be easily developed. The covalent attachment of any uorescent dye, if it contains at least one functional group for the reaction with hydroxyl groups or a spacer moiety, is conceivable. By varying the ratio of sensor and reference uorescence dextran derivatives, the dynamic range of the sensor nanoparticles can be adapted to the sensitivity of the setup.

The authors gratefully acknowledge J.H. Clement of the Dept. for Internal Medicine II, University Clinic Jena, for his support in cell experiments, and the nancial support by Carl-Zeiss Stiftung and Stiftung fu r Technologie, Innovation und Forschung Thu ringen. A.G. and G.J.M. were supported by Deutsche Forschungsgemeinschaft within project MO 1062/2-2, C.B. and G.J.M. were supported by the EU within the Marie Curie project MTKDCT-2005-029554.

Notes and references

1 W. F. Boron and E. L. Boulpaep, Medical Physiology, Elsevier Saunders, Philadelphia, 2003. 2 N. Demaurex, News Physiol. Sci., 2002, 17, 15. 3 T. J. Rink, R. Y. Tsien and T. J. Pozzan, J. Cell Biol., 1982, 95, 189 196. 4 A. M. Paradiso, R. Y. Tsien and T. E. Machen, Proc. Natl. Acad. Sci. U. S. A., 1984, 81, 74367440. 5 J. E. Whitaker, R. P. Haugland and F. G. Prendergast, Anal. Biochem., 1991, 194, 330344. 6 R. Y. Tsien, Nature, 1981, 290, 527528. , F. Sureau and J. Bolard, Anal. 7 O. Seksek, N. Henry-Toulme Biochem., 1991, 193, 4954. 8 H. A. Clark, R. Kopelman, R. Tjalkens and M. A. Philbert, Anal. Chem., 1999, 71, 48374843. 9 F. Wang, W. B. Tan, Y. Zhang, X. Fan and M. Wang, Nanotechnology, 2006, 17, R1. 10 K. P. McNamara, T. Nguyen, G. Dumitrascu, J. Ji, N. Rosenzweig and Z. Rosenzweig, Anal. Chem., 2001, 73, 32403246. 11 O. Kreft, A. M. Javier, G. B. Sukhorukov and W. J. Parak, J. Mater. Chem., 2007, 17, 44714476. 12 A. Burns, P. Sengupta, T. Zedayko, B. Baird and U. Wiesner, Small, 2006, 2, 723726. 13 T. Heinze, T. Liebert, B. Heublein and S. Hornig, Adv. Polym. Sci., 2006, 205, 199291. 14 X. Shi, S. Wang, X. Chen, S. Meshinchi and J. R. Baker Jr, Mol. Pharmacol., 2006, 3, 144151. 15 T. Liebert, S. Hornig, S. Hesse and T. Heinze, J. Am. Chem. Soc., 2005, 127, 1048410485. 16 K. Wagner, A. Kautz, M. Ro der, M. Schwalbe, K. Pachmann, J. H. Clement and M. Schnabelrauch, Appl. Organomet. Chem., 2004, 18, 514519. 17 T. Heinze, T. Liebert and A. Koschella, Esterication of Polysaccharides, Springer, Heidelberg, 2006. 18 S. Hornig, T. Heinze, S. Hesse and T. Liebert, Macromol. Rapid Commun., 2005, 26, 19081912. 19 S. Hornig and T. Heinze, Biomacromolecules, 2008, DOI: 10.1012/ bm8000155. 20 N. Klonis and W. H. Sawyer, Photochem. Photobiol., 2000, 72, 179 185. 21 S. Hornig and T. Heinze, Carbohydr. Polym., 2006, 68, 280286.

1172 | Soft Matter, 2008, 4, 11691172

This journal is The Royal Society of Chemistry 2008