- Email: [email protected]
Fluorescence imaging of the activity of glucose oxidase using a hydrogen-peroxide-sensitive europium probe
ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 340 (2005) 66–73 www.elsevier.com/locate/yabio
Fluorescence imaging of the activity of glucose oxidase using a hydrogen-peroxide-sensitive europium probe Meng Wu, Zhihong Lin, Michael Scha¨ferling, Axel Du¨rkop, Otto S. Wolfbeis * Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg, D-93040 Regensburg, Germany Received 16 November 2004
Abstract A method for optical imaging of the activity of glucose oxidase (GOx) using a fluorescent europium(III) tetracycline probe for hydrogen peroxide is presented. A decay time in the microsecond range and the large Stokes shift of 210 nm of the probe facilitate intensity-based, time-resolved, and decay-time-based imaging of glucose oxidase. Four methods for imaging the activity of GOx were compared, and rapid lifetime determination imaging was found to be the best in giving a linear range from 0.32 to 2.7 mUnit/mL. The detection limit is 0.32 mUnit/mL (1.7 ng mL1) which is similar to that of the time-resolved (gated) imaging using a microtiterplate reader. Fluorescent imaging of the activity of GOx is considered to be a useful tool for GOx-based immunoassays with potential for high-throughput screening, immobilization studies, and biosensor array technologies. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Glucose oxidase; Enzyme imaging; Fluorescent probe; Time-resolved imaging; Enzyme activity assay
Glucose oxidase (GOx, EC 1.1.3.4)1 is a flavoenzyme that catalyzes specifically the oxidation of glucose into gluconate and hydrogen peroxide Eq. (1) [1]. It is one of the most widely used enzymes due to its availability and stability. As a representative of oxidases, it has been studied in solution and in the immobilized state on solid surfaces [2] or encapsulated in microspheres [3] for applications in biosensors [4], industrial bioreactors [5], and bio-fuel cells [6]. Apo-glucose oxidase has also been utilized in studies for ‘‘nanowiring’’ [7] of the electron transfer. GOx has been applied not only to the electrochemical and optical detection of glucose [8] but also *
Corresponding author. Fax: +49 941 943 4064. E-mail address: [email protected] (O.S. Wolfbeis). 1 Abbreviations used: GOx, glucose oxidase; EuTc, europium–tetracycline complex; EuTc-HP, europium–tetracycline–hydrogen peroxide; HP, hydrogen peroxide; Tc, tetracycline; FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging; POx, peroxidase; CCD, charge-coupled device; LED, light-emitting diode. 0003-2697/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2005.01.050
in GOx-labeled enzyme amplification analysis for immunoassays [9]. GOx-based enzyme immunoassays have been used for the detection and screening of, e.g., steroids, drugs, environmental pollutants, and special peptides [10]. Recently, GOx has been applied to microchip or microarray technologies for the development of protein chips for proteomics [11].
ð1Þ Numerous methods for the determination of GOx have been developed. A set of assays have emerged based on the measurement of oxygen [12], pH [13], or H2O2 [14]. However, methods based on the measurement of H2O2 produced by GOx are of particular interest because H2O2 does not form a background but is produced only during enzymatic reaction. Since H2O2 cannot be easily visualized, it is usually converted into a colored or fluorescent product using peroxidase (POx).
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
In recent years, fluorescent imaging technologies have attracted substantial attention because of the multitude of information on both the spatial and the temporal characterization of the target analyte [15,16]. Confocal [17], multiphoton excitation [18], near-field [19], and decaytime-based imaging [20,21] technologies have been developed among others. Existing methods for imaging GOx are restricted mainly to scanning electrochemical microscopy [22,23] and scanning chemiluminescence imaging [24]. This is largely due to the fact that almost all fluorescence imaging methods for GOx need a second enzyme (POx) to visualize the H2O2 produced by the GOx-catalyzed reaction. There have been reports on H2O2-based imaging using dihydrorhodamine 123 [25], homovanillic acid [26], scopoletin [27], dichlorodihydrofluorescein diacetate [28] or others [29], all using POx along with a second (chromogenic or fluorogenic) substrate [30]. Based on our previous reports on the europium–tetracycline (EuTc) fluorescent probe for determination of either H2O2 [31] or glucose either in solution [32] or using biosensor membranes [33,34] we present here a scheme for imaging GOx at neutral pH and without the need for a second enzyme. It also is shown that, in addition to intensity-based imaging of GOx, the long luminescence decay time of the probe enables time-resolved and lifetime-based imaging.
Materials and methods Materials and reagents
67
Austria; www.tecan.com). The excitation filter was set to 405 nm and the emission filter to 612 nm. All experiments were performed at a programmed temperature of 30.0 ± 0.1 °C. Either black Fluotrac200 microtiterplates or black plates with a flat transparent bottom (Greiner bio-one, Frickenhausen, Germany; www.greiner-lab.com) were used. Imaging device The device used in this study has been previously reported [35,36] and was used with minor modifications. As shown schematically in Fig. 1, it is composed of a fast gateable CCD camera (SensiMod; from PCO, Kelheim, Germany; www.pco.de), a pulsable LED array with 96 UV light-emitting diodes (kmax 405 nm, Roithner Laser Technik, Vienna, Austria; www.roithner-laser.com), a pulse generator (Scientific Instruments DG 535, Sunnyvale, CA, USA; www.srsys.com; not shown in Fig. 1), an optical emission filter (KV 550; Schott, Mainz, Germany; www.schott.com), and an optical excitation filter (BG 12; Schott), with a light-guiding adapter consisting of 96 optical fibers (diameter 3 mm) for matching the focus of the CCD camera. A computer was used for control and visualization of the experiments that were programmed in Interactive Data Language (IDL; from Research Systems, Boulder, CO, USA; www.rsinc.com). The manipulation and calculation of the images, such as the rotation and cropping of the images, the subtraction of the dark image (blank, without illumination) from the fluorescent image, and the integration of the
All solutions were prepared in a 10 mmol L1 3-(Nmorpholino)propanesulfonate (Mops) buffer of pH 6.9. Glucose oxidase (from Aspergillus niger; 185,000 Unit/ g; used without further purification) was from Sigma– Aldrich (Steinheim, Germany) and tetracycline hydrochloride was from Serva (Heidelberg, Germany; www.serva.de). The activity of GOx is defined by the provider as follows: one unit will oxidize 1.0 lmol of b-D-glucose to D-gluconolactone and H2O2 per min at pH 5.1 and 35 °C. EuCl3 hexahydrate was from Alfa Products (Danvers, MA, USA; www.alfa.com). The EuTc stock solution was prepared by mixing 10 mL of a 6.3 mmol L1 EuCl3 solution with 10 mL of a 2.1 mmol L1 tetracycline solution and diluting it to 100 mL with Mops buffer (the molar ratio of Eu3+ to Tc being 3:1). The glucose stock solution (0.277 mol L1) was stored overnight before use to warrant the equilibration of a and b anomers. Instruments Fluorescence measurements and time-resolved (gated) intensity detection were performed on a GENios+ microtiter plate reader (Tecan, Gro¨dig, Salzburg,
Fig. 1. Setup of the imaging system. (A) Fast gateable CCD camera; (B) optical emission filter (KV 550); (C) light-guiding adapter consisting of 96 optical fibers (diameter 3 mm); (D) 96-well microtiterplate (black with transparent bottom); (E) optical excitation filter (BG 12); (F) pulsable LED array with 96 UV light-emitting diodes (kmax = 405 nm).
68
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
single images and their ratioing, were done by a self-developed program based on Matlab (6.1; Mathwork, Natick, MA, USA; www.mathwork.com). The images of the ratio from decay-time-based detection were inverted to their negatives for a more facile comparison of different methods. Imaging of the activity of GOx To each well of a 96-well microtiter plate with transparent bottom was added 100 lL of the EuTc standard solution, 15 lL of a 277 mmol L1 glucose solution, and enough Mops buffer to make up the total volume to 200 lL. Phosphate buffer should be avoided since it quenches the fluorescence of the EuTc–H2O2 complex (refered to as EuTc-HP). The samples (50 lL) containing GOx with an activity between 1.2 and 5.6 mUnit/ mL were added simultaneously. The images of the plates were taken after a 30-min incubation at 30 °C. Blank values (F0) were obtained by adding buffer in place of GOx. In the time-resolved (gated) nonimaging determination using a commercially available microwell plate fluorescence reader, a lag time of 60 ls and an integration time of 40 ls were employed [33].
Fig. 2. Kinetic response of the glucose/EuTc system to increasing activities of glucose oxidase; 100 lL of a the EuTc stock solution, 15 lL of a 277.2 mmol L1 glucose solution, and Mops buffer made up to a total volume of 200 lL. Glucose oxidase activities (from A to G): 0.0, 0.27, 0.54, 1.4, 2.7, 5.4, and 13.5 mUnit/mL, respectively.
bate, uric acid, phosphate, and citrate interfere [33]. The baseline has a slight downward drift which is due to a slight increase in temperature at the beginning of the experiment (following addition of enzyme solution). The luminescence decay profiles of EuTc and EuTc-HP
Results The scheme The assay for GOx activity via H2O2 is based on the response of EuTc to enzymatically produced H2O2 to form the EuTc-HP complex according to ð2Þ It is interesting to note that the stoichiometry (Eu:Tc) giving the largest signal changes with H2O2 is 3:1. Like others [37], we were unable to further elucidate the chemical structure of EuTc since it cannot be isolated in solid form. It is known, however, that HP rather weakly binds to EuTc and that binding is highly pH dependent [38]. It therefore may be more appropriate to formulate the equilibrium as [Eu3Tc + n H2O2 M Eu3Tc-HP + (n 1) HP]. The kinetic response of the glucose/EuTc system to increasing activities of GOx is shown in Fig. 2. With higher activities of GOx, and therefore increasing production of H2O2, the fluorescence (more correctly the luminescence) of the Eu complex increases due to the formation of more strongly fluorescent EuTc-HP. The intensity of the fluorescence of EuTc-HP is around 15fold higher than that of EuTc, and its quantum yield is around 4%. Glucose and GOx themselves have no significant effect on either EuTc or EuTc-HP, while ascor-
The probe EuTc is converted into EuTc-HP by enzymatically produced hydrogen peroxide, and this results in an up to 15-fold increase in fluorescence intensity. At the same time the decay times (s) undergo a significant change [32,33]. The rather complex decay profile of EuTc 3:1 complex in buffer solution of pH 6.9 is composed of three components, with decay times of 8.7 ls (58% contribution), 30.4 ls (40%), and 174 ls (2%). The EuTc-HP complex decays in a similar fashion, the decay times being 13.2 ls (34%), 59 ls (64%), and 158 ls (2%). Details have been reported and discussed in recent papers [31,35]. The multiple decay is likely to be due to three activation channels of the same species rather than to several chemical species. First, only EuTc is detected in this optical system (at an excitation wavelength of 400 nm and the emission being collected at 616 nm). Second, the decay profiles do not strongly vary with the molar ratio of Eu:Tc (checked for ratios between 3:1 and 1:3). The v2 values after triple exponential fit are 1.55 for the EuTc complex and 1.46 for EuTc-HP which in fact is excellent. The data reveal that time-resolved measurements are best performed after a lag time of >60 ls to selectively detect the 59-ls component of the EuTc-HP complex (with its stronger intensity) and—at the same time—suppress the shorter components (of 8.7 and 30.4 ls) of the EuTc complex and the 13.2-ns component of EuTc-HP. Lag times of 60 ls can be adjusted in most present day fluorescence (and microplate) readers. The difference
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
also can be seen by calculating the average decay times of EuTc (30 ls) and EuTc-HP (60 ls). Both time-resolved and decay-time-based determination and imaging become possible. We have performed assays in both formats. Time-resolved determination of GOx using a microplate reader First, a time-gated fluorescent assay for GOx activities was worked out for microtiterplate readers. Gated detection with a lag time of 60 ls and an integration time of 40 ls is adequate to overcome the background fluorescence of the samples and plates which usually have nanosecond lifetimes. The reader assay resulted in a linear range for the GOx assay from 0.32 to 14 mUnit/mL, with a limit of detection (S/N = 3) of 0.32 mUnit/mL. The reader assay usually has a broader dynamic range and smaller standard deviations compared to those of the CCD imaging method. Detection schemes for imaging Next, we transferred the previous results to an imaging assay. The four imaging schemes shown in Fig. 3 were tested for their suitability for quantitative imaging of GOx activity. Their features are summarized in Table 1. In conventional fluorescence intensity imaging (FII) [29] one detection window is opened while excitation is
69
on. In contrast to this often applied method, in time-resolved imaging (TRI) [39] the detection window is opened only after the excitation pulse. TRI imaging technology is most useful to eliminate (by gating) the autofluorescence of samples and plates with their nanosecond decay times. It should be kept in mind that both FII and TRI are intensity-based detection schemes. Methods for luminescence decay-time-based imaging (in contrast to FII and TRI) have certain advantages since they are independent of (a) light source and detector fluctuations, (b) concentration of the fluorophore, Table 1 Time programs of the four imaging schemes studied in this work (all times in ls) Methoda
FII
TRI
PDR
RLD
Excitation ti tt
0 50
0 50
0 50
0 50
Length of window 1 (=W1) ti tt
0 50
80 130
0 50
80 130
Length of window 2 (= W2) ti tt
— —
— —
100 150
180 300
b
a FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging. b ti, initiation time; tt, terminating time (all in ls); see Fig. 3.
Fig. 3. Schematic of the principles of the four imaging schemes. Timelines (i.e., the gating of windows W1 and W2) are indicated in Table 1.
70
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
and (c) light scatter. The commonly used techniques for decay-time imaging are utilized here with a two-window detection scheme (see Fig. 3) and ratiometric quantification scheme as shown in Eq. (3): R¼
W 1 W dark W 2 W dark
image
:
ð3Þ
here. It should be noted, though, that for multiexponentially decaying systems such as EuTc-HP the precision in the determination of the average lifetimes strongly depends on the selection of the detection windows [40]. Quantitative aspects of GOx imaging
image
Two schemes were tested. In phase delay rationing imaging (PDR) [36,37] two windows are acquired, one in the excitation phase and one in the phase after excitation. The ratio of the two windows can be used for the quantification of the activities of GOx. Rapid lifetime determination imaging (RLD), in contrast, is based on the acquisition of two windows after excitation [40,41], the ratio of intensities of the two windows enabling a straightforward determination of decay time. Moreover, RLD—in being a gated method also—can efficiently suppress background ns fluorescence. The resulting intensity-based and lifetime-based GOx images are displayed in Fig. 4. They indicate that the fluorescent probe EuTc is applicable to all four schemes. The intensity-based images (FII, see Fig. 4A; TRI, see Fig. 4B) are affected by flash-to-flash variations in intensity and by variations in the brightness from LED to LED in the battery of light sources (Fig. 1). This results in high well-to-well standard deviations. The high inhomogeneities in the single wells result from scattering effects. The lifetime-based images (PDR and RLD, see Figs. 4C and 4D), in contrast, reveal good homogeneity due to their independence of intensity variations of the excitation light, the quantity of the fluorophore present in the wells, and that of scattered light. Cutoff optical filters and a common ns pulse generator are adequate for the optimal time-resolved imaging of GOx. Previous studies have shown that ratiometric methods such as RLD are suitable for determination of fluorescence lifetime [42], especially when using fluorescent probes displaying decay times in the ls range. While the RLD method is intended for use in monoexponentially decaying systems [43], it is found also to work quite well in the multiexponential decay system used
The RLD scheme was optimized with respect to the concentration of glucose and the concentration of EuTc. The optimal response (i.e., the largest change in fluorescence intensity) was obtained when using 20 lL of the glucose stock solution and 95 lL of the EuTc stock solution, with 200 lL total volume in the well of the microplate. The steady state (endpoint) method was used since the reaction reaches the endpoint after a 30-min incubation, as shown in Fig. 2. The results show that different activities of GOx clearly can be differentiated and result in the different intensities (and decay times) of EuTc-HP. It should be noted that kinetic imaging of GOx also is conceivable. However, it is less favored because it is critically dependent on precise timing. The rapid lifetime determination scheme was further exploited for quantitative imaging of the activity of GOx. The normalized ratio of the two images is shown
Fig. 5. Rapid lifetime determination imaging of the activity of glucose oxidase. The pictures reflect the normalized ratio of the two images according to Eq. (3) (in pseudo colors; see the color bar). All experiments were performed in triplicate (rows). The cocktails in wells (from 1 to 12) had the following compositions: GOx activities 0 (blank), 135, 54.1, 27.1, 13.5, 5.4, 2.7, 1.35, 0.54, 0.27, 0.14, 0.05 mUnit/mL, respectively; plus 100 lL of the EuTc stock solution, 15 lL of a 277.2 mmol L1 glucose solution; total volume made up to 200 lL with Mops buffer.
Fig. 4. Comparison of the images obtained for the activity of glucose oxidase by the four methods (after data processing). (A) Fluorescence intensity imaging (FII); (B) time-resolved imaging (TRI); (C) phase delay imaging (PDR); (D) rapid lifetime imaging (RLD). Experimental parameters are specified in Table 1; those for data processing are specified under Materials and methods. The solutions in the wells (from left to right in duplicate) contained GOx (135, 54.1, 27.1, 13.5, 5.4, 2.7, 1.35, 0.54, 0.27, 0.14, 0.05, and 0 mUnit/mL, respectively), 100 lL of the EuTc stock solution, and 15 lL of a 277.2 mmol L1 glucose solution. The total volume of 200 lL was made up with Mops buffer.
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
71
Fig. 6. Calibration plot for the determination of the activity of GOx by rapid lifetime determination imaging. The right chart is the calibration for the linear (= initial) range. Experimental conditions as in Fig. 5.
in pseudo colors in Fig. 5. The results show that GOx can be quantified easily via the averaged values of the gray scale images as shown in Fig. 6 which gives a quantitative plot. The response is linear from 0.32 to 2.7 mUnit/mL of GOx, with a detection limit of about 0.32 mUnit/mL (1.7 ng mL1). The GOx activity has been validated independently via determination of H2O2 by the standard peroxidase/o-dianisidine method (Sigma) and gave excellent agreement. Under our experimental conditions saturation was observed at around 20 mUnit/mL which is due to complete consumption of glucose and not to consumption of oxygen. The assay was optimized for low enzyme levels. If higher levels are to be determined, the concentration of glucose has to be increased accordingly. Figures of merit of the methods employed in this work are summarized in Table 2. It shows that RLD is the optimal choice for the quantitative imaging of GOx activity since it has the largest linear range at comparable limits of detection for GOx. In addition, RLD is a fairly straightforward method that requires a moderately sophisticated instrumental effort. Thus, the method of choice for bioanalytical applications is the RLD method because it is best with regard to sensitivity, dynamic range, and reproducibility. It also is Table 2 Figures of merit for the determination of GOx by the four imaging schemes studied in this work Methoda b
Limit of detection Linear rangeb Standard deviationc Remarks
FII
TRI
PDR
RLD
12.8 12.8–27.1 49
3.1 3.1–27.1 33
0.35 0.35–0.54 6
0.32 0.32–2.7 7
d
d
a FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging. b mUnit/mL. c % at 2.7 Unit/mL. d The black/white images were reversed for better comparison of all images.
capable of suppressing interference by stray light and ambient light due to internal signal referencing. This makes RLD superior to both intensity-based imaging schemes. It should be noted that, to make data comparable, the pictures have to be processed first. This results from the fact that in FII and TRI an increase in GOx activity results in an increase in luminescence intensity, while in PDR and RLD it results in a decrease of the ratios of the two windows. Since such images cannot be compared, those of PDR and RLD were inversed (i.e., converted into their ‘‘negatives’’) to make them comparable.
Discussion To the best of our knowledge a decay-time-based scheme for detection or imaging of oxidases via H2O2 has not been reported so far. The activity of GOx has been determined in solution via oxygen consumption [44] by measurement of the phase shift of the luminescence of an oxygen-quenchable probe. The scheme here is more promising insofar as the activity is measured via the production of H2O2which—unlike oxygen—does not form a strong and, in the worst case, varying background, but rather is completely absent in most samples. Lanthanide-based labels and microspheres have been used in time-resolved (gated) immunoassays [45–47] due to long decay times which are on the order of ls to ms. In the scheme reported here, dissolved EuTc is used as an H2O2-sensitive probe. There is no need for reagents other than glucose and EuTc. The methods for imaging GOx activity reported here include both intensity-based and decay-time-based approaches. The probe used (EuTc) is easily accessible, has fairly high sensitivity to H2O2, a large StokesÕ shift (210 nm), and a line-like emission peaking at 616 nm, is compatible with commercial shortwave diode laser light sources (with lines at 375 and 405 nm, respectively) and stable toward oxygen, and has a working pH of 7.
72
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
The methods presented here may be applied to enzyme assays and to enzyme-based (and amplification) immunoassays. While demonstrated here for GOx only, it most likely can be extended to image other oxidases such as uricase. Another application may be in highthroughput screening. Furthermore, imaging of GOx (and, conceivably, other oxidases) offers the possibility for analyzing multienzyme arrays or multisubstrate arrays [48]. The capability of spatial discrimination also makes imaging of interest to assess enzyme activity in immobilization studies [49,50] such as those testing the quality of glucose biosensors based on GOx.
Acknowledgments We acknowledge the help of Jo¨rg Enderlein of Forschungszentrum Ju¨lich, Germany and of Henrik Bauer of PicoQuant GmbH, Berlin in some of the decay time measurements. M.W. and Z.L. acknowledge financial support from Chromeon GmbH (Regensburg).
References [1] J. Raba, H.A. Mottola, Glucose oxidase as an analytical reagent, Crit. Rev. Anal. Chem. 25 (1995) 1–42. [2] M. Delvaux, S. Demoustier-Champagne, Immobilisation of glucose oxidase within metallic nanotubes arrays for application to enzyme biosensors, Biosens. Bioelectron. 18 (2003) 943–951. [3] S. Slomkowski, B. Miksa, New types of microspheres and microsphere-related materials for medical diagnostics, Polymers Adv. Tech. 13 (2002) 906–918. [4] L.Q. Chen, X.E. Zhang, W.H. Xie, Y.F. Zhou, Z.P. Zhang, A.E.G. Cass, Genetic modification of glucose oxidase for improving performance of an amperometric glucose biosensor, Biosens. Bioelectron. 17 (2002) 851–857. [5] M. Traeger, G.N. Qazi, U. Onken, C.L. Chopra, Contribution of endocellular glucose oxidase and exocellular glucose oxidase to gluconic acid production at increased dissolved-oxygen concentrations, J. Chem. Tech. Biotechnol. 50 (1991) 1–11. [6] S. Tsujimura, K. Kano, T. Ikeda, Glucose/O-2 biofuel cell operating at physiological conditions, Electrochem. 70 (2002) 940–942. [7] Y. Xiao, F. Patolsky, E. Katz, J.F. Hainfeld, I. Willner, Plugging into enzymes: nanowiring of redox enzymes by a gold nanoparticle, Science 299 (2003) 1877–1881. [8] S. Reiter, K. Habermu¨ller, W. Schuhmann, A reagentless glucose biosensor based on glucose oxidase entrapped into osmiumcomplex modified polypyrrole films, Sens. Actuat. B 79 (2001) 150–156. [9] E. Kopetzki, K. Lehnert, P. Buckel, Enzymes in diagnostics— achievements and possibilities of recombinant-DNA technology, Clin. Chem. 40 (1994) 688–704. [10] A. Tsuji, M. Maeda, H. Arakawa, S. Shimizu, T. Ikegami, Y. Sudo, H. Hosoda, T. Nambara, Fluorescence and chemiluminescence enzyme immunoassays of 17-alpha-hydroxyprogesterone in dried blood spotted on filter-paper, J. Steroid Biochem. Mol. Biol. 27 (1987) 33–40. [11] K. Kojima, A. Hiratsuka, H. Suzuki, K. Yano, K. Ikebukuro, I. Karube, Electrochemical protein chip with arrayed immunosensors with antibodies immobilized in a plasma-polymerized film, Anal. Chem. 75 (2003) 1116–1122.
[12] D.B. Papkovsky, T.C. OÕRiordan, G.G. Guilbault, An immunosensor based on the glucose oxidase label and optical oxygen detection, Anal. Chem. 71 (1999) 1568–1573. [13] K. Tohda, M. Gratzl, A microscopic, continuous, optical monitor for interstitial electrolytes and glucose, ChemPhysChem 4 (2003) 155–160. [14] G.G. Guilbault, Paul Brignac Jr., Mary Zimmer, Homovanillic acid as a fluorometric substrate for oxidative enzymes, analytical applications of the peroxidase, glucose oxidase, and xanthine oxidase systems, Anal. Chem. 40 (1968) 190. [15] N.J. Emptage, Fluorescent imaging in living systems, Curr. Opin. Pharmacol. 1 (2001) 521–525. [16] R.J. Gillies, In vivo molecular imaging, J. Cell. Biochem. (Suppl.) 39 (2002) 231–238. [17] K.S. Rastogi, R.L. Cooper, Z.Q. Shi, M. Vranic, Quantitative measurement of islet glucagon response to hypoglycemia by confocal fluorescence imaging in diabetic rats: effects of phlorizin treatment, Endocrine 7 (1997) 367–375. [18] Q.Q. Ruan, Y. Chen, E. Gratton, M. Glaser, W.W. Mantulin, Cellular characterization of adenylate kinase and its isoform: twophoton excitation fluorescence imaging and fluorescence correlation spectroscopy, Biophys. J. 83 (2002) 3177–3187. [19] J.M. Kim, T. Ohtani, S. Sugiyama, T. Hirose, H. Muramatsu, Simultaneous topographic and fluorescence imaging of single DNA molecules for DNA analysis with a scanning near-field optical/ atomic force microscope, Anal. Chem. 73 (2001) 5984–5991. [20] H. Szmacinski, J.R. Lakowicz, Fluorescence lifetime-based sensing and imaging, Sens. Actuat. B 29 (1995) 16–24. [21] R. Cubeddu, D. Comelli, C. DÕAndrea, P. Taroni, G. Valentini, Time-resolved fluorescence imaging in biology and medicine, J. Phys. D 35 (2002) R61–R76. [22] M. Niculescu, S. Gaspar, A. Schulte, E. Cso¨regi, W. Schuhmann, Visualization of micropatterned complex biosensor sensing chemistries by means of scanning electrochemical microscopy, Biosens. Bioelectron. 19 (2004) 1175–1184. [23] C. Zhao, G. Wittstock, An SECM detection scheme with improved sensitivity and lateral resolution: detection of galactosidase activity with signal amplification by glucose dehydrogenase, Angew. Chem. Int. Ed. 43 (2004) 4170–4172. [24] S. Kasai, H.F. Zhou, T. Matsue, Chemiluminescence imaging of localized enzymes by scanning chemiluminescence microscopy, Chem. Lett. (2000) 200–201. [25] S. Szucs, G. Vamosi, R. Poka, A. Sarvary, H. Bardos, M. Balazs, J. Kappelmayer, L. Toth, J. Szo¨llo¨si, R. Adany, Single-cell measurement of superoxide anion and hydrogen peroxide production by human neutrophils with digital imaging fluorescence microscopy, Cytometry 33 (1998) 19–31. [26] M. Pazdzioch-Czochra, A. Widenska, Spectrofluorimetric determination of hydrogen peroxide scavenging activity, Anal. Chim. Acta 452 (2002) 177–184. [27] K.S. Gould, J. McKelvie, K.R. Markham, Do anthocyanins function as antioxidants in leaves? Imaging of H2O2 in red and green leaves after mechanical injury, Plant Cell Environ. 25 (2002) 1261–1269. [28] T. Johansson, M. Petersson, J. Johansson, S. Nilsson, Real-time imaging through optical fiber array-assisted laser-induced fluorescence of capillary electrophoretic enantiomer separations, Anal. Chem. 71 (1999) 4190–4197. [29] J.W. Borst, M.A. Uskova, N.V. Visser, A.J.W.G. Visser, in: R. Kraayenhof (Ed.), Springer Series on Fluorescence, vol. 2, Springer, Berlin-Heidelberg, 2002, pp. 337–348. [30] L.M. Henderson, J.B. Chappell, Diydrorhodamine 123—a fluorescent-probe for superoxide generation, Eur. J. Biochem. 217 (1993) 973–980. [31] O.S. Wolfbeis, A. Du¨rkop, M. Wu, Z. Lin, A Europium-ionbased luminescent sensing probe for hydrogen peroxide, Angew. Chem. Int. Ed. 41 (2002) 4495–4498.
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73 [32] M. Wu, Z. Lin, A. Du¨rkop, O.S. Wolfbeis, Time-resolved enzymatic determination of glucose using a fluorescent europium probe for hydrogen peroxide, Anal. Bioanal. Chem. 380 (2004) 619–626. [33] O.S. Wolfbeis, M. Scha¨ferling, A. Du¨rkop, Reversible optical sensor membrane for hydrogen peroxide using an immobilized fluorescent probe, and its application to a glucose biosensor, Microchim. Acta 143 (2003) 221–227. [34] M. Scha¨ferling, M. WU, O.S. Wolfbeis, Time-resolved fluorescent imaging of glucose, J. Fluoresc. 14 (2004) 561–568. [35] M. Scha¨ferling, M. WU, J. Enderlein, H. Bauer, O.S. Wolfbeis, Time-resolved luminescence imaging of hydrogen peroxide using sensor membranes in a microwell format, Appl. Spectrosc. 57 (2003) 1386–1392. [36] G. Liebsch, I. Klimant, C. Krause, O.S. Wolfbeis, Fluorescent imaging of pH with optical sensors using time domain dual lifetime referencing, Anal. Chem. 73 (2001) 4354–4363. [37] Y. Rakicioglu, J.H. Perrin, S.G. Schulman, Increased luminescence of the tetracycline-europium(III) system following oxidation by hydrogen peroxide, J. Pharm. Biomed. Anal. 20 (1999) 397– 399. [38] L.M. Hirschy, T.F. Van Geel, J.D. Winefordner, R.N. Kelly, S.G. Schulman, Characteristics of the binding of europium (III) to tetracycline, Anal. Chim. Acta 166 (1984) 207–219. [39] P. Hartmann, W. Ziegler, Lifetime imaging of luminescent oxygen sensors based on all-solid-state technology, Anal. Chem. 68 (1996) 4512–4514. [40] K.K. Sharman, A. Periasamy, H. Ashworth, J.N. Demas, N.H. Snow, Error analysis of the rapid lifetime determination method for double-exponential decays and new windowing schemes, Anal. Chem. 71 (1999) 947–952. [41] R.J. Woods, S. Scypinski, L.J.C. Love, Transient digitizer for the determination of microsecond luminescence lifetimes, Anal. Chem. 56 (1984) 1395–1400.
73
[42] D.J. Desilets, J.T. Coburn, D.A. Lantrip, P.T. Kissinger, F.E. Lytle, On-the-fly determination of fluorescence lifetimes from twopoint decay measurements, Anal. Chem. 58 (1986) 1123–1128. [43] S.P. Chan, Z.J. Fuller, J.N. Demas, B.A. DeGraff, Optimized gating scheme for rapid lifetime determinations of single-exponential luminescence lifetimes, Anal. Chem. 73 (2001) 4486–4490. [44] D.B. Papkovsky, G.V. Ponomarev, W. Trettnak, P. OÕLeary, Phosphorescent complexes of porphyrin ketones: optical properties and application to oxygen sensing, Anal. Chem. 71 (1999) 1568–1573. [45] A.M. Pelkkikangas, S. Jaakohuhta, T. Lo¨vgren, H. Harma, Simple, rapid, and sensitive thyroid-stimulating hormone immunoassay using europium(III) nanoparticle label, Anal. Chim. Acta 517 (2004) 169–176. [46] M.E. Waris, N.J. Meltola, J.T. Soini, E. Soini, O.J. Peltola, P.E. Hanninen, Two-photon excitation fluorometric measurement of homogeneous microparticle immunoassay for C-reactive protein, Anal. Biochem. 309 (2002) 67–74. [47] A. Beeby, S.W. Botchway, I.M. Clarkson, S. Faulkner, A.W. Parker, D. Parker, J.A.G. Williams, Luminescence imaging microscopy and lifetime mapping using kinetically stable lanthanide(III) complexes, J. Photochem. Photobiol. B 57 (2000) 83– 89. [48] S. Kasai, Y. Hirano, N. Motochi, H. Shiku, M. Nishizawa, T. Matsue, Simultaneous detection of uric acid and glucose on a dual-enzyme chip using scanning electrochemical microscopy/ scanning chemiluminescence microscopy, Anal. Chim. Acta 458 (2002) 263–270. [49] M. Delvaux, S. Demoustier-Champagne, Immobilisation of glucose oxidase within metallic nanotubes arrays for application to enzyme biosensors, Biosens. Bioelectron. 18 (2003) 943–951. [50] S. Gaspar, W. Schuhmann, T. Laurell, E. Cso¨regi, Design, visualization, and utilization of enzyme microstructures built on solid surfaces, Rev. Anal. Chem. 21 (2002) 245–266.
Fluorescence imaging of the activity of glucose oxidase using a hydrogen-peroxide-sensitive europium probe Meng Wu, Zhihong Lin, Michael Scha¨ferling, Axel Du¨rkop, Otto S. Wolfbeis * Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg, D-93040 Regensburg, Germany Received 16 November 2004
Abstract A method for optical imaging of the activity of glucose oxidase (GOx) using a fluorescent europium(III) tetracycline probe for hydrogen peroxide is presented. A decay time in the microsecond range and the large Stokes shift of 210 nm of the probe facilitate intensity-based, time-resolved, and decay-time-based imaging of glucose oxidase. Four methods for imaging the activity of GOx were compared, and rapid lifetime determination imaging was found to be the best in giving a linear range from 0.32 to 2.7 mUnit/mL. The detection limit is 0.32 mUnit/mL (1.7 ng mL1) which is similar to that of the time-resolved (gated) imaging using a microtiterplate reader. Fluorescent imaging of the activity of GOx is considered to be a useful tool for GOx-based immunoassays with potential for high-throughput screening, immobilization studies, and biosensor array technologies. Ó 2005 Elsevier Inc. All rights reserved. Keywords: Glucose oxidase; Enzyme imaging; Fluorescent probe; Time-resolved imaging; Enzyme activity assay
Glucose oxidase (GOx, EC 1.1.3.4)1 is a flavoenzyme that catalyzes specifically the oxidation of glucose into gluconate and hydrogen peroxide Eq. (1) [1]. It is one of the most widely used enzymes due to its availability and stability. As a representative of oxidases, it has been studied in solution and in the immobilized state on solid surfaces [2] or encapsulated in microspheres [3] for applications in biosensors [4], industrial bioreactors [5], and bio-fuel cells [6]. Apo-glucose oxidase has also been utilized in studies for ‘‘nanowiring’’ [7] of the electron transfer. GOx has been applied not only to the electrochemical and optical detection of glucose [8] but also *
Corresponding author. Fax: +49 941 943 4064. E-mail address: [email protected] (O.S. Wolfbeis). 1 Abbreviations used: GOx, glucose oxidase; EuTc, europium–tetracycline complex; EuTc-HP, europium–tetracycline–hydrogen peroxide; HP, hydrogen peroxide; Tc, tetracycline; FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging; POx, peroxidase; CCD, charge-coupled device; LED, light-emitting diode. 0003-2697/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2005.01.050
in GOx-labeled enzyme amplification analysis for immunoassays [9]. GOx-based enzyme immunoassays have been used for the detection and screening of, e.g., steroids, drugs, environmental pollutants, and special peptides [10]. Recently, GOx has been applied to microchip or microarray technologies for the development of protein chips for proteomics [11].
ð1Þ Numerous methods for the determination of GOx have been developed. A set of assays have emerged based on the measurement of oxygen [12], pH [13], or H2O2 [14]. However, methods based on the measurement of H2O2 produced by GOx are of particular interest because H2O2 does not form a background but is produced only during enzymatic reaction. Since H2O2 cannot be easily visualized, it is usually converted into a colored or fluorescent product using peroxidase (POx).
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
In recent years, fluorescent imaging technologies have attracted substantial attention because of the multitude of information on both the spatial and the temporal characterization of the target analyte [15,16]. Confocal [17], multiphoton excitation [18], near-field [19], and decaytime-based imaging [20,21] technologies have been developed among others. Existing methods for imaging GOx are restricted mainly to scanning electrochemical microscopy [22,23] and scanning chemiluminescence imaging [24]. This is largely due to the fact that almost all fluorescence imaging methods for GOx need a second enzyme (POx) to visualize the H2O2 produced by the GOx-catalyzed reaction. There have been reports on H2O2-based imaging using dihydrorhodamine 123 [25], homovanillic acid [26], scopoletin [27], dichlorodihydrofluorescein diacetate [28] or others [29], all using POx along with a second (chromogenic or fluorogenic) substrate [30]. Based on our previous reports on the europium–tetracycline (EuTc) fluorescent probe for determination of either H2O2 [31] or glucose either in solution [32] or using biosensor membranes [33,34] we present here a scheme for imaging GOx at neutral pH and without the need for a second enzyme. It also is shown that, in addition to intensity-based imaging of GOx, the long luminescence decay time of the probe enables time-resolved and lifetime-based imaging.
Materials and methods Materials and reagents
67
Austria; www.tecan.com). The excitation filter was set to 405 nm and the emission filter to 612 nm. All experiments were performed at a programmed temperature of 30.0 ± 0.1 °C. Either black Fluotrac200 microtiterplates or black plates with a flat transparent bottom (Greiner bio-one, Frickenhausen, Germany; www.greiner-lab.com) were used. Imaging device The device used in this study has been previously reported [35,36] and was used with minor modifications. As shown schematically in Fig. 1, it is composed of a fast gateable CCD camera (SensiMod; from PCO, Kelheim, Germany; www.pco.de), a pulsable LED array with 96 UV light-emitting diodes (kmax 405 nm, Roithner Laser Technik, Vienna, Austria; www.roithner-laser.com), a pulse generator (Scientific Instruments DG 535, Sunnyvale, CA, USA; www.srsys.com; not shown in Fig. 1), an optical emission filter (KV 550; Schott, Mainz, Germany; www.schott.com), and an optical excitation filter (BG 12; Schott), with a light-guiding adapter consisting of 96 optical fibers (diameter 3 mm) for matching the focus of the CCD camera. A computer was used for control and visualization of the experiments that were programmed in Interactive Data Language (IDL; from Research Systems, Boulder, CO, USA; www.rsinc.com). The manipulation and calculation of the images, such as the rotation and cropping of the images, the subtraction of the dark image (blank, without illumination) from the fluorescent image, and the integration of the
All solutions were prepared in a 10 mmol L1 3-(Nmorpholino)propanesulfonate (Mops) buffer of pH 6.9. Glucose oxidase (from Aspergillus niger; 185,000 Unit/ g; used without further purification) was from Sigma– Aldrich (Steinheim, Germany) and tetracycline hydrochloride was from Serva (Heidelberg, Germany; www.serva.de). The activity of GOx is defined by the provider as follows: one unit will oxidize 1.0 lmol of b-D-glucose to D-gluconolactone and H2O2 per min at pH 5.1 and 35 °C. EuCl3 hexahydrate was from Alfa Products (Danvers, MA, USA; www.alfa.com). The EuTc stock solution was prepared by mixing 10 mL of a 6.3 mmol L1 EuCl3 solution with 10 mL of a 2.1 mmol L1 tetracycline solution and diluting it to 100 mL with Mops buffer (the molar ratio of Eu3+ to Tc being 3:1). The glucose stock solution (0.277 mol L1) was stored overnight before use to warrant the equilibration of a and b anomers. Instruments Fluorescence measurements and time-resolved (gated) intensity detection were performed on a GENios+ microtiter plate reader (Tecan, Gro¨dig, Salzburg,
Fig. 1. Setup of the imaging system. (A) Fast gateable CCD camera; (B) optical emission filter (KV 550); (C) light-guiding adapter consisting of 96 optical fibers (diameter 3 mm); (D) 96-well microtiterplate (black with transparent bottom); (E) optical excitation filter (BG 12); (F) pulsable LED array with 96 UV light-emitting diodes (kmax = 405 nm).
68
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
single images and their ratioing, were done by a self-developed program based on Matlab (6.1; Mathwork, Natick, MA, USA; www.mathwork.com). The images of the ratio from decay-time-based detection were inverted to their negatives for a more facile comparison of different methods. Imaging of the activity of GOx To each well of a 96-well microtiter plate with transparent bottom was added 100 lL of the EuTc standard solution, 15 lL of a 277 mmol L1 glucose solution, and enough Mops buffer to make up the total volume to 200 lL. Phosphate buffer should be avoided since it quenches the fluorescence of the EuTc–H2O2 complex (refered to as EuTc-HP). The samples (50 lL) containing GOx with an activity between 1.2 and 5.6 mUnit/ mL were added simultaneously. The images of the plates were taken after a 30-min incubation at 30 °C. Blank values (F0) were obtained by adding buffer in place of GOx. In the time-resolved (gated) nonimaging determination using a commercially available microwell plate fluorescence reader, a lag time of 60 ls and an integration time of 40 ls were employed [33].
Fig. 2. Kinetic response of the glucose/EuTc system to increasing activities of glucose oxidase; 100 lL of a the EuTc stock solution, 15 lL of a 277.2 mmol L1 glucose solution, and Mops buffer made up to a total volume of 200 lL. Glucose oxidase activities (from A to G): 0.0, 0.27, 0.54, 1.4, 2.7, 5.4, and 13.5 mUnit/mL, respectively.
bate, uric acid, phosphate, and citrate interfere [33]. The baseline has a slight downward drift which is due to a slight increase in temperature at the beginning of the experiment (following addition of enzyme solution). The luminescence decay profiles of EuTc and EuTc-HP
Results The scheme The assay for GOx activity via H2O2 is based on the response of EuTc to enzymatically produced H2O2 to form the EuTc-HP complex according to ð2Þ It is interesting to note that the stoichiometry (Eu:Tc) giving the largest signal changes with H2O2 is 3:1. Like others [37], we were unable to further elucidate the chemical structure of EuTc since it cannot be isolated in solid form. It is known, however, that HP rather weakly binds to EuTc and that binding is highly pH dependent [38]. It therefore may be more appropriate to formulate the equilibrium as [Eu3Tc + n H2O2 M Eu3Tc-HP + (n 1) HP]. The kinetic response of the glucose/EuTc system to increasing activities of GOx is shown in Fig. 2. With higher activities of GOx, and therefore increasing production of H2O2, the fluorescence (more correctly the luminescence) of the Eu complex increases due to the formation of more strongly fluorescent EuTc-HP. The intensity of the fluorescence of EuTc-HP is around 15fold higher than that of EuTc, and its quantum yield is around 4%. Glucose and GOx themselves have no significant effect on either EuTc or EuTc-HP, while ascor-
The probe EuTc is converted into EuTc-HP by enzymatically produced hydrogen peroxide, and this results in an up to 15-fold increase in fluorescence intensity. At the same time the decay times (s) undergo a significant change [32,33]. The rather complex decay profile of EuTc 3:1 complex in buffer solution of pH 6.9 is composed of three components, with decay times of 8.7 ls (58% contribution), 30.4 ls (40%), and 174 ls (2%). The EuTc-HP complex decays in a similar fashion, the decay times being 13.2 ls (34%), 59 ls (64%), and 158 ls (2%). Details have been reported and discussed in recent papers [31,35]. The multiple decay is likely to be due to three activation channels of the same species rather than to several chemical species. First, only EuTc is detected in this optical system (at an excitation wavelength of 400 nm and the emission being collected at 616 nm). Second, the decay profiles do not strongly vary with the molar ratio of Eu:Tc (checked for ratios between 3:1 and 1:3). The v2 values after triple exponential fit are 1.55 for the EuTc complex and 1.46 for EuTc-HP which in fact is excellent. The data reveal that time-resolved measurements are best performed after a lag time of >60 ls to selectively detect the 59-ls component of the EuTc-HP complex (with its stronger intensity) and—at the same time—suppress the shorter components (of 8.7 and 30.4 ls) of the EuTc complex and the 13.2-ns component of EuTc-HP. Lag times of 60 ls can be adjusted in most present day fluorescence (and microplate) readers. The difference
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
also can be seen by calculating the average decay times of EuTc (30 ls) and EuTc-HP (60 ls). Both time-resolved and decay-time-based determination and imaging become possible. We have performed assays in both formats. Time-resolved determination of GOx using a microplate reader First, a time-gated fluorescent assay for GOx activities was worked out for microtiterplate readers. Gated detection with a lag time of 60 ls and an integration time of 40 ls is adequate to overcome the background fluorescence of the samples and plates which usually have nanosecond lifetimes. The reader assay resulted in a linear range for the GOx assay from 0.32 to 14 mUnit/mL, with a limit of detection (S/N = 3) of 0.32 mUnit/mL. The reader assay usually has a broader dynamic range and smaller standard deviations compared to those of the CCD imaging method. Detection schemes for imaging Next, we transferred the previous results to an imaging assay. The four imaging schemes shown in Fig. 3 were tested for their suitability for quantitative imaging of GOx activity. Their features are summarized in Table 1. In conventional fluorescence intensity imaging (FII) [29] one detection window is opened while excitation is
69
on. In contrast to this often applied method, in time-resolved imaging (TRI) [39] the detection window is opened only after the excitation pulse. TRI imaging technology is most useful to eliminate (by gating) the autofluorescence of samples and plates with their nanosecond decay times. It should be kept in mind that both FII and TRI are intensity-based detection schemes. Methods for luminescence decay-time-based imaging (in contrast to FII and TRI) have certain advantages since they are independent of (a) light source and detector fluctuations, (b) concentration of the fluorophore, Table 1 Time programs of the four imaging schemes studied in this work (all times in ls) Methoda
FII
TRI
PDR
RLD
Excitation ti tt
0 50
0 50
0 50
0 50
Length of window 1 (=W1) ti tt
0 50
80 130
0 50
80 130
Length of window 2 (= W2) ti tt
— —
— —
100 150
180 300
b
a FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging. b ti, initiation time; tt, terminating time (all in ls); see Fig. 3.
Fig. 3. Schematic of the principles of the four imaging schemes. Timelines (i.e., the gating of windows W1 and W2) are indicated in Table 1.
70
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
and (c) light scatter. The commonly used techniques for decay-time imaging are utilized here with a two-window detection scheme (see Fig. 3) and ratiometric quantification scheme as shown in Eq. (3): R¼
W 1 W dark W 2 W dark
image
:
ð3Þ
here. It should be noted, though, that for multiexponentially decaying systems such as EuTc-HP the precision in the determination of the average lifetimes strongly depends on the selection of the detection windows [40]. Quantitative aspects of GOx imaging
image
Two schemes were tested. In phase delay rationing imaging (PDR) [36,37] two windows are acquired, one in the excitation phase and one in the phase after excitation. The ratio of the two windows can be used for the quantification of the activities of GOx. Rapid lifetime determination imaging (RLD), in contrast, is based on the acquisition of two windows after excitation [40,41], the ratio of intensities of the two windows enabling a straightforward determination of decay time. Moreover, RLD—in being a gated method also—can efficiently suppress background ns fluorescence. The resulting intensity-based and lifetime-based GOx images are displayed in Fig. 4. They indicate that the fluorescent probe EuTc is applicable to all four schemes. The intensity-based images (FII, see Fig. 4A; TRI, see Fig. 4B) are affected by flash-to-flash variations in intensity and by variations in the brightness from LED to LED in the battery of light sources (Fig. 1). This results in high well-to-well standard deviations. The high inhomogeneities in the single wells result from scattering effects. The lifetime-based images (PDR and RLD, see Figs. 4C and 4D), in contrast, reveal good homogeneity due to their independence of intensity variations of the excitation light, the quantity of the fluorophore present in the wells, and that of scattered light. Cutoff optical filters and a common ns pulse generator are adequate for the optimal time-resolved imaging of GOx. Previous studies have shown that ratiometric methods such as RLD are suitable for determination of fluorescence lifetime [42], especially when using fluorescent probes displaying decay times in the ls range. While the RLD method is intended for use in monoexponentially decaying systems [43], it is found also to work quite well in the multiexponential decay system used
The RLD scheme was optimized with respect to the concentration of glucose and the concentration of EuTc. The optimal response (i.e., the largest change in fluorescence intensity) was obtained when using 20 lL of the glucose stock solution and 95 lL of the EuTc stock solution, with 200 lL total volume in the well of the microplate. The steady state (endpoint) method was used since the reaction reaches the endpoint after a 30-min incubation, as shown in Fig. 2. The results show that different activities of GOx clearly can be differentiated and result in the different intensities (and decay times) of EuTc-HP. It should be noted that kinetic imaging of GOx also is conceivable. However, it is less favored because it is critically dependent on precise timing. The rapid lifetime determination scheme was further exploited for quantitative imaging of the activity of GOx. The normalized ratio of the two images is shown
Fig. 5. Rapid lifetime determination imaging of the activity of glucose oxidase. The pictures reflect the normalized ratio of the two images according to Eq. (3) (in pseudo colors; see the color bar). All experiments were performed in triplicate (rows). The cocktails in wells (from 1 to 12) had the following compositions: GOx activities 0 (blank), 135, 54.1, 27.1, 13.5, 5.4, 2.7, 1.35, 0.54, 0.27, 0.14, 0.05 mUnit/mL, respectively; plus 100 lL of the EuTc stock solution, 15 lL of a 277.2 mmol L1 glucose solution; total volume made up to 200 lL with Mops buffer.
Fig. 4. Comparison of the images obtained for the activity of glucose oxidase by the four methods (after data processing). (A) Fluorescence intensity imaging (FII); (B) time-resolved imaging (TRI); (C) phase delay imaging (PDR); (D) rapid lifetime imaging (RLD). Experimental parameters are specified in Table 1; those for data processing are specified under Materials and methods. The solutions in the wells (from left to right in duplicate) contained GOx (135, 54.1, 27.1, 13.5, 5.4, 2.7, 1.35, 0.54, 0.27, 0.14, 0.05, and 0 mUnit/mL, respectively), 100 lL of the EuTc stock solution, and 15 lL of a 277.2 mmol L1 glucose solution. The total volume of 200 lL was made up with Mops buffer.
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
71
Fig. 6. Calibration plot for the determination of the activity of GOx by rapid lifetime determination imaging. The right chart is the calibration for the linear (= initial) range. Experimental conditions as in Fig. 5.
in pseudo colors in Fig. 5. The results show that GOx can be quantified easily via the averaged values of the gray scale images as shown in Fig. 6 which gives a quantitative plot. The response is linear from 0.32 to 2.7 mUnit/mL of GOx, with a detection limit of about 0.32 mUnit/mL (1.7 ng mL1). The GOx activity has been validated independently via determination of H2O2 by the standard peroxidase/o-dianisidine method (Sigma) and gave excellent agreement. Under our experimental conditions saturation was observed at around 20 mUnit/mL which is due to complete consumption of glucose and not to consumption of oxygen. The assay was optimized for low enzyme levels. If higher levels are to be determined, the concentration of glucose has to be increased accordingly. Figures of merit of the methods employed in this work are summarized in Table 2. It shows that RLD is the optimal choice for the quantitative imaging of GOx activity since it has the largest linear range at comparable limits of detection for GOx. In addition, RLD is a fairly straightforward method that requires a moderately sophisticated instrumental effort. Thus, the method of choice for bioanalytical applications is the RLD method because it is best with regard to sensitivity, dynamic range, and reproducibility. It also is Table 2 Figures of merit for the determination of GOx by the four imaging schemes studied in this work Methoda b
Limit of detection Linear rangeb Standard deviationc Remarks
FII
TRI
PDR
RLD
12.8 12.8–27.1 49
3.1 3.1–27.1 33
0.35 0.35–0.54 6
0.32 0.32–2.7 7
d
d
a FII, fluorescence intensity imaging; TRI, time-resolved imaging; PDR, phase delay rationing imaging; RLD, rapid lifetime determination imaging. b mUnit/mL. c % at 2.7 Unit/mL. d The black/white images were reversed for better comparison of all images.
capable of suppressing interference by stray light and ambient light due to internal signal referencing. This makes RLD superior to both intensity-based imaging schemes. It should be noted that, to make data comparable, the pictures have to be processed first. This results from the fact that in FII and TRI an increase in GOx activity results in an increase in luminescence intensity, while in PDR and RLD it results in a decrease of the ratios of the two windows. Since such images cannot be compared, those of PDR and RLD were inversed (i.e., converted into their ‘‘negatives’’) to make them comparable.
Discussion To the best of our knowledge a decay-time-based scheme for detection or imaging of oxidases via H2O2 has not been reported so far. The activity of GOx has been determined in solution via oxygen consumption [44] by measurement of the phase shift of the luminescence of an oxygen-quenchable probe. The scheme here is more promising insofar as the activity is measured via the production of H2O2which—unlike oxygen—does not form a strong and, in the worst case, varying background, but rather is completely absent in most samples. Lanthanide-based labels and microspheres have been used in time-resolved (gated) immunoassays [45–47] due to long decay times which are on the order of ls to ms. In the scheme reported here, dissolved EuTc is used as an H2O2-sensitive probe. There is no need for reagents other than glucose and EuTc. The methods for imaging GOx activity reported here include both intensity-based and decay-time-based approaches. The probe used (EuTc) is easily accessible, has fairly high sensitivity to H2O2, a large StokesÕ shift (210 nm), and a line-like emission peaking at 616 nm, is compatible with commercial shortwave diode laser light sources (with lines at 375 and 405 nm, respectively) and stable toward oxygen, and has a working pH of 7.
72
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73
The methods presented here may be applied to enzyme assays and to enzyme-based (and amplification) immunoassays. While demonstrated here for GOx only, it most likely can be extended to image other oxidases such as uricase. Another application may be in highthroughput screening. Furthermore, imaging of GOx (and, conceivably, other oxidases) offers the possibility for analyzing multienzyme arrays or multisubstrate arrays [48]. The capability of spatial discrimination also makes imaging of interest to assess enzyme activity in immobilization studies [49,50] such as those testing the quality of glucose biosensors based on GOx.
Acknowledgments We acknowledge the help of Jo¨rg Enderlein of Forschungszentrum Ju¨lich, Germany and of Henrik Bauer of PicoQuant GmbH, Berlin in some of the decay time measurements. M.W. and Z.L. acknowledge financial support from Chromeon GmbH (Regensburg).
References [1] J. Raba, H.A. Mottola, Glucose oxidase as an analytical reagent, Crit. Rev. Anal. Chem. 25 (1995) 1–42. [2] M. Delvaux, S. Demoustier-Champagne, Immobilisation of glucose oxidase within metallic nanotubes arrays for application to enzyme biosensors, Biosens. Bioelectron. 18 (2003) 943–951. [3] S. Slomkowski, B. Miksa, New types of microspheres and microsphere-related materials for medical diagnostics, Polymers Adv. Tech. 13 (2002) 906–918. [4] L.Q. Chen, X.E. Zhang, W.H. Xie, Y.F. Zhou, Z.P. Zhang, A.E.G. Cass, Genetic modification of glucose oxidase for improving performance of an amperometric glucose biosensor, Biosens. Bioelectron. 17 (2002) 851–857. [5] M. Traeger, G.N. Qazi, U. Onken, C.L. Chopra, Contribution of endocellular glucose oxidase and exocellular glucose oxidase to gluconic acid production at increased dissolved-oxygen concentrations, J. Chem. Tech. Biotechnol. 50 (1991) 1–11. [6] S. Tsujimura, K. Kano, T. Ikeda, Glucose/O-2 biofuel cell operating at physiological conditions, Electrochem. 70 (2002) 940–942. [7] Y. Xiao, F. Patolsky, E. Katz, J.F. Hainfeld, I. Willner, Plugging into enzymes: nanowiring of redox enzymes by a gold nanoparticle, Science 299 (2003) 1877–1881. [8] S. Reiter, K. Habermu¨ller, W. Schuhmann, A reagentless glucose biosensor based on glucose oxidase entrapped into osmiumcomplex modified polypyrrole films, Sens. Actuat. B 79 (2001) 150–156. [9] E. Kopetzki, K. Lehnert, P. Buckel, Enzymes in diagnostics— achievements and possibilities of recombinant-DNA technology, Clin. Chem. 40 (1994) 688–704. [10] A. Tsuji, M. Maeda, H. Arakawa, S. Shimizu, T. Ikegami, Y. Sudo, H. Hosoda, T. Nambara, Fluorescence and chemiluminescence enzyme immunoassays of 17-alpha-hydroxyprogesterone in dried blood spotted on filter-paper, J. Steroid Biochem. Mol. Biol. 27 (1987) 33–40. [11] K. Kojima, A. Hiratsuka, H. Suzuki, K. Yano, K. Ikebukuro, I. Karube, Electrochemical protein chip with arrayed immunosensors with antibodies immobilized in a plasma-polymerized film, Anal. Chem. 75 (2003) 1116–1122.
[12] D.B. Papkovsky, T.C. OÕRiordan, G.G. Guilbault, An immunosensor based on the glucose oxidase label and optical oxygen detection, Anal. Chem. 71 (1999) 1568–1573. [13] K. Tohda, M. Gratzl, A microscopic, continuous, optical monitor for interstitial electrolytes and glucose, ChemPhysChem 4 (2003) 155–160. [14] G.G. Guilbault, Paul Brignac Jr., Mary Zimmer, Homovanillic acid as a fluorometric substrate for oxidative enzymes, analytical applications of the peroxidase, glucose oxidase, and xanthine oxidase systems, Anal. Chem. 40 (1968) 190. [15] N.J. Emptage, Fluorescent imaging in living systems, Curr. Opin. Pharmacol. 1 (2001) 521–525. [16] R.J. Gillies, In vivo molecular imaging, J. Cell. Biochem. (Suppl.) 39 (2002) 231–238. [17] K.S. Rastogi, R.L. Cooper, Z.Q. Shi, M. Vranic, Quantitative measurement of islet glucagon response to hypoglycemia by confocal fluorescence imaging in diabetic rats: effects of phlorizin treatment, Endocrine 7 (1997) 367–375. [18] Q.Q. Ruan, Y. Chen, E. Gratton, M. Glaser, W.W. Mantulin, Cellular characterization of adenylate kinase and its isoform: twophoton excitation fluorescence imaging and fluorescence correlation spectroscopy, Biophys. J. 83 (2002) 3177–3187. [19] J.M. Kim, T. Ohtani, S. Sugiyama, T. Hirose, H. Muramatsu, Simultaneous topographic and fluorescence imaging of single DNA molecules for DNA analysis with a scanning near-field optical/ atomic force microscope, Anal. Chem. 73 (2001) 5984–5991. [20] H. Szmacinski, J.R. Lakowicz, Fluorescence lifetime-based sensing and imaging, Sens. Actuat. B 29 (1995) 16–24. [21] R. Cubeddu, D. Comelli, C. DÕAndrea, P. Taroni, G. Valentini, Time-resolved fluorescence imaging in biology and medicine, J. Phys. D 35 (2002) R61–R76. [22] M. Niculescu, S. Gaspar, A. Schulte, E. Cso¨regi, W. Schuhmann, Visualization of micropatterned complex biosensor sensing chemistries by means of scanning electrochemical microscopy, Biosens. Bioelectron. 19 (2004) 1175–1184. [23] C. Zhao, G. Wittstock, An SECM detection scheme with improved sensitivity and lateral resolution: detection of galactosidase activity with signal amplification by glucose dehydrogenase, Angew. Chem. Int. Ed. 43 (2004) 4170–4172. [24] S. Kasai, H.F. Zhou, T. Matsue, Chemiluminescence imaging of localized enzymes by scanning chemiluminescence microscopy, Chem. Lett. (2000) 200–201. [25] S. Szucs, G. Vamosi, R. Poka, A. Sarvary, H. Bardos, M. Balazs, J. Kappelmayer, L. Toth, J. Szo¨llo¨si, R. Adany, Single-cell measurement of superoxide anion and hydrogen peroxide production by human neutrophils with digital imaging fluorescence microscopy, Cytometry 33 (1998) 19–31. [26] M. Pazdzioch-Czochra, A. Widenska, Spectrofluorimetric determination of hydrogen peroxide scavenging activity, Anal. Chim. Acta 452 (2002) 177–184. [27] K.S. Gould, J. McKelvie, K.R. Markham, Do anthocyanins function as antioxidants in leaves? Imaging of H2O2 in red and green leaves after mechanical injury, Plant Cell Environ. 25 (2002) 1261–1269. [28] T. Johansson, M. Petersson, J. Johansson, S. Nilsson, Real-time imaging through optical fiber array-assisted laser-induced fluorescence of capillary electrophoretic enantiomer separations, Anal. Chem. 71 (1999) 4190–4197. [29] J.W. Borst, M.A. Uskova, N.V. Visser, A.J.W.G. Visser, in: R. Kraayenhof (Ed.), Springer Series on Fluorescence, vol. 2, Springer, Berlin-Heidelberg, 2002, pp. 337–348. [30] L.M. Henderson, J.B. Chappell, Diydrorhodamine 123—a fluorescent-probe for superoxide generation, Eur. J. Biochem. 217 (1993) 973–980. [31] O.S. Wolfbeis, A. Du¨rkop, M. Wu, Z. Lin, A Europium-ionbased luminescent sensing probe for hydrogen peroxide, Angew. Chem. Int. Ed. 41 (2002) 4495–4498.
Fluorescence imaging of the activity of glucose oxidase / M. Wu et al. / Anal. Biochem. 340 (2005) 66–73 [32] M. Wu, Z. Lin, A. Du¨rkop, O.S. Wolfbeis, Time-resolved enzymatic determination of glucose using a fluorescent europium probe for hydrogen peroxide, Anal. Bioanal. Chem. 380 (2004) 619–626. [33] O.S. Wolfbeis, M. Scha¨ferling, A. Du¨rkop, Reversible optical sensor membrane for hydrogen peroxide using an immobilized fluorescent probe, and its application to a glucose biosensor, Microchim. Acta 143 (2003) 221–227. [34] M. Scha¨ferling, M. WU, O.S. Wolfbeis, Time-resolved fluorescent imaging of glucose, J. Fluoresc. 14 (2004) 561–568. [35] M. Scha¨ferling, M. WU, J. Enderlein, H. Bauer, O.S. Wolfbeis, Time-resolved luminescence imaging of hydrogen peroxide using sensor membranes in a microwell format, Appl. Spectrosc. 57 (2003) 1386–1392. [36] G. Liebsch, I. Klimant, C. Krause, O.S. Wolfbeis, Fluorescent imaging of pH with optical sensors using time domain dual lifetime referencing, Anal. Chem. 73 (2001) 4354–4363. [37] Y. Rakicioglu, J.H. Perrin, S.G. Schulman, Increased luminescence of the tetracycline-europium(III) system following oxidation by hydrogen peroxide, J. Pharm. Biomed. Anal. 20 (1999) 397– 399. [38] L.M. Hirschy, T.F. Van Geel, J.D. Winefordner, R.N. Kelly, S.G. Schulman, Characteristics of the binding of europium (III) to tetracycline, Anal. Chim. Acta 166 (1984) 207–219. [39] P. Hartmann, W. Ziegler, Lifetime imaging of luminescent oxygen sensors based on all-solid-state technology, Anal. Chem. 68 (1996) 4512–4514. [40] K.K. Sharman, A. Periasamy, H. Ashworth, J.N. Demas, N.H. Snow, Error analysis of the rapid lifetime determination method for double-exponential decays and new windowing schemes, Anal. Chem. 71 (1999) 947–952. [41] R.J. Woods, S. Scypinski, L.J.C. Love, Transient digitizer for the determination of microsecond luminescence lifetimes, Anal. Chem. 56 (1984) 1395–1400.
73
[42] D.J. Desilets, J.T. Coburn, D.A. Lantrip, P.T. Kissinger, F.E. Lytle, On-the-fly determination of fluorescence lifetimes from twopoint decay measurements, Anal. Chem. 58 (1986) 1123–1128. [43] S.P. Chan, Z.J. Fuller, J.N. Demas, B.A. DeGraff, Optimized gating scheme for rapid lifetime determinations of single-exponential luminescence lifetimes, Anal. Chem. 73 (2001) 4486–4490. [44] D.B. Papkovsky, G.V. Ponomarev, W. Trettnak, P. OÕLeary, Phosphorescent complexes of porphyrin ketones: optical properties and application to oxygen sensing, Anal. Chem. 71 (1999) 1568–1573. [45] A.M. Pelkkikangas, S. Jaakohuhta, T. Lo¨vgren, H. Harma, Simple, rapid, and sensitive thyroid-stimulating hormone immunoassay using europium(III) nanoparticle label, Anal. Chim. Acta 517 (2004) 169–176. [46] M.E. Waris, N.J. Meltola, J.T. Soini, E. Soini, O.J. Peltola, P.E. Hanninen, Two-photon excitation fluorometric measurement of homogeneous microparticle immunoassay for C-reactive protein, Anal. Biochem. 309 (2002) 67–74. [47] A. Beeby, S.W. Botchway, I.M. Clarkson, S. Faulkner, A.W. Parker, D. Parker, J.A.G. Williams, Luminescence imaging microscopy and lifetime mapping using kinetically stable lanthanide(III) complexes, J. Photochem. Photobiol. B 57 (2000) 83– 89. [48] S. Kasai, Y. Hirano, N. Motochi, H. Shiku, M. Nishizawa, T. Matsue, Simultaneous detection of uric acid and glucose on a dual-enzyme chip using scanning electrochemical microscopy/ scanning chemiluminescence microscopy, Anal. Chim. Acta 458 (2002) 263–270. [49] M. Delvaux, S. Demoustier-Champagne, Immobilisation of glucose oxidase within metallic nanotubes arrays for application to enzyme biosensors, Biosens. Bioelectron. 18 (2003) 943–951. [50] S. Gaspar, W. Schuhmann, T. Laurell, E. Cso¨regi, Design, visualization, and utilization of enzyme microstructures built on solid surfaces, Rev. Anal. Chem. 21 (2002) 245–266.