Genetically Encoded Stress Indicator for Noninvasively Imaging

A functional magnetic resonance. (fMRI) technology is useful for imaging neural activity associated with emotions such as neutral, sad, and anxious st...
1 downloads 0 Views 347KB Size
Download PDF
Anal. Chem. 2005, 77, 6588-6593

Genetically Encoded Stress Indicator for Noninvasively Imaging Endogenous Corticosterone in Living Mice Sung Bae Kim,† Takeaki Ozawa,†,‡ and Yoshio Umezawa*,†

Department of Chemistry, School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-Ku, Tokyo 113-0033, Japan, and Department of Molecular Structure, Institute for Molecular Science, Myodaiji, Okazaki, Aichi 444-8585, and PREST, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama, Japan

Physical and emotional stress is one of the major controllers of physiological reactions and homeostasis in living animals. A stress hormone, corticosterone, is secreted from the adrenal cortex into the blood vessels when animals sense the stress. The quantitative evaluation of corticosterone in living animals has been limited because of the unavailability of suitable methods in vivo. For a noninvasive molecular imaging of the stress, we developed a method for detecting physiological increases in the endogenous corticosterone caused by exo- and endogenous stress in living animals. We constructed a pair of genetically encoded indicators composed of cDNAs of glucocorticoid receptor (GR), split Renilla luciferase (RLuc), and a Synechocystis sp. DnaE intein. The GR fused with C-terminal halves of RLuc and DnaE is localized in the cytosol, whereas a fusion protein of N-terminal halves of RLuc and DnaE is localized in the nucleus. If corticosterone induces GR translocation into the nucleus, the C-terminal RLuc meets the N-terminal one in the nucleus, and full-length RLuc is reconstituted by protein splicing with DnaE. Cell-based methods provided a quantitative bioluminescence assay of the extent of GR translocation into the nucleus. We further demonstrated that the indicator enabled noninvasive imaging against two different types of imposed stress: a forced swimming and metabolic perturbation caused by 2-deoxy-D-glucose. This stress indicator should be valuable for screening pharmacological compounds and for tools to study the mechanism of physiological stress. Physiological responses to physical and emotional stress pass through two major systems: the neural circuits of the nervous system and the endocrine system. The activation of the endocrine system results in stimulation of synthesis of corticosterone, one of so-called stress hormones, in the adrenal cortex. Secretion of corticosterone into the blood vessel influences a variety of physiological reactions such as gluconeogenesis, deposition of * To whom correspondence should [email protected]. † The University of Tokyo. ‡ Institute for Molecular Science and PREST.

be

addressed

E-mail:

6588 Analytical Chemistry, Vol. 77, No. 20, October 15, 2005

liver glycogen, and antiinflammatory responses.1,2 These physiological reactions occur as a result of the signal transduction of corticosterone via the nuclear transport of the glucocorticoid receptor (GR) in live cells. Upon binding corticosterone, GR is transported from the cytosol into the nucleus, in which GR activates the transcription of specific target genes.3 Such a dynamic nuclear transport of GR is an initial step for regulating the magnitude and specificity of stress-related gene expressions in response to physical and emotional stress imposed on living subjects. Determination of the stress levels in living subjects relied on collection of blood samples and subsequent enzyme immunoassays.4 Such in vitro immunoassays are effective for quantifying the corticosterone levels with an automated optical-detecting system. Technical progress is now reducing laborious handling, but complex assay procedures limit spatial and temporal analysis of the corticosterone levels. A functional magnetic resonance (fMRI) technology is useful for imaging neural activity associated with emotions such as neutral, sad, and anxious states.5 The fMRI imaging gives magnetic contrasts of blood oxygen noninvasively in human brain. However, the magnetic contrasts depend on changes in blood flow and concentrations of hemoglobin, and therefore, fMRI does not allow for a direct measure of stress in living subjects.6 If a method for quantitative and noninvasive imaging of the corticosterone levels is implemented, it is therefore an important advance. To detect a stress level in living mice, we adapted an imaging method from our previous studies using reconstitution of a split Renilla luciferase (RLuc) by protein splicing.7 We newly constructed a pair of genetically encoded indicators for the noninvasive determination of a corticosterone level (Figure 1). The indicator consists of N- and C-terminal halves of a Synechocystis sp. DnaE intein and split Rluc. When the RLuc is split into two (1) de Kloet, E. R. Ann. N. Y. Acad. Sci. 2004, 1018, 1-15. (2) Sandi, C. Nat. Rev. Neurosci. 2004, 5, 917-930. (3) Htun, H.; Barsony, J.; Renyi, I.; Gould, D. L.; Hager, G. L. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 4845-4850. (4) Brooke, S. M.; Bliss, T. M.; Franklin, L. R.; Sapolsky, R. M. Neurosci. Lett. 1999, 267, 21. (5) Perlstein, W. M.; Elbert, T.; Stenger, V. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4753-4753. (6) Massoud, T. F.; Gambhir, S. S. Genes Dev. 2003, 17, 545-580. (7) Kim, S. B.; Ozawa, T.; Watanabe, S.; Umezawa, Y. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11542-11547. 10.1021/ac0510078 CCC: $30.25

© 2005 American Chemical Society Published on Web 09/20/2005

Figure 1. Basic strategy for stress imaging. (A) The structures of cDNA constructs of the indicators. pcDRc-GR consists of the C-terminal half of DnaE (dnaEc), C-terminal half of RLuc (rLucC; 230-311 aa), GC linker (GGGGSG), and full-length GR. pcRDn-NLS consists of FLAG epitope (DYKDDDDK), N-terminal half of RLuc (rLucN; 1-229 aa), N-terminal half of DnaE (dnaEn), and nuclear localization signal (NLS; (DPKKKRKV)3). Additional sequences at the boundaries between RLuc and dnaE are shown under the bars. (B) A schematic illustration of the translocation steps of GR. When GR binds a stress hormone, it translocates into the cellular nucleus. Interaction between DnaE-C and DnaE-N causes protein splicing, and thereby, RLuc-N and RLuc-C are linked by a peptide bond. The reconstituted RLuc recovers its bioluminescence activity, which is measured with a luminometer or with a cooled CCD camera.

fragments, the bioluminescence activity is completely lost. The N-terminal RLuc connected with the N-terminal DnaE is localized in the nucleus, whereas the C-terminal RLuc and DnaE joined to GR is in the cytosol. Upon binding to corticosterone, the GR translocates into the nucleus, in which protein splicing with DnaE undergoes producion of a full-length RLuc. Thus reconstituted RLuc produces bioluminescence with its substrate, coelenterazine. We describe that the genetically encoded indicators work to detect quantitatively and sensitively stress hormones, and the cells carrying a pair of the indicators enabled imaging of the endogenous corticosterone as an index of stress intensity sensed by living mice. EXPERIMENTAL SECTION Construction of Plasmids. The plasmid pcRDn-NLS encoding the N-terminal domains of Renilla luciferase (RLuc-N; 1-229 aa) and DnaE (DnaE-N; 1-123 aa) was described in our previous report.7 The plasmid pcDRc-GR was prepared from pcDRc-AR by replacing the cDNA of AR by GR as follow: The cDNA encoding a full-length GR (1-777 aa) was modified by PCR to add unique enzyme sites, NotI and XhoI, at the N- and C-terminal ends, respectively. The cDNA fragment was subcloned into the expres-

sion vector pcDRc-AR at NotI/XhoI sites, named pcDRc-GR. The PCR product was sequenced to ensure fidelity with a BigDye Terminator Cycle Sequencing kit and a genetic analyzer ABI Prism310 (PE Biosystems). Western Blot. NIH 3T3 cells were cultured on 10-cm culture dishes to 90% confluence. The cells were cotransfected with pcRDn-NLS and pcDRc-GR, and incubated at 37 °C with 5% CO2 and 95% filter-cleaned air for 4 h. Twelve hours after the medium change with Dulbecco’s modified eagle’s medium (DMEM; Sigma) supplemented with 10% steroid-free newborn calf serum (NCS) and 1% penicillin-streptomycin (P/S), the cells were washed with PBS and lysed in 200 µL of lysis buffer (1% SDS, 10% glycerol, 10% 2-mercaptoethanol, 0.001% bromophenol blue, 50 mM Tris-HCl, pH 6.8). Each 7 µL of samples was electrophoresed in 8% acrylamide gel, transferred to a nitrocellulose membrane, and blotted with a mouse anti-FLAG antibody (Sigma) or mouse anti-β-actin antibody (Sigma). The blots were incubated with an alkaline phosphatase-conjugated secondary antibody (Jackson) and visualized by a chemiluminescence reagent (New England Biolabs) and a luminescence image analyzer (LAS-1000, Fuji Film). Immunocytochemistry. NIH 3T3 cells were cultured on thin cover glasses (0.7 × 105 cells/slide) and transfected with pcRDnAnalytical Chemistry, Vol. 77, No. 20, October 15, 2005

6589

NLS or pcDRc-GR. The cells were incubated for 12 h and fixed on the cover glasses with a 3% paraformaldehyde solution. The cells on the cover glasses were blocked with 0.2% fish skin gelatin and then incubated with a mouse anti-GR antibody (Santa Cruz) or a mouse anti-FLAG antibody (Sigma). The cells were incubated with Cy-5-conjugated secondary antibody (Jackson) for 30 min, and then the nuclei of the cells were stained with Sytox Green (Molecular Probes). The fluorescence images were recorded using a confocal laser-scanning microscope (LSM510, Zeiss) fitted with a band-pass filter (514-535 nm) for Sytox Green and a longpass filter (665 nm) for Cy-5. Cell-Based in Vitro Assay. NIH 3T3 cells were maintained in DMEM supplemented with 10% NCS and 1% P/S and plated in 12-well plates. The cells in the plates were cotransfected with pcRDn-NLS and pcDRc-GR using a lipofection reagent, LipofectAMINE 2000 (Invitrogen), in an antibiotic-free medium for 4 h. The medium was then replaced with DMEM supplemented with 10% steroid-free NCS and 1% P/S and incubated at 37°C with 5% CO2 for 12 h. The cells on each well of the plate were stimulated with differing concentrations of steroids for 2 h. The cells were then harvested, and their luciferase activities were evaluated by using a Renilla luciferase assay kit (Promega) and a luminometer (Minilumat LB9506; Berthold) with an integration time of 20 s. For the cell fractionation assay, NIH 3T3 cells cotransfected with pcRDn-NLS and pcDRc-GR were grown in an incubator (5% CO2, 37 °C) for 12 h. The cells were then stimulated with corticosterone (10-6 M) or 0.1% DMSO (vehicle) for 2 h. The cells were harvested and suspended in a homogenization buffer solution (0.25 M sucrose, 5 mM EDTA, 20 mM Tris-HCl, pH 7.4). The plasma membranes of cells were then crushed with a tip sonicator, and the homogenate was centrifuged at 600g for 15 min to separate the nucleus from the other cell components. The luminescence intensities of each fraction were measured with the luminometer. Optical in Vivo Imaging of Forced-Swimming Mice. The NIH 3T3 cells transfected with pcRDn-NLS, pcDRc-GR, or both were harvested from each plate and suspended in DMEM. An aliquot of the suspended cells (1 × 106) was implanted in four different sites on the back of anesthetized BALB/c nude mice (5 week old, ∼15 g, female, Sankyo Labo Services, Tokyo, Japan). The mice were kept in a temperature-controlled chamber for 12 h and thereafter fasted for 4 h. The mice were separated into two groups: one group is for a stress imaging of forced-swimming mice and the other is for its reference. The mouse group for a stress imaging was dropped into a cylindrical chamber (height, 25 cm; diameter, 10 cm) filled with a 10-cm-height water at 2325 °C and left swimming there for 5 min (Porsolt forced swim test).8 The mice were then saved from the water, towel-dried, and returned to the cage for the following tests. Two hours after the swimming, mice were injected intraperitoneally (ip) with an aliquot of an RLuc substrate, coelenterazine (0.025 mg/body of weight). The luminescence was obtained every 2-min interval using an in vivo imaging system (IVIS100; Xenogen) with a cooled CCD camera, and the obtained images were analyzed with LIVING IMAGE software (Xenogen). To quantify the measured light, regions of interest were drawn over the cell-implanted area, and the mean luminescence intensities (photons per second per cm2 per steradian) were evaluated. 6590

Analytical Chemistry, Vol. 77, No. 20, October 15, 2005

Optical in Vivo Imaging of Metabolically Stressed Mice. The NIH 3T3 cells were transfected with pcRDn-NLS, pcDRc-GR, or both and implanted in four different sites on the back of mice, and then the mice were kept in a temperature-controlled chamber for 12 h. The mice were separated into two groups: one group is for a metabolic stimulation (glucoprivation) with 2-deoxy-D-glucose (2DG) and the other is for a reference. The reference group was injected ip with 100 µL of PBS, whereas the other group was injected ip with 100 µL of 2DG solution (0.17 mg/body of weight). Both groups of mice were kept in a soundproof chamber before imaging. After the coelenterazine injection (0.025 mg/body of weight), the mice were imaged by the In Vivo Imaging System. ELISA with Mouse Blood Samples. For blood collection, one group of mice was kept in a soundproof cage without any stress-causing stimulation, whereas the other group of mice was given stress, which was induced by a forced swimming as described above. A 20-µL aliquot of blood samples was collected every 1 h from the tail veins of the mice, and the blood samples were immediately freezed for stock at -80 °C. An ELISA was operated for the quantitative analysis of corticosterone concentrations using a 96-well enzyme immunoassay kit (Assay Designs Inc.). RESULTS Expression of Fusion Proteins and Confirmation of Their Localization. To examine the intracellular localization of the fusion proteins, localization of each expressed fusion protein was probed with its specific primary antibody and a Cy-5-conjugated secondary antibody. The fusion proteins were imaged with Cy-5, and the nucleus regions were stained with a specific marker, Sytox Green (Figure 2A). In the absence of stimulation with a specific ligand, the C- and N-terminal fusion proteins were localized in the cytosol and nucleus, respectively, whereas in the presence of corticosterone, the two fusion proteins were colocalized in the nucleus as a result of the ligand-dependent nuclear transport of GR. To confirm that the protein splicing occurred upon colocalization of the fusion proteins, we analyzed the size of expressed fusion proteins by the western blot. Both plasmids, pcRDn-NLS and pcDRc-GR, were transfected into the COS-7 cells, and crude extracts were prepared from the COS-7 cells after 16-h incubations. In the absence of corticosterone, the anti-FLAG antibody recognized only a specific band of 45 kDa, the size of which was the same as that of an unspliced N-terminal fusion protein, N-terminal halves of RLuc and DnaE. In the presence of corticosterone, the anti-FLAG antibody recognized a specific polypeptide band of 120 kDa in addition to the 45 kDa of an unspliced peptide band. Electrophoretic mobility of the 120 kDa band was consistent with the predicted size of the spliced product (Figure 2B), indicating that the protein splicing reaction proceeded upon addition of corticosterone. To check that the reconstitution of the split RLuc occurred in the nucleus, we measured the luminescence intensities of nuclear and cytosolic fractions. The fractions were obtained by a sucrose density gradient fractionation of the cotransfected NIH 3T3 cells (Figure 2C). The luminescence intensity from the nucleus fraction (8) Dailly, E.; Chenu, F.; Renard, C. E.; Bourin, M. Fundam. Clin. Pharmacol. 2004, 18, 601.

Figure 2. In vitro characterization of the indicators. (A) Immunocytochemical analysis of the localizations of fusion proteins. Two fusion proteins expressed from pcDRc-GR and pcRDn-NLS were recognized by anti-GR and anti-Flag antibodies, respectively, and stained with Cy-5-conjugated secondary antibodies (first column from left). The nuclei were stained with Sytox Green (second columun from left). The transmission images were taken (third column from the left), and their merged images are shown in the fourth column. In the first and second rows, the localizations of the GR-connected indicator are shown, while the third and forth rows showed the localization of the NLS-connected indicator. “corti-” and “corti+” represent the absence and presence of corticosterone stimulation, respectively. (B) Western blot of lysates obtained from intact COS-7 cells (lane 1) and from the cells carrying pcRDn-NLS and pcDRc-GR in the absence (lane 2) or presence (lane 3) of 1 µM corticosterone. As a reference for the amount of proteins electrophoresed, β-actin was stained with its specific antibody. (C). Determination of luminescence intensities from each cell fraction. The cellular fractions were obtained from cotransfected cellls with pcRDn-NLS and pcDRc-GR in the presence and absence of 1 µM corticosterone (n ) 3).

of the cells stimulated by corticosterone was found to be ∼5 times larger than that from the cytosol fraction of the cells stimulated with corticosterone. The results indicate that RLuc folded correctly and recovered its luminescence activity in the nucleus. From the above results, we concluded that corticosterone stimulation induced the nuclear translocation of the C-terminal fusion protein, resulting in the protein splicing of the split RLuc and recovery of the activity of RLuc in the nucleus. To quantitate the extent of GR translocation, the cells transfected with pcDRc-GR and pcRDn-NLS were stimulated with

differing concentrations of corticosterone for 2 h. The cells were harvested, and the luminescence intensities were monitored with a luminometer. Upon corticosterone stimulation, the luminescence intensity from the cotransfected cells was found strong enough to be discriminated from that of the mock-transfected cells (Figure 3). Compared with the luminescence intensity from the cotransfected cells, the luminescence intensity from the cells transfected with either pcDRn-NLS or pcDRc-GR did not show any noticeable increase in luminescence intensities in the tested concentration ranges of corticosterone. The luminescence intensities from the Analytical Chemistry, Vol. 77, No. 20, October 15, 2005

6591

Figure 3. Determination of the extent of the nuclear transport of GR after stimulation with differing concentrations of corticosterone or cortisol. The NIH 3T3 cells were cotransfected with pcRDn-NLS and pcDRc-GR, and the activities of the reconstituted RLuc were tested 2 h after the addition of each specified concentration of corticosterone (red) or cortisol (yellow) (n ) 3). Open circles mean the results of corticosterone stimulation to the cells carrying only pcDRc-GR.

cotransfected cells increased with corticosterone concentrations from 10-8 to 10-6 M. Cortisol also induced considerably high luminescence intensities in the concentration range from 10-8 to 10-5 M. Corticosterone provided stronger luminescence intensities than cortisol at all concentration ranges, suggesting that corticosterone is a higher potent inducer of the nuclear transport of GR than cortisol. In Vivo Imaging of Exteroceptive Stress in Living Mice. The NIH 3T3 cells transfected with the indicators were applied for investigating the endogenous stress hormone activities in an interesting site of living mice. Four different types of cells were prepared and implanted in the back of mice: The cells were transiently transfected with pcRDn-NLS (site 2), pcDRc-GR (site 3), and pcRDn-NLS and pcDRc-GR (site 4) and of mock transfected cells (site 1; Figure 4). The mice were forced to swim for 5 min in a water pool and were injected ip 2 h after the swimming with an aliquot of the substrate for RLuc (coelenterazine; 0.025 mg/ body of weight). The resulting CCD images showed a significant increase in the luminescence intensities from the sites implanted with the cells carrying both pcRDn-NLS and pcDRc-GR (Figure 4A). The groups of mice with swimming showed higher intensities compared to the other without swimming. The average luminescence intensities with or without swimming were (22.3 ( 0.9) × 103 and (16.8 ( 1.0) × 103 photons s-1 cm-2 sr-1, respectively (n ) 2). Mice under the swimming stress showed 32.7% higher luminescence intensities than those without the stress. Sites 1-3 showed luminescence intensities overlapped with the background ((∼9.0 ( 0.8) × 103 photons s-1 cm-2 sr-1), which originated from the body heat of living mice and nonspecific decomposition of the substrate, coelenterazine. The results demonstrated that the stress indicators enabled us to noninvasively image the extent of GR nuclear translocation, which was stimulated by physiologically relevant concentrations of stress hormones in mice blood vessels. We also confirmed the elevation of corticosterone concentrations in the blood vessels with a conventional enzyme-linked immunosorbent assay (ELISA) (n ) 3). The results showed that the 6592 Analytical Chemistry, Vol. 77, No. 20, October 15, 2005

Figure 4. In vivo imaging of stress in living mice. (A) Visualization of forced-swimming stress in living mice. Translocation of GR into the nucleus was induced by an endogenously secreted stress hormone, corticosterone. NIH 3T3 cells carrying pcRDn-NLS (site 2), pcDRc-GR (site 3), and pcRDn-NLS and pcDRc-GR (site 4) and of mock transfected cells (site 1) were implanted on the back of mice. The luminescence intensities from the mice with and without swimming were imaged using a cooled CCD camera after the injection of coelenterazine. The pseudocolor images of bioluminescence intensities were superimposed on photographic images of mice (n ) 2). (B) Determination of the concentration of plasma corticosterone secreted in the blood of mice after a 5-min forced swim. The 20-µL blood samples were obtained from mice every 1-h interval, and their corticosterone levels were determined by the ELISA method (n ) 3). Red circles show blood samples of mice with swimming stress, whereas blue circles show those without swimming stress.

corticosterone levels increased up to 192 ng/mL 1 h after the forced swimming, the value of which was 7.1 times higher than that of stress-free mice (27 ng/mL). From these results, we concluded that the elevated corticosterone caused a selective and strong luminescence emission from the cell-implanted site of mice. In Vivo Imaging of an Interoceptive Stress in Living Mice. We also imaged an interoceptive stress through the elevation of blood corticosterone caused by a metabolic perturbation owing to 2DG (Figure 5).9 The mice were implanted with the cells carrying pcRDn-NLS and pcDRc-GR. Of the two groups, one group of mice was stimulated with 2DG and the other with PBS. The former group stimulated with 2DG showed higher luminescence (9) Ritter, S.; Watts, A. G.; Dinh, T. T.; Sanchez-Watts, G.; Pedrow, C. Endocrinology 2003, 144, 1357-1367.

Figure 5. Quantitative imaging of a metabolic stress induced by 2DG in living mice. NIH 3T3 cells transiently transfected with pcRDnNLS (site 2), pcDRc-GR (site 3), and pcRDn-NLS and pcDRc-GR (site 4) and of mock transfected cells (site 1) were implanted on the back of mice. One group was injected ip with 2DG and the other with PBS. After 2 h, the luminescence intensities were imaged with a cooled CCD camera after the injection of coelenterazine (n ) 2).

intensities ((31.2 ( 0.5) × 103 photons s-1 cm-2 sr-1) than PBSinjected group ((26.0 ( 1.7) × 103 photons s-1 cm-2 sr-1) (n ) 2). Mice stimulated with 2DG exhibited 20.0% higher luminescence intensities compared to those without 2DG. The luminescence intensities from sites 1-3 were overlapped with those of the background from the mouse skin. Their maximum intensities were obtained at 10 min after injection of coelenterazine. The results indicate that the strong luminescence intensity upon injection of 2DG is caused by an increase in the corticosterone concentrations as a result of glucoprivation by 2DG. DISCUSSION Herein, we described the development of a pair of new genetically encoded indicators for a quantitative imaging of stress sensed in living mice based on the physiological elevation of corticosterone. The extent of GR translocation was estimated in an in vitro assay with the cells carrying the indicators. We also demonstrated that the present in vivo imaging technology without need for gene transcription is sensitive enough to detect the magnitude of physical and emotional stress sensed by mice. The results indicate that the detection limit is low enough to determine the physiological fluctuation of corticosterone. Previous studies have categorized the stressors into two groups; interoceptive (systemic, physiological, or homeostatic) and exteroceptive (neurogenic, psychological, or emotional) stressors.9 We visualized the extent of the both stress types sensed by a mouse in the form of luminescence intensities under basal conditions. The results demonstrate that both interoceptive and (10) Edmonds, B. K.; Edwards, G. L. Brain Res. 1998, 801, 21-28. (11) Sanders, N. M.; Ritter, S. Diabetes 2001, 50, 2831-2836.

exteroceptive stressors can be imaged with this imaging technology. In both cases, the signals are based on the elevation of a physiological stress transducer, corticosterone, which activates and translocates GR into the nucleus, resulting in the reconstitution of RLuc at the implanted site of a living mouse. The molecular cellular event, called stress hormone signaling in eukaryotic cells, was thus quantitatively visualized in the scale of a whole body with reflecting the physiological circumstance. This in vivo technology can also provide tissue- or site-specific information about the activities or side effects of drugs because we can selectively implant the cells carrying the indicators in mouse tissues or organs of interest. Blood component abnormalities by bioactive xenochemicals are highly concerned, and a lot of data provides critical information for diagnosing diseases and related studies. As an example, a glucose analogue of 2DG is known to activate the hypothalamicpituitary-adrenal axis, resulting in a significant elevation of plasma corticosterone level in living subjects.10 Corticosterone, which is elevated by glucoprivation, has been suspected in the pathogenesis of hypoglycemia-associated autonomic failure (HAAF), a lifethreatening manifestation.11 The present imaging study provides an efficient method for screening such metabolic stresses and thus can help people exclude the metabolically risky stressors on human body in animal models. In summary, we developed a novel molecular imaging system for physical and emotional stress based on split RLuc and its signal-dependent reconstitution by protein splicing. This is a new nontranscriptional approach to evaluate the stress effects in living subjects. We evaluated the extent of nuclear transport of GR activated by corticosterone in the whole body level. We demonstrated that this technology is applicable for quantitative imaging of two representative stress types, interoceptive and exteroceptive, in living mice. With the present imaging technique, we visualized noninvasively the level of intravascular stress hormones secreted as a stress reaction. The present method may provide a wide variety of applications for pharmacological or toxicological purposes such as testing various endo- and exogenous risk factors and identifying specific molecular events responsible for diseases in living subject, e.g., HAAF and coronary artery disease. It can also facilitate the development of transgenic animals that express the indicators in the tissues or organs with controllable promoters. ACKNOWLEDGMENT This work was supported by grants from Japan Science and Technology Agency (JST), the Precursory Research for Embryonic Science and Technology (PREST), Japan Society for the Promotion of Science (JSPS), and the Ministry of Education, Science, and Culture of Japan. Received for review June 8, 2005. Accepted August 12, 2005. AC0510078

Analytical Chemistry, Vol. 77, No. 20, October 15, 2005

6593