FluoQ: A Tool for Rapid Analysis of Multiparameter Fluorescence

Jul 24, 2013 - FluoQ, a new macro for the image analysis software ImageJ, which enables fast analysis of multiparameter time-lapse fluorescence micros...
2 downloads 6 Views 2MB Size
Download PDF
Letters pubs.acs.org/acschemicalbiology

FluoQ: A Tool for Rapid Analysis of Multiparameter Fluorescence Imaging Data Applied to Oscillatory Events Frank Stein, Manuel Kress, Sabine Reither, Alen Piljić, and Carsten Schultz* European Molecular Biology Laboratory (EMBL), Cell Biology and Biophysics Unit, Meyerhofstraße 1, 69117 Heidelberg, Germany S Supporting Information *

ABSTRACT: The number of fluorescent sensors and their use in living cells has significantly increased in the past years. Yet, the analysis of data from single cells or cell populations usually remains a very time-consuming enterprise. Here, we introduce FluoQ, a new macro for the image analysis software ImageJ, which enables fast analysis of multiparameter time-lapse fluorescence microscopy data with minimal manual input. FluoQ provides statistical analysis of all measured parameters and delivers the results in multiple graphic and numeric displays. We demonstrate the power of FluoQ by applying the macro to data analysis in the development and optimization of novel FRET reporters for monitoring the performance of calcium/calmodulin-binding inositol trisphosphate kinases A and B (ITPKA and ITPKB) in HeLa cells. We find that conformational changes in the ITPKA-based sensor follow receptor-mediated calcium oscillations. This indicates that ITPKA contributes to the regulation of intracellular calcium transients by limiting inositol trisphosphate levels. The development of fluorescent sensors and their use in single living cells has altered our understanding of how cellular events develop and progress.1−3 Unlike other techniques, real-time monitoring provides single cell or even subcellular resolution and teaches us that cultured cells usually react individually rather than in synchrony. This is particularly true for complex signaling patterns such as oscillations of ion concentration or enzyme activity.4 While calcium oscillations are frequently tracked by small molecule-based or genetically encoded fluorescent reporters, much less is investigated how the oscillatory behavior is translated into downstream signaling events at the single cell level. It is of special interest to know how the activity of the many calcium/calmodulin-binding proteins 5 transmits calcium oscillations in the signaling network.6,7 One calcium/calmodulin-binding protein directly involved in the modulation of calcium oscillations is inositol trisphosphate kinase (ITPK).8−10 It is an enzyme that specifically phosphorylates the second messenger Ins(1,4,5)P3 to Ins(1,3,4,5)P4 and thereby limits the opening frequency of the Ins(1,4,5)P3 receptor, a calcium channel in the endoplasmic reticulum that permits calcium release from internal stores. The latter is the main regulator of calcium oscillations and it has been demonstrated in the past that Ins(1,4,5)P3 levels oscillate in synchrony with calcium oscillations.11−13 Therefore, we decided to create genetically encoded FRET sensors based on ITPKA and ITPKB to investigate in how far the activity of these enzymes follows calcium oscillations at an equivalent pace. Although time-consuming optimization of FRET reporters in the wet-lab is already speeded up by using FRET frame backbone cassettes,14,15 the analysis of the obtained image data is still very laborious and challenging. Some recent software © 2013 American Chemical Society

solutions simplified automated image analysis,16 but there is still a growing need for new image analysis tools.17 We therefore developed FluoQ, an extensive macro extending the popular image analysis software ImageJ 18 for fast and reproducible analysis of imaging data. The macro processes most life sciences image file formats and analyzes any number of experiments in a batch-mode fashion. Image processing and cell segmentation works either interactively or in an automatic fashion and therefore covers any application between low and high throughput analysis. As a final step, FluoQ automatically summarizes the analysis in multiple plots and generates data tables that can be imported into any data analysis software such as MS Excel, Origin, or R. This significantly accelerates data analysis and makes large amounts of data manageable. In addition, FluoQ saves a full documentation of all performed manipulation steps making the whole process transparent and reproducible. Figure 1 shows the overall workflow we pursued to develop new FRET sensors based on ITPKA and ITPKB. To obtain the ITPK FRET sensor constructs, truncated sequences encoding ITPKA and ITPKB fragments consisting only of the active kinase and the calmodulin binding domain (Figure 1a) were cloned into a library of 30 vector backbones. This set of 30 plasmids was described earlier for the development of Camk2a and DAPK1 sensors 14 (Figure 1b). The set contained the donor fluorophore mTurquoise 19 and the acceptor fluorophore Venusd 15 in various combinations of orientations through Received: May 15, 2013 Accepted: July 24, 2013 Published: July 24, 2013 1862

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology

Letters

circular permutations and linker lengths. These different combinations lead to a different Förster resonance energy transfer (FRET) efficiency for each backbone when expressed in cells. HeLa Kyoto cells were transfected with the reporter plasmid DNA in 8-chamber Lab-Tek dishes (Figure 1c). Transfected cells were stimulated with the ionophore ionomycin and the fluorescent intensity was followed over time. The raw microscopy data were analyzed with the newly developed ImageJ macro FluoQ allowing fast identification of the best performing FRET sensor (Figure 1d, e). Initial attempts were made with the vector F40, since this backbone yielded robust ratio changes in previous screens for DAPK and Camk2a.14 Cell transfection and subsequent analysis by fluorescence microscopy revealed good transfection efficiency and cytoplasmic localization of the sensors. Cells were monitored by fluorescence microscopy before and after increasing intracellular calcium levels and activation of calmodulin with ionomycin (10 μmol L−1). Binding of calmodulin to the ITPK insert led to a change in the orientation and/or distance of the donor and acceptor fluorophores and a difference in FRET. To extract this information from the raw microscopy data, images were processed and the acceptor/donor (Venusd/mTurquoise) ratio images were calculated (see Supporting Information (SI) Figure 1). Subsequently, cell images were segmented and the mean pixel intensity for each region of interest (ROI) over time was extracted from all channels (Figure 2a). For each time-lapse trace, the change in amplitude was calculated according to formula 1 (Figure 2a). FluoQ was initially applied to automate and standardize the image processing steps and the extraction and evaluation of time-traces from single cells expressing the F40-ITPKA construct. FluoQ performed all of the above-mentioned image processing and evaluation steps while also summarizing all extracted data in various spreadsheets, text files, and statistical plots (see SI Figure 2). The initial user interface (see SI Figure 3) offers various options for each step to ensure customized data handling (see SI Table 1). Furthermore, all processed images, dialogue choices, and a full record of the analysis are saved automatically together with the statistical summaries into a new folder to ensure maximal transparency and reproducibility (see SI Table 2). FluoQ gives the user the option to interfere with any part of the image analysis during the process. Being able to manually correct for potential errors of the algorithm is a crucial feature for reliable data analysis. However, FluoQ offers also fully automated analysis. In any mode of FluoQ, each process is displayed on the screen to enable full control of how the data appear after each manipulation and what data will be extracted for statistical summaries (see SI Figure 4). FluoQ analysis of ITPKA-F40 revealed an amplitude change of 25% in the ratio channel upon ionomycin stimulation. This ratio change is due to an intensity increase in the donor and a decrease in the acceptor (sometimes called transfer) channel (Figure 2b and c). The ITPKB-F40 sensor was analyzed in the same manner and a ratio change of about 10% was observed (see Figure 2c). To further increase the performance of the sensors based on ITPKA and ITPKB, the inserts were cloned into the whole library of 30 vectors. The constructs were treated and analyzed in the same manner as the F40 construct. We performed at least six independent experiments for each construct to evaluate technical and biological deviation.

Figure 1. Work-flow overview for the development of new FRET reporters. (a) Domain structure of ITPKA and ITPKB and insert design (black, protein sequence; dark blue, substrate binding residues (both according to UniProtKB P23677 (ITPKA) and similarity (ITPKB)); light blue, residues for ATP binding; light orange, calmodulin binding site; red, amplified regions; and gray (IPK), kinase domain). (b) The ITPK insert was cloned into the FRET backbone library that contained donor and acceptor pairs in different orientation and linker lengths. (c) Cells were seeded in 8-chamber Lab-Teks, transfected with the plasmid DNA and imaged by fluorescence microscopy. (d) Microscopy data were analyzed with the newly developed ImageJ macro FluoQ for rapidly identifying the best performing sensor (e). 1863

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology

Letters

Figure 2. Analysis of ITPKA- and ITPKB-based sensors. (a) The mean pixel intensity for each time point of one segmented cell was extracted. The amplitude change after ionomycin (10 μmol L−1) stimulation was calculated according to formula 1. Cyan bar indicate unused time points during ionomycin treatment and signal equilibration . (b) Averaged time-traces for all cells in the field of view ordered by channel (Error bars: standard deviation). (c) Mean of ratio changes for exploratory experiments for ITPKA-F40 and ITPKB-F40 (Error bars: standard deviation).

changes can be explained with the different fluorophores and their orientations. The three best performing backbones for ITPKA were F41 (−31.2 ± 5.2%), F49 (−28.5 ± 6.0%) and F43 (−26.5 ± 4%), whereas for ITPKB the three best performing sensors were F46 (−19.6 ± 1.4%), F59 (−18.5 ± 4.6%), and F49 (−16.1 ± 2.2%). Treating cells with ionomycin lead to a major increase in the intracellular calcium concentration way beyond the physiological level. To investigate whether the sensors also report on more subtle calcium changes, we stimulated cells with ATP (100 μmol L−1). This is known to lead to calcium oscillations via activation of purinergic receptors.21 We measured calcium levels with the genetically encoded calcium sensor R-GECO 22 in parallel to the ratiometric ITPK probes. The resulting multiparameter imaging data set was analyzed simultaneously with FluoQ, since the macro can handle any number of intensiometric and/or ratiometric channels. Cells not expressing any sensor for ITPKA or ITPKB showed calcium oscillations upon ATP stimulation (Figure 4, green bar). However, calcium transients were not observed in cells expressing the kinase active ratiometric probes (Figure 4, orange bar, or compare averaged time-traces in SI Figure 6). Also, no FRET changes of the kinase probes were detected. Ionomycin treatment at the end of each experiment showed that the sensors were functional. This indicated that overexpression of unregulated and constitutively active ITPK sensors interfered with the calcium signaling machinery after ATP stimulation. Presumably, Ins(1,4,5)P3 was rapidly phosphorylated to Ins(1,3,4,5)P4, decreasing Ins(1,4,5)P3 levels for Ins(1,4,5)P3 receptor binding and calcium release from internal stores.23 To further prove that the kinase activity of the probes was responsible for the loss of calcium oscillations upon

FluoQ is able to analyze the raw data of the whole library by processing multiple experimental files in a row. This batch mode is implemented in a very user-friendly manner. FluoQ takes only a single directory as input and searches automatically for experiments in any subfolder that matches the conditions selected in the initial user interface. The macro makes use of the LOCI Bio-Formats importer 20 and is therefore able to handle most of the life-sciences image file formats. Also, multiple experiments saved in a single file are recognized. The search results in a list where the user ticks the experiments to analyze. In addition to the analysis of single experiments, FluoQ summarizes averaged time-traces and amplitude changes of all experiments to immediately judge the overall outcome of the experiment set and to simplify further data processing and analysis (see SI Figure 4). The analysis of the library screen (the whole library is depicted in Figure 3a) by FluoQ revealed the ratio changes shown in Figure 3b (see SI Figure 5 for averaged time-traces). On average, sensors with the ITPKA insert gave a higher dynamic range (higher ratio changes) than sensors with the ITPKB insert. In addition, comparison between ITPKA and ITPKB showed differences in dynamics and a different FRETframe gave the best performance (F41 for ITPKA and F46 for ITPKB). Additionally, the influence of the linker length (2, 4, or 8 amino acids) was highly variable. In some cases (ITPKA F34−F36 and ITPKB F49−F51) different linker lengths did not influence the dynamic range of the sensors, whereas we observed a dramatic impact in other cases (e.g. ITPKA F37− F39 and ITPKB F58−F60). Also, it was not always the shortest linker (2 amino acids) that gave the largest ratio change demonstrating the benefit of including different linker lengths in the library screen. However, most of the difference in ratio 1864

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology

Letters

Figure 3. Overview of the FRET construct library and performance of ITPKA and ITPKB sensors. (a) List of constructs (aa = amino acids, mTurq = Turquoise,19 mTurqD = Turquoise lacking C-terminal 11 aa, Ven = Venusd,15 ITPK = truncated ITPK insert (see Figure 2a) with different linker lengths as indicated, cp = circular permutated). (b) Mean and standard deviation of ratio changes derived from averaged cell traces of independent experiments. Experiment numbers are given in each bar.

performing candidates to study how far the activity of these kinases respond to ATP induced calcium oscillations. We found that changes in the conformation of ITPKA follow receptormediated calcium oscillations at an equivalent pace. This exemplifies that oscillatory calcium signaling can be directly translated into transient downstream effects. We thereby unraveled an additional component of the signaling network regulating calcium oscillations. Although, the function of the ITPK product Ins(1,3,4,5)P4 is to date not fully understood, the rapid degradation of Ins(1,4,5)P3 seems to contribute to the steep decay in each spike of the Ins(1,4,5)P3/Ca2+ signal and therefore promotes the sharp appearance of calcium oscillations. To make full use of the FRET frame library and their rapid access to FRET reporter candidates, we developed FluoQ. This ImageJ-based macro facilitates the analysis of the resulting imaging data at a single cell level in a reproducible and standardized fashion. Using FluoQ, the time required for the analysis of the example shown here was reduced from days to hours (see SI Figure 9). The macro is compatible with most bioimage file formats and processes image data sets either in a very interactive or completely automatic way. The incorporated batch-mode automatically finds experiments in the file system with similar setups to simplify high-throughput analysis. Compared to other existing software solutions (e.g. Cell

ATP stimulation, we created kinase inactive mutants of the best performing probes by introducing a mutation at the ATP binding site 24 (see Figure 1a for mutation sites). By introducing the mutations into the ITPKs, we were able to fully recover the ATP-induced calcium oscillations (Figure 4a and SI Figure 7) indicating that the sensors were now unable to interfere with the signaling network. We also observed that ITPK mutants responded to the calcium increase upon ATP stimulation with a synchronous (within the time resolution of 3s) ratio change (see SI Figure 6). The best performing probes (ITPKA-F41, F44, and F49) could follow transient calcium changes in an oscillatory manner (see Figure 4b and SI Figure 8 for more example traces). These results indicated that ITPKA was able to translate very transient calcium signals into direct kinase activity and thereby it becomes conceivable that ITPKA plays a key role in modulating calcium oscillations by limiting the oscillatory behavior of Ins(1,4,5)P3 levels.4 We were not able to verify if this is also true for ITPKB. For ITPKA, only sensors with a ratio change of at least 30% were able to resolve calcium oscillations. Since for ITPKB, maximum ratio changes were below 20%, we can only speculate if ITPKB also translates transient calcium signals into kinase activity. For the underlying study, we applied the FRET frame library of Piljić et al.14 to screen for novel genetically encoded FRET sensors based on ITPKA and ITPKB. We used the three best 1865

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology

Letters

Figure 4. Performance of the three best sensors upon ATP stimulation (100 μmol L−1). (a) Effect of kinase active and inactive ITPK-based sensors on calcium transients upon ATP stimulation. Cells were transfected with R-GECO and with or without constructs encoding for kinase active (ITPKA-WT) or kinase inactive mutants. Number of calcium transients within 720 s after ATP stimulation was counted. Shown is a BoxWhisker plot with different vectors and conditions on the x-axis and the number of calcium transients on the y-axis. Cell numbers for each condition are given under each bar. (b) Single cell time trace comparison of calcium and ITPK sensor activity response upon ATP (after 90 s) and ionomycin (after 810 s) stimulation. Traces were rescaled and y-scales were removed for clarity. Note that only ITPKA-based sensors are suitable to follow calcium oscillations.

Profiler 25), FluoQ automatically generates multiple plots and summaries. This renders not only the possibility for an immediate evaluation of multiple experiments but also eases further data processing with almost any statistical software. FluoQ is designed to visualize, quantify, and summarize cellular responses to any stimulus one might add to cells in a fluorescent microscopy setup. We are convinced that these features will have a huge impact on acceleration and reproducibility of time-lapse image analysis of many chemical biology tools with a fluorescent readout. In conclusion, by combining automated data analysis by FluoQ with rapid cloning using the FRET construct cassette and potentially automated microscopy for analysis of the constructs,14 another step has been made toward the rapid and efficient production of semioptimized FRET sensors at a very large scale. The example of FRET sensors based on ITPKA and ITPKB demonstrated not only the usefulness of the combined approaches, but it also highlighted ITPK as an additional component to the signaling network regulating calcium oscillations.

■

GTTCT CTCAG CCAGG CTGGC CAGGA TG (536−1454 bp according to NM 00220), ITPKB: ITPKB-F AGCGC TACCG GTCTG GACCA GCAGA AACCT AGAGT GAGC, ITPKB-R AATAT TACGC GTGGC GAGTG GGGCA TCCTG GGACA TC (2249−3178 bp according to NM 00221). The restriction sites AgeI and MluI were introduced through primers in the course of PCR. Backbones and inserts were digested with AgeI and MluI, and purified using the Qiagen Gel Extraction Kit. The designated FRET construct-insert combinations were ligated using T4 Ligase from Fermentas. Kinase dead mutations of ITPKA and ITPKB were K246Aand K745A, respectively. They were introduced by site directed mutagenesis using the following primers and the QuikChange Lightning Site-Directed mutagenesis kit. (ITPKA K246A-F GTGCT CGACT GCGCC ATGGG CGTCA GG, ITPKA K246A-R CCTGA CGCCC ATGGC GCAGT CGAGC AC, ITPKB K745A-F TGTGT GATGG ACTGC GCCAT GGGAA TCAGG ACC, ITPKB K745A-R GGTCC TGATT CCCAT GGCGC AGTCC ATCAC ACA). Cell Culture and Transfection. All cell experiments were performed in HeLa Kyoto cells. The cells were passaged and maintained in DMEM consisting of 1 g L−1 glucose without pyruvate (Gibco), 10% FBS (Gibco), and 0.1 mg mL−1 Primocin (Invitrogen). For imaging, eight chambered LabTek dishes were used. (4−5) × 105 cells per chamber were seeded and incubated for 20−24 h. Transfection was performed at around 50% confluence in Opti-MEM (Gibco) with 1 μL Fugene HD (Promega) for 100 ng of DNA. In case of double transfection, 75 ng of each construct was used with 2 μL of Fugene HD.

METHODS

Cloning. FRET constructs where generated using the previously described FRET sensor library of Piljić et al. 14 human ITPKA (amplified from IRAT p970 H0522D clone purchased from Source Bioscience ImaGenes) and ITPKB (amplified from plasmid 32528 GFP-IP3KB26 purchased from addgene) were amplified by PCR using the following primers. ITPKA: ITPKA-F CGATC TACCG GTGAA GCGGG CGAGG ACGTG GGTCA G, ITPKA-R GGTAC CACGC 1866

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology Twenty to twenty-four hours after transfection, the cells were gently washed and incubated in imaging buffer (20 mmol−1 Hepes, pH 7.4, 115 mmol−1 NaCl, 1.2 mmol−1 CaCl2, 1.2 mmol−1 MgCl2, 2.4 mmol−1 K2HPO4, 2 g L−1 D-glucose) at 37 °C with 5% CO2 for at least 15 min before imaging. Cells were stimulated with either 10 μmol−1 ionomycin or 100 μmol−1 ATP. For imaging of calcium, the calcium sensor R-GECO (CMV-R-GECO) was used. The calcium sensor was cotransfected with the FRET construct. Microscopy. All experiments were performed on a Leica TCS SP2 AOBS microscope (Leica Microsystems) at room temperature (RT). For imaging an HCX PL APO lbd.BL 40.0 ×/1.40 oil objective was used. The excitation and emission settings were kept identical for all experiments (excitation wavelength, 405 nm; emission detection of acceptor, 520−540 nm; donor emission, 470−510 nm). Images were taken with fully opened pinhole, in 8 bit mode and with 10 s between two scans. The cells were stimulated after 100 s with ionomycin at a final concentration of 10 μmol−1. For calcium imaging R-GECO was excited with 532 nm and emission was detected between 580 and 640 nm. Sequenced recording was used in order to prevent the 532 nm laser to excite the fluorophores of the FRET constructs. ATP was added to a final concentration of 100 μmol−1 followed by ionomycin addition. Image Processing and Analysis. Images were analyzed using FluoQ and FIJI,27 a distribution of ImageJ (FIJI is the recommended ImageJ distribution for FluoQ, since the macro makes use of several plug-ins that come with FIJI, but not with plain ImageJ). The following processing options were chosen within the macro: the background was subtracted using ImageJ’s built in function. A median filter (radius size = 2) was used to smooth the images. Before calculating the ratio channel, images were transformed to 32-bit float and a threshold was applied to remove low value pixels from analysis. Cells were segmented automatically by FluoQ using the histogram-based “Triangle” threshold algorithm to create a binary cell mask, the watershed algorithm to separate cell clumps, and finally the particle analyzer plug-in to define ROIs. FluoQ measured the mean pixel intensity of each ROI over time and saved all measured data and calculated parameter in a text file (EXPNAME_data set.txt, see SI Table 2) that was subsequently loaded into the R program 28 in order to do the data analysis. Experiments that showed no or a delayed FRET response to ionomycin due to a pipetting error where excluded from the analysis. Plots were produced using the gglot2 R package.29

■

ACKNOWLEDGMENTS

■

REFERENCES

We thank K. Sauer (Scripps Research Institute) for ITPKB constructs, G. Reither, A. Nadler, V. Laketa, and D. Höglinger for beta-testing of FluoQ and K. Miura for support and advice with the ImageJ macrolanguage. We are grateful for expert support by the EMBL Advanced Light Microscopy Facility in maintaining the used microscopes. This work was partially funded by Transregio 83 of the DFG and the EU-funded Integrated Project LIVIMODE. FluoQ Macro is available free of charge from the Supporting Information or from https:// github.com/fstein/FluoQ.

(1) Grynkiewicz, G., Poenie, M., and Tsien, R. Y. (1985) A new generation of Ca2+ indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440−3450. (2) Miyawaki, A. (2003) Visualization of the spatial and temporal dynamics of intracellular signaling. Dev. Cell 4, 295−305. (3) Newman, R. H., Fosbrink, M. D., and Zhang, J. (2011) Genetically encodable fluorescent biosensors for tracking signaling dynamics in living cells. Chem. Rev. 111, 3614−3666. (4) Putyrski, M., and Schultz, C. (2011) Switching heterotrimeric G protein subunits with a chemical dimerizer. Chem. Biol. 18, 1126− 1133. (5) Swulius, M. T., and Waxham, M. N. (2008) Ca(2+)/calmodulindependent protein kinases. Cell. Mol. Life Sci. 65, 2637−2657. (6) Berridge, M. J., Bootman, M. D., and Roderick, H. L. (2003) Calcium signalling: Dynamics, homeostasis, and remodelling. Nat. Rev. Mol. Cell Biol. 4, 517−529. (7) Kholodenko, B. N. (2006) Cell-signalling dynamics in time and space. Nat. Rev. Mol. Cell Biol. 7, 165−176. (8) Pattni, K., and Banting, G. (2004) Ins(1,4,5)P3 metabolism and the family of IP3−3Kinases. Cell Signal 16, 643−654. (9) Xia, H. J., and Yang, G. (2005) Inositol 1,4,5-trisphosphate 3kinases: functions and regulations. Cell Res. 15, 83−91. (10) Wen, B. G., Pletcher, M. T., Warashina, M., Choe, S. H., Ziaee, N., Wiltshire, T., Sauer, K., and Cooke, M. P. (2004) Inositol (1,4,5) trisphosphate 3 kinase B controls positive selection of T cells and modulates Erk activity. Proc. Natl. Acad. Sci. U.S.A. 101, 5604−5609. (11) Harootunian, A. T., Kao, J. P., Paranjape, S., and Tsien, R. Y. (1991) Generation of calcium oscillations in fibroblasts by positive feedback between calcium and IP3. Science 251, 75−78. (12) Thomas, A. P., Bird, G. S., Hajnoczky, G., Robb-Gaspers, L. D., and Putney, J. W., Jr. (1996) Spatial and temporal aspects of cellular calcium signaling. FASEB J. 10, 1505−1517. (13) Politi, A., Gaspers, L. D., Thomas, A. P., and Hofer, T. (2006) Models of IP3 and Ca2+ oscillations: Frequency encoding and identification of underlying feedbacks. Biophys. J. 90, 3120−3133. (14) Piljic, A., de Diego, I., Wilmanns, M., and Schultz, C. (2011) Rapid development of genetically encoded FRET reporters. ACS Chem. Biol. 6, 685−691. (15) Nagai, T., Yamada, S., Tominaga, T., Ichikawa, M., and Miyawaki, A. (2004) Expanded dynamic range of fluorescent indicators for Ca(2+) by circularly permuted yellow fluorescent proteins. Proc. Natl. Acad. Sci. U.S.A. 101, 10554−10559. (16) Eliceiri, K. W., Berthold, M. R., Goldberg, I. G., Ibanez, L., Manjunath, B. S., Martone, M. E., Murphy, R. F., Peng, H., Plant, A. L., Roysam, B., Stuurman, N., Swedlow, J. R., Tomancak, P., and Carpenter, A. E. (2012) Biological imaging software tools. Nat. Methods 9, 697−710. (17) Carpenter, A. E., Kamentsky, L., and Eliceiri, K. W. (2012) A call for bioimaging software usability. Nat. Methods 9, 666−670. (18) Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9, 671−675. (19) Goedhart, J., van Weeren, L., Hink, M. A., Vischer, N. O., Jalink, K., and Gadella, T. W., Jr. (2010) Bright cyan fluorescent protein

ASSOCIATED CONTENT

S Supporting Information *

Additional figures and tables as described in the text. FluoQ Macro. This material is free via the Internet at http://pubs.acs. org.

■

■

Letters

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. 1867

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868

ACS Chemical Biology

Letters

variants identified by fluorescence lifetime screening. Nat. Methods 7, 137−139. (20) Linkert, M., Rueden, C. T., Allan, C., Burel, J. M., Moore, W., Patterson, A., Loranger, B., Moore, J., Neves, C., Macdonald, D., Tarkowska, A., Sticco, C., Hill, E., Rossner, M., Eliceiri, K. W., and Swedlow, J. R. (2010) Metadata matters: Access to image data in the real world. J. Cell Biol. 189, 777−782. (21) Dubyak, G. R., and el-Moatassim, C. (1993) Signal transduction via P2-purinergic receptors for extracellular ATP and other nucleotides. Am. J. Physiol. 265, C577−606. (22) Zhao, Y., Araki, S., Wu, J., Teramoto, T., Chang, Y. F., Nakano, M., Abdelfattah, A. S., Fujiwara, M., Ishihara, T., Nagai, T., and Campbell, R. E. (2011) An expanded palette of genetically encoded Ca(2)(+) indicators. Science 333, 1888−1891. (23) Berridge, M. J., and Irvine, R. F. (1989) Inositol phosphates and cell signalling. Nature 341, 197−205. (24) Gonzalez, B., Schell, M. J., Letcher, A. J., Veprintsev, D. B., Irvine, R. F., and Williams, R. L. (2004) Structure of a human inositol 1,4,5-trisphosphate 3-kinase: Substrate binding reveals why it is not a phosphoinositide 3-kinase. Mol. Cell 15, 689−701. (25) Carpenter, A. E., Jones, T. R., Lamprecht, M. R., Clarke, C., Kang, I. H., Friman, O., Guertin, D. A., Chang, J. H., Lindquist, R. A., Moffat, J., Golland, P., and Sabatini, D. M. (2006) CellProfiler: Image analysis software for identifying and quantifying cell phenotypes. Genome Biol. 7, R100. (26) Dewaste, V., Moreau, C., De Smedt, F., Bex, F., De Smedt, H., Wuytack, F., Missiaen, L., and Erneux, C. (2003) The three isoenzymes of human inositol-1,4,5-trisphosphate 3-kinase show specific intracellular localization but comparable Ca2+ responses on transfection in COS-7 cells. Biochem J. 374, 41−49. (27) Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J. Y., White, D. J., Hartenstein, V., Eliceiri, K., Tomancak, P., and Cardona, A. (2012) Fiji: An open-source platform for biologicalimage analysis. Nat. Methods 9, 676−682. (28) R Core Team (2012) R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing. (29) Wickham, H. (2009) ggplot2: Elegant Graphics for Data Analysis, Springer: New York.

1868

dx.doi.org/10.1021/cb4003442 | ACS Chem. Biol. 2013, 8, 1862−1868