BRIEF COMMUNICATIONS NATURE METHODS | VOL.6  NO.9  | SEPTEMBER 2009  | 643 organisms change position and shape because of tissue growth and morphogenesis. Thus, the cell membranes move constantly in the detection volume and are not sufficiently stable for FCS meas- urements. Moreover, membrane movements result in a false posi- tive cross-correlation, as proteins with different fluorescent labels appear to be moving together even in the absence of binding. (ii) There is a mismatch between the refractive index of water and of the tissue, leading to a distortion of the detection volume in dependence of the focusing depth. This prohibits a calibration necessary for quantitative measurements. (iii) Spectral cross-talk can result in false positive cross-correlation signals. In this study on fibroblast growth factor receptor-ligand inter- action in zebrafish embryos, we combined the strengths of sev- eral recently developed FCS techniques in a modular setup to overcome the above-mentioned limitations (Supplementary Fig. 1). In particular, we used scanning FCS 11 , in which the detec- tion volume is repeatedly scanned through a vertical membrane perpendicularly (Fig. 1, Supplementary Notes 1–4 and Online Methods). Here (i) we process data off-line to correct for mem- brane movements; (ii) we implement two adjacent foci to circum- vent the necessity for a calibration of the detection area; (iii) we combine alternating dual-color excitation with scanning FCS to prevent spectral cross-talk and artifacts induced by membrane movements. In combination with static FCS, this allowed us to assess the mobilities of the receptors Fgfr1 and Fgfr4 and to quan- tify their in vivo binding affinities to their ligand Fgf8. Fibroblast growth factors are important secreted proteins with diverse roles in regulating cell proliferation, migration and differentiation in vertebrate development 12 . They act through fibroblast growth factor receptors to initiate signaling events. In vertebrates, 27 fibroblast growth factors and 5 fibroblast growth factor receptors (Fgfr1–4 and Fgfr1-like) have been described, but the receptor-ligand binding specificity in vivo has not been resolved. Here we concentrate on Fgf8, a key signaling molecule important for dorsal-ventral patterning, appendage and brain formation during vertebrate development 12,13 . In gastrulating zebrafish embryos, Fgf8 activates its downstream target genes spry4, pea3 and erm in successively broader domains away from its source 13,14 (Supplementary Fig. 2a). Both Fgfr1 and Fgfr4 are expressed in the responding tissue (Supplementary Fig. 2b) and are putative candidates for transducing Fgf8 signals. Previous in vitro experiments suggest that Fgfr4 has 20-fold higher activity for Fgf8 than Fgfr1 (ref. 2), whereas in vivo experiments indicate that Fgfr1 is likely to be a major receptor for Fgf8 (refs. 4–7). For example, Fgfr1 knockout mice have severe gastrulation defects and mesodermal patterning impairment 4,5 , but Fgfr4 knock- out mice have no apparent phenotype 6 . In zebrafish embryos, functional knockdown of Fgfr1 using morpholino oligos results Modular scanning FCS quantifies receptor-ligand interactions in living multicellular organisms Jonas Ries 1,3 , Shuizi Rachel Yu 1–3 , Markus Burkhardt 1 , Michael Brand 1,2 & Petra Schwille 1,2 Analysis of receptor-ligand interactions in vivo is key to biology but poses a considerable challenge to quantitative microscopy. Here we combine static-volume, two-focus and dual-color scanning fluorescence correlation spectroscopy to solve this task at cellular resolution in complex biological environments. We quantified the mobility of fibroblast growth factor receptors Fgfr1 and Fgfr4 in cell membranes of living zebrafish embryos and determined their in vivo binding affinities to their ligand Fgf8. Receptor-ligand interactions have essential roles in many biologi- cal systems and are responsible for activating intracellular sign- aling cascades. Various in vitro assays have been developed that determine binding affinities with high precision 1 . However, the measured data do not necessarily reflect the complexity of a living organism and can sometimes produce inconsistent results 2,3 . In contrast, in vivo data from loss-of-function experiments or co- localization studies provide mainly indirect or qualitative infor- mation 4–7 . The direct measurement of receptor-ligand affinities in the living tissue is challenging, and thus it is difficult to obtain a comprehensive picture of signaling processes. A technique with great potential to solve this task is fluores- cence correlation spectroscopy (FCS). It is based on detecting fluorescence fluctuations from a small confocal volume (~0.5 fl). Statistical analysis by autocorrelation of these fluctuations pro- vides quantitative information on local concentrations and diffu- sion coefficients of fluorescent molecules present (Supplementary Note 1). By cross-correlating fluorescence fluctuations in two spectral channels, bimolecular binding can be inferred because only co-diffusing binding partners lead to a considerable cross- correlation. Conventional FCS uses a static detection volume. It is an established technique for studying molecules in solution and living cells 8 , and its implementation in living organisms has recently been developed 9 . However, its application to receptor- ligand interactions in vivo, albeit frequently suggested, has been limited. The main reasons are as follows 10 . (i) Cells of living 1 Biotechnology Center and 2 Center for Regenerative Therapies, Technical University of Dresden, Dresden, Germany. 3 These authors contributed equally to this work. Correspondence should be addressed to M.Brand (michael.brand@biotec.tu-dresden.de) or P.S. (petra.schwille@biotec.tu-dresden.de). RECEIVED 23 APRIL; ACCEPTED 9 JUNE; PUBLISHED ONLINE 2 AUGUST 2009; DOI:10.1038/NMETH.1355 © 2009 Nature America, Inc. All rights reserved.