DOI: 10.1002/cphc.200500632 Direct Observation of Single Protein Molecules in Aqueous Solution David Grünwald, [a] Andreas Hoekstra, [a] Thomas Dange, [a] Volker Buschmann, [b] and Ulrich Kubitscheck* [a] Single molecule tracking (SMT) by fluorescence video micro- scopy has matured to be a well established method in the last few years. [1] SMT has mainly been applied to molecules on in- terfaces, whereas applications to biophysical or biological problems within axially extended cell biological systems are still relatively rare. [2] In order to yield sufficiently good signal- to-noise ratios cameras equipped with image intensifiers or back-illuminated slow-scan charge-coupled devices (CCDs) were used as spatial detectors. This allows for the imaging of processes such as the movement of viruses or large proteins, and to analyze binding events within the cellular interior. [3] It still remains questionable, however, whether the achievable time resolution is really high enough to follow the trajectories of single protein molecules within solutions like physiological buffers or within the cellular interior. Therefore it is not yet clear, how reliable the results of SMT really are under these im- portant conditions. In this report we demonstrate the imaging and tracking of single protein molecules in a phosphate/Hepes buffer with high sensitivity at frame rates of  350 Hz. The analysis of the mean–square distances extracted from the SMT data yielded the same diffusion constants as obtained from flu- orescence correlation spectroscopy (FCS). This demonstrates that the state-of-the-art single molecule microscopy is capable of resolving the trajectories of single molecules in physiological buffer and that the tracking of molecules inside living cells is possible if frame rates greater than 300 frames per second are achieved. Results and Discussion Streptavidin labeled with five fluorophores Cy5 (SAv; Amer- sham, Biosciences Europe, Freiburg), and a-mouse monoclonal IgG (mAb) with a labeling ratio of four dyes (Alexa Fluor 635, Molecular Probes, Leiden, The Netherlands) were used as mo- bility probes. The diffusion of these probes was analyzed by SMT and FCS in independent measurements. Furthermore we examined the mobility of streptavidin-conjugated Qdots 655- SAv (Quantum Dot Corp., Hayward, CA) by video microscopy. For SMT experiments a user-constructed setup as described in the experimental section was used. FCS measurements were carried out with a ConfoCor2 (Carl Zeiss, Jena, Germany). All samples were highly diluted in a phosphate/Hepes buffer (for detailed conditions, see experimental section). SMT data were processed and analyzed using Diatrack 3.0 from Semasopht to identify, localize and track single molecule signals (Figure 1). [4] The software correlator of the ConfoCor2 was used to correlate the FCS data. All further data processing was performed using Origin 7.5 (OriginLab Corp., Northampton) and ImageJ (Wayne Rasband, NIH). From single particle positions observed in successive frames the mean square displacements (MSD) were determined as a function of time and plotted in Figure 2. A linear fit encom- passing the first three data points of SAv-Cy5 and the first five data points of mAb were used to calculate the diffusion coeffi- cient, D SMT , since MSD = 4 D*t. For SAv-Cy5 we thus determined D SMT = 80  5 mm 2 s(& Figure 2; see also Table 1), while the the- oretically expected D for a 60 kDa protein is 81 mm 2 s. For the mAb we measured D SMT = 42  5 mm 2 s by SMT (* Figure 2). This agrees well with the value of D = 39 mm 2 s reported by Jos- sang and co-workers. [5] Theoretically, a diffusion constant of 57.4 mm 2 s is expected if the mAb is approximated by a sphere, with the molecular weight of 170 kDa. However, this theoreti- cal estimate yielded a slightly greater value, since a mAb re- sembles a trianglar shape rather than a compact sphere. Con- sidering the diffusion coefficient of the smaller green fluores- [a] D. Grünwald, Dr. A. Hoekstra, T. Dange, Prof. Dr. U. Kubitscheck Department of Physical and Theoretical Chemistry Wegeler Str. 12 Rheinische Friedrich-Wilhelms-Universität Bonn 53115 Bonn (Germany) Fax: (+ 49)228 739424 E-mail: u.kubitscheck@uni-bonn.de [b] Dr. V. Buschmann Max-Delbrück-Centrum Robert Rçssle-Str. 10, 13092 Berlin (Germany) Supporting information for this article is available on the WWW under http://www.chemphyschem.org or from the author. Figure 1. Single molecule images of SAv-Cy5 (1-10), mAb-Alexa635 (11-20) and Qdots655-SAv (21-30). Raw and corresponding filtered images are shown on top of each other. Single frame integration times were 2.16 ms (SAv) and 2.7 ms (mAb, Qdots), movies (2000 frames) were acquired at frame rates of 440 and 340 Hz, respectively. (see supporting information for movie data). 812 # 2006 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2006, 7, 812 – 815