Multicolor super-resolution fluorescence imaging via multi-parameter fluorophore detection

Abstract

Understanding the complexity of the cellular environment will benefit from the ability to unambiguously resolve multiple cellular components, simultaneously and with nanometer-scale spatial resolution. Multicolor super-resolution fluorescence microscopy techniques have been developed to achieve this goal, yet challenges remain in terms of the number of targets that can be simultaneously imaged and the crosstalk between color channels. Herein, we demonstrate multicolor stochastic optical reconstruction microscopy (STORM) based on a multi-parameter detection strategy, which uses both the fluorescence activation wavelength and the emission color to discriminate between photo-activatable fluorescent probes. First, we obtained two-color super-resolution images using the near-infrared cyanine dye Alexa 750 in conjunction with a red cyanine dye Alexa 647, and quantified color crosstalk levels and image registration accuracy. Combinatorial pairing of these two switchable dyes with fluorophores which enhance photo-activation enabled multi-parameter detection of six different probes. Using this approach, we obtained six-color super-resolution fluorescence images of a model sample. The combination of multiple fluorescence detection parameters for improved fluorophore discrimination promises to substantially enhance our ability to visualize multiple cellular targets with sub-diffraction-limit resolution.

Figure 1

Photoswitching behavior of Alexa 750 and comparison with Alexa 647. a) Top: Blinking of fluorescence emission from Alexa 750 under continuous 752 nm illumination. Bottom: Reversible light-driven switching of the Cy3-Alexa 750 dye pair under continuous 752 nm illumination (indicated by the red bar) and periodic pulses of 532 nm green illumination (indicated by the green bars). The 532 nm light efficiently re-activates Alexa 750 after it is switched off by the 752 nm light. b) Characterization of the number of photons per switching cycle (red bars) and equilibrium blinking fraction (or on-off duty cycle, blue bars), for Alexa 750 in comparison with Alexa 647. Switching characteristics were measured for buffers containing one of two different primary thiols: β-mercaptoethanol (BME) or β-mercaptoethylamine (MEA).

Figure 2

Comparison of two-color STORM images taken with the dual-emission channel scheme and the dual-activation channel scheme. a–c) Dual-emission channel STORM imaging of microtubules and mitochondria in BSC-1 cells immuno-stained with Alexa 750 and Alexa 647, respectively (no activator fluorophore was used). d–f) Dual-emission channel STORM imaging of microtubules and mitochondria in BSC-1 cells immuno-stained with the Cy3-Alexa 750 dye pair and the Cy3-Alexa 647 dye pair, respectively. g–i) Dual-activation channel STORM imaging of microtubules and mitochondria in BSC-1 cells immuno-stained with the Alexa 405-Alexa 647 pair and the Cy3-Alexa 647 pair, respectively. The STORM data shown in panels g–i have been corrected for color crosstalk (see Supporting Information). Scale bars 500 nm.

Figure 3

Six-color STORM imaging using multi-parameter detection. a) Six-color STORM image of streptavidin molecules on glass. Streptavidin was dual-labeled with one of three activator fluorophores (Cy3, Cy2, or Alexa 405) and a photo-switchable reporter fluorophore (Alexa 647 or Alexa 750), and bound to a glass coverslip coated with biotin (inset). Each fluorophore localization event was color coded according to the wavelength of the activation pulse which preceded it and the fluorescence channel in which it was detected. Scale bar 500 nm. b–e) Detailed views of the boxed regions in (a) showing the localizations plotted as crosses. Tightly clustered groups of localizations were predominantly composed of a single color, revealing the labeling of each streptavidin molecule. Scale bars 25 nm.