2.1. Multiphoton Microscope

The multiphoton microscope used in this study is a home-built, inverted system. The 780-nm output of a diode-pumped solid-state laser (Millennia, Spectra Physics, Santa Clara, California) titanium-sapphire laser (Tsunami, Spectra Physics) was used as the excitation source. Upon reflecting from a galvanometer-driven x−y mirror system, the laser is guided into an inverted microscope (TE2000U, Nikon, Japan). A pair of beam expanders at the input of the microscope is used to enlarge the laser spot in order to ensure overfilling of the focusing objective. The expanded beam is then reflected into the focusing objective by a primary dichroic mirror (720DCLR, Chroma Technology, Bellows Falls, Vermont). The objective used was an air lens (Nikon S Fluor 20×/NA0.75) with a working distance of 1 mm, and the on-sample power was 50 mW. The signal generated at the focal volume is collected by the focusing objective in the epi-illuminated geometry and separated into a generic four-channel photomultiplier detection system. The use of additional dichroic mirrors (550DCXR, 495DCXR, and 420DCLR) and bandpass filters (red: 610/75 nm, green: 525/50 nm, blue: 460/50 nm, and second-harmonic generation: 390/18 nm) allow the separation of the signal into three fluorescence and one backward second-harmonic generation channels. In this study, we only used the blue and red channels for the detection of Hoechst 33342 labeled nuclei and CellTracker^™ Orange CMTMR labeled cytoplasm, respectively. Each optical image plane is 253×253μm2 in size. To achieve a 3-D image, the specimens are placed on a three-axis specimen translation stage (ProScan^®, Prior Scientific, United Kingdom). After an optical image is acquired in the image plane, the specimen is translated to the next axial position for further image acquisition. In this manner, we acquired 3-D images separated by 1μm. A 3-D image stack is typically composed of 20 to 30 images.