大鼠原代脂肪细胞中单个GLUT4囊泡的成像——全内反射荧光显微成像技术的应用

Real-time and dynamic tracking of single GLUT4 vesicles in primary rat adipocytes——application of total internal reflection fluorescence microscopy

本研究采用高斯拟合算法与全内反射荧光显微技术(TIRFM),通过实时动态的单粒子追踪拍摄来分析原代培养大鼠脂肪细胞基础条件及胰岛素刺激下细胞膜附近绿色荧光蛋白(GFP)标记的GLUT4囊泡的运动变化.使用高斯拟合来消除光学成像系统中发生的原始样品中单像素点荧光强度值的污染,重建出囊泡真实的中心点荧光强度值.通过对序列图像中的囊泡点应用高斯拟合和相应的搜索算法,得到所有囊泡的高斯中心点荧光强度值序列,继而描绘出囊泡的动态转运路径.结果表明,基础条件下膜附近的多数GLUT4囊泡处于沿同一路径的快速长距离转运状态;胰岛素刺激后,长距离转运的GLUT4囊泡数量减少,多数GLUT4囊泡固定成团或者在原地摆动呈限制性的运动.当GLUT4囊泡锚定至膜上后,即迅速与膜融合.说明高斯拟合算法及全内反射荧光显微技术的结合可以很好地实现囊泡纳米尺度的三维实时追踪.

Combining of Gauss distribution regression fitting method with total internal refection fluorescence microscopy （TIRFM）, the real-time and dynamic intermediate movements of the green fluorescent protein （GFP） tagged single GLUT4 vesicle in the primary rat adipocytes in basal and insulin stimulated states were studied. The original fluorescence intensity of single pixel in sample will be tainted and diffuse to the pixels in the vicinity. The nature of this kind of artifact is almost alike to Gauss distribution. Therefore, Gauss distribution regression fitting was utilized to eliminate this artifact, and as a result to obtain the real fluorescence intensity of the central pixel of a vesicle, the width of the fluorescence vesicle and this pixel＇s coordinate. Similarly utilizing this fitting to the time-lapse images the coordinates of central pixels of the vesicles were gained and ultimately their moving trajectories were depicted. The results indicated that in the basal state, most of the GLUT4 vesicles were in rapidly long-range movements along the predefined trajectories, but insulin reduced the long-range movements and made the vesicles to cluster or in restricted movements. When docked, the GLUT4 vesicles were fused to the PM quickly. The primary function of insulin to GLUT4 translocation is likely to mobilize, dock and fuse the GLUT4 vesicles to specific sites in the PM. Consequently, the nanometer wide single particle 3D-tracking could be achieved by the combination of Gauss distribution regression fitting method with TIRFM.