Abstract:
Ice thickness is a critical parameter in single particle cryo-EM - too thin ice can break during imaging or exclude the sample of interest, while ice that is too thick contributes to more inelastic scattering that precludes obtaining high resolution reconstructions. Here we present the practical effects of ice thickness on resolution, and the influence of energy filters, accelerating voltage, or detector mode. We collected apoferritin data with a wide range of ice thicknesses on three microscopes with different instrumentation and settings. We show that on a 300 kV microscope, using a 20 eV energy filter slit has a greater effect on improving resolution in thicker ice; that operating at 300 kV instead of 200 kV accelerating voltage provides significant resolution improvements at an ice thickness above 150 nm; and that on a 200 kV microscope using a detector operating in super resolution mode enables good reconstructions for up to 200 nm ice thickness, while collecting in counting instead of linear mode leads to improvements in resolution for ice of 50-150 nm thickness. Our findings can serve as a guide for users seeking to optimize data collection or sample preparation routines for both single particle and in situ cryo-EM. We note that most in situ data collection is done on samples in a range of ice thickness above 150 nm so these results may be especially relevant to that community.
摘要:
冰厚度是单粒子低温EM的关键参数-太薄的冰可能在成像过程中破裂或排除感兴趣的样品。而太厚的冰会导致更多的非弹性散射,从而无法获得高分辨率的重建。在这里,我们介绍了冰厚度对分辨率的实际影响,以及能量过滤器的影响,加速电压,或检测器模式。我们在具有不同仪器和设置的三个显微镜上收集了各种冰厚度的脱铁蛋白数据。我们证明在300千伏显微镜上,使用20eV的能量过滤器狭缝对提高较厚的冰的分辨率有更大的影响;在300kV而不是200kV的加速电压下操作可以在150nm以上的冰厚度下提供显著的分辨率提高;并且在200kV显微镜上使用以超分辨率模式操作的检测器可以对高达200nm的冰厚度进行良好的重建,而以计数而不是线性模式收集导致50-150nm厚度的冰的分辨率提高。我们的发现可以为寻求优化单粒子和原位冷冻EM的数据收集或样品制备程序的用户提供指导。我们注意到,大多数原位数据收集是在150nm以上的冰厚度范围内的样品上完成的,因此这些结果可能与该社区特别相关。