LOW TEMPERATURE ELECTRON MICROSCOPY

Abstract

In the investigation of biological ultrastructure by electron microscopy the most significant observation to date is a 1 nm repeat in the purple membrane of Halobacterium halobium (73). With a high degree of certainty, the spacing can be ascribed to the inter a-helical separation within bacteriorhodopsin molecules. This resolution is only made possi­ ble because of the unique properties of the purple membrane and is due in particular to its two-dimensional crystallinity, which results in a repetition of the structural data. For a nonperiodic object, the meaning­ ful resolution often lies between 3 and 5 nm, as illustrated by comparing two models that have been proposed independently for the structure of ribosomes (45, 82). More frequently, it is in even larger structures that uncertainties become apparent. For example, there are good arguments for questioning the validity of the unit membrane as it is seen in most ultrathin sections (62), and electron microscopy has not yet been able to settle the question of whether chromatin fibres are 10, 20 or 30 nm in diameter. In short, the resolution obtained with biological structures is disappointing when compared with the resolving power of modern electron microscopes of between 0.2 and 0.5 nm. It is thus not the performance of the electron microscope that is the limiting factor, but the damage caused during specimen preparation and irradiation by the electron beam (4). Cryoelectron microscopy has long been considered as a possible avenue to overcome both of these limitations. On the one hand it was anticipated that the beam would have a less damaging effect on cooled specimens (i.e. cryostabilization), whereas on the other hand, freezing