Standard EM methods yield two-dimensional projections. Thus, researchers such as Gary Borisy (Northwestern University, Chicago, IL) have studied systems that approximate two dimensionality, such as the flattened lamellipodia at the front of a moving cell. But, says Borisy, “there is no substitute for three-dimensional data. Even in the systems we have analyzed I would love to have three-dimensional data.”
The tomogram provides these data by rotating the sample between sampling runs. After each rotation the sample must be refocused and realigned—a process that the German group automated to minimize beam damage to the specimen. The sample itself was prepared by quick freezing. This eliminates fixation artifacts and leaves membrane systems intact.
The group observed linkages of actin filaments both to the membrane and, at a wide variety of angles, to each other. Assigning these linkages to specific actin-binding proteins will require either immunological labeling (via microinjection of gold-conjugated antibodies) or pattern recognition of the linking proteins. The 2.5-MD proteasome was sufficiently hefty and distinctive to be recognized by the group, but identifying diminutive actin-binding proteins will be a far greater challenge.
To meet that challenge, the Baumeister laboratory has recently shipped in a new microscope cooled by liquid helium rather than liquid nitrogen. The lower temperature should allow longer exposures to electrons with less damage, thus giving the laboratory a shot at increasing resolution from the current 5–6 nm to a projected 2 nm. “In two years we will know if this is the answer,” says Medalia.
But even with the existing resolution there is plenty to do. Borisy, for one, wants to know how lamellipodia keep themselves flat by restricting upwards growth of actin filaments. “This [current study] is to show that indeed this technique works,” says Medalia. “Now it will spread and people can do as much as their imagination allows.” ▪