Evenly Dispersed Gold Beads on TEM Grids

It is often useful to coat the surfaces of electron microscope grids with an irregular electron-dense coating consisting of gold subunits in the range of 5-50 nm. For example, gold beads are useful in providing fiducial markers during alignment of a tomographic series of images (Koster et al., 1992; Fung et al., 1996). They also add some structural detail to the background substrate of a grid which can act as an aid in focusing, particularly for automated focusing algorithms which rely on beam tilt induced image shifts (Koster and Ruijter, 1992). The protocols described here have been developed with this latter application in mind.

We have developed a system, called Leginon, which provides for automated acquisition of cryo electron micrographs (Potter et al., 1999). In the current implementation of the system the automated focusing is usually performed at a magnification of 66,000x. The focusing algorithm uses cross correlation techniques to measure the displacement between two images acquired with different beam tilts. The frame size of the digitally acquired images is 256x256 and at this magnification each frame corresponds to an area of approximately 700 A². To facilitate the cross correlation algorithm it is desirable for a small number of the gold subunits to be present in each frame captured by the digital camera. These provide large strong features and thus provide a strong, sharp peak in the cross correlation map. The gold beads used in this protocol are appropriate in their size and electron density to this task.

The overall density of the gold beads is critical to the task of automated focusing. If the density of the gold beads is too low, only a limited number of beads will be imaged in each acquired (256x256) frame. This may result in the gold beads present in one frame moving completely out of the view upon beam tilting. This causes problems in the cross correlation algorithms and occasionally results in incorrect results for the focus measurement. Ideally, the gold beads should be spread across the surface at a density that results in a 1:1 ratio of naked carbon and gold beads. The protocols described below result in a variety of final densities on the carbon film.

Protocol

Gold beads were purchased from Sigma (20 nm colloidal gold, concentration approximately 0.01% as HAuCl₄ ; catalog # G1652).

1. Concentrate the beads by spinning them in a microfuge at 6K for about 10 min. The concentrated solution becomes dark red. Concentrate 3 mL beads into about 50 mL.
2. To the dark side of the quantifoil grid (this procedure doesn't work on the side with the shiny rim) add a few mL 0.2 mg/mL polylysine and wait 15 min. The fastest way to make a batch of grids is to spread a large piece of parafilm on the bench, put drops of polylysine down, and lay the grids on these drops, dark side down. For less concentrated beads, us 0.1 mg/mL polylysine and incubate 5 min.
3. Transfer the grid to a drop of water sitting on parafilm. Wait 30sec. Transfer the grid sequentially to two more drops of water, waiting 30sec in between.
4. Lift the grid with forceps and wick dry with filter paper. Allow the grid to dry completely.
5. For very concentrated beads, add 10 mL gold beads to parafilm in a humid chamber and set the grid on top. Incubate a few hours. For less concentrated beads, 10 min is sufficient (more beads stick with longer incubations).
6. Transfer the grid to a drop of water, wait a few seconds, then lift with forceps and drop it face down onto a piece of filter paper. This procedure blots the excess gold beads off evenly, in a way in which wicking the beads off from the side does not achieve. For less concentrated beads, wicking from the side with filter paper is fine.