I am purifying a membrane protein in DDM/CHS buffer and freezing cryoEM grids using standard parameters. However, I notice the ice in the middle of the holes is extremely thin, and many holes are broken or break during data collection due to how thin the ice is (image below). I have tried to alter Vitrobot parameters (reducing the blot force and reducing blot time) but ice thickness in the middle of the holes does not significantly improve. By contrast, same protein sample in different detergent (e.g. LMNG/CHS) using the same freezing conditions does not lead to this hole issue. DDM/CHS is the more suitable biochemical condition for my sample, so I am trying to fix this freezing issue with DDM/CHS.
I’ve seen many papers where groups successfully solve structures in DDM/CHS, which makes me think this issue is not inherent to the detergent and should be fixable.
Specifically, when preparing my sample, I use 0.03% DDM/ 0.003% CHS in my SEC buffer. Then I concentrate ~50-100 fold using a 100kDa concentrator to my desired concentration. Finally, I apply 3ul to R1.2/1.3 Quantifoil Au cryoEM grids using a Vitrobot Mark 4 (standard freezing protocol).
I am hoping someone experienced in this issue can advise.
I’ve never had that issue with similar compositions (quite the opposite), however I would try reducing the detergent further (as it tends to concentrate along with the protein). Typically the less detergent there is, the less protein concentration is required as well. If necessary a small amount of a lower CMC detergent can be supplemented (LMNG or GDN) to keep the protein in solution at low % total detergent. Adding 1-5% glycerol may also help.
I would also review the total dose you are using and check at what frame in the movies your holes are melting - perhaps you can use less exposure.
You can also try using more volume - perhaps the most precise element of vitrification - such as 3.5 - 5µl. It can also help to use “wet” filter papers that have sat at 100% humidity for a few minutes and to ensure high humidity by placing one or two small squares of sponges, inundated with water, into the humid box. When you add the sample to the grid, I also recommend using a “dragging” motion to spread the droplet.
If these are 200 mesh grids, you could also get thicker ice using 300 or 400 mesh. (I prefer 300 mesh for single particle application due to the good combination of high collection rate/grid real estate and resilience). And of course using smaller holes (0.6/1.0) as well.
