CRYOGENIC ELECTRON MICROSCOPY

1. Particle mass/size

A) Recommended minimum size of a particle is around 80kDa

for Cryo-EM. Preferred particle size is > 150 kDa. Lower sizes may be possible but likelihood of achieving resolutions better than 4Å drops off significantly as mass drops below 80 kDa.

B) The metrics above refer to contiguous ordered mass, i.e. mass that behaves roughly as a single rigid body.

2. Symmetry

Internal symmetry within the target particle is highly beneficial.

3. Purity

Can tolerate somewhat lower purity than crystallization, especially with large particles. Small particles < 150 kDa should still be at similar purity as for crystallography (at least 90% purity).

4. Particle properties

Unfolded/disordered regions are particularly prone to destructive interactions with the air-water interface. Strongly recommend construct design and/or protein engineering to remove these if possible. If there are no related experimental structures available, AlphaFold is useful for this analysis.

5. Air-water interface

At least 75% of projects we see are not well behaved when frozen on normal holey grids, due to destructive interactions at the surface of the thin film, during the hundreds of milliseconds between blotting and vitrification. Amelioration requires sub-CMC detergent to act as surfactant (preferred, a normal part of our Milestone 1 screening below), or adsorption to a continuous surface prior to blotting (unpredictable, grids are more difficult to work with, only used as necessary as an extension to Milestone 1 grid screening).

Sample Requirements:

1. Preferred particle concentration and volume

for Cryo-EM work is 100 μl of 3 to 4mg/ml. Lower concentrations are not a problem in principle, however lower concentrations mean fewer particles per imaged area, which can (a) limit resolution or (b) drive up costs if we need to compensate by adding extra days of data collection to reach a target resolution.

2. Sample buffer

A) Excess sample buffer must be provided for making dilutions during screening.

B) Salts (e.g., NaCl) concentration > 0.5 M results in contrast reduction, avoid if at all possible.

C) Must keep any ingredients that increase solution viscosity at low concentration. Notable examples are glycerol or sucrose. Can tolerate up to ~ 4% glycerol if needed, however this carries the risk of extending timelines if this negatively impacts grid preparation reproducibility (i.e.., if we need to prepare many duplicate grids in order to catch one that freezes nicely). If shipped at very high sample concentration, dilution into a glycerol-free buffer has been a successful strategy; otherwise recommend gel-filtration polishing into a more compatible buffer.

Milestone 1: Grid screening

A) Sample is vitrified over 2-3 dilutions and a detergent screen

B) A small number (3-5 per grid square) of images are taken from

3 grid squares representing the range of apparent ice thickness available on the grid

C) Evaluation by eye for optimal particle monodispersity, density, and morphology

D) Deliverables:

i. The images taken, in .mrc and .png formats

ii. Written report on expert opinion and next steps

E) If prognosis is poor but there is high confidence in the input material: recommend extended milestone 1 to examine behavior on continuous surface grids (amorphous carbon, graphene oxide)

Milestone 2: Pilot dataset and processing

A) Collect ~ 500 exposure movies, using the same conditions/setup as would be used for a full dataset. From either the best milestone 1 grid or a new grid based on optimization from milestone 1 results.

B) Microscope used (Krios versus Glacios) will depend on project needs and/or microscope availability.

C) Data is also processed like a full dataset (see notes for milestone 3), as far as it will go. Goals are to determine:

i. Are the visible particles/material capable of being aligned and averaged (2D classification)?

ii. Do the 2D class averages meet the expectations for the target particle (e.g., is a subunit missing, is the conformation as desired)?

iii. Is there a diversity of particle views, or is preferred orientation problem?

iv. If all is well in 2D, attempt to jump to 3D for better insight into

(i) and (ii) above.

D) Deliverables:

i. Written report with expert analysis

ii. Images of 2D classes if successful (.mrc and .png)

iii. 3D refined volume or classes (.mrc)

iv. Raw data available upon request

Milestone 3: Full dataset collection and processing

A) Collect ~ 4000 to 5000 exposure movies using a 24 hour time slot on a Titan Krios microscope. When possible, grid is prepared fresh, incorporating any optimizations indicated by milestones 1 and 2.

B) Data is processed to its full potential -- when refinement (including CTF refinement, per-particle motion correction, and multi-body refinement, where appropriate) and classification cease to yield resolution improvements.

C) Data processing pipeline is based primarily but not exclusively on the Relion software package. Exact pipeline is proprietary and varies with the needs of the target particle, but if Helix is included in journal publications, full assistance will be provided including figures, methods text, etc.

D) Deliverables:

i. Written report with expert analysis, including key facts and figures (final particle count, resolution, FSC curves, particle angular distribution, sharpening and local resolution information)

ii. Images of 2D classes from final particle set (.mrc and .png)

iii. Final 3D refined volume(s) (.mrc):

• Half maps

• Full map

• Sharpened map

• Sharpened map, filtered to local resolution

iv. Raw data available upon request

Milestone 4: Atomic model building

A) Requires starting model/fragment, or high enough resolution to ‘read’ sequence directly from the map.

B) Deliverables:

i. “Table 1” summary of model geometry statistics and map-model fit

ii. Atomic in .pdb and .cif format

• Data processing pipeline is based primarily but not exclusively on the Relion software package. Exact pipeline is proprietary and varies with the needs of the target particle.

• CryoSparc industrial license is prohibitively expensive. In the hands of experienced users, Relion provides all the utility and effectiveness of Cryosparc, albeit without the professionally designed user interface.

Helix relies on ThermoFisher Krios and Glacios microscopes. All 24-hour data acquisitions are completed on a Krios with K3 energy filter and camera, while most screening is performed on Glacios or the Krios K3 based on availability. Helix maintains primary scopes as well as contingency options to assure data screening and collection can occur without significant interruption.