Showing surfaces of molecule complexes in membranes

Tom Goddard
June 18, 2024

We go through how to show surfaces of molecular complexes seen on membrane surfaces in cryoEM tomograms. Because the tomograms are noisy we will look at just a layer near a segmented membrane and smooth the density and hide small connected surfaces that are noise. An alternative for visualizing membrane proteins is to use orthogonal slices.

Lysosome with 5 membrane complexes on bottom side.

Segmentation mask of membranes shown in yellow.

Example: V-ATPases on lysosomes

We try to display V-ATPases sticking out from lysosome membrane surfaces. Here is a part the tomogram lysosome.mrc from the Chan-Zuckerberg Imaging Institute.

We will use a segmentation lysosome_membranes.mrc of the membranes in the tomogram created with membrain-seg using command

      membrain segment --tomogram-path lysosome.mrc
               --ckpt-path MemBrain_seg_v10_alpha.ckpt --out-folder .
    

Open these in ChimeraX (using version 1.8)

      open lysosome.mrc
      open lysosome_membranes.mrc
    

Smooth lysosome surface

Lysosome surface colored.

Lysosome surface separated.

The segmentation of the membrane is mask with 0 and 1 values so the surface of this mask has stair-step artifacts. First we'll make a smooth version of this surface in ChimeraX.

      volume #2 step 1 level 0.5
      surface smooth #2 iterations 100
    

To separate just the lysosome surface we use the Right Mouse Toolbar tab and click the Blob mouse mode, then click with the right mouse button on the lysosome membrane to give it a new color (blue). Then we split the surface into the blue and yellow part with the "surface splitbycolor" command. We can save the lysosome surface to a file lysosome_surface.stl

      surface splitbycolor #3
      save lysosome_surface.stl model #4.1
    

Invert, mask and smooth tomogram

We invert the map so the density for membrane and complexes have positive values. Then we mask to a range of about 30 nm from the membrane (60 grid points since pixel size is 0.5 nm). Then smooth the map. Then hide small surface blobs. Using the Blob mouse mode we can color the membrane complexes different colors by clicking on each.

      volume scale #1 factor -1e5
      volume mask #5 surface #4.1 extend 60
      vol gaussian #6 sdev 25
      surf dust #7 size 200
    

Tomogram inverted.

Masked to layer near membrane.

Smoothed density.

Dust hidden, particles colored.

Observations

1. Connected surfaces. This example was an easy case in a few ways. The lysosome membrane surface was not touching any other membranes so the Blob mouse mode that colors connected surfaces could separate it. Likewise the membrane molecular complexes were not connected to other density after smoothing so the Blob mouse mode worked to color those also.
   
   
2. Masking crops too much. The "volume mask" command produces a minimal box containing the surfaces and did not increase the box size when I used the "extend 60" option. I fixed that bug in ChimeraX daily build June 19, 2024 or newer. In older ChimeraX you can add the "fullMap true" option to the volume mask command so it does not use a minimal box.
   
   
3. Orthoplane visualization. There are other ways to look at membrane complexes in ChimeraX with orthogonal slices. But I think if the surfaces can be shown without being obscured by crowding and noise they give a clearer few of the particles.
   
   
4. Missing wedge artifacts. The membrane complexes look stretched vertically because of tomogram missing wedge artifacts.
   
   
5. Mouse mode. Would be useful to have a mouse mode that allowed clicking on a particle in a tomogram plane to show its surface with automatic smoothing and hiding dust and disconnecting weakly connected density. This would allow assessing particle shape and size in a simpler way then looking at slices.