Averaging of Viral Envelope Glycoprotein Spikes from Electron Cryotomography Reconstructions using Jsubtomo

The overall goal of this procedure is to study the structures of glycoprotein complexes present on the surfaces of enveloped viruses. This is accomplished by first locating the virus particles in the tomographic volumes that have been previously calculated from electron cryo microscopy data. The second step is to create two statistically independent initial models of the glycoprotein complexes using a manually defined subset of data.

Next, these models are refined and then used to automatically detect all glycoprotein complexes on the virus particles. The final step is to refine the structure of the glycoprotein complex using the full data set. Ultimately, molecular visualization tools are used to interpret the final structure.

Three dimensional structural analysis of these glycoprotein spikes is invaluable for understanding viral pathogenesis as well as in drug design. The main advantage of this technique over existing methods, such as x-ray crystallography, is that this technique allows solving structures of membrane proteins in the natural membrane environment. This technique can help to address key topics in structural biology.

For example, the structure of many membrane containing viruses. Although this technique can provide insight into glycoprotein complexes in these viruses, it can also be applied to other systems such as membrane proteins incorporated into liposomes. The procedure will be demonstrated by two members of my group, Slee and David Bittel.

To begin open a autogram file in B show. Pick virus particles manually in the grams by defining the center coordinate of a virus particle using the particle picking tool. Repeat this step until all variants have been processed.

Save virus coordinates in a star file. Proceed to repeat this process for all tomos. Next run J sub tomo py in extract mode To extract vari on sub volumes into individual volume files, use the saved star files as input files to generate two independent initial models.

First, open ovarian sub volume map file. In B show. Pick a subset of spikes in the Varian sub volumes by defining the center coordinate of a spike using the particle picking tool accessible via the toolbox window.

Repeat this step until all clearly distinct spikes have been processed. Save spike coordinates in a star file. Repeat this step until approximately 200 glycoprotein spikes have been processed.

After assigning views to the varians as described in the text protocol, create a real space mask using J sub tomo. Create mask py. Then generate a reciprocal space mask using J sub tomo create wedge mask py.

The reciprocal space mask is used to exclude regions in the missing wedge region, resulting from single axis tomographic data collection. Next, create two initial averages using J.Sub tomo create averages. PY using a saved selection file as the input file.

Use ization to reduce noise in the initial averages is iteratively align and average the two previously generated models with J sub tomo iterate gold PY.Details for both stages of this process can be found in the text protocol. Assess the degree of symmetry in the structure by examining the resulting low-pass filtered map file in kyira. For example, if the spike is a trimeric complex, threefold symmetry should be evident.

Continue refinement of the structure as detailed in the text protocol. Next, generate seeds located evenly on the Varian surface for template matching and assign an initial view vector to the seeds by running JVs py. Use the previously generated star files as input files generate approximately 1.5 times more seeds than the expected number of spikes.

This view vector approximates the direction of the spike closest to each seed point. Next, generate two independent averages of the virus surface as before. Then refine the position of the seeds as detailed in the text protocol.

Finally, generate com marker files of the refined seed star files using jvs.py. Examine the seeds by opening the CMM files and the associated virion map files. In kyira.

Make sure that the refined seeds are aligned correctly relative to the virus membrane. The refined markers can be colored based on their cross correlation Coefficients automatically locate all the spikes in the Varian sub volumes using local template matching around refined seeds align and average the located spikes use the averages generated from a subset of manually picked spikes as initial templates. Additional details for this process can be found in the text protocol to visualize the results, open the refined structure of the spike in kyira for visualization.

Proceed to fit the atomic structures. Then create a composite model of the virion using j Subo Create model DO py. Finally open the composite model for visualization in kyira.

The initial model was refined using 205 manually picked spikes.Threefold. Symmetry of the centermost spike was evident without applying any symmetry and was imposed on the subsequent rounds of refinement for detecting all the spikes manually on the Varian surfaces. 106 seeds were generated for each vion at the radius of 43 nanometers and spacing of 20 degrees, and their positions were iteratively refined relative to the membrane.

The best correlating glycoprotein spike patches were used to calculate the final average. The average was resolved to 35 angstrom resolution. It revealed a trimeric spike structure in the middle in addition to some contribution from six neighboring spikes.

Composite models of the varis calculated by placing the structure in the known positions revealed placement of spikes on the Varian surface. Occasionally, locally ordered patches of spikes were evident. While attempting this procedure, it's important to use only the best data and to carefully check all the intermediate results.

Following this procedure, x-ray crystallography structures can be fitted into the final average to get a more accurate view of protein-protein interactions. After watching this video, you should have a good understanding of how to average viral envelope glycoprotein structures using J Omo.