Effects of chameleon dispense-to-plunge speed on particle concentration, complex formation, and final resolution: A case study using the Neisseria gonorrhoeae ribonucleotide reductase inactive complex

Ribonucleotide reductase (RNR) is an essential enzyme that converts ribonucleotides to deoxyribonucleotides and is a promising antibiotic target, but few RNRs have been structurally characterized. We present the use of the chameleon, a commercially-available piezoelectric cryogenic electron microscopy plunger, to address complex denaturation in the Neisseria gonorrhoeae class Ia RNR. Here, we characterize the extent of denaturation of the ring-shaped complex following grid preparation using a traditional plunger and using a chameleon with varying dispense-to-plunge times. We also characterize how dispense-to-plunge time influences the amount of protein sample required for grid preparation and preferred orientation of the sample. We demonstrate that the fastest dispense-to-plunge time of 54 ms is sufficient for generation of a data set that produces a high quality structure, and that a traditional plunging technique or slow chameleon dispense-to-plunge times generate data sets limited in resolution by complex denaturation. The 4.3 Å resolution structure of Neisseria gonorrhoeae class Ia RNR in the inactive α4β4 oligomeric state solved using the chameleon with a fast dispense-to-plunge time yields molecular information regarding similarities and differences to the well studied Escherichia coli class Ia RNR α4β4 ring.

Particle distributions within the ice of the 54 ms data set 2 from a chameleon-plunged grid. (c) Particle distributions within the ice of the Vitrobot-plunged grid. All scale bars indicate the ice thickness measured at the edge of the hole. (d) Average ice thicknesses of the center (stripes) and edge (solid) of holes found at the edge (blue), halfway (orange), or center (teal) of grid squares in two fast-plunged chameleon grids. For location in grid square, the ratio of center hole thickness to edge hole thickness was 1 : 1.1-1.5 whereas the ratio of halfway hole thickness to edge hole thickness was 1 : 1.1-1.3 (range indicates variation in comparing thickness of hole edges versus hole centers). (e) Average ice thicknesses of holes from the Vitrobot-plunged grid. For location in grid square, the ratio of center hole thickness to edge hole thickness was 1 : 0.9-1.0 whereas the ratio of halfway hole thickness to edge hole thickness was 1 : 1.0-1.1 (range indicates variation in comparing thickness of hole edges versus hole centers). Number of holes in each sample is indicated above the respective bars. Tomography data processed using IMOD (Kremer et al., 1996). Figure 5. Representative example of determination of fraction of intact rings in a data set. (a) All 2D classes (b) All rings (c) Intact rings. The fraction of intact rings was determined by dividing the number of particles contributing to the 2D classes present in (c) by the number of particles contributing to the 2D classes in (b). The data set displayed is from the grid plunged using the chameleon with a 390 ms dispense-to-plunge time. 2D classification was completed using Relion 3.0 (Scheres, 2012). Figure 6. 3D classification results for all analyzed data sets. Six classes and 100 rounds of classification were specified in Relion for each data set. Particle inputs to the 3D classification were those that contributed to 2D classes consisting of intact rings or ambiguous side views in each of two rounds of 2D classification. All classes are visualized at an RMSD value of 6. 3D classification was completed using Relion 3.0 (Scheres, 2012). Figure 7. Ice thickness varies among the grids used in this study. Atlases taken of each grid that was used to collect a data set. The chameleon-plunged grids have a strip of vitrified sample (indicated with a white rectangle), whereas the Vitrobot-plunged grid has an ice thickness gradient across the grid that is typical of a traditional blotting-based grid preparation technique.

Supplementary Figure 8. Preferred orientation analysis for all data sets analyzed in this study. (a)
Angular distribution plots generated in the Relion 3D autorefine program corresponding to data sets from each chameleon grid and the Vitrobot grid visualized in Chimera (Pettersen et al., 2004). Each processed data set contains 7492 particles. The width of the cylinders varies due to the use of different angular sampling parameters for lower resolution versus higher resolution reconstructions. A larger or redder cylinder indicates more particles contributing to that view of the reconstruction. Each map was generated through the Relion postprocess step and is shown at an RMSD value of 6 (Scheres, 2012). (b) 3D FSC plots indicating the Fourier shell correlation as a function of spatial frequency for each data set in the x-direction (blue), y-direction (green), and z-direction (red), as well as the overall global FSC curve (goldenrod) and the average cosine phase (black). Plots were generated using the 3DFSC web server (3dfsc.salk.edu) (Tan et al., 2017).

Supplementary Figure 9. Postprocessed maps of the untruncated data sets colored by local resolution.
Local resolution map of the data set generated from the chameleon-plunged grid (54 ms dispense-toplunge time, data set 2; top) and the Vitrobot-plunged grid (bottom). Local resolution was calculated using the Relion LocalRes program and was visualized in Chimera (Pettersen et al., 2004;Scheres, 2012). The two images in each panel on the left and right are a surface view and a central slice of the molecules, respectively. Figure 10. A data set collected using a grid plunged on the Vitrobot with a high concentration of glycerol does not process to a high resolution. (a) All 200 2D classes generated from the extra glycerol data set. 2D classes were generated using 25 iterations in Relion 3.0 (Scheres, 2012). (b) 3D classes generated from the extra glycerol data set. Six classes and 100 rounds of classification were specified in Relion for classification. Particle inputs to the 3D classification were those that contributed to 2D classes consisting of intact rings or ambiguous side views in each of two rounds of 2D classification. All classes are visualized using Chimera at an RMSD value of 6 (Pettersen et al., 2004). (c) Postprocessed map generated from the data from the extra glycerol Vitrobot-plunged grid that processed to 8.3 Å resolution. Map is visualized at an RMSD value of 6.

Supplementary
Supplementary Figure 11. Density for α/β interface and β tail residues in the N. gonorrhoeae ribonucleotide reductase structure. (a) The density for some residues in the interface region can be seen. Residues H25 and S41, which have been implicated in organism fitness and drug resistance, are labeled. Interface residues are shown as sticks. (b) There is little density present for the N. gonorrhoeae β tail residues (short orange loop) despite the tail being structurally visible in E. coli crystal structures (long orange helix and sheet). The density in both (a) and (b) is contoured to 6σ. All images generated using Chimera (Pettersen et al., 2004).  Table 3. Glow discharge times used for each chameleon grid analyzed in this study. All glow discharging was completed at -12 mA. 619  60  390  60  150 140 then an additional 50 54 (data set 1) 350 54 (data set 2) 250

Dispense-to-plunge time (ms) Glow discharge time (s)
Supplementary  Table 9. Ice thickness calculations for various holes in the Vitrobot-plunged and two fast-plunged chameleon grids as measured by tomography. Ice thickness was measured manually using IMOD by visually identifying the upper and lower faces of the ice and/or particles or contamination on either surface (Kremer et al., 1996).