Baited reconstruction for visualization of Beta galactosidase ligand binding pocket at high resolution.

(A) Reconstruction of Bgal from 2DTM coordinates using images from a previously published dataset (Saur et al., 2020) using a Bgal crystal structure (PDB: 1DP0) (Juers et al., 2000) as a template, with a 10 Å sphere around the PETG ligand omitted. (B) A 2D slice through the reconstruction in (A) including the region deleted from the density shows no obvious discontinuity in the density. (C) A view of the density in (A) indicated with a red box, with regions within 1.8 Å of the template model highlighted in red. Gray indicates density of Bgal outside of the template, purple indicates density consistent with the position of PETG and blue indicates additional density that likely represent water molecules. (D) A stick diagram showing the locations of the atoms in the template used for template matching. (E) Published density from (Saur et al., 2020) aligned and scaled as in (A). (F) As in (B), showing a region of the published density in (E). (G) As in (C), showing the same region of the published density in (E). (H) As in (D), showing all atoms annotated in the crystal structure, including those omitted before generating the 2DTM template.

Visualizing drugs and small molecules bound to the ribosome in vivo.

(A) A reconstruction of the ribosome from 2DTM coordinates identified in the cytoplasm of FIB-milled Sacchromyces cerevisiae cell sections showing clear density for both the 60S (part of the template) and the 40S (outside of the template). (B) A slice of the reconstruction in (A), indicating the local resolution using the indicated color coding. The arrow indicates the P-site tRNA. (C) Regions of the density >3 Å from the template model are indicated in pink. The crystal structure PDB: 4U3U was aligned with the template model and the position of cycloheximide (CHX) was not altered. (D) As in C, showing density corresponding to a spermidine (PDB: 7R81) and unaccounted for density outside of the template (black arrow), which may also represent a polyamide.

Baited reconstruction reveals high-resolution features in vivo without template bias.

(A) Slice of a reconstruction using 2DTM coordinates identified with a template lacking the protein L7A. Color coding indicates the local resolution as indicated in the key. (B) As in (A), pink indicates the 2DTM template used to identify the targets used in the reconstruction. (C) The model PDB: 6Q8Y is shown in the density. Red corresponds to the protein L7A, which was omitted from the template used to identify targets for the reconstruction. Blue corresponds to model features that were present in the template. (D) Single nucleotide omit template and (E) reconstruction showing emergence of density outside of the template, including a phosphate bulge, black arrow. Single amino acid omit templates lacking Phe (F), Arg (H) or Ser (J) and density (G, I, K), respectively showing emergence of features consistent with each amino acid.

Baited reconstruction provides quantitative metric for template bias.

(A) Observed template bias (Ω) calculated using the cisTEM program measure_template_bias as a function of the 2DTM SNR threshold used to select targets from images of yeast lamellae. Blue arrows indicate the reconstructions shown in Figure 4C. (B) Plot showing a comparison of the predicted false positive rate and the observed Ω. The plotted straight line indicates the best fit linear function y = 0.96x - 0.05. (C) Images showing the same region of maps resulting from reconstruction using targets identified with the indicated template at the indicated 2DTM SNR threshold. Red indicates the location of the omitted residue in the omit template.

Preparation and simulation of the 3D templates used in this study

(A) Example micrograph from (Saur et al., 2020). Scale bar indicates 200 Å. (B) Maximum intensity projection showing the maximum 2DTM SNR at each pixel after a 2DTM search of the image in (A). (C) 2D projections indicating the best matching 2D templates of Bgal in the image in (A). (D) Estimated local resolution of the reconstruction in Figure 1A, as indicated by the color key. (E) Region of the reconstruction in (D) shown in Figure 1B. (F) FSC evaluated between the half-maps as a function of spatial frequency.

(A) Area of an example micrograph showing the yeast cytoplasm from (Lucas et al., 2023). Scale bar indicates 200 Å. (B) Maximum intensity projection showing the highest 2DTM SNR at each pixel after a 2DTM search of the image in (A) with an LSU template. (C) 2D projections indicating the best matching 2D templates in the image in (A). (D) Plot showing the FSC of the two half maps of the reconstruction shown in Figure 2. (E) Density showing the region of the map in Figure 2D color coded by the local density.

Cycloheximide stalls the ribosome in a non-rotated PRE- translocation state in vivo.

(A) Local resolution filtered map shown in Figure 2, only showing features outside of the template shows density consistent with A/A and P/P tRNAs. (B) The same region of the map rotated to show weak density in the polypeptide exit tunnel consistent with the nascent polypeptide attached to the A-site tRNA, consistent with the non-rotated PRE- translocation state.

Plot showing the average relative number of LSUs detected in 7 images of ∼150 nm thick lamellae using templates generated from subsections of the LSU of the indicated molecular mass. The x-intercept indicates the smallest particle that was detectable on average, which in this case is 300 kDa.