Plunge-freezing versus High-Pressure Freezing (HPF) for the moss Physcomitrium patens.

The top and bottom row show schemes of the samples and corresponding examples of TEM overview images, respectively. a, Plunge-freezing. Patterns introduced by crystalline ice are indicated by asterisks (*). b,c, High-pressure freezing. Samples prepared by the “Waffle” Method or in-carrier HPF in (b) and (c), respectively. TEM overviews of lamellae from HPF samples show no indication of crystalline ice. In (a)-(c), ice crystals contaminating the lamella surface which were introduced during transfers between microscopes are marked by blue transparent pentagons.

Sample screening and targeted lamella preparation

a,b, Sample screening using a cryo-fluorescence light microscope (FLM). An overview of a “waffle” grid is shown in (a). The white rectangle indicates the position of the grid squares in (b) containing a target area. Displayed in (a) and (b) are the transmitted light channel (gray) and the cellular autofluorescence (green). c, Lateral targeting. Overlay of a FIB image (gray) and cellular autofluorescence (green) recorded with the integrated Fluorescence Light Microscope (int-FLM). Both images (Extended Data Fig. 3g) were acquired with the stage in trench milling orientation resulting in sample top views. d,e, axial targeting. d, SEM image of the block-face exposed through trench milling. e, Overlay of FIB (gray) and int-FLM FIB-view image (green) of the cellular autofluorescence. Both images were recorded with the stage in lamella milling orientation. Cell junctions and locations of cells are marked with white arrowheads and dashed lines, respectively, in (b-e).

Serial Lift-Out with double sided attachment from below and section trimming.

a-c, Double-sided attached lamellae with the attachment from below. FIB and SEM images are shown in (a) and (b), respectively. The attachment sites are marked in blue. The white rectangle in (b) marks the field of view in (c). c, TEM overview image of the lamella. d,e, Single slices of denoised tomograms. Chl indicates Chloroplasts, Mito Mitochondria, ER the cortical Endoplasmatic Reticulum and CW the cell wall. f, Four int-FLM top views of successive Serial Lift-Out sections (reflected light in gray and cellular autofluorescence in green). The junction between cells (white arrowhead), is only contained in three of them. g, Overlay of FIB (gray) and int-FLM top view images (green) of the top right section in (f) before and after section trimming in the left and right panel, respectively.

Subtomogram averaging of rubisco.

a, Slice of a denoised tomogram recorded on a chloroplast (Supplementary Movie 9). b, Three-dimensional rendered representation with segmented chloroplast membranes (gray) and the rubisco subtomogram average (green) pasted at the positions determined after template matching and subtomogram classification. c, Rubisco subtomogram average in isosurface representation (top row) and an octameric model of large rubisco subunits predicted by AlphaFold 3 docked into the average (bottom row). d, Close-up view of the average shown in (c) revealing that predicted α-helices fit well into the experimental density map. e, The local density of rubisco particles as a function of the distance from the reference particle. Green line and area are the average of experimental density ± one standard deviation, respectively, obtained from 10 tomograms. Black line and gray area are the mean ± one standard deviation obtained for a simulated random particle distribution.

The cryo-ET workflow applied to other tissues and species.

a,b, Physcomitrium patens phyllid sample. a, TEM overview image of a single-sided attached lamella. b, a single slice of a tomogram recorded on the nuclear envelope (Supplementary Movie 10). c,d, Arabidopsis thaliana embryo sample. c, Low magnification TEM image of a lamella. d, Single tomographic slice of the cell wall between cells (Supplementary Movie 11). e,f Salt gland of Limonium bicolor. e, TEM overview image of a double-sided attached lamella. f, Single slice of a tomogram. Chl indicates Chloroplasts, Mito Mitochondria, OB oil bodies, Nuc the nucleus, CW the cell wall and ER the Endoplasmatic reticulum. All tomographic slices are from computationally denoised tomograms (Supplementary Movie 12).

Plunge-freezing of P. patens with cryoprotectants

List of the tested plunge-freezing methods with the employed concentrations of cryoprotectants.

Plunge-frozen P. Patens tissues and the use of cryoprotectants

a-c, Plunge frozen protonemata. a, SEM overview image of a grid. b,c, FIB image before and after lamella milling, respectively. d,e, SEM and FIB images of plunge frozen phyllids, respectively. f, Protonemata grown on medium and plunged in medium containing 100 mM proline as cryoprotectant. The cells are dried out.

The different stage orientations used during lamella preparation

Top and bottom row show the trench milling and lamella milling orientation, respectively, for the geometry in FIB/SEM instruments produced by Thermo Fisher Scientific using a shuttle with 45° pre-tilt. a, The shuttle with respect to stage tilt, stage rotation and the relative location of the FIB and SEM beams. b, Use of the integrated fluorescence light microscope (int-FLM). In trench milling orientation (top panel), the int-FLM records sample top views for lateral targeting and in lamella milling orientation (bottom panel) it can be used in FIB-view mode for axial targeting. c, FIB milling. In trench milling orientation (top panel), the ion beam is normal to the sample surface and used to mill trenches above and below the target area. In Lamella milling orientation (bottom panel), the FIB is used to thin the lamella to a thickness of about 100-300 nm from a shallow angle. d, Lift-out. In trench milling orientation (top panel), the lift-out results in sectioning planes roughly normal to the sample surface and cross-sections of protonemata can be obtained. In lamella milling orientation (bottom panel), sectioning planes are roughly parallel to the sample surface resulting in longitudinal sections of protonemata.

Targeted trench milling aided by the int-FLM

a, SEM and FIB images of a “waffle” grid. b-d, FLM images of a “waffle” grid (transmitted light in gray and cellular autofluorescence in green). Grid overview in (b). The squares mark the field of views in (c) and (d) which are grid squares containing single and multiple cell layers, respectively. The cells in (d) were likely damaged during HPF. e, Scheme summarizing all trench milling steps. All trench milling steps are performed in trench milling orientation (Extended Data Fig. 2). The milling direction is always towards the target area (indicated by black arrowheads) using regular cross-sections. f-h FIB (gray) and int-FLM top view images of cellular autofluorescence (green) for successive trench milling steps in the left and right panel, respectively. Int-FLM top views were recorded between trench milling steps to assure the target area was properly centered. First, a marker pattern was milled (yellow arrowhead) and identified in int-FLM images. From there the distance to the target area was measured (white arrowhead) in (f). This information was used to place the patterns for milling step I in (e). The results are shown in (g). After trench milling the target area is centered within the block between trenches (h). In (f)-(g) white rectangles indicate the field of view in the second image.

Fine milling of “waffle” lamellae

a-c, Scheme of the milling patterns used for “waffle” lamella preparation. The block after trench milling and rotating the stage into the lamella milling orientation is shown in (a). First the lamella is thinned to a thickness of 6 µm. Then a broad notch is milled (b) and the lamella further thinned down (c). All thinning steps in (a) and (c) were milled with regular cross-section with milling direction towards the target area indicated by black arrowheads. Fine milling was then performed with rectangle patterns milled in parallel. d, FIB and SEM image of a “waffle” lamella.

Lift-out in different stage orientations

The left panel shows data for the lift-out with the stage in lamella milling orientation, the right panel for the lift-out in trench milling orientation. a,b, FIB images with the needle attached to the lift-out block. c,d, Int-FLM top views. The left image and the right image were recorded after trench milling and after sectioning, respectively (reflected light in gray and cellular autofluorescence in green). e,f, TEM overviews of the final lamellae obtained from the sections in (c) and (d). The outlines of protonemata cross-sections are marked by dashed lines and cell junctions by white arrowheads.

Serial Lift-Out with single-sided attachment

a, Alignment of the lift-out block to the marker line (blue line and arrowhead) in FIB and SEM. b, Attachment of the lower part of the lift-out block on one side. The patterns for redeposition milling are indicated in blue. c, Int-FLM top views of single-sided attached Serial Lift-Out sections acquired in trench milling orientation (reflected light in gray, cellular autofluorescence in green). The junction between cells is only contained the three right-most sections, marked by white arrowheads. d, Section trimming in trench milling orientation. A section before and after trimming is shown in the top and bottom panel, respectively. e, Single-sided attached lamella before and after fine milling in the top and bottom panel, respectively. The milling patterns are indicated in blue. A 1.5×1.5 µm block is left opposite of the attachment site. f, Single-sided attached lamellae often bend during or after fine milling.

Serial Lift-Out with double-sided attachment from the sides

a, Milling of marker lines for the alignment of the lift-out block. Grid bars before and after marker line milling in the left and right panel, respectively. Marker lines are indicated by a blue arrowhead. b, Alignment of the lift-out block to the marker lines. SEM and FIB images are shown in the left and right panel, respectively. The bottom front edge of the lift-out block is indicated by a blue line and the marker line by a blue arrowhead. c, Attachment of the lift-out block on both sides. Patterns for redeposition milling are indicated in blue. d, FIB image of a lift-out section. e, FIB and SEM image of the section in the left and right panel, respectively, after the top layer was removed. The SEM image reveals the location of the cell junction as target area (white arrowheads) which was, subsequently, used to minimize the width of the milling patterns for lamella fine milling.

Serial Lift-Out with double-sided attachment from below

a, Preparation of 8×90 µm cavities in trench milling orientation. b, Preparation of the plateaus in lamella milling orientation. Left panel: FIB image of cavities before milling. Middle panel: after milling regular cross-section patterns (xyz: 9 x 8 x 20 µm) generating the rough shape of the plateaus. Due to stage drift, they can have different heights. Right panel: After rectangle patterns were milled in parallel (xyz: 9 x 1 x 20 µm) to even out height differences. This allows the attachment of the bottom of the lift-out block to both plateaus from below. Milling patterns are indicated by blue rectangles. c, Alignment of the lift-out block with respect to the plateaus in the SEM. The center of the block should be roughly on top of the front edge of the plateaus (sketch in right panel). Precision of the placement, however, is not imperative. d, the lift-out block placed on the plateaus, after attachment from below and after sectioning in the left, middle and right panel, respectively. Patterns for redeposition milling are depicted in blue.

Preparation of in-carrier frozen HPF samples for lift-out

a, Screening of in-carrier frozen HPF sample using the int-FLM. FIB and a montage of int-FLM top view images in left and right panel, respectively. The field-of-view in (b) is marked by a white rectangle. b, Int-FLM images before and after trench milling in the left and right panel, respectively. In (a) and (b), marker patterns are labelled with yellow arrowheads and yellow rectangles. c, Milling of the undercut, aided by the int-FLM for in-carrier frozen samples. Overlay of FIB images (gray) and int-FLM FIB-view image of cellular autofluorescence (green) before and after milling the undercut in the left and right panel, respectively. Both images were recorded in lamella milling orientation. The undercut is marked by a blue rectangle. In (a)-(c) cell junctions are labelled by white arrowheads.

HPF workflows applied to other tissues and species

a-c, Physcomitrium patens phyllids. a, FLM overview of a phyllid “waffle” grid. b, Grid square containing seemingly intact tissue. c, Grid square containing damaged tissue. d,e Arabidospsi thaliana embryos. d, Image of embryos placed in the squares of a 50 mesh grid. e, Schematic of a “waffle” grid with and without a spacer ring. Embryos fit into grid squares but are thicker than the grid bars and would therefore be damaged during freezing (top panel). The addition of a spacer ring, as described in18, prevents squeezing damage and can allow high-pressure freezing of thicker samples. f-h, Salt gland of Limonium bicolor. f, Dissection of the first true leaf (FL). The top panel shows a scheme of the FL location in 2 weeks old plants. The bottom panel are photographs of a whole two weeks old plant before and the dissected FL in the left and right panel, respectively. g, int-FLM top view image of the salt gland of in-carrier frozen FL. h, int-FLM image top view image of a Serial Lift-Out section containing a fraction of the salt gland. In (g) and (h) autofluorescence of the salt gland is shown in blue, the cellular autofluorescence in green and the reflected light in gray.