Long shelf-life streptavidin support-films suitable for electron microscopy of biological macromolecules.

Lipids The biotinylated lipid used here is 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-(cap biotinyl), supplied as a 10 mg/mL solution in chloroform/methanol/water (Avanti Polar Lipids). This was diluted to 1.0 mg/mL with a solution of chloroform/methanol/water and aliquoted into small volumes intended for a single usage. The aliquots were sealed under nirogen gas and stored at −80 °C. No deterioration as a function of time was observed in the ability of such aliquots to produce high-quality streptavidin monolayer crystals. Nevertheless, as a precaution, we prepare new aliquots after a period of about 6 months. Streptavidin Streptavidin (SA) was purchased from New England Biolabs (catalog number N7021S). This sample is provided at a concentration of ~1 mg/mL, dissolved in 10 mM sodium phosphate pH 7.2 with 0.15 M NaCl. This was aliquoted in quantities intended for single use, frozen in liquid nitrogen, and stored at −80 °C. Similar to what we do for the lipid, as a precaution, we prepare new aliquots of streptavidin after a period of about 6 months. Protocol for growing monolayer crystals directly on holey-carbon EM grids A lipid monolayer, cast on an air-water interface, is first picked up by touching the lipid from above with a hydrophobic, holey-carbon EM grid. This results in Langmuir-Schaefer transfer of patches of the monolayer that span the holes of the carbon film, as was discovered by ( Kubalek et al., 1991 ). We presume that an additional, unwanted lipid monolayer is also transferred to the air-water surface of the small volume of buffer that adheres to the (now hydrophilic) face of the EM grid. Thus, to remove as much lipid as possible from the surface of the adhering drop, we touch the grid to three successive, 50 μL drops of subphase buffer sitting on parafilm. Next, SA is added to the small droplet of buffer that adheres to the face of the EM grid. The grid is then incubated long enough to allow binding and subsequent crystallization of SA. We refer to this technique as the “on-grid” crystallization method. Figure S1 shows two photographs that further illustrate the steps just described. We cast the above-mentioned lipid monolayer on a small trough holding ~5 mL of subphase buffer consisting of 50 mM HEPES (pH 7.5), 0.15 M KCl, and 10 % trehalose (Swanson Health Products). Including trehalose in the subphase buffer serves only to eliminate the need for a buffer-exchange “wash” before air-drying (see the following section). This simplification became possible after we observed that adding 10% trehalose has no effect on the quality of monolayer crystals of streptavidin. Before applying lipid, however, we first “dust” the air-water interface lightly with unscented talcum powder, and then we apply a small droplet (~10 μL) of castor oil at the center. As the castor oil spreads, the advancing front of oil sweeps away contaminants at the air-water interface and compresses them to the perimeter of the trough. A Hamilton syringe is then used to deliver ~0.5 μL of previously aliquoted lipid to the center of the trough. At this point the thin film of castor oil also serves as a “piston” to control the surface pressure of the lipid monolayer ( Langmuir, 1917 ). We believe that the castor-oil piston is a desirable but not necessary element of our protocol. Using a trough, rather than smaller, individual wells, facilitates use of the castor-oil piston, and allows one to prepare multiple grids from a single lipid monolayer. We wash Quantifoil grids by first dipping them into chloroform and then into 95% ethanol. We also apply an additional ~5 nm of evaporated carbon to the top side of Quantifoil grids and then allow these to “age” for at least three days to make the freshly evaporated carbon more hydrophobic. Just before use, we again wash these grids by dipping them into 95% ethanol. We have less experience with C-flat grids, but we have successfully used them as received, i.e. without depositing additional carbon. As is mentioned in the Supplemental Material , in the section “ Issues still to be addressed ”, we prefer to use gold or molybdenum grids to copper grids. In our current crystallization protocol, we dilute an aliquot of SA to a concentration of 0.2 mg/mL with subphase buffer, and we use 4 μL of diluted SA for each grid. After applying SA to the EM grid, the crystals are grown within a humidity chamber. Care is taken to minimize evaporation by placing crushed ice on the tweezers, with the intent to cool the grid slightly below the ambient dew point. Whereas an incubation time of 1–2 minutes appears to be too short to ensure full crystal formation, we observed crystallization to be completed within 10 minutes. Following an incubation time that, for convenience, is often about half an hour when doing many grids at a time, most of the unbound SA is washed away by placing the grid on top of a 200 μL drop of wash buffer. The composition of the wash buffer is 10% trehalose in 10 mM HEPES (pH 7.5), as before, but with the KCl concentration now reduced to 50 mM. After waiting several seconds, the floating grid is caught with a tweezers and lifted vertically until it separates from the wash drop. As noted further in the Supplemental Material , we believe that this step in the protocol may be the one in which the monolayer crystals of SA are at greatest risk of becoming fragmented or even lost completely. Even when crystals have been severely damaged (at this stage, as we believe), most, if not all, holes in the holey carbon film still remain covered by a monolayer of SA bound to biotinylated lipid. Streptavidin crystals are then embedded in trehalose and backed with evaporated carbon After washing the grids to remove unbound SA, excess trehalose solution is gently “wicked off” by touching the edge of the grid to a piece of filter paper. The grid then is left on a filter paper with the wet side facing the air, and any remaining liquid on the grid is allowed to dry. After the trehalose solution has dried, a thin layer (~5 nm or less) of evaporated carbon is deposited on the back side (lipid-tail side) of the EM grid. Grids are placed ~20 cm below a carbon-arc source, and evaporation is done at a vacuum of 10 −5 torr or lower. We use carbon rods with a tip width of 1 mm (Ted Pella catalog number 62–107), which require less power to evaporate, in order to minimize heating of the trehalose-embedded SA “target”. To further minimize the risk of radiant heating, the carbon rod is heated very rapidly, resulting in breakage of the thin tip and a concomitant flash of evaporated carbon. A cartoon showing the structure of the resulting carbon-backed, trehalose-embedded streptavidin support films is presented in Supplemental Figure S2 . We store the carbon-backed, trehalose embedded SA crystals at room temperature in a sealed container. We prefer to store grids over silica gel that is pink (but not white) to maintain a relatively constant value of humidity. Just before use, the grid is rehydrated by touching to two successive, 50 μL drops of a solution of 10 mM HEPES (pH 7.5) with 150 mM KCl without trehalose and then left on a 100 μL drop of the same solution for 10 minutes. This same process is repeated a second time, with the intent being to rinse away all remaining trehalose. After that, the grid is further washed with whatever buffer is optimal for the macromolecular sample under investigation.

Preparation of grids and electron microscopy of 70S ribosomal particles

Ribosomes were purified from E. coli strain MRE600 using sucrose gradient centrifugation, as previously described ( Blaha et al., 2000 ). Ribosome complexes were formed by incubating 1.5 μM deacylated tRNA Phe and 3 μM mRNA of sequence 5′-GGCAAGGAGGUAAAAUUCUACAAA-3′ (Thermo Scientific) with 0.5 μM ribosomes at 37 °C for 15 minutes in the buffer A: 20 mM HEPES, pH 7.5/ 70 mM KCl/ 6 mM MgOAc/ 1 mM TCEP. The antibiotic spectinomycin (Sigma-Aldrich) at a concentration of 20 μM was added to the pre-formed complex and incubated an additional 10 minutes at 37 °C. Ribosomes were biotinylated by adding 5-fold excess of biotin-labeling reagent (Solulink Catalog No. B-1007-110) and incubated 20 minutes at room temperature. Excess unreacted biotin was removed by using 1 mL S300 Sephacryl (GE Healthcare) gel-filtration spin column. The column was equilibrated with Buffer A, and usually 50 μL of ribosome complex was loaded on the column and spun on a table centrifuge at 3200 rpm for 1 minute. Grids were first washed 3 times with 50 μL drops of cold ribosome buffer, and then a 4 μL aliquot of ribosomes was applied at a concentration in the range of 20–40 nM. After incubating for 20 minutes, chilled and in a humidity box to minimize evaporation, as described above for growth of monolayer crystals, unbound ribosomes were washed away by touching to three successive, 50 μL drops of buffer. To prepare negatively stained specimens, we added 4 μL of 2% uranyl acetate to the lens of buffer adhering to the rinsed grid, which then was mixed by repeated, gentle pipetting while on the grid. This was followed by two cycles in which 3 μL was removed from the grid and 4 μL of 2% uranyl acetate was again added. After the second cycle, the excess uranyl acetate solution that remained was removed by blotting with filter paper. Images were recorded with a Gatan CCD camera on a JEOL 1200 electron microscope. To prepare cryo-EM samples, the grid is transferred to the Vitrobot tweezers after washing away unbound ribosomes with ribosome buffer. The tweezers and grid were then loaded into the Vitrobot Mark IV chamber, which was previously equilibrated at a temperature of 15 °C and a relative humidity setting of 100%. In order to standardize the volume of liquid on the grid before blotting, excess liquid was first wicked off by touching the bottom edge of the grid with filter paper, which was brought in through a side port of the Vitrobot chamber. Following this, 1.2 μL of sample buffer was then added to the wet face of the grid. The blotting time used was 5 s, with a force setting of 8, a blotting time of 3 seconds, and zero wait/drain time. For routine evaluation of cryo-grids, images were recorded with a Gatan CCD camera on a Philips CM 200. High-resolution images were obtained with a Gatan K2 camera on an FEI low-base Titan, using a Gatan cryo-holder. In this latter case, images were recorded as dose-fractionated movies consisting of twenty 300 ms frames, each with an exposure of 1.2 electrons/Å 2 at the specimen. The pixel size in these images was 1.3 Å, referred to the specimen. The movie frames were aligned and summed with the motion-correction software developed by ( Li et al., 2013 ). Data processing The “2dx” software package ( Gipson et al., 2007 ) was used to unbend the SA lattice in a few of the images. This was done to sharpen the Bragg spots in the Fourier transforms, thereby increasing the resolution at which spots could be detected with a good signal-to-noise ratio. In all other cases, Fourier filtration of Bragg peaks was used to remove the image of the SA crystal, without unbending, before particles were boxed for further analysis. To do this, a script, available upon request, was written in MATLAB to identify all pixels where the magnitude of the Fourier transform of an image was higher than a user-defined value. The Fourier-transform magnitudes in such pixels, plus those in several adjacent pixels, were replaced by the average value of the surrounding background, and the phases were replaced by random values. The threshold value was adjusted manually, while looking at a display of the Fourier transform of an image, until all visible diffraction spots were removed. Candidate ribosome particles were automatically boxed with a software tool provided in RELION ( Scheres, 2012 ). Images, with candidate particles outlined, were edited manually, using the BOXER tool provided in EMAN ( Tang et al., 2007 ) to remove initial candidates that were obviously aggregates or other undesired material. Three-dimensional classification of particles and subsequent refinement again used tools provided in RELION.

Protocol for growing monolayer crystals directly on holey-carbon EM grids A lipid monolayer, cast on an air-water interface, is first picked up by touching the lipid from above with a hydrophobic, holey-carbon EM grid. This results in Langmuir-Schaefer transfer of patches of the monolayer that span the holes of the carbon film, as was discovered by ( Kubalek et al., 1991 ). We presume that an additional, unwanted lipid monolayer is also transferred to the air-water surface of the small volume of buffer that adheres to the (now hydrophilic) face of the EM grid. Thus, to remove as much lipid as possible from the surface of the adhering drop, we touch the grid to three successive, 50 μL drops of subphase buffer sitting on parafilm. Next, SA is added to the small droplet of buffer that adheres to the face of the EM grid. The grid is then incubated long enough to allow binding and subsequent crystallization of SA. We refer to this technique as the “on-grid” crystallization method. Figure S1 shows two photographs that further illustrate the steps just described. We cast the above-mentioned lipid monolayer on a small trough holding ~5 mL of subphase buffer consisting of 50 mM HEPES (pH 7.5), 0.15 M KCl, and 10 % trehalose (Swanson Health Products). Including trehalose in the subphase buffer serves only to eliminate the need for a buffer-exchange “wash” before air-drying (see the following section). This simplification became possible after we observed that adding 10% trehalose has no effect on the quality of monolayer crystals of streptavidin. Before applying lipid, however, we first “dust” the air-water interface lightly with unscented talcum powder, and then we apply a small droplet (~10 μL) of castor oil at the center. As the castor oil spreads, the advancing front of oil sweeps away contaminants at the air-water interface and compresses them to the perimeter of the trough. A Hamilton syringe is then used to deliver ~0.5 μL of previously aliquoted lipid to the center of the trough. At this point the thin film of castor oil also serves as a “piston” to control the surface pressure of the lipid monolayer ( Langmuir, 1917 ). We believe that the castor-oil piston is a desirable but not necessary element of our protocol. Using a trough, rather than smaller, individual wells, facilitates use of the castor-oil piston, and allows one to prepare multiple grids from a single lipid monolayer. We wash Quantifoil grids by first dipping them into chloroform and then into 95% ethanol. We also apply an additional ~5 nm of evaporated carbon to the top side of Quantifoil grids and then allow these to “age” for at least three days to make the freshly evaporated carbon more hydrophobic. Just before use, we again wash these grids by dipping them into 95% ethanol. We have less experience with C-flat grids, but we have successfully used them as received, i.e. without depositing additional carbon. As is mentioned in the Supplemental Material , in the section “ Issues still to be addressed ”, we prefer to use gold or molybdenum grids to copper grids. In our current crystallization protocol, we dilute an aliquot of SA to a concentration of 0.2 mg/mL with subphase buffer, and we use 4 μL of diluted SA for each grid. After applying SA to the EM grid, the crystals are grown within a humidity chamber. Care is taken to minimize evaporation by placing crushed ice on the tweezers, with the intent to cool the grid slightly below the ambient dew point. Whereas an incubation time of 1–2 minutes appears to be too short to ensure full crystal formation, we observed crystallization to be completed within 10 minutes. Following an incubation time that, for convenience, is often about half an hour when doing many grids at a time, most of the unbound SA is washed away by placing the grid on top of a 200 μL drop of wash buffer. The composition of the wash buffer is 10% trehalose in 10 mM HEPES (pH 7.5), as before, but with the KCl concentration now reduced to 50 mM. After waiting several seconds, the floating grid is caught with a tweezers and lifted vertically until it separates from the wash drop. As noted further in the Supplemental Material , we believe that this step in the protocol may be the one in which the monolayer crystals of SA are at greatest risk of becoming fragmented or even lost completely. Even when crystals have been severely damaged (at this stage, as we believe), most, if not all, holes in the holey carbon film still remain covered by a monolayer of SA bound to biotinylated lipid.

Supplementary Material 1 2