Characterization of Fully Recombinant Human 20S and 20S-PA200 Proteasome Complexes.

Key Resources Table REAGENT or RESOURCE SOURCE IDENTIFIER

Bacterial and Virus Strains

DH10EmbacY cells Geneva Biotech N/A Pir1 competent cells ThermoFisher scientific Cat# C101010 Top10 competent cells ThermoFisher scientific Cat# C404003 Chemicals, Peptides, and Recombinant Proteins FuGENE® HD Transfection Reagent Promega Cat# E2311 Human 20S Proteasome Enzo Cat# BML-PW8720-0050 Superose 6 increase 10/300 GL GE Healthcare Cat# 29-0915-96 Streptactin Superflow plus 5ml cartridge QIAGEN Cat# 30060 Suc-LLVY-AMC Boston Biochem Cat# S-280 Boc-LRR-AMC Boston Biochem Cat# S-300 Z-LLE-AMC Boston Biochem Cat# S-230 Tris-HCl Sigma-Aldrich N/A Glycerol VWR Cat# 56-81-5 NaCl Sigma-Aldrich TEV protease N/A N/A DTT Thermo Scientific Cat# R0862 EDTA Sigma-Aldrich Cre recombinase NEB Cat# M0298S Qiaprep Spin Miniprep kit reagents QIAGEN Cat# 27106 InsectXPRESS media Lonza Cat# BE12-730Q FBS GIBCO Cat# 10270106 30 kDa cut-off Vivaspin 20 concentrators Superdex 200 Increase 10/300 GL GE Healthcare Cat# 28-9909-44 HEPES Sigma-Aldrich Cat# H3375 DMSO Sigma-Aldrich Cat# BP231-100 Deposited Data r20S EM map This study EMD-4877 r20S atomic coordinates This study PDB: 6RGQ r20S-PA200 EM map This study EMD-4860 r20S-PA200 atomic coordinates This study PDB: 6REY Experimental Models: Cell Lines Sf9 insect cells N/A N/A Recombinant DNA cDNAs for all subunits (Gene synthesis) Epoch N/A pACEBAc1 Geneva Biotech N/A pIDC Geneva Biotech N/A pIDS Geneva Biotech N/A Software and Algorithms EPU Automated Data Acquisition Software for Single Particle Analysis v1.9.1 ThermoFisher https://www.fei.com/software/epu/ Prism 8 software GraphPad https://www.graphpad.com/scientific-software/prism/ ResMap1.1.4 Kucukelbir et al., 2014 http://resmap.sourceforge.net cisTEM Grant et al., 2018 RRID: SCR_016502 ; https://cistem.org Coot v0.8.9 Emsley et al., 2010 RRID: SCR_014222 ; https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/ PHENIX v1.13 Mustyakimov et al., 2012 RRID: SCR_014224 ; http://www.phenix-online.org UCSF Chimera Pettersen et al., 2004 RRID: SCR_004097 ; https://www.cgl.ucsf.edu/chimera/ Phyre2 Kelley et al., 2015 http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index MolProbity Chen et al., 2010 RRID: SCR_014226 ; http://molprobity.biochem.duke.edu/index.php?MolProbSID=s5mqsj74op6r8o3u1j5jkeasr2&eventID=2 Pymol Molecular Graphics System Schrödinger, LLC RRID: SCR_000305 ; https://pymol.org/2/ LigPlot+ Laskowski and Swindells, 2011 https://www.ebi.ac.uk/thornton-srv/software/LigPlus/ I-Tasser Yang et al., 2015 RRID: SCR_014627 ; https://zhanglab.ccmb.med.umich.edu/I-TASSER/ Other QUANTIFOIL R1.2/1.3 grids 300 mesh Quantifoil Micro Tools R1.2/1.3 Lead Contact and Materials Availability Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Paula da Fonseca ( pauladf@mrc-lmb.cam.ac.uk ).

Experimental Model and Subject Details Cell Lines

Sf9 insect cells were cultivated in InsectXPRESS media (Lonza), at 27°C.

Bacterial Strains

Constructs were transformed and amplified in either Escherichia coli TOP10 or Pir1 chemically competent cells (Invitrogen). The DH10EMBacY Escherichia coli strain (Geneva Biotech) containing a baculovirus shuttle vector (bacmid) and a constitutively expressing YFP expression cassette was used to create all expression bacmids containing the cloned genes of interest. Method Details Cloning and generation of baculovirus cDNAs encoding each of the human 20S proteasome subunits, proteasome assembly chaperones and PA200 were synthesized as codon-optimized genes for expression in E. coli (Epoch). These genes were subsequently cloned in Multibac Turbo vectors pACEBac1, pIDS and/or pIDC using a strategy adapted from the IVA method ( García-Nafría et al., 2016 ). All proteasome α subunits were cloned into a pACEBac1 vector, the β subunits into a pIDS, the chaperones in both pIDC and pACEBac1, and PA200 in pACEBac1. The proteasome β subunits were cloned with their N-terminal pre-peptides, which are cleaved at the final stages of proteasome assembly and maturation resulting in proteolytic active complexes ( Budenholzer et al., 2017 ). We added a TwinStrep-tag at the C terminus of β7 to assist on the purification of mature homogeneous proteasome complexes, as this is known to be the last subunit to be incorporated during the proteasome assembly pathway ( Li et al., 2007 , Marques et al., 2007 ). We also created untagged and N-terminal Twin-Strep tagged versions of PA200, to be co-expressed with the 20S proteasome or expressed on its own, respectively. The pACEBac1 vector, containing the α subunits, and the pIDC vector, containing the β subunits, were Cre-Loxed ( Fitzgerald et al., 2006 ) to create a single vector (20S proteasome plasmid). Attempts were made to Cre-Lox the chaperones containing pIDC vector to either the protesome α or β subunit vectors, but due to the size of the plasmids, the high number of genes and gene repetitions many random recombinant events occur, and for that reason we decided to clone the proteasome assembly chaperones in a separate pACEBac1 vector ( Figure S1 A). All assembled plasmids were transposed to DH10EmBacY cells (Geneva Biotech), and Bacmid DNA purified following published protocols ( O’Reilly et al., 1994 ). Bacmids were transfected (P1 virus), after which Sf9 cells, at 1.5x10 6 cells/ml in InsectXPRESS media (Lonza), were infected and incubated for approximately 72 hours until viability was around 80%. The P2 virus was then filtered and 2% Hi-FBS serum added. This was kept as a stock at 4°C, protected from light.

Expression of proteasome complexes

Each individual baculovirus was initially amplified (P3 amplification) by infecting Sf9 cells at 1.5 x10 6 cells/ml with 1:100 of P2 virus, followed by a 72 hour incubation at 27°C with shaking, or until cell viability was about 80%. The P3 virus were then filtered and used to co-infect Sf9 cells with a density of roughly 2x10 6 cells/ml (12 mL of each P3 virus for each 500 mL of cells). Expression of the human 20S proteasome was accomplished by co-infection of Sf9 cells with two baculoviruses, one containing the 20S proteasome the other the assembly chaperones. The 20S-PA200 complex was expressed by triple co-infection with baculoviruses containing (1) the 20S proteasome, (2) the chaperones and (3) the PA200. Interestingly, we observed that human 20S proteasomes can assemble when expressing only its 14 subunits, but with more than a 3-fold decrease in yield of mature complexes compared with that obtained with chaperone co-expression (data not shown). This suggests that endogenous insect proteasome assembly chaperones can, at least partially, assist in the assembly of the human complex. After infection, the Sf9 cells were incubated at 27°C for 48 hours and harvested by centrifugation at 3,000 g , 20 min, 4°C. Cell pellets were washed with cold PBS, frozen in LN2 and stored at −80°C. Purification proteasome complexes The purification strategy was the same for both reconstituted 20S proteasome and 20S-PA200 complexes. Briefly, thawed cells were resuspended in 3-4 times volume of buffer W (50 mM Tris pH 7.5, 150 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, 1 mM EDTA), lysed by sonication and the lysate cleared by centrifugation at 48,400 g for 30 minutes. The clear lysate was then passed through a 5 μm filter and loaded onto tandem Streptactin Superflow Plus columns (QIAGEN), equilibrated in buffer W, and eluted in buffer E (50 mM Tris pH7.5, 150 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, 1mM EDTA, 2.5 mM d-desthiobiotin). Proteasome containing fractions were pooled together and the TwinStrep-tag cleaved by overnight dialysis against buffer W containing TEV protease, at a ratio of 1:50 (TEV to protein). The sample was filtered again with a 0.4 μm filter and loaded onto Streptactin Superflow Plus columns, where the flow-through was collected. The protein was concentrated using 30 kDa cut-off Vivaspin 20 concentrators (Sartorius) and loaded onto a Superose 6 Increase 10/300 gel filtration column (GE Healthcare) equilibrated with 50 mM Tris pH 7.4, 100 mM NaCl, 1 mM EDTA ( Figure S1 B). Typical yields for the purification of reconstituted 20S proteasomes and 20S-PA200 complexes are 2-3 mg and 1 mg of protein purified from 1 L of Sf9 cultures, respectively.

Expression and purification of PA200

Expression of the human

PA200 was accomplished by infection of Sf9 cells with the TwinStrep-Tev-PA200 containing baculovirus. The purification protocol was the same as described for the recombinant 20S proteasome, except that the last size exclusion chromatography step was performed in a Superdex 200 Increase 10/300 (GE Healthcare). Proteasome activity assay 20S proteasome proteolytic activities were measured by fluorescence spectroscopy using the β1, β2 and β5 specific substrates (from Boston Biochem) carboxybenzyl-Leu-Leu-Glu-7-amino-4-methylcoumarin (Z-LLE-AMC), tert-butyloxycarbonyl-Leu-Arg-Arg-7-amino-4-methylcoumarin (Boc-LRR-AMC) and N-succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (Suc-LLVY-AMC), respectively. Recombinant human 20S proteasomes, endogenous human 20S proteasomes (Enzo) and recombinant human 20S-PA200 complexes, all at 7.5 nM in 25 mM HEPES, pH 7.5 and 0.5 mM EDTA, were incubated with 50 μM substrate (stock at 5 mM in DMSO) for 30 minutes at room temperature. Fluorescence intensity, generated by the release of AMC fluorophores, was measured in triplicate with a Pherastar (BMG Labtech) with λ excitation = 350 nm and λ emission = 450 nm. PA200 titration experiments were done by mixing 60 nM of either endogenous or recombinant 20S proteasomes with increasing concentrations of recombinant PA200 (0 nM, 15 nM, 20 nM, 30 nM, 60 nM, 120 nM, 180 nM and 240 nM), followed by incubation for 1 hour at room temperature. The samples were then diluted 20 times and incubated with 50 μM of each of the three proteasome fluorogenic substrates. Measurements were done in Pherastar (BMG Labtech) after incubation at room temperature for 30 minutes. All experiments were done in triplicate.

Cryo-electron microscopy Quantifoil

R1.2/1.3 electron microscope grids, with a 300 gold mesh, were coated with a thin layer of carbon freshly floated from mica, following the procedures we previously optimized for the preparation of grids with endogenous human proteasomes ( da Fonseca and Morris, 2015 , Morris and da Fonseca, 2017 ). We find that the continuous carbon layer favors an even particle distribution, while its electron scattering facilitates the assessment of the information in the recorded images, as judged by the recovery of Thon rings to high resolution in their power spectra, which also contributes to an accurate defocus estimation for the correction of the contrast transfer function associated effects. The grids used for imaging the recombinant human 20S proteasome were glow-discharged in the presence of pentylamine, as we previously described for the endogenous complex ( da Fonseca and Morris, 2015 , Morris and da Fonseca, 2017 ). The grids used for the 20S-PA200 complex were glow-discharged in atmospheric air. 3 μL of sample were loaded on the grids and flash frozen by plunging the grids into liquid ethane using a Vitrobot Mark IV (FEI), operated at 22°C, 95% humidity, 20 s waiting and 5 s blotting times. The grids were transferred into a FEI Titan Krios electron microscope, operated at 300 keV and a nominal magnification of x95,000, resulting in a calibrated sampling of 0.81 Å per pixel at the image level. Images were recorded with EPU software using a Falcon III direct electron detector operating in counting mode, at an electron dose per pixel of ∼0.5 e - /s, with 60 s exposures saved as 75 frame movies with evenly distributed electron dose ( Figures S2 A and S2B).

Single particle analysis

The cryo-EM images were processed using the cis TEM software ( Grant et al., 2018 ). For the analysis of the recombinant human 20S proteasome, the frames of each of the 483 movies recorded were aligned and summed into single images. Upon inspection, 411 images were selected for further analysis, based on the ice quality, image contrast and the recovery of isotropic Thon rings to high resolution. The selected images had a defocus range from −0.8 μm to −3.4 μm. A total of 58,281 particles were picked from the selected images and subsequently classified. The resulting 2D class averages reveal a very homogeneous sample and were used to exclude particle picking false positives. No proteasome top-views, corresponding to projections along the proteasome long axis, were selected since the proteasome side-views show a tomographic distribution of Euler angles, evenly distributed around one of the main axis, providing a complete and even information recovery in Fourier space. The resulting selection of 50,885 particles were used for Auto Refine, with C2 symmetry imposed, using our previous 3.5 Å cryo-EM map of the endogenous human 20S proteasome (EMD-2981) ( da Fonseca and Morris, 2015 ) as starting reference. The overall image processing workflow is shown in Figure S7 A. The resolution of the final map was estimated at 2.6 Å by Fourier shell correlation, as implemented in cisTEM ( Figure S2 C). The analysis of the cryo-EM images of the recombinant human 20S-PA200 followed the same procedures as described for the recombinant 20S proteasome. From the 1,114 movies recorded, 717 were selected for further analysis and had a defocus range from −1.2 μm to −2.9 μm. 29,434 particles were picked from the selected images and classified. The resulting class averages revealed the presence of some uncapped 20S proteasomes, which were excluded from the particle data-set together with particle picking false positives. The 17,356 particles from the selected class-averages were subjected to another round of 2D classification into 10 classes that showed no evidence for significant heterogeneity. Therefore, all 17,356 particles were used for 3D Auto Refine, with C2 symmetry imposed, using our cryo-EM map of the recombinant human 20S proteasome as starting reference, followed by a round of local refinement. An additional 3D classification into 2 classes, using C1 symmetry, indicated that 28% of the particles in the final map are from single capped complexes. The overall image processing workflow is shown in Figure S7 B.The resolution of the final map was estimated at 3.0 Å by Fourier shell correlation, as implemented in cisTEM ( Figure S2 D).

Molecular modeling

The model of the human recombinant 20S proteasome was built based on the X-ray crystal structure of the endogenous human 20S proteasome (PDB: 5LE5 ) ( Schrader et al., 2016 ) using real-space refinement in Coot ( Emsley et al., 2010 ) and Phenix ( Afonine et al., 2012 ). The model of the recombinant human 20S-PA200 complex was built based on our cryo-EM model of the recombinant human 20S proteasome and models for the human PA200 derived from Phyre2 ( Kelley et al., 2015 ) and I-Tasser ( Yang et al., 2015 ). These PA200 models were used as initial guides for the assignment of secondary structure using real-space refinement in Coot ( Emsley et al., 2010 ) and Phenix ( Afonine et al., 2012 ), followed by correction of the sequence register using the amino acid side chain densities clearly resolved in our cryo-EM map. Connecting loops and all other regions were built ab initio from the cryo-EM density. The final atomic models of the recombinant human 20S proteasome and 20S-PA200 complexes were validated using MolProbity ( Chen et al., 2010 ) ( Figures S2 G and S2H). Graphic representations The representations of our structures as shown in Figures 2 A, 2D, S2 E, and S2F were created using UCSF Chimera ( Pettersen et al., 2004 ), while all the remaining structure representations in the manuscript were created using the Pymol Molecular Graphics System, Schrödinger, LLC. The cryo-EM maps shown in Figures S2 E and S2F were colored according to local resolution as estimated using ResMap ( Kucukelbir et al., 2014 ). The diagrams of the protein-protein interaction networks shown in Figure S4 were created with LigPlot+ ( Laskowski and Swindells, 2011 ).

Quantification and Statistical Analysis

Cell density was measured in a Countess II (Life technology). Protein concentrations were measured in a Nanodrop 2000c (Life technologies). Proteasome activity was measured in Pherastar a (BMG Labtech) with λ excitation = 350 nm and λ emission = 450 nm. Data analysis, including the calculation of Error bars as mean with standard deviation derived from the triplicate measurements represented in Figures 1 and S1 , was done in Prism 8 software (GraphPad).

Data and Code Availability

The cryo-EM density maps of 20S and 20S-PA200 have been deposited into the EMDataBank with accession codes EMD-4877 and EMD-4860, respectively. The corresponding atomic coordinates have been deposited in the Protein Data Bank with accession codes PDB: 6RGQ and PDB: 6REY .

Lead Contact and Materials Availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Paula da Fonseca ( pauladf@mrc-lmb.cam.ac.uk ).

Experimental Model and Subject Details Cell Lines

Sf9 insect cells were cultivated in InsectXPRESS media (Lonza), at 27°C.

Bacterial Strains

Constructs were transformed and amplified in either Escherichia coli TOP10 or Pir1 chemically competent cells (Invitrogen). The DH10EMBacY Escherichia coli strain (Geneva Biotech) containing a baculovirus shuttle vector (bacmid) and a constitutively expressing YFP expression cassette was used to create all expression bacmids containing the cloned genes of interest.

Method Details Cloning and generation of baculovirus cDNAs encoding each of the human 20S proteasome subunits, proteasome assembly chaperones and PA200 were synthesized as codon-optimized genes for expression in E. coli (Epoch). These genes were subsequently cloned in Multibac Turbo vectors pACEBac1, pIDS and/or pIDC using a strategy adapted from the IVA method ( García-Nafría et al., 2016 ). All proteasome α subunits were cloned into a pACEBac1 vector, the β subunits into a pIDS, the chaperones in both pIDC and pACEBac1, and PA200 in pACEBac1. The proteasome β subunits were cloned with their N-terminal pre-peptides, which are cleaved at the final stages of proteasome assembly and maturation resulting in proteolytic active complexes ( Budenholzer et al., 2017 ). We added a TwinStrep-tag at the C terminus of β7 to assist on the purification of mature homogeneous proteasome complexes, as this is known to be the last subunit to be incorporated during the proteasome assembly pathway ( Li et al., 2007 , Marques et al., 2007 ). We also created untagged and N-terminal Twin-Strep tagged versions of PA200, to be co-expressed with the 20S proteasome or expressed on its own, respectively. The pACEBac1 vector, containing the α subunits, and the pIDC vector, containing the β subunits, were Cre-Loxed ( Fitzgerald et al., 2006 ) to create a single vector (20S proteasome plasmid). Attempts were made to Cre-Lox the chaperones containing pIDC vector to either the protesome α or β subunit vectors, but due to the size of the plasmids, the high number of genes and gene repetitions many random recombinant events occur, and for that reason we decided to clone the proteasome assembly chaperones in a separate pACEBac1 vector ( Figure S1 A). All assembled plasmids were transposed to DH10EmBacY cells (Geneva Biotech), and Bacmid DNA purified following published protocols ( O’Reilly et al., 1994 ). Bacmids were transfected (P1 virus), after which Sf9 cells, at 1.5x10 6 cells/ml in InsectXPRESS media (Lonza), were infected and incubated for approximately 72 hours until viability was around 80%. The P2 virus was then filtered and 2% Hi-FBS serum added. This was kept as a stock at 4°C, protected from light.

Expression of proteasome complexes

Each individual baculovirus was initially amplified (P3 amplification) by infecting Sf9 cells at 1.5 x10 6 cells/ml with 1:100 of P2 virus, followed by a 72 hour incubation at 27°C with shaking, or until cell viability was about 80%. The P3 virus were then filtered and used to co-infect Sf9 cells with a density of roughly 2x10 6 cells/ml (12 mL of each P3 virus for each 500 mL of cells). Expression of the human 20S proteasome was accomplished by co-infection of Sf9 cells with two baculoviruses, one containing the 20S proteasome the other the assembly chaperones. The 20S-PA200 complex was expressed by triple co-infection with baculoviruses containing (1) the 20S proteasome, (2) the chaperones and (3) the PA200. Interestingly, we observed that human 20S proteasomes can assemble when expressing only its 14 subunits, but with more than a 3-fold decrease in yield of mature complexes compared with that obtained with chaperone co-expression (data not shown). This suggests that endogenous insect proteasome assembly chaperones can, at least partially, assist in the assembly of the human complex. After infection, the Sf9 cells were incubated at 27°C for 48 hours and harvested by centrifugation at 3,000 g , 20 min, 4°C. Cell pellets were washed with cold PBS, frozen in LN2 and stored at −80°C. Purification proteasome complexes The purification strategy was the same for both reconstituted 20S proteasome and 20S-PA200 complexes. Briefly, thawed cells were resuspended in 3-4 times volume of buffer W (50 mM Tris pH 7.5, 150 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, 1 mM EDTA), lysed by sonication and the lysate cleared by centrifugation at 48,400 g for 30 minutes. The clear lysate was then passed through a 5 μm filter and loaded onto tandem Streptactin Superflow Plus columns (QIAGEN), equilibrated in buffer W, and eluted in buffer E (50 mM Tris pH7.5, 150 mM NaCl, 5% (v/v) glycerol, 1 mM DTT, 1mM EDTA, 2.5 mM d-desthiobiotin). Proteasome containing fractions were pooled together and the TwinStrep-tag cleaved by overnight dialysis against buffer W containing TEV protease, at a ratio of 1:50 (TEV to protein). The sample was filtered again with a 0.4 μm filter and loaded onto Streptactin Superflow Plus columns, where the flow-through was collected. The protein was concentrated using 30 kDa cut-off Vivaspin 20 concentrators (Sartorius) and loaded onto a Superose 6 Increase 10/300 gel filtration column (GE Healthcare) equilibrated with 50 mM Tris pH 7.4, 100 mM NaCl, 1 mM EDTA ( Figure S1 B). Typical yields for the purification of reconstituted 20S proteasomes and 20S-PA200 complexes are 2-3 mg and 1 mg of protein purified from 1 L of Sf9 cultures, respectively.

Expression and purification of PA200

Expression of the human

PA200 was accomplished by infection of Sf9 cells with the TwinStrep-Tev-PA200 containing baculovirus. The purification protocol was the same as described for the recombinant 20S proteasome, except that the last size exclusion chromatography step was performed in a Superdex 200 Increase 10/300 (GE Healthcare). Proteasome activity assay 20S proteasome proteolytic activities were measured by fluorescence spectroscopy using the β1, β2 and β5 specific substrates (from Boston Biochem) carboxybenzyl-Leu-Leu-Glu-7-amino-4-methylcoumarin (Z-LLE-AMC), tert-butyloxycarbonyl-Leu-Arg-Arg-7-amino-4-methylcoumarin (Boc-LRR-AMC) and N-succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (Suc-LLVY-AMC), respectively. Recombinant human 20S proteasomes, endogenous human 20S proteasomes (Enzo) and recombinant human 20S-PA200 complexes, all at 7.5 nM in 25 mM HEPES, pH 7.5 and 0.5 mM EDTA, were incubated with 50 μM substrate (stock at 5 mM in DMSO) for 30 minutes at room temperature. Fluorescence intensity, generated by the release of AMC fluorophores, was measured in triplicate with a Pherastar (BMG Labtech) with λ excitation = 350 nm and λ emission = 450 nm. PA200 titration experiments were done by mixing 60 nM of either endogenous or recombinant 20S proteasomes with increasing concentrations of recombinant PA200 (0 nM, 15 nM, 20 nM, 30 nM, 60 nM, 120 nM, 180 nM and 240 nM), followed by incubation for 1 hour at room temperature. The samples were then diluted 20 times and incubated with 50 μM of each of the three proteasome fluorogenic substrates. Measurements were done in Pherastar (BMG Labtech) after incubation at room temperature for 30 minutes. All experiments were done in triplicate.

Cryo-electron microscopy Quantifoil

R1.2/1.3 electron microscope grids, with a 300 gold mesh, were coated with a thin layer of carbon freshly floated from mica, following the procedures we previously optimized for the preparation of grids with endogenous human proteasomes ( da Fonseca and Morris, 2015 , Morris and da Fonseca, 2017 ). We find that the continuous carbon layer favors an even particle distribution, while its electron scattering facilitates the assessment of the information in the recorded images, as judged by the recovery of Thon rings to high resolution in their power spectra, which also contributes to an accurate defocus estimation for the correction of the contrast transfer function associated effects. The grids used for imaging the recombinant human 20S proteasome were glow-discharged in the presence of pentylamine, as we previously described for the endogenous complex ( da Fonseca and Morris, 2015 , Morris and da Fonseca, 2017 ). The grids used for the 20S-PA200 complex were glow-discharged in atmospheric air. 3 μL of sample were loaded on the grids and flash frozen by plunging the grids into liquid ethane using a Vitrobot Mark IV (FEI), operated at 22°C, 95% humidity, 20 s waiting and 5 s blotting times. The grids were transferred into a FEI Titan Krios electron microscope, operated at 300 keV and a nominal magnification of x95,000, resulting in a calibrated sampling of 0.81 Å per pixel at the image level. Images were recorded with EPU software using a Falcon III direct electron detector operating in counting mode, at an electron dose per pixel of ∼0.5 e - /s, with 60 s exposures saved as 75 frame movies with evenly distributed electron dose ( Figures S2 A and S2B).

Single particle analysis

The cryo-EM images were processed using the cis TEM software ( Grant et al., 2018 ). For the analysis of the recombinant human 20S proteasome, the frames of each of the 483 movies recorded were aligned and summed into single images. Upon inspection, 411 images were selected for further analysis, based on the ice quality, image contrast and the recovery of isotropic Thon rings to high resolution. The selected images had a defocus range from −0.8 μm to −3.4 μm. A total of 58,281 particles were picked from the selected images and subsequently classified. The resulting 2D class averages reveal a very homogeneous sample and were used to exclude particle picking false positives. No proteasome top-views, corresponding to projections along the proteasome long axis, were selected since the proteasome side-views show a tomographic distribution of Euler angles, evenly distributed around one of the main axis, providing a complete and even information recovery in Fourier space. The resulting selection of 50,885 particles were used for Auto Refine, with C2 symmetry imposed, using our previous 3.5 Å cryo-EM map of the endogenous human 20S proteasome (EMD-2981) ( da Fonseca and Morris, 2015 ) as starting reference. The overall image processing workflow is shown in Figure S7 A. The resolution of the final map was estimated at 2.6 Å by Fourier shell correlation, as implemented in cisTEM ( Figure S2 C). The analysis of the cryo-EM images of the recombinant human 20S-PA200 followed the same procedures as described for the recombinant 20S proteasome. From the 1,114 movies recorded, 717 were selected for further analysis and had a defocus range from −1.2 μm to −2.9 μm. 29,434 particles were picked from the selected images and classified. The resulting class averages revealed the presence of some uncapped 20S proteasomes, which were excluded from the particle data-set together with particle picking false positives. The 17,356 particles from the selected class-averages were subjected to another round of 2D classification into 10 classes that showed no evidence for significant heterogeneity. Therefore, all 17,356 particles were used for 3D Auto Refine, with C2 symmetry imposed, using our cryo-EM map of the recombinant human 20S proteasome as starting reference, followed by a round of local refinement. An additional 3D classification into 2 classes, using C1 symmetry, indicated that 28% of the particles in the final map are from single capped complexes. The overall image processing workflow is shown in Figure S7 B.The resolution of the final map was estimated at 3.0 Å by Fourier shell correlation, as implemented in cisTEM ( Figure S2 D).

Molecular modeling

The model of the human recombinant 20S proteasome was built based on the X-ray crystal structure of the endogenous human 20S proteasome (PDB: 5LE5 ) ( Schrader et al., 2016 ) using real-space refinement in Coot ( Emsley et al., 2010 ) and Phenix ( Afonine et al., 2012 ). The model of the recombinant human 20S-PA200 complex was built based on our cryo-EM model of the recombinant human 20S proteasome and models for the human PA200 derived from Phyre2 ( Kelley et al., 2015 ) and I-Tasser ( Yang et al., 2015 ). These PA200 models were used as initial guides for the assignment of secondary structure using real-space refinement in Coot ( Emsley et al., 2010 ) and Phenix ( Afonine et al., 2012 ), followed by correction of the sequence register using the amino acid side chain densities clearly resolved in our cryo-EM map. Connecting loops and all other regions were built ab initio from the cryo-EM density. The final atomic models of the recombinant human 20S proteasome and 20S-PA200 complexes were validated using MolProbity ( Chen et al., 2010 ) ( Figures S2 G and S2H). Graphic representations The representations of our structures as shown in Figures 2 A, 2D, S2 E, and S2F were created using UCSF Chimera ( Pettersen et al., 2004 ), while all the remaining structure representations in the manuscript were created using the Pymol Molecular Graphics System, Schrödinger, LLC. The cryo-EM maps shown in Figures S2 E and S2F were colored according to local resolution as estimated using ResMap ( Kucukelbir et al., 2014 ). The diagrams of the protein-protein interaction networks shown in Figure S4 were created with LigPlot+ ( Laskowski and Swindells, 2011 ).

Supplemental Information Document S1. Figures S1–S7 Document S2. Article plus Supplemental Information