Construction of AC40-tagged S. pombe strain

A construct for genomic insertion of a 10xHis/Flag tag was ordered as plasmid (Gene Art). The construct was amplified and genomically inserted into the haploid S. pombe strain 972h-: A 100 ml YPD culture was started at optical density (OD₆₀₀) of 0.25 from an over night culture at 30 °C. After 5–6 h, OD₆₀₀ was at 1.0 and cells were harvested in 250 ml conical tubes (1361 g, 5 min). Cells were resuspended in 25 ml sterile water by vortexing and again centrifuged. Cells were resuspended in 1 ml of sterile 100 mM Li₂Ac solution. The suspension was then transferred into a 1.5 ml reaction tube and centrifuged for 15 s (tabletop centrifuge, full speed). Supernatant was removed and the pellet resuspended in 400 µl of fresh 100 mM Li₂Ac. In parallel, 500 µl of salmon sperm DNA (2 mg/ml) were boiled at 95 °C for 5 min and quickly chilled on ice. The cells were then split into 100 µl aliquots, pelleted and the supernatant removed. To a pellet, the following transformation mix was added in the following order: (1) 240 µl sterile PEG3350 (50% w/v), (2) 36 µl 1 M Li₂Ac, (3) 50 µl salmon sperm DNA (2 mg/ml), and (4) 34 µl PCR product of the insertion construct. Tubes were vigorously vortexed for more than 1 min and incubated at 30 °C for 30 min under shaking. Subsequently, reactions were transferred to 42 °C and incubated under shaking for 25 min. Cells were then pelleted (tabletop centrifuge at 6000 g for 15 s), the supernatant removed and cells were resuspended in 1 ml YPD medium. Cells were transferred into 15 ml conical tubes and shaken at 30 °C for 3 h. After centrifugation at 1361 g for 5 min, the pellet was resuspended in 500 µl sterile water and plated on YPD plates with Kanamycin/G418. The plates were incubated at 30 °C for 3-4 days, single colonies picked and re-plated on fresh plates. For verification of correct genomic insertion, the respective regions were amplified by PCR and sequenced.

Fermentation of S. pombe

S. pombe cells were plated on YPD plates and grown at 30 °C for 48–72 h. A preculture of 500 ml was started and grown over night in YPD at 30 °C under shaking. Cells were inspected for contaminations via light microscopy and secondary cultures of 2 l each were inoculated at a starting OD₆₀₀ of 0.3–0.5. After 10–12 h, cells were inspected visually and transferred into the 200 l fermenter at a starting OD₆₀₀ of 0.30–0.35. YPD medium was prepared in the fermenter, but pH was not adjusted and was therefore at ~6.0 initially. The medium was autoclaved and Ampicillin and Tetracycline were added to final concentrations of 100 µg/µl and 12.5 µg/µl, respectively. Antifoam reagent was added to reduce foaming during the fermentation. The fermenter was operated at 22 Nl/min (normal litres per minute) air influx and with 250 rpm stirring at 30 °C. After 11–13 h, an OD₆₀₀ of 6.0 to 7.5 was reached and cells were harvested with a continuous-flow centrifuge, resuspended in freezing buffer (150 mM HEPES pH 7.8, 60 mM MgCl₂, 20% v/v glycerol, 5 mM DTT, 1 mM PMSF, 1 mM Benzamidine, 60 µM Leupeptin, 200 µM Pepstatin; 0.5 ml buffer for each g of cells) and flash-frozen in liquid nitrogen for storage at −80 °C.

Pol I purification

The protocol for the purification of Sc Pol I^5,45 was slightly modified to be applicable for 10x His tagged S. pombe Pol I:

Frozen fermenter pellets (=150 g cells in a total volume of 225 ml) were thawed and ammonium sulfate concentration adjusted to 400 mM. Cells were lysed after adding 3 ml PI (100x) and 200 ml glass beads (diameter 0.5 mm) by bead beating for 90 min (30 s mixing, 60 s break) under constant cooling. After cell lysis glass beads were removed by filtering and washed with dilution buffer (100 mM HEPES pH 7.8, 20 mM MgCl₂, 400 mM (NH₄)₂SO₄). The crude cell extract was then centrifuged (4 °C; 8,600 g; JLA 16.250) for 60 min to remove the cell debris. The supernatant was afterwards ultracentrifuged (4 °C, 167,424 g; 45Ti rotor) for 90 min. The top fat layer was carefully removed using a 25-ml pipette, the mid-layer was subsequently collected without disturbing the viscous bottom DNA-pellet. The aspired mid-layer was dialysed overnight (16 h + ) at 4 °C against dialysis buffer (50 mM KAc, 20 mM HEPES pH 7.8, 1 mM MCl₂, 10 % v/v glycerol, 10 mM ß-Mercaptoethanol, 1x PI (Benzamidine & PMSF)). The dialysed extract was ultracentrifuged for 2 h (4 °C; 41,856 g; 45Ti rotor). The Pol I containing pellet was resuspended and pellets pooled in Res/W1 buffer (1.5 M KAc, 20 mM HEPES pH 7.8, 1 mM MgCl₂, 10 mM Imidazole, 10 % v/v glycerol, 10 mM ß-Mercaptoethanol, 0.5 PI). After 2 h incubation on a rotating wheel (4 °C; 10 rpm) 4 ml equilibrated Ni-NTA beads were added to the suspension and further incubated for 4 h (4 °C, 7 rpm). After incubation the suspension was decanted into gravity columns, the Pol I binding Ni-NTA beads were subsequently washed with Res/W1 buffer (5 CV) and W2 buffer (300 mM KAc, 20 mM HEPES pH 7.8, 1 mM MgCl₂, 25 mM Imidazole, 10 % v/v glycerol, 10 mM ß-Mercaptoethanol) (5 CV). Pol I was then eluted with 20 ml total volume of E200 buffer (300 mM KAc, 20 mM HEPES pH 7.8, 1 mM MgCl₂, 200 mM Imidazole, 10 % v/v glycerol, 10 mM ß-Mercaptoethanol).

The eluate was therefore ultracentrifuged (4 °C; 46,378 g; 45Ti rotor) for 20 min and loaded on a MonoQ 10/100 column (GE Healthcare) equilibrated with 15% B (Mono-Buffer A: 20 mM HEPES pH 7.8, 1 mM MgCl₂, 10% v/v glycerol, 5 mM DTT; Mono-Buffer B: 2 M KAc, 20 mM HEPES pH 7.8, 1 mM MgCl₂, 10% v/v glycerol, 5 mM DTT). Pol I was eluted with a linear gradient of 13 CVs from 0.3 M to 1.4 M KAc (elution at around 0.9 M KAc). Pol I containing fractions were pooled and diluted 200 mM KAc with Buffer A and again centrifuged (4 °C; 16,696 g; 45Ti rotor). Next, the sample was loaded on a MonoS 5/50 column (GE Healthcare) equilibrated with 200 mM KAc. Pol I was eluted with a linear gradient from 0.2 M to 0.7 M KAc with a plateau of 5 CV at 0.35 M (elution at around 0.5 M KAc). The peak fractions were analyzed on a gel, pooled, concentrated (Amicon; 100 kDa Molecular weight cut-off), flash-frozen in liquid nitrogen, and stored at −80 °C.

RNA elongation and cleavage assays

Purified Sc or Sp Pol I (1, 0.5 or 0.25 pmol) was pre-incubated with 0.25 pmol of pre-annealed minimal nucleic acid scaffold (template DNA: 5′-CGAGGTCGAGCGTGTCCTGGTCTAG-3′, non-template DNA: 5′-CGCTCGACCTCG-3′; RNA: 5′-FAM-GACCAGGAC-3′) in transcription buffer (20 mM HEPES pH 7.8, 60 mM (NH₄)₂SO₄, 8 mM MgSO₄, 10 µM ZnCl₂, 10% (v/v) glycerol, 10 mM DTT) for 20 min at 20 °C. For RNA elongation, NTPs (1.4 mM end concentration each) were added and the reaction was incubated for 30 min at 28 °C. To examine cleavage activity, the pre-incubated reaction with a twofold molar excess of Pol I compared to scaffold was incubated for 30 min at 28 °C without the addition of NTPs. To stop the reaction an equal amount of 2x RNA loading dye (8 M Urea, 2× TBE, 0.02% bromophenol blue, 0.02% xylene cyanol) was added and the sample was heated to 95 °C for 5 min. As control 0.25 pmol of scaffold was treated identically, without the addition of polymerase and NTPs. 0.125 pmol of FAM-labeled RNA product (as well as a marker containing 9 nt, 15 nt and 21 nt long FAM-labeled RNAs: 5′-FAM-GACCAGGAC-3′, 5′-FAM-AACGGAGACCAGGAC-3′, 5′-FAM-UGUUCUUCUGGAAGUCCAGTT-3′) was separated by gel electrophoresis (20% polyacrylamide gel containing 7 M Urea) and visualized with a Typhoon FLA9500 (GE Healthcare).

Preparation of Pol I elongation complex

Synthetic DNA (IDT) and RNA (Qiagen) oligonucleotides were designed and assembled as described¹⁵, with the scaffold sequence for the template DNA (5′-AAGCTCAAGTACTTAAGCCTGGTCATTACTAGTACTGCC-3′), non-template DNA (5′-GGCAGTACTAGTAAACTAGTATTGAAAGTACTTGAGCTT-3′), and RNA (5′-UAUCUGCAUGUAGACCAGGC-3′; for the underlined nucleotides a methylene bridge between the 2′-O and the 4′-C of the ribose ring has been formed, thus creating a locked nucleic acid, LNA). Annealing was achieved by equimolar mixing (40 µM), then heating to 95 °C, and gradually reducing the temperature to 20 °C over 90 min. Pol I (1 mg/ml) was incubated with a 1.35-fold molar excess of pre-annealed EC-scaffold for 30 min at room temperature.

Crosslinking

Purified Pol I (Mono S Eluate at concentration 1.0–1.3 mg/ml) was incubated with BS3 (1 mM final concentration) for 30 min (30 °C, 300 rpm), the reaction was stopped by adding Asp-Lys (9 mM final; 25 °C, 300 rpm) for 20 min followed by ammonium hydrogen carbonate (60 mM final; 25 °C; 300 rpm) for 20 min.

Cryo-EM grid preparation

The samples were centrifuged (4 °C; 21,130 g; Eppendorf tabletop centrifuge) for 5 min, to remove aggregates, and the supernatant carefully transferred into a fresh tube. The sample was then applied to a Superose 6 Increase 3.2/300 column in Solo4 buffer (5 mM HEPES pH 7.8, 1 mM MgCl₂, 10 µM ZnCl₂, 150 mM KCl, 5 mM DTT). The Pol I containing fraction was again centrifuged (4 °C; 21,130 g) for 5 min, and concentration was adjusted to approximately 100 µg/ml. Four µl of sample was applied to a glow discharged (2x; 0.4 mbar 15 mA; 100 s) R1.2/1.3 Cu #300 grid (Quantifoil) and plunge frozen in liquid ethane (Vitrobot Mark IV, Thermo Fisher Scientific; 100 % humidity; 4 °C; 5 s wait time; 5 s blotting time; blot force 12).

Single-particle cryo-EM

Images were collected on a Titan Krios Electron Microscope (Thermo Fisher Scientific) at 300 keV. Movies of 40 frames were acquired on a Falcon III direct electron detector at 75,000x magnification (pixel size 1.0635 Å). The movies were recorded in linear mode with a dose rate of ~19 e⁻/px/s and a total dose of around 86 e⁻/Ų. The defocus span from −1.4 µm to −2.4 µm alternating in 0.2 µm intervals with a total of four exposures per hole.

Data processing

The EC dataset was processed using the RELION 3.0 suite⁴⁶ (Supplementary Fig. 2). Movie frames were aligned and dose weighted using Relion’s own implementation of MotionCor and Contrast Transfer Function (CTF) parameters were estimated using GCTF. A total of 3,598 movies were chosen based on accumulated motion, visual inspection and CTF values, astigmatism, defocus and maximal resolution. A set of 100 randomly picked micrographs throughout the dataset was chosen for reference-free auto-picking using the Laplacian-of-Gaussian (LoG) routine and yielding 2,829 particles. Two-dimensional classification resulted in templates for reference-based auto-picking yielding 299,038 particles. Two-fold binned particles (128 pixel boxes) were subjected to reference-free 2D classification (250 Å mask). Following removal of contaminants, a total of 156,493 unbinned particles were selected and aligned in 3D using an initial model generated in RELION as reference. These particles then underwent CTF refinement, bayesian polishing, followed by another round of CTF refinement. Masked Auto-refinement resulted in a reconstruction at 3.89 Å overall resolution (0.143 FSC). Removal of particles showing increased flexibility of the Jaw and Clamp subdomains were removed by 3D Classification, resulting in 61,954 particles that allow reconstruction of an Sp Pol I EC at 4.00 Å resolution.

The ‘monomer’ dataset was processed using the RELION 3.0 suite⁴⁶ unless stated otherwise (Supplementary Fig. 1). After importing pre-averaged movie frames (sums) the Contrast Transfer Function (CTF) parameters were estimated with the embedded Gctf program. Pre-processing was performed as described for the EC dataset. LoG picking resulted in a total of 11,594 particles from which 1,142 were chosen for template-based auto-picking following 2D classification. Removal of ~90% of initially picked particles can be attributed to contamination and damage resulting from stress on the air-buffer interface. Initial auto-picking identified 874,753 particles from 4,333 micrographs, of which ~50% were discarded as contamination based on 2D-Classification. A total of 477,791 particles were subjected to 3D classification using PDB 5M3M as reference. This allowed the removal of damaged particles and Pol I particles with highly flexible subdomains. The remaining 79,313 particles were subjected to CTF-refinement and bayesian polishing as for EC particles. The final 3D-reconstruction of monomeric Sp Pol I at a nominal resolution of 3.84 Å shows an even orientational distribution of particles and some flexibility in the peripheral jaw, clamp and stalk regions.

In initial 2D classifications, minor dimer-classes were noticed. Thus, auto-picked particles were re-extracted in larger boxes of 360 pixels and analyzed in a second, independent processing tree. From 510,315 Pol - like particles selected by 2D classification, a class of 17,552 particles could be attributed to well-defined dimers. Particles were centered on the interface between both Pols, re-extracted and another 450 poor particles removed based on 2D classification without sampling. Final 3D auto refinement imposing C2 symmetry yielded a reconstruction of Sp Pol I dimers at an overall resolution of 4.5 Å.

Model building

At nominal resolutions of 3.8–4.0 Å, we derive near-atomic models for most regions of Sp Pol I monomers and the EC. To commence model interpretation, we constructed homology models of subunits A190, A135, AC40, AC19, A43, A14 (ker1 in Sp) and A12.2 based on sequence comparison with their Sc homologs, alignment of actual and predicted secondary structures, and domain searches using HHPRED⁴⁷. To construct homology models, the MODELLER⁴⁸ implementation of UCSF Chimera was used⁴⁹. Structures of the general subunits Rpb5, Rpb6, Rpb8, Rpb10 and Rpb12 were imported from the crystal structure of Sp Pol II⁵⁰. Subdomain boundaries were defined based on Sc homology (Fig. 1 and Supplementary Fig. 4). Subdomains were then rigid body fitted into EC densities (which were first obtained) using COOT⁵¹. While many regions allowed accurate fitting of sidechain orientations in the sharpened cryo-EM map, others suffered from poor main-chain tracing. Hence, density-guided modeling was performed in the clamp head, dock-insertion, foot-insertion and part of the jaw regions in subunit A190, as well as the subunit A12.2 and the toe domain of subunit AC40. Modeling of stalk-subunits A43 and A14 was limited to rigid body fitting of trimmed homology model domains in the unsharpened cryo-EM density. As a final step, real-space refinement was carried out using phenix.refine⁵². The Sp Pol I monomer was built by placement of the EC model and adjustment of subdomains in COOT, followed by manual inspection and real-space refinement using phenix.refine. The dimer model was constructed by placement of two monomers, rigid body fitting of subdomains and refinement in phenix-refine using NCS restraints.

Concentration-dependent dimerization using dynamic light scattering

Frozen Sp Pol I was thawed and diluted (1.5 µM, 1.0 µM, 0.75 µM, 0.5 µM) in S600 buffer (10 mM HEPES pH 7.8, 1 mM MgCl₂, 0.01 mM ZnCl₂, 5 mM DTT, 1.5 % (v/v) glycerol, and 0.6 M KAc) to a final volume of 20 µl. Technical duplicates of 10 µl of each Pol I concentration were loaded into glass capillaries (Prometheus NT.48 Series nanoDSF Grade Standard Capillaries) and each capillary mounted into a Prometheus Panta (NanoTemper Technologies GmbH). Fourty consecutive DLS measurements of each capillary were taken at 75 % LED power, 100% Laser power and 15 °C (total measurements per condition: n = 80). Calculation and visualization were carried out using GraphPad Prism version 8.0.1 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com. The boxes extend from the 25th to 75th percentiles⁵³. The whiskers in Fig. 2b are drawn down to the 10th percentile and up to the 90th. Points below and above the whiskers are drawn as individual points.

Analytical size exclusion chromatography

A total of 50 µg of frozen Sp Pol I was thawed and diluted to 2.93 µM with SEC buffer (5 mM HEPES pH 7.8, 1 mM MgCl₂, 10 μM ZnCl₂, 5 mM DTT, and 1.5 M KAc (S1500), 600 mM (S600), or 300 mM (S300)) to a total volume of 30 µl. The sample was centrifuged (4 °C; 21,130 g; Eppendorf tabletop centrifuge) for 5 min, to remove aggregates, and the supernatant carefully transferred into a fresh tube. The sample was then applied to a Superose 6 Increase 3.2/300 column (GE Healthcare; flow 0.035 ml/min; 50 µl fractions) in the respective buffer (S1500, S600, or S300). Peak fractions (Fig. 2) were diluted and negatively stained. After each run, a 30 µl mixture of marker proteins (Thyroglobulin (669 kDa), ß-Amylase (223 kDa), Albumin (66.5 kDa), Calmodulin (29 kDa) was applied to the column for calibration in each buffer.

Negative staining, EM data collection and image processing

Analytical SEC peak fractions were diluted to 20% (v/v) and 10% (v/v) in their respective buffers and were centrifuged (4 °C; 21,130 g; Eppendorf tabletop centrifuge) for 5 min. Five µl of the samples were then applied to 400-mesh copper grids (G2400C; Plano) with a self-made carbon film of ~7 nm thickness (self-made). After 30 s, grids were washed in 200 µl ddH₂O for 30 s, and stained three times in 20 µl saturated uranyl formate solution (30 s). After each step, excess liquid was removed with a filter paper. Images were collected on a JEOL 2100-F Transmission Electron Microscope operated at 200 keV and equipped with TVIPS-F416 (4kx4k) CMOS-detector at 40,000x magnification (pixel size 2.7 Å) with alternating defocus (−2.5 to −4.5 µm).

The images were processed using RELION 3.1 (see above). For the high-salt SEC peak fraction (S1500), a total of 90 out of 98 collected micrographs were analyzed. A set of 10 randomly chosen images was used to train and optimize the reference-free auto-picking using Laplacian-of-Gaussian (LoG) routine. These settings were then applied on all 90 images yielding 40,544 particles that were applied to reference-free 2D classification (380 Å mask). After removal of junk, a total of 20,054 particles were classified into 16 classes.

For the shown low-salt SEC peak fraction (S300), 129 micrographs were analyzed. A set of 10 randomly chosen images was used to train and optimize the reference-free auto-picking using Laplacian-of-Gaussian (LoG) routine. These particles underwent selection based on 2D classification and 3D centering yielding 3,532 particles that were subsequently used as a template for a reference-based auto-picking from all 129 images resulting in 36,733 particles. After 3D centering using the filtered density of PDB 5M3M as reference and removal of junk particles by 2D classification (380 Å mask), 24,462 particles remained. The outcome of a 2D classification into 16 classes was then compared to the high-salt particles (Supplementary Fig. 5).

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.