Structural Characterisation of Mitochondrial Macromolecular

Sorbhi Rathore Structural characterisation of mitochondrial macromolecular complexes using cryo-EM Structural characterisation of mitochondrial macromolecular complexes using cryo-EM Mitoribosome biogenesis and respiratory chain supercomplex

Sorbhi Rathore

ISBN 978-91-7911-156-4

Department of Biochemistry and Biophysics

Doctoral Thesis in Biochemistry at Stockholm University, Sweden 2020

Structural characterisation of mitochondrial macromolecular complexes using cryo-EM Mitoribosome biogenesis and respiratory chain supercomplex Sorbhi Rathore Academic dissertation for the Degree of Doctor of Philosophy in Biochemistry at Stockholm University to be publicly defended on Monday 8 June 2020 at 13.00 in Magnélisalen, Kemiska övningslaboratoriet, Svante Arrhenius väg 16 B.

Abstract Mitochondria, popularly known as powerhouse of the cell, contain specialised mitoribosomes that synthesise essential membrane proteins. These essential proteins are required to form enzyme complexes, which carry out the process of oxidative phosphorylation (OXPHOS). OXPHOS is carried out by five enzyme complexes (Complex I-V), out of which complex I, III and IV pump protons during electron transfer from NADH to O2 and complex V uses the generated proton gradient to synthesise ATP. Cryo-EM, as a revolutionary technique in structural biology made it possible to determine the structures of mitoribosome assembly intermediates and respiratory chain supercomplexes. These structures have allowed us to investigate the mitoribosome biogenesis pathway in human and yeast and to gain deeper insights into the architecture of supercomplexes. In the first area of research, using cryo-EM we were for the first time able to capture mitoribosomes in different late stages of assembly and to determine their high-resolution structures with novel factors bound. Investigation of this process was previously unreachable due to lack of techniques to trap these mitoribosome complexes in different states of assembly. The structures of these assembly intermediates establish the role of assembly factors such as MALSU1, LOR8F8, mt-ACP, MTG1 and mitoribosomal proteins (MRPs) in mitoribosome biogenesis and to ensure proper maturation of each subunit, reflecting their role in regulating translation. Furthermore, genetic deletion studies of MTG1 and uL16m in yeast show the importance of transiently acting factors and MRPs in the mitoribosome assembly process and their effects on translation. The assembly pathway of mitoribosomes is critical for protein synthesis since defects in the translation process causes inherited human pathologies. Therefore, elucidation of mitoribosomal biogenesis pathways may also contribute to the development of potential new therapeutic opportunities. In the second research area, structures of the respiratory chain supercomplex from yeast were determined. These are the first near-atomic resolution structures that show organization of complex III and complex IV into two distinct classes that form higher order assemblies (III2IV1and III2IV2). Moreover, the architecture of the supercomplex structures differs from the previously determined respirasomes (I1III2IV1) structures in mammals. We obtained a near-atomic resolution structure of the yeast complex IV, revealed core protein-protein and protein-lipid interactions that hold the supercomplex together. Moreover we found novel subunits required for supercomplex formation in S. cerevisiae. The last part of my study focuses on cryo-EM sample method development where we could successfully demonstrate the usefulness of a simple pressure-assisted sample preparation method for microcrystals, proteins and mitochondria. Our findings show great resolution improvements of selected area electron diffraction patterns of microcrystals, a significant reduction in needed sample concentration for single particle studies and an enrichment of gold nano-particles for tomographic studies.

Keywords: Mitochondria, mitoribosome biogenesis, mitoribosome assembly factors, yeast respiratory supercomplexes, single particle electron cryo-microscopy, cryo-EM sample preparation.

Stockholm 2020 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-180776

ISBN 978-91-7911-156-4 ISBN 978-91-7911-157-1

Department of Biochemistry and Biophysics

Stockholm University, 106 91 Stockholm

STRUCTURAL CHARACTERISATION OF MITOCHONDRIAL MACROMOLECULAR COMPLEXES USING CRYO-EM

Sorbhi Rathore

Structural characterisation of mitochondrial macromolecular complexes using cryo-EM

Mitoribosome biogenesis and respiratory chain supercomplex

Sorbhi Rathore ©Sorbhi Rathore, Stockholm University 2020

ISBN print 978-91-7911-156-4 ISBN PDF 978-91-7911-157-1

Printed in Sweden by Universitetsservice US-AB, Stockholm 2020 To my mum, papa and husband

List of publications

I. Brown A, Rathore S, Kimanius D, Aibara S, Bai X.C, Rorbach J, Amunts A & Ramakrishnan V., (2017). Structures of the human mitochondrial ribosome in the native states of assembly. Nature Structural and Molecular Biology, 24, 866-869.

II. Rathore S*, Conrad J*, Ott M & Barrientos A., (2020). Cryo- EM reveals different mitoribosome assembly intermediates in yeast knockout strains of Mtg1 and uL16m. (Manuscript)

III. Rathore S*, Berndtsson J*, Marin-Buera L*, Conrad J, Carroni M, Brzezinski P & Ott M., (2019). Cryo-EM structure of the yeast respiratory supercomplex. Nature Structural and Molecu- lar Biology, 26, 50–57.

IV. Zhao J, Hongyi Xu, Carroni M, Lebrette H, Wallden K, Moe A, Matsuoka R, Conrad J, Rathore S, Högbom M & Zou X., (2020). A simple pressure-assisted method for cryo-EM specimen preparation. (Submitted)

Additional publication

V. Berndtsson J*, Aufschnaiter A*, Rathore S, Marin-Buera L, Dawitz H, Diessl J, Kohler V, Barrientos A, Büttner S, Fontanesi F & Ott M., (2020). Disruption of mitochondrial respiratory supercomplexes impairs bioenergetics and competitive fit- ness. (Submitted)

* These authors contributed equally

Abbreviations

Mitoribosome(s) mitochondrial ribosomes OXPHOS oxidative phosphorylation MRP(s) mitoribosomal protein(s) mt-LSU large subunit of the mitoribosome mt-SSU small subunit of the mitoribosome mt-tRNA mitochondrial transfer RNA mt-rRNA mitochondrial ribosomal RNA CP central protuberance ES rRNA expansion segments PET polypeptide exit tunnel 5′ UTRs 5′ untranslated regions Mrh4 4th putative Mitochondrial DEAD-Box RNA Helicase Mtg1 Mitochondrial Ribosome Associated GTPase 1 MALSU1 Mitochondrial Assembly Of Ribosomal Large Subunit 1 L0R8F8 MIEF1 upstream open reading frame protein MIEF1 mitochondrial dynamics protein MID51 mt-ACP mitochondrial acyl carrier protein MRM Mitochondrial rRNA Methyltransferase MTERF Mitochondrial Transcription Termination Factor DDX28 DEAD-Box Helicase 28 AAA ATPases associated with diverse cellular activities ADP adenosine-5'-diphosphate ATP adenosine-5'-triphosphate GTP guanosine-5'-triphosphate NADH nicotinamide adenine dinucleotide ETC electron transport chain CI-CIV complex I, II, III and IV SC supercomplex Q ubiquinone QH2 ubiquinol Cryo-EM electron cryo-microscopy Cryo-ET electron cryo-tomography SPA single particle analysis 2D two-dimensional 3D three-dimensional

iii

Contents

List of publications ...... i Abbreviations ...... iii Introduction ...... 1 Background ...... 1 Structures of the 55S human and 74S yeast mitoribosomes ...... 3 Mitoribosome biogenesis ...... 7 Assembly factors involved in mitoribosome biogenesis ...... 8 Role of mitoribosome assembly factors involved in ...... 9 mt-LSU biogenesis ...... 9 Other important known mitoribosome assembly factors ...... 10 Importance of mitoribosome biogenesis in understanding related human pathologies ...... 11 OXPHOS and the respiratory chain complexes in mitochondria ...... 13 Organisation of respiratory chain complexes into supercomplexes ...... 15 Significance of phospholipids for supercomplex structure and function ...... 17 Towards understanding the functional role of supercomplexes ...... 17 Principles of cryo-EM ...... 19 Sample Preparation ...... 19 Imaging with the electron microscope ...... 20 Calculation of a 3D structure ...... 24 Aims ...... 27 Summary of papers ...... 29 Conclusion and future perspectives ...... 33 Sammanfattning på svenska ...... 35 Popular science summary in English ...... 37 Acknowledgement ...... 39 References ...... 40

Introduction

Background The life of all eukaryotes depends on mitochondria that generate their energy supply in the form of ATP. Inside the mitochondria, ATP is synthesised by coordination of several protein complexes (I-V) that carry out the process of oxidative phosphorylation (OXPHOS). Some of the essential, mitochondria- encoded OXPHOS proteins are first synthesised by specialised mitochondrial ribosomes (mitoribosomes) and subsequently incorporated into “respiratory chain protein complexes” that are embedded in the inner mitochondrial membrane (Fig. 1). These respiratory chain protein complexes (I-V) regulate the cell’s ATP production in response to changes in their cellular environment. Although the assembly of respiratory chain complexes (I-V) and their organisation into supercomplexes has been well established, the reason why supercomplexes exist still remains to be understood1,2. Supercomplexes are present in various species such as yeast, plants, some bacteria and also include mammals where they form respirasomes3,4. Mitochondria of Saccha- romyces cerevisiae lack complex I 2, and their complex III and IV form supercomplexes (III2/IV1 and III2/IV2). Despite variation in their structure and composition amongst species, the conserved in situ arrangement of supercomplex assemblies signifies to have an important role in electron transfer 5 from NADH to O2 .

Mitoribosomes co-localise with the inner mitochondrial membrane during the synthesis of OXPHOS proteins, thereby facilitating co-translational insertion of highly hydrophobic nascent polypeptides required for the formation of respiratory chain complexes (Fig. 1). Structural complementarity between mitochondrial ribosomes and OXPHOS complexes6 can be seen between species; from bacteria to eukaryotes such as fungi and mammals. Mitoribosomes share common features between bacterial and cytosolic ribosomes of eukaryotes such as fungi and mammalian. But differences between human and yeast genetic systems make the structure and composition of mitoribosomes distinct from each other. Human mitoribosomes, for example, are almost exclusively involved in synthesising hydrophobic membrane proteins that are part of the respiratory chain, whereas yeast mitoribosomes synthesise a soluble protein (Var1) along with essential membrane proteins of the OXPHOS. Over the past years, several structural and biochemical studies revealed the organisation and regulation of mitochondrial protein synthesis. Various steps of protein synthesis, such as initiation, elongation, termination, recycling and co-translational insertion of the nascent proteins into the inner mitochondrial membrane have, been well described7. Never-

1 theless, there is no detailed structural information available on the assembly of mitoribosomes in yeast and human mitochondria. Also, our knowledge about auxiliary factors that facilitate mitoribosome assembly is currently very limited.

In order to study the structures for these flexible macromolecular assemblies (mitoribosomes and supercomplexes), electron cryo-microscopy (cryo-EM) and electron cryo-tomography (cryo-ET) have proven to be indispensable tools to determine and visualise structures in their native states5,8,9,10,11,12. This has provided a wealth of information about their structural organisation and for comprehending their functions.

Fig. 1. Important components of a mitochondrion. Mitochondria are ubiquitous, semi- autonomous cellular organelles, which are separated from the cytoplasm by the outer and inner mitochondrial membrane. The inner membrane is folded into cristae, which are site of protein synthesis and OXPHOS activity. The schematic of mitochondria shows its important components: mt-DNA, mitoribosomes, OXPHOS complexes (I-V), protein insertion machinery (insertase), protein transport machinery (TIM: Translocase of the inner membrane; TOM: Translocase of the outer membrane). Mitoribosome assembly factors, MRPs and several other factors required for various processes inside the mitochondria are encoded by the nuclear genome and synthesised by cytosolic ribosomes that are imported to the mitochondria by a translocase of the outer and inner membrane. However, the outer membrane is porous through which small, uncharged molecules and ions can freely diffuse through pore-forming membrane proteins (porins).

2 Structures of the 55S human and 74S yeast mitoribosomes: Architecture, composition, polypeptide exit tunnel, mRNA exit channel and translational activators Determining the structures of mitoribosomes to near-atomic resolutions, by using cryo-EM, has provided exceptional insights into their structure and function. Despite existing similarities between mitoribosomes (human and yeast) and bacterial ribosomes, they are considerably different from each other in their composition, structure and functional requirements (Table 1 and Fig. 2A-D)12,7. Additionally, differences in the size, genetic content and organisation of their mitochondrial genomes make mitoribosomes highly specialised by employing efficient translation mechanisms.

Human mitoribosomes are specialised to synthesise 13 essential membrane proteins involved in OXPHOS. They are composed of 82 nuclear encoded proteins and mitochondria-encoded mt-rRNA that has been reduced by a factor of two when compared to the bacterial counterpart. A fully assembled, mature human mitoribosome has a sedimentation coefficient of 55S and a molecular mass of ~2.7 MDa. It is composed of a 39S large subunit (mt- LSU) and a 28S small subunit (mt-SSU). These subunits contain 16S rRNA and 12S rRNA respectively along with mt-tRNAVal in the central protuberance (CP)11,13. Human mt-LSU is composed of 52 proteins, of which 22 are specific to mitochondria, whereas mt-SSU is composed of 30 proteins with 14 specific to mitochondria12 (Fig. 2D).

Mitoribosomes of Saccharomyces cerevisiae synthesise 7 essential membrane proteins involved in OXPHOS including a soluble protein of the mt- SSU called Var1 (uS3m). The 74S yeast mitoribosome has an overall molecular weight of ~3-3.3 MDa. It is composed of a 54S large subunit (1.9 MDa) and a 37S small subunit (1.1 MDa) which contain 21S and 15S rRNA respectively and lacks the 5S rRNA and also the 5S RNA-binding proteins (L18 and L25) in its central protuberance region14. Yeast mt-LSU is composed of 39 proteins, of which 13 proteins are specific to mitochondria. The mt-SSU is composed of 34 proteins out of which 14 are mitochondria- specific proteins. Out of these proteins 7 have their homologs in the mammalian mitoribosome and 7 are yeast mitoribosome specific10 (Fig. 2C).

In addition, yeast mitoribosomes contain 11 unique mitochondria specific rRNA expansion segments (ES) with 392 additional nucleotides14, when compared to their bacterial counterpart. Mammalian mitoribosomes, however, do not contain rRNA expansion segments and have an overall reduced rRNA content, which is confined to the inner ribosome core that has more

3 protein coverage. Although the functional role of ESs in yeast mitoribosome is yet to be understood, it may have an architectural role similarly to that of CP tRNA in mammals or 5S rRNA in bacteria. The CP is an important fea- ture of the large subunit because it facilitates intersubunit contacts to the small subunit and to the bound tRNAs. Moreover, it acts as a structural scaf- fold for the binding of ribosomal proteins15,16,17.

Fig. 2. Comparison of ribosome structures of bacteria, eukarya (cytosolic) and mitochondria. (A-D). Structures highlight differences in their overall architecture, size, protein and rRNA content. Bacterial ribosomes contain 5S rRNA in the CP region, which is substi- tuted by an rRNA expansion segment in yeast and in human by tRNAVal. The rRNA of the LSU and SSU is coloured in blue and yellow respectively. (PDB: 5KCR18; 6SGC19; 5MRC10; 3J9M11 used for comparison)

It is also interesting to see how the path of the polypeptide exit tunnel (PET) differs between 55S and 74S mitoribosomes. Although the PET path in mammalian mitoribosomes (55S) and bacterial ribosomes (70S) show similarities to each other, the presence of more proteins and two rRNA deletions in the mammalian mt-LSU (39S) makes the polypeptide exit tunnel more proteinaceous than in bacteria17,13. In the yeast mitoribosome, the PET is extensively remodelled by mitoribosomal proteins (MRPs) and its entrance is more constricted due to base pairing between rRNA residues (U2877 and A1958). In addition, two aromatic residues of uL22 and uL24 further con- strict the top of the polypeptide tunnel to a diameter of about 9 Å. Presuma- bly, these alterations in the yeast tunnel may provide mechanism for re- sistance to certain antibiotics11.

Although, the structures of the 55S and the 74S have given unprecedented insights into the architecture, composition and functionality of these macromolecular machines, yet the mechanisms that regulate mitochondrial translation remain elusive. Yeast mitochondrial mRNAs have long 5′ untranslated regions (5′ UTRs) and transcript-specific translational activators that initiate and regulate translation. Whereas mammalian mitochondrial mRNAs lack 5’UTRs or the bacterial Shine-Dalgarno sequences and unlike yeast mitochondrial mRNAs are polyadenylated at the 3′ end20,21,22. Still, translational activation does take place in mammals and the translational activator of

4 Cox1 (TACO1) is the first reported mitochondrial-specific translational activator in mammals23.

Cryo-EM structures have revealed species-specific translation requirements in mammalian and yeast mitochondria that show two distinct mRNA exit channels in the mt-SSU between these organisms. In mammals, the mRNA exit channel is made by three proteins (bS1m, bS21m, and mS37)11 in comparison to S. cerevisiae, where the mRNA channel exit is formed by seven proteins (mS42, mS43, bS6m, uS15m, uS17m, bS18m and bS21m). Thus the extended channel platform likely allows interaction of translational activators with the mt-SSU and the mRNA10. Furthermore, in yeast, the extended binding platform of the mRNA channel exit with bound translational activators might protect the emerging mRNA from oxidative damage and degradation in the mitochondrial matrix9,8,24. Translational activators are present in small quantity inside mitochondria, which regulate translation efficiency according to the change in their environment, e.g., in yeast, switching to a non-fermentable substrate results in increase of translational activators25.

In yeast, an extensive research on translational activators has revealed their role in facilitating translation by interacting with mitochondria-specific mRNAs26,20. Many translational factors have been identified so far; however, it is believed that they might have different modes of functioning. Studies have shown that nuclear encoded translational activators interact with either the 5′ UTR of the mRNA to promote translation or stabilise mRNA, while some translational activators bind to mitoribosomes27. The COX1 mRNA is the target of four translational activators: Pet309, Mam33, Mss51 and Mss116. Pet111, which has recently been classified as a PPR protein, serves as membrane-associated translational activator of COX228. Synthesis of Cox3 depends on the translational activators Pet54, Pet122 and Pet494 that interact together with COX3 mRNA, proteins of the mt-SSU and the inner mitochondrial membrane29,27. Synthesis of COB mRNA is regulated by Cbs1, Cbs2, Cbp1, Cbp3 and Cbp6. Similarly, translational activators exist for other mitochondrial encoded genes such as ATP6/8, ATP9 and Var120, which thereby regulate translation of mitochondrial mRNAs linking it to the assembly of mitochondrial respiratory chain complexes.

5 Table 1. Comparison of composition from bacteria and yeast cytosolic ribosomes with mitoribosomes (yeast and mammalian) along with their genomes. (Adapted from Greber and Ban et al., 201412 and updated7,10).

*Coeff. (Coefficient); Sediment. (Sedimentation); ribo. (ribosomal); nt (nucleotides) Organism E. coli Cytosolic Mitoribosome Mitoribosome (S. cerevisiae) (S. cerevisiae) (Human and porcine) Sediment. coeff. 70S 80S 74S 55S Molecular weight 2.3MDa 3.3-4.3MDa 3-3.3MDa 2.7MDa ribo. proteins 54 79-80 73 82 rRNA molecules 3 4 2 3 LARGE SUBUNIT Sediment. coeff. 50S 60S 54S 39S LSU proteins 33 46-47 39 52 rRNA 23S 26S–28S 21S 16S (1,569 nt) (2904nt) (3,396– (3,296nt) CP tRNA and 5S 5,034nt) (73–75nt) (120nt) 5.8S (156- tRNAVal -Human 158nt) tRNAPhe-Porcine 5S (120–121nt) SMALL SUBUNIT Sediment. coeff. 30S 40S 37S 28S SSU proteins 21 33 34 30 rRNA 16S 18S 15S 12S (1,534 nt) (1,800– (1,649 nt) (962 nt) 1,870nt) GENOME Size 4.7Mb ~12Mb 75Kb 16.5Kb (mtDNA) (mtDNA) RNAs encoded 46 tRNAs 42 tRNAs30 2 rRNAs, 2 rRNAs, 24 tRNAs, 22 tRNAs, and 7 mRNAs, and 11 mRNAs 9S RNA of RNase P OXPHOS - - Cox1, Cox2, COX1, COX2, proteins Cox3, CytB, COX3, CYB, Atp6, Atp8, ND1, ND2, Atp9 ND3, ND4, ND4L, ND5, ND6, ATP6, ATP8 MRPs - - Var1 None

6 Mitoribosome biogenesis In both mammals and yeast, the mitoribosomal proteins are encoded by the nuclear genome and imported into the mitochondrial matrix where they are assembled with newly synthesised mt-rRNA. Therefore, mitoribosomal biogenesis or assembly has an additional level of complexity over the cytoplasmic system, as it requires high level of synchronisation between the nuclear and mitochondrial genomes.

Unlike bacterial or cytoplasmic ribosomes, the assembly of mitoribosomes is an extremely intricate and dynamic process. The biogenesis of mitoribosomes can be described by the following strictly coordinated and overlapping steps, which are driven by numerous trans acting assembly factors31: (a) synthesis of mt-rRNAs and post-processing steps; (b) expression of nuclear genes for mitoribosomal proteins (MRPs); (c) gradual maturation of each subunit, including post-transcriptional and post-translational modifications; (d) assembly of the mature mt-SSU and mt-LSU to form functional mitoribosomes (Fig. 3).

In situ structures of yeast and human mitoribosomes have shown their association with the inner membrane in translation-competent mitochondria8,9, where several inner membrane proteins and MRPs at the tunnel exit accom- plish the membrane attachment of the mitoribosome. These include interaction of Mba1, Mdm38 and Oxa1 in yeast32,9 and mL45 and OXA1L in hu- mans8,33. It has also been suggested that the assembly of the mt-LSU occurs while being tethered to the inner mitochondrial membrane31 while the mt- SSU follows a separate assembly pathway (Fig. 3).

The assembly pathways of mt-LSU and mt-SSU remain elusive in both mammals and yeast and most of the clues to understand them are provided from in vitro reconstitution experiments of bacterial ribosomal proteins and their assembly factors34,35,36,37,38. In yeast, additional insights were gained from biochemical and mass-spectrometry studies39.

Fig. 3. Mammalian mitoribosome biogenesis and the translation cycle. Mitoribosome biogenesis occurs by precise coordination of the nuclear and mitochondrial genome. All the essential components needed for the mitoribosome assembly (12S and 16S rRNA, MRPs, assembly factors) come together in a concerted manner to form a mature mitoribosome. One of the crucial steps in the assembly process is post-translational modifications of some proteins and posttranscriptional modification of rRNAs. Various assembly factors of mt-SSU and mt-LSU drive and regulate the mitoribosome assembly pathway. During this process, the mt- LSU is tethered to the inner mitochondrial membrane, whereas mt-SSU assembly follows a separate pathway. The schematic also displays the translation cycle, which occurs in three universal steps: initiation, elongation, and termination followed by recycling, and each step is facilitated by translation factors. In yeast, mitoribosome biogenesis and translation follows an overall similar pathway. (Adapted from Hallberg et al., 201431 and updated40,41).

Assembly factors involved in mitoribosome biogenesis Numerous proteins (auxiliary assembly factors) have been suggested to be implicated in the complex mitoribosomal biogenesis pathway, including GTPases, DEAD-box helicases, nucleases, rRNA modifying enzymes, and chaperones42,38. However, considering the lack of structural information on the assembly stages, little is known about interaction of auxiliary factors with the mitoribosome (Fig. 3). In humans, several biogenesis assembly factors have been characterised for mt-LSU: MRM2, MRM3 MALSU1 (C7orf30), MTERF4, MTERF3, FASTKD2, DDX28, DHX30, GTPBP1043,44,45,46,47,48 and for mt-SSU: NOA1 (c4orf14), GRSF1, ERAL149,47,50,51. NOA1 is the only factor that was shown to function in the early stages of assembly, whereas the rest of the characterised assembly factors are involved in late stages of the mitoribosome biogenesis pathway. A new factor called NGRN is also speculated to be involved in mitoribosome biogenesis52. Early and late stages of biogenesis also differ in their compartmental localisation. While the former takes place at

8 the mt-DNA nucleoids where mitoribosomal proteins are assembled onto mt-rRNAs during its transcription42, the latter occurs at the inner mitochondrial membrane, which is the ultimate site of translational activity53.

Importantly, in yeast, many known homologs of human mitoribosome assembly factors exist, such as Mrm1, Mrm2, Pus5 required for rRNA methyl- ation; Mtg1, Mtg2, Mtg3 for GTPase activity and a DEAD box protein, Mrh4 for remodelling of the 21S rRNA-protein interactions54,55,38. This makes S. cerevisiae an excellent model organism to study the effect of genetic deletion or mutation of these auxiliary factors that cause metabolic defects and accumulation of subunit precursors. Moreover, in studies similar done to those in E. coli56, systematic deletion of 37 mt-LSU proteins have shown effects on protein translation and assembly process of S. cerevisiae mt-LSU39. This provides strong evidence for the role of MRPs at various stages of mitoribosome assembly.

The number of the mitochondrial assembly factors characterised so far represents only a small set of proteins in comparison to bacterial and cytosolic ribosomes. Moreover, there are only a few biochemical studies reporting on factors involved in mt-SSU biogenesis, making it difficult to completely understand mitoribosome assembly. Notably, most of the current structural information comes from cryo-EM studies of pre-60S and pre-40S cytoplasmic ribosomes during assembly that were purified from S. cerevisiae57,58,59,60 or from in vitro studies on E. coli ribosomes34,34, 35,36. Thus, further research in this field is required to identify novel mitoribosomal assembly factors and to establish their function.

Role of mitoribosome assembly factors involved in mt-LSU biogenesis Formation of a mature mt-LSU or mt-SSU requires several trans-acting assembly factors that interact with the rRNAs and MRPs during several steps of mitoribosome assembly. As described in the section above, only a handful of these factors have their role defined in mitoribosome biogenesis.

GTPases: GTPases are a universally conserved class of regulatory proteins and are involved in various cellular functions. GTPases that associate with the ribosomes are composed of Obg type G-domains with a characteristic fold and 4–5 highly conserved motifs (G1-G5) and additional N- and C- terminal domains. Importantly, GTPases constitute a large and diverse class of ribosome assembly factors in bacteria and eukaryotes, such as yeast, where they bind to in the GTP-bound form and dissociate after GTP hydrol-

9 ysis. As described above, a few GTPases have been characterised in mitochondria of yeast and human that play similar roles to those observed in E. coli. For example, GTPases involved in bacterial ribosome maturation are RsgA, Era and YqeH for the 30S subunit and YihA, RbgA, Der and Obg for the 50S subunit61. Depletion of GTPases in bacteria leads to accumulation of stable assembly intermediates, which have been purified using sucrose gra- dients and analysed for their protein content. In S. cerevisiae, assembly of 60S LSU requires five essential GTPases: Nog1, Nug1, Nug2 (Nog2), Ria1, and Lsg1, while Bms1 is the only known GTPase involved in assembly of the yeast 40S small subunit62.

RNA helicases: DEAD-box proteins are superfamily-2 RNA helicases that are universally involved in RNA-mediated processes. These enzymes clamp the RNA substrate in an ATP-dependent manner, either to form an RNA- binding complex or to unwind the double-stranded RNA63. In human mitochondria, the DHX30 helicase participates in mt-LSU assembly and knock- down of DHX30 in human cells produces a severe translation defect and reduced monosome formation. Similarly, Mrh4 in yeast and DDX28 in humans also interact with mt-LSU and monosomes during late stages of mitoribosome assembly. They are also involved in promoting remodelling of the 21S rRNA-protein interactions64,54. However, a possibility exists that these helicases could have more defined roles during mitoribosome assembly and translation.

Methyltransferases: Modification of rRNA is an important part of ribosome assembly, which occurs during or after transcription or during the assembly process. MRM1, MRM2 and RNMTL1 are responsible for modifying human mitochondrial 16S rRNA at residues G1145, U1369, and G1370, respectively65. However, it is known that only MRM2 and MRM3 are required for biogenesis and function of the large subunit of the mitochondrial ribosome43. In yeast, Mrm1 and Mrm2 play similar roles in modifying the 21S rRNA, however their roles in mitoribosome assembly are not yet de- fined66,38.

Other important known mitoribosome assembly factors MALSU1 (c7orf30): MALSU1 belongs to the family of well-conserved proteins which has the YbeB (E.coli)/iojap (Zea mays) domain (also called the DUF143 family, universally conserved ribosomal silencing factors). Biochemical studies have shown that MALSU1 is particularly necessary for mt-LSU biogenesis and stability. siRNA results have shown reduced mitochondrial translation and respiratory deficiency due to loss of fully

10 assembled mitoribsomes and unstable mt-LSU44. Moreover, the conserved interaction between C7orf30 and uL14m promotes mt-LSU biogenesis by sterically inhibiting the association of the small and large ribosomal subu- nits67,68.

MTERF4/3: MTERF proteins play an important role in coordinating mitochondrial transcription69. Mitochondrial transcription termination factors have been reported to be involved in regulating mt-LSU biogenesis. MTERF4 along with NSUN4, a methyltransferase, promotes monosome assembly45. However, if MTERF3 is essential for mitoribosome biogenesis is still unclear46.

Chaperones: It has been reported that some chaperones, such as the m- AAA protease might be involved in regulating mitoribosome assembly and translation by controlling proteolytic maturation of a ribosomal subunit70.

Importance of mitoribosome biogenesis in understanding related human pathologies Mutations in mt-DNA and defects in the mitochondrial transcription and translation machinery are the major cause of mitochondrial dysfunction that results in severe human pathologies71,72. For example deficiencies in OXPHOS comprise the largest collection of inborn errors of metabolism52. Specifically for mitoribosomes, mutations in proteins uL3m73, uL12m74, mL4475, mS2276, bS16m77, and uS7m78 affect protein stability and formation of mitoribosomal subunits75, resulting in OXPHOS deficiencies and fatal disorders such as dysmorphism, lactic acidosis, neurodegeneration and car- diomyopathies. Since involvement of assembly factors is critical for the correct mitoribosome maturation, their functioning is also medically rele- vant79. Knockout studies of assembly factors in mouse models have shown impairment of mt-LSU assembly and mitochondrial translation46,80.

Defects in mitochondrial translation result in mitochondrial diseases that primarily affect tissues with a high metabolic demand, such as the nervous, muscular, cardiac, ocular and endocrine systems81. In addition, upregulation of mitochondrial translation was found to occur in a subset of human malig- nancies, and its inhibition showed anti-cancer activity in preclinical stud- ies82. This is likely due to metabolic reprogramming of incipient cancer cells that results in extremely high dependencies of tumorogenesis on the mitochondrial activity and intrinsic mitochondrial biogenesis83. Therefore, mitoribosomes and their assembly pathways are now recognised as novel tar-

11 gets for therapeutic intervention84,79.

To address the challenges associated with the characterisation of assembly factors in human cell lines, yeast as a model organism has provided solu- tions to counter the problems associated with genetic manipulations of human mt-DNA. S. cerevisiae is a facultative anaerobic organism allows us to use its machinery to understand mitochondrial biogenesis and molecular mechanisms of human diseases resulting from impaired mitochondrial metabolism. Considering the similarities that exist between yeast and human mitochondria, application of systematic functional screening using a whole- genome pool of yeast deletion mutants has led to the identification of various human nuclear genes involved in mitochondrial disorders85,86. Hence, characterisation of assembly factors in the context of human mitoribosome biogenesis will be important from a medical point of view.

12 OXPHOS and the respiratory chain complexes in mitochondria In cells, mitochondria are the site for oxidative phosphorylation, which is the prime source for ATP generation. In addition, several other vital processes such as lipid and steroid synthesis, biosynthesis of iron-sulphur clusters and iron metabolism also occur inside mitochondria87,88.

In OXPHOS, energy is generated in the form of ATP by transferring electrons from oxidisable substrates such as NADH and FADH2 to molecular oxygen (O2). This involves electron transfer along an electron transport chain (ETC), which is composed of four major enzyme complexes (I-IV) that are embedded within the inner mitochondrial membrane. During this process, protons are pumped from the mitochondrial matrix into the inter- membrane space (IMS) by complexes I, III and IV. This creates a proton gradient across the inner mitochondrial membrane. The proton gradient is subsequently used by the F1Fo-ATP synthase (complex V) to produce ATP 89,90 from ADP and Pi (Fig. 4) and also enables mitochondrial import and export of metabolites and proteins.

The individual respiratory chain complexes are described below (Fig. 4):

Complex I (NADH: ubiquinone oxidoreductase) is the largest membrane- bound enzyme of the ETC with an overall size of 970 kDa. It contributes to ATP production by catalysing electron transfer from NADH to ubiquinone, which is coupled to proton translocation across the membrane90,91. In the redox reaction catalysed by complex I, NADH is oxidised by flavin mono- nucleotide, and the resulting two electrons are carried sequentially along seven FeS clusters to bound ubiquinone in order to reduce ubiquinone (Q) to ubiquinol (QH2). This redox reaction releases potential energy, which is utilised to transport protons across the mitochondrial inner membrane, contributing to the proton-motive force (Δp). Mammalian complex I contains 45 subunits, comprising 14 conserved core subunits involved in catalytic activity, and 31 mammalian-specific supernumerary subunits. Seven subunits of the hydrophobic arm are encoded by the mitochondrial genome and the other seven hydrophilic subunits are encoded by the nuclear genome92. Complex I is absent in S. cerevisiae and NADH dehydrogenases (Ndi1, Nde1 and Nde2) are present, instead93. These enzymes perform oxidation of NADH and transfer electrons to the mobile electron carrier ubiquinone (Q), yet they do not perform proton transport across the inner mitochondrial membrane.

13 Complex II (Succinate: ubiquinone oxidoreductase) is an integral membrane protein complex. In mammals, complex II consists of four subunits (SDHA, SDHB, SDHC, and SDHD) that are encoded by nuclear DNA. Ac- cording to the crystal structure of CII, the complex comprises three prosthetic groups: FAD, [Fe–S] clusters, and heme94. During OXPHOS, SDH transfers electrons from succinate via its [Fe–S] clusters to ubiquinone95. Com- plex II does not pump protons but contributes indirectly to the ∆p via reduction of the ubiquinone pool96. Biochemical studies suggest similarities between yeast and mammalian CII97.

Complex III (ubiquinol-cytochrome c oxidoreductase or cytochrome bc1 complex) catalyses the electron transfer from ubiquinol (QH2) to cyt c, which is coupled to the net translocation of protons across the membrane, thereby oxidizing ubiquinol and reducing cytochrome c98. A transient interaction between the soluble protein cytochrome c (cyt c) with the cytochrome c1 (Cyt1) subunit of the cytochrome bc1 complex transports electrons to cytochrome c oxidase (CIV). Complex III functions via a Q-cycle mechanism that couples electron transfer to proton translocation across the inner mitochondrial membrane that ultimately drives ATP synthesis. In the Q cycle, ubiquinol is oxidised at the Qo site where one electron is transferred to heme c1 via the [2Fe − 2S] cluster, which donates the transferred electron to cyt c for its reduction. CIII uses the second electron to reduce Q bound at the Qi site. This releases protons into the inter membrane space. After a second round of electron transfer, the fully reduced Q leaves the binding pocket98. In yeast, complex III forms a homodimer, with each monomer containing ten subunits. Three of its subunits form the catalytic core, namely cytochrome b (Cyt b) with two non-covalently attached heme b groups, a Rieske iron–sulphur protein (Rip1), a [2Fe-2S] iron-sulphur cluster and cyto- 98 chrome c1 (Cyt1) containing a covalently attached heme c group .

Complex IV (cytochrome c oxidase) is the ultimate enzyme of the electron transport chain. It is a multi-subunit enzyme complex that catalyses the reduction of molecular oxygen to water, which is coupled to proton translocation across the membrane. Yeast cytochrome c oxidase consists of twelve subunits99,100. Its catalytic core is composed of mitochondria encoded subunits Cox1, Cox2, Cox3 that are conserved across species101. The complex 102 contains two hemes (a and a3), and two copper centres, i.e. CuA and CuB . Complex IV catalyses a redox reaction by accepting electrons from the reduced and soluble cytochrome c (cyt. c) at the CuA site that are subsequently transferred to the prosthetic heme cluster. Electrons then reach the catalytic site comprising of CuB and heme a3, where molecular oxygen (O2) receives 4 electrons and 4 protons and is thereby reduced to 2 molecules of water (H2O). In the process of reducing one molecule of O2 to two molecules of

14 H2O, four protons are taken up from the matrix, while four protons are pumped across the membrane to the IMS, contributing to the proton gradient. Proton pumping occur through the D and K pathways which form proton channels in cytochrome c oxidase, however if H-channels contribute to proton transfer in complex IV is still a topic of debate103,104.

Fig. 4. Components of the OXPHOS system. The respiratory complexes of the mitochondrial OXPHOS-ETC are shown, namely NADH-ubiquinone oxidoreductase (complex I; PDB: 5XTD105), Succinate-ubiquinone oxidoreductase (complex II, PDB: 1ZOY94), ubiquinol-cytochrome c oxidoreductase (complex III; PDB: 5XTE105) and cytochrome c oxidase (complex IV, PDB: 5Z62106) which transfer electrons generated by oxidation of NADH and succinate to molecular O2 via mobile electron carriers (Q, ubiquinone; QH2, ubiquinol and soluble cytochrome c (PDB: 2B4Z107). Protons are pumped by CI, CIII and CIV from the matrix across the inner mitochondrial membrane to the inter membrane space (IMS). The electrochemical energy stored in ∆p is harvested by ATP synthase (complex V, PDB: 5ARA108) to synthesise ATP. (Adapted from Letts et al., 201796).

Organisation of respiratory chain complexes into supercomplexes Analysis of mitochondrial membrane proteins by blue native polyacrylamide gel electrophoresis (BN-PAGE) has provided one of the main evidences for the association of complex III and IV in yeast (S. cerevisiae) and of complex I, III, IV in mammals (bovine) into supramolecular assemblies called supercomplexes (SCs)3. A series of cryo-EM studies led to the determination of near atomic resolution structures of SCs from bovine, porcine and ovine, that revealed their precise arrangement in the inner mitochondrial membrane109,110,111 and was also confirmed by in situ cryo-ET studies5,112 (Fig. 5A-D). In mammals, CI forms a supercomplex with CIII and CIV (SC I1III2IV1) together with ubiquinol and cytochrome c (cyt c), which transfers electrons from NADH to O2 and hence has been called the respirasome. Moreover, SCs can organise themselves into various assemblies of different stoichiometries such as (SC I1III2), (SC III2IV1) and (SC III2IV2) depending

15 on the species or tissue type96 (Fig. 5A, C). In addition, the highest-order assembly of the human respiratory chain megacomplex (MC I2III2IV2), which was determined by cryo-EM, comprises of 140 subunits and associated cofactors105 (Fig. 5B). The study suggested the role of respiratory mega- complexes in responding efficiently to energy-requiring conditions inside the mitochondria. However, high-resolution structures of yeast SCs were only available one year ago (2019) along with prior low resolution cryo-EM studies confirming their existence in S. cerevisiae113.The difficulties in structural characterization of SCs result from structural heterogeneity and inher- ent flexibility of CIV to form SCs with CIII and limitations of cryo-EM grid preparation, such as exposing the sample to a large air-water interface, thereby destabilizing or disintegrating these weakly bound assemblies of supercomplexes.

In addition, establishing the regulatory role of supercomplex assembly factors (RCF1, RCF2, RCF3 in yeast; COX7A2L/COX7RP/SCAFI in mammals) has been a pending issue. Studies by two research groups have shown contradicting results on the functioning of COX7A2L assembly factor in the formation and stabilization of mitochondrial SCs114,115.

Fig. 5. Comparison of the organisation of respiratory supercomplexes in different organisms. The individual complexes (CI, CIII, CIV) assemble into supercomplexes of different dimensions. In humans the respiratory complexes form supercomplex or bigger assemblies called megacomplex. However, in bacteria such as M. smegmatis smaller supercomplexes are formed. (PDB: 5XTH105; PBD: 5XTI105; PDB: 6QC3116; PDB: 6HWH117)

16 Significance of phospholipids for supercomplex structure and function Nearly 30% of the mitochondrial inner membrane is composed of phospho- lipids118. The weakly interacting supercomplexes are stabilised in this extremely protein-rich environment with the help of these phospholipids. In both yeast and mammals, cardiolipin is the unique phospholipid of mitochondria that is present in high amounts in the inner mitochondrial mem- brane119,120, where it is required to stabilise SCs by interacting with various subunits. Contrary to cardiolipin, phosphatidylethanolamine (PE), another inner membrane phospholipid, has been reported to destabilise SCs121, thus keeping the correct balance of phospholipid species between the complexes. Depletion of PE impairs the activity of respiratory chain complexes, in par- ticular the cytochrome c oxidase in yeast and human. However, depletion of phosphatidylcholine does not affect the activity and stability of respiratory chain supercomplexes in yeast122. These studies demonstrate the dynamic molecular mechanisms that exist between these phospholipids and SCs, which together control mitochondrial respiration. Further evidence is provided by the cryo-EM maps of mammalian repira- somes showing densities between interfaces of contacting complexes which are filled with phospholipids that facilitate the arrangement of SCs123.

Towards understanding the functional role of supercomplexes Despite gained insight through structural and biochemical studies, the assembly mechanism, the structural organisation and the physiological role of SCs remain elusive2. A number of hypotheses have been made that explain why SCs exist. One of those is substrate channelling, which enhances catal- ysis within SCs. However, this theory has been refuted by both structural and biophysical results. Initially, two theories existed that explained the formation of SCs. The first is the fluid state theory, where individual complexes diffuse freely in the membrane, and Q and cyt c diffuse randomly124. The second theory is the solid state theory where SC formation lead to closed channels that result in enclosed exchange pathways of ubiquinone/ubiquinol between CI and CIII, and for cyt c between CIII and CIV, thereby preventing mobile electron carriers from diffusing to the outside pool2,125. Recent cryo-EM structural analysis has not shown substrate channels, or barriers to free diffusion, that connect the substrate-binding sites of CI, CIII, and CIV2,126. Another possible reason suggested to describe the existence and function of SCs is decreased ROS production, since CI and CIII constitute the main redox centres responsible for oxygen reduction to

17 superoxide127,128. In addition, SCs may also be involved in preventing protein aggregation in the protein rich environment of the mitochondrial inner membrane by stabilising the assembly of individual complexes and in regulating respiratory chain activity2,129.

It is important to consider that none of the structures published so far have been able to reveal factors that play a role in supercomplex assembly or formation. This could be due to the weak, dynamic interactions with the supercomplexes or lack of abundance of respiratory complex factors in the mitochondria. In addition, using mild detergents for membrane solubilisation could also lead to partial disruption of weak interactions within these complexes. Hence, to better understand the assembly and function of supercomplexes, there is an urgent need of finding assembly intermediates and assembly factors bound to the supercomplex adopting several states.

18 Principles of cryo-EM

Sample Preparation Proteins need to be shielded from the vacuum in the electron microscope by a support medium. In cryo-EM130, plunge freezing of the protein solution into liquid ethane at liquid nitrogen temperature produces amorphous ice, which provides such a support medium. This freezing method preserves the protein in its natural buffer environment (Fig. 6A). For this purpose, the sample is applied onto a glow-discharged metal grid such as copper or gold that is overlaid with a foil of perforated carbon. The excess buffer is removed from the grid by blotting with a cellulose filter paper, forming a thin liquid film of ~10-90 nm131, which is then immediately plunge frozen. To achieve high resolution the ice film needs to be as thin as possible, since thick ice increases the noise level in images and reduces particle alignability. The production of too thin films, however, causes proteins to interact with the air-water interface, where they denature132 (Fig. 6B). The remaining intact proteins are concentrated on the layer of denatured protein, where they bind to in a preferred orientation133. Both, the effect of denaturation causing proteins to dissociate and the problem of preferential orientation hamper the 3D reconstruction procedure. This can be solved by either stabilising the protein complex using detergent or simply by immobilising the particles on a carbon or graphene support film that shields the proteins from interacting with the air-water interface132 (Fig. 6C).

Advanced sample preparation methods reduce particle interaction with the air-water interface by removing the blotting process altogether and instead spray the protein solution onto self-blotting grids134. Another disadvantage of blotting is that 99% of the protein sample is removed by the filter paper demanding sample concentrations above 1 µM. Therefore, novel freezing methods that allow low sample concentrations would enable the study of proteins that are difficult to isolate in high quantity or that cannot be sufficiently concentrated.

Although cryo-EM has opened up the structural characterisation of many macromolecular complexes that could previously not be crystallised135 and thus not be studied by X-ray crystallography, there are also important sample requirements that have to be met. Firstly, the protein complex needs to be stable, enriched and sufficiently purified136. Secondly, the protein complex has to be big enough (>60-100 kDa)137, since bigger mass means more elastic scattering, which in turn means better contrast and better alignability. Alignability also increases the more asymmetric a complex is. It is also easi- er to achieve higher resolution from rigid proteins, e.g. apoferritin, which

19 has been resolved to 1.5Å138. Yet, most biological complexes contain more flexible regions to function well, such as DNA replication139 or mitoriboso- mal140 complexes and are thus difficult to resolve to higher resolutions. An- other important requirement is sufficient complex stability to ensure that individual subunits do not dissociate during grid preparation136,139. In such cases, it is crucial to optimise buffer composition. Buffers cannot contain too high salt and glycerol concentrations, since those additives have similar relative contrast to that of the proteins and would lead to extremely noisy images, which makes particle alignment impossible.

For tomographic studies of organelles and cells the sample also needs to be supplemented with nanogold particles that are required for tilt series align- ment141.

Fig. 6. A. Schematic overview of cryo-EM sample preparation. (A-1) (a) A drop of sample is applied to glow-discharged copper grids overlaid with a foil of perforated carbon. (b) Ex- cess buffer is blotted away by a cellulose filter paper. (c) The resulting thin protein-water film is plunge-frozen into liquid ethane at liquid nitrogen temperature. (A-2) A coherent electron beam in the electron microscope illuminates the frozen-hydrated specimen. By operating the microscope in underfocus, the resulting scattering events produce phase contrast images that can be captured as 2D projections by a direct electron detector. (A-3) Using computational methods, a 3D structure can be calculated from experimental 2D projections. (B) Hypothesis of air-water interface effects during sample preparation. The protein denatures at the air- water interface, thereby forming a layer of denatured protein to which intact proteins bind to in a preferred manner. (C) Usage of a continuous carbon film can shield the specimen from the air-water interface, which yields intact particles. (Fig. 6B and C) adapted from Glaeser, 2018132.

Imaging with the electron microscope Imaging of macromolecular structures with a light microscope is impossible, since the wavelength of visible light (400-700nm) is much bigger than the dimensions of macromolecules (10-20nm). Since the wavelength of electrons is about 1000 times smaller than that of light (0.02 Å at 300kV), they

20 can overcome the resolution limit of the light wavelength and are well-suited to resolve atoms (1Å) of metals142 and structures of biological specimens to near-atomic resolution143. Electrons possess wave-particle duality144. The wave nature of electrons describes the spatial (parallel beam) and temporal coherency (unique wavelength) of an electron beam and accounts for phase interference in image formation. The particle nature allows electro-magnetic lenses145 to change the path of negatively charged electrons in the electron microscope as well as explains the elastic scattering events that are important for image formation (see below).

The imaging system (Fig. 7A) of an electron microscope146 consists of (a) an electron gun that emits electrons, (b) 2-3 electro-magnetic lenses (and aper- tures) that form the condenser lens system, which can change beam size and produces a coherent beam, an (c) objective lens that after interaction of the electron beam with the sample combines the unaltered and elastically scattered electron wave to form an image, (d) an objective aperture that produces amplitude contrast and prevents beam charging, (e) the projector lens system that magnifies the image and (f) the fluorescence screen or detector that displays or records the image.

In addition to the imaging system, modern electron microscopes have the following functionalities: (a) an effective pumping system that produces a good vacuum thereby preventing electrons from scattering air molecules, (b) nitrogen dewars that keep the sample below -150°C and thus preserve the amorphous ice, (c) an autoloader, that can automatically load and unload grids and (d) a compustage that is extremely stable and allows fine position- ing of the sample with respect to the beam as well as stage tilting for alignment and tomography purposes.

Fig. 7B illustrates the interaction of electrons with the specimen146. Because electrons in a transmission electron microscope are extremely fast, most of them do not interact and simply pass through the specimen thereby forming the unaltered wave. Some electrons however are elastically scattered by the nuclei of the specimen that alter their direction and change their phase: this forms the scattered wave. The scattering intensity is proportional to the elec- trostatic potential of the specimen. Interference of scattered and unscattered wave produces a phase contrast that is strongly enhanced by operating the objective lens in underfocus. An objective aperture that removes highly scattered electrons produces additional amplitude contrast. Operating the microscope in underfocus, however, multiplies the image transforms with the contrast transfer function146 (CTF), which introduces frequency dependent contrast reversals in Fourier space (Fig. 7C-4) and image displacements in real space (Fig. 7C 2, 3, 5 and 6). Lower defocus gives a relatively low contrast,

21 yet those images look sharper since they suffer less image displacements and preserve their high frequency information better. Practically, in order to avoid spatial frequencies with zero contrast, one collects images at different defocus values were the specimen is still visible (e.g., between -0.7 and -2.5 µm). The effects of the CTF are corrected for during data processing by applying a Wiener filter147.

The scattering potential of biological specimens is almost similar to that of water and that is why they produce very little relative contrast. Long exposure times are also not an option since inelastic electron scattering with the atom’s electron cloud deposits energy into the sample that gradually de- grades it: a process that is called radiolysis148. Inelastic scattering adds additional noise to the images, but can be removed by imaging filters149, which significantly improve amplitude contrast. Low-dose imaging via specialised automatic acquisition software150 prevents pre-exposure of acquisition areas by using very low electron fluxes for area mapping and by performing auto- focus on nearby carbon areas. In addition, novel image acquisition software uses aberration-free beam-image shift151 that ensures fast throughput while greatly reducing stage movement, thus preserving high frequency components.

The recent introduction of direct electron detectors devices (DDD) that are based on monolithic active pixel sensors152 have revolutionised cryo-EM. On the one hand, DDDs have higher detective quantum efficiencies153 (DQE) than conventional photographic film. On the other hand DDDs enable the collection of movie frames that allow restoration of sample motion as well as radiation damage dependent dose weighting154. Practically, use of higher magnifications improves the DQE for low frequencies components and thus enhances alignability. In addition, the imaging magnification also determines the resolution limit of the obtainable 3D structure, since according to the Nyquist theorem the sampling frequency should be double the frequency of highest frequency component of the underlying signal, which is 2x the pixel size.

Fig. 7. A. Schematic diagram of a transmission electron microscope. Electrons generated in the gun are formed into a coherent beam by the condenser lenses C1 and C2 before interacting with the sample located in the objective lens. Scattered and unscattered wave fronts are combined by the objective lens, forming an image that is subsequently magnified by the projector lens system and finally visualised by a screen or detector. Adapted from Williams and Carter, 1996146. B. Interaction of the electron beam with the specimen. An electron can interact with an isolated atom in three different ways: the fast travelling electrons pass through the specimen (unaltered wave), some electrons are elastically scattered by the nuclei of the atoms (scattered wave), thereby changing their direction and experiencing a phase shift, or the electrons interact with the atom’s electron shell, where they lose energy and pass it to the sample. Adapted from Williams and Carter, 1996146. C. Effects of the CTF on image contrast. Images show simulated particles for different scenarios: C-1: no CTF and no noise applied; C-2: CTF at 1 µm underfocus without noise applied; C-3: CTF at 1 µm underfocus with noise applied; C-5: CTF at 2.5 µm underfocus without noise applied; C-6: CTF at 2.5 µm underfocus with noise applied. C-4 shows the CTF profile for defoci 1.5 and 2.5.

23 Calculation of a 3D structure The electron microscope generates thousands of micrographs containing many 2D projections of the underlying 3D structure. Data processing ex- tracts those experimental projections from the micrographs to calculate a 3D structure. The relationship between 2D projections and 3D structure can be described by the central-slice theorem155 and is depicted in Fig. 8A. The Fourier transform of an individual experimental projection represents a slice in the 3D Fourier transform and interpolating many slices that adopt different views of the underlying structure fills the 3D Fourier space. The real space 3D structure can then be calculated by an inverse Fourier transfor- mation.

The main difficulty for calculating a 3D structure is that the relative orientations of the experimental 2D projections are unknown and that images contain high levels of noise thereby preventing accurate orientation determination. In order to solve this problem, 3D structures are calculated via an iterative expectation-maximisation procedure156. During the expectation step, experimental projections are compared with computer-generated projections of a low-pass filtered initial model (Fig. 8C), where probability distributions over all possible orientations are assigned to each experimental projection. During the maximisation step, the experimental projections are combined with the prior information into a smooth 3D reconstruction, where signal-to- noise ratios for different spatial frequencies are estimated and lower signal- to-noise components are subsequently down-weighted. The generated 3D structure can be improved in a subsequent iteration step (Fig. 8C).

Sample heterogeneity due to insufficient sample preparation may hamper the 3D reconstruction process, since the resulting 2D projections do not repre- sent a unique underlying 3D structure. Therefore the experimental projections need to be sorted into different classes and iteratively aligned against multiple initial models157 (Fig. 8D). Aligning random subsets of the data against a low-pass filtered consensus model generates the needed initial models. Yet, structural heterogeneity may also exist within a complex, for example in ribosomes, where the ratcheting-like motion between the large and small subunit is needed for the process of protein translation. Consensus refinement of those particles aligns the particles with respect to the biggest body, whereas smaller bodies that are displaced in the alignment suffer resolution loss. Initial studies improved the resolution of the different bodies by aligning the 2D projections against reference projections of the individual bodies obtained by masking of the 3D reference158. Those comparisons are however inconsistent since the experimental images still contain information of the surrounding bodies, which decreases image alignment of smaller bodies. Consistent comparison can be achieved by first subtracting the experi-

24 mental projections with the signal of the particle environment and then com- paring the subtracted projections with reference projections of the individual bodies158,159.

In electron cryo-tomography one collects a series of images projections from the same specimen area, while tilting the stage from -60° to +60° (Fig. 8B). The tilt series is then aligned with the aid of nano-gold fiducials and the CTF is estimated and compensated24. Using weighted back-projection160, a 3D volume can be reconstructed from the aligned tilted images (Fig. 8B).

Fig. 8. Principles of data processing adapted from Nogales & Scheres, 2015155. A. The central Fourier slice theorem. B. Principle of electron cryo-tomography: a series of tilted image projections is back-projected to generate a tomogram. C. Iterative refinement of experimental 2D projections against projections of an initial model. D. Principle of 3D classifica- tion to calculate two different structures from the same data set.

26 Aims

The research presented here is divided into three different topics and so are the aims:

1. Mitoribosome biogenesis -The main aim of this research was to understand how mitoribosomes are assembled and which factors are involved during the assembly process. As described in the introduction section, the molecular details of the mitoribosome assembly pathway largely remain elusive and only a few assembly factors are biochemically characterised. Numerous assembly factors are involved in both yeast and human mitochondria suggesting that the mitoribosome assembly pathway in mitochondria of eukaryotic cells is more complex than that of bacteria. The aim was to obtain the structure of the assembly factor MALSU1 with the mitoribosome in order to visualise and structurally characterise the interaction of a mitoribosome assembly intermediate. The second aim was to identify new assembly factor(s) that associate with MALSU1. The third aim of this project was to use yeast as a model organism to study the effect of genetic deletions of the assembly factor Mtg1 and the mi- toribsomal protein uL16m on the mitoribosome assembly pathway. These aims were addressed by firstly preparing and biochemically characterising samples suitable for cryo-EM. Finally, structures of these late–stage mitoribosome assemblies in their native states were determined the cryo-EM and structurally characterised.

2. Respiratory supercomplex- The main aim of this study was to obtain a high-resolution structure of respiratory supercomplexes (III2IV1 and III2IV2) in S. cerevisiae and to visualise the inter- complex (CIII-CIV) interactions in order to understand their organisation in S. cerevisiae. To address these aims cryo-EM sample preparation was optimised and structures of two classes of the yeast supercomplex were determined by cryo-EM. Comparison of these supramolecular structures with mammalian respirasomes provided further insights into their organisation where complex I is absent.

3. Cryo-EM sample preparation method development- The main aim of this study was to investigate how micro crystals, proteins and mitochondria are suitable for a novel pressure-assisted sample preparation method (PREASSIS). To address this aim, comparison to the established vitrobot method were undertaken: (a) in case of micro crystals, the highest resolution of the diffraction spots was com-

27 pared; (b) in case of protein solution, needed concentrations and overall resolution of the resulting structures were compared and (c) in case of mitochondria, gold bead quantity and distribution, membrane integrity, as well as mitochondria location within the hole were analysed.

28 Summary of papers

Paper I - Structures of the human mitochondrial ribosome in native states of assembly Mammalian mitoribosomes synthesise 13 essential subunits of the oxidative phosphorylation machinery. Coordination between the nuclear and mitochondrial genetic systems generates additional complexity for the assembly of mitoribosomes when compared to their bacterial counterpart. While mt- rRNA is encoded by the mitochondrial genome, all 82 mitoribosomal proteins along with numerous assembly factors that drive the assembly of the mitoribosomes are imported from the cytoplasm.

In this study we determined cryo-EM structures of two different assemblies of the mt-LSU in their native state (Fig. 9). Both classes were globally resolved to ~3Å. Comparison of the two structures revealed insights into the timing of rRNA folding and protein incorporation of bL36m during late stages of assembly. Furthermore, we found three novel assembly factors- MALSU1, L0R8F8, mt-ACP that bind to the mt-LSU. This study further establishes the role of L0R8F8 and mt-ACP in mitoribosomal biogenesis. These structures suggest that a module formed by MALSU1, L0R8F8, mt- ACP prevents premature subunit assembly thus redefining the role of ribosome silencing factor (RsfS)/DUF143 family as multi-functional biogenesis factor in human mitochondria.

Fig. 9. Structures of human mitochondrial mt-LSU intermediates in native states of assembly. (A-B). Assembly factors (MALSU1, LOR8F8, mt-ACP) bind to uL14m and 16S rRNA. (PDB: 5OOL, 5OOM)

29 Paper II - Cryo-EM reveals different mitoribosome assembly intermediates in yeast knockout strains of Mtg1 and uL16m Yeast mitoribosomes synthesise 7 essential subunits of the oxidative phosphorylation. Similar to the human mitoribosomes coordination between the two genetic systems generate additional level of complexity for the assembly of mitoribosomes since all the mitoribosomal proteins except Var1 have to be imported into the mitochondria along with various auxiliary factors required for mitoribosome assembly. In this study, we analysed the effect of genetic deletion of assembly factor Mtg1 and the mitoribosomal protein uL16m on the mitoribosome assembly pathway. For the knockout strain of Mtg1, cryo-EM analysis showed accumulation of mitoribosome assembly intermediates at various stages, likely representing the late-stages of mitoribosome assembly (Fig. 10). Comparison of these intermediates provide a deeper understanding of incorporation of several mitoribosomal proteins, rRNA folding events and effects on the maturation of ribosomes and hence translation. A marked reduction of mature ribosomes was observed in the Δmtg1 cell line. We also discovered extra density in our map, indicating the binding of factor(s) at the mRNA channel exit in these mitoribosomes.

These results demonstrate interdependence of uL16m and uL6m on the Mtg1 assembly factor that is crucial for the incorporation of these proteins during assembly. Moreover, in these structures aberration in the assembly pathway could be observed upon genetic deletion of Mtg1 and uL16m, further corroborating their role in mitoribosome biogenesis. The structures reported in this study signify the role of GTPases in mitoribosome maturation, incorporation of MRPs and proper rRNA folding events during assembly in eukaryotes.

Fig. 10. Accumulated mitoribosome assembly intermediates upon Mtg1 deletion. A. Mature mitoribsome B. subassembly-1 C. subassembly-2.

30 Paper III - Cryo-EM structures of the yeast respiratory supercomplex In cells, mitochondria are the site for oxidative phosphorylation, which is the prime source for ATP generation. S. cerevisiae lacks the conventional complex I which is replaced by three non-proton translocating enzymes: Ndi1, Nde1 and Nde2 that oxidise NADH. Similar to respirasomes in mammals, which are formed by complex I-IV, in yeast complex III and complex IV associate to form supramolecular assemblies called supercomplexes.

In this study we solved the first high-resolution structures of the yeast respiratory supercomplex; III2IV1 and III2IV2 to ~3.3 Å and 3.5Å respectively. This allowed us to build a model of the III2IV1 supercomplex assembly (Fig. 11). We were able to identify two novel subunits: Qcr 10 in complex III region and an unidentified helix in the complex IV region. We could further visualise protein-protein interactions formed by Cor1 and Cox5a subunits and protein-lipid interactions that hold together the complexes in the inner mitochondrial membrane.

Comparison of the organisation of mammalian respirasomes (I1III2IV) with yeast supercomplexes (III2IV1) revealed that yeast has a shorter cytochrome c diffusion distance likely reflecting that organisms need to be able to adapt efficiently in different growth conditions. The structural data will further enable us to carry out biochemical studies that address the formation and action of supercomplexes.

Fig. 11. Structure of the yeast supercomplex (III2IV1). Complex III (yellow) is a homodimer composed of 10 subunits each and complex IV (red) contains 12 subunits. Phospholipids are shown in green. (PDBs: 6YMX, 6YMY)

31 Paper IV - A simple pressure-assisted method for cryo- EM specimen preparation Upon successful purification of protein complexes, many parameters for sample freezing, such as grid type, glow discharge conditions, sample concentration and blotting time have to be optimised. In some studies, therefore, cryo-EM sample preparation is a major bottleneck, since micrographs with low particle number, dissociated protein complexes or particles adopting preferred orientations cannot be considered for further data analysis. In order to produce samples more reliable, blot-free methods have been developed that are an alternative to the conventional vitrobot blotting method.

In this study we developed a simple pressure-assisted sample preparation method (PREASSIS) that can be used to freeze micro-ED crystals, protein complexes for single particle analysis (SPA) and organelles for tomographic studies (Fig. 12A). For microED, our PREASSIS results show more effective removal of a highly viscous mother liquid from SaR2lox crystals resulting in great resolution improvements. For SPA studies, PREASSIS greatly reduced required sample concentrations of 10 to 20 times. For tomographic studies of mitochondria, we showed that PREASSIS yields much higher numbers of gold nanoparticles when compared to vitrobot blotting and that mitochondria do not touch the edge of the carbon holes: both conditions are necessary for accurate tilt series alignment and defocus estimation. Tilt- series acquisition, subsequent tilt series alignment and tomogram generation showed the intactness of mitochondrial membranes (Fig. 12B).

Fig. 12. PREASSIS set up B for cryo-EM specimen preparation. A. PREASSIS setup B. A drop of sample is applied onto a glow-discharged grid and the remaining liquid is directly removed by suction. B. Tomogram of a yeast mitochondrion; grids were prepared by using Preassis set up. Outer mitochondrial membrane (OM), inner mitochondrial membrane (IMM) and cristae are clearly visible, suggesting the intactness of the mitochondrion.

32 Conclusion and future perspectives

The research reported in this thesis presents cryo-EM as an indispensable tool to determine the structures of flexible macromolecular mitochondrial complexes that allowed me to investigate mitoribosomes and respiratory supercomplexes in their near native states, detect novel proteins and understand their functions.

Our understanding of the human mitoribosome assembly pathway and its molecular mechanisms still remain elusive and largely unexplored, despite its high significance to medical research and related pathologies. The main challenge in this field is to prepare samples suitable for cryo-EM studies. Among the causes that impede structural characterization of mitoribosome assembly intermediates are their fast turnover, low abundance, high heterogeneity and fragility. However, during the last seven years, a string of tech- nological breakthroughs has turned single particle cryo-EM into a rapid and precise method for high-resolution structure determination for low abundant and heterogeneous samples, providing a detailed description of previously intractable molecular pathways.

My studies described in this thesis could show how various assembly factors interact with mitochondrial ribosomes and I could capture different stages of mitoribosome assembly. The studies further establish the role of assembly factors in facilitating mitoribosome subunit maturation, e.g., by preventing premature subunit joining and possibly connecting other pathways in mitochondria or controlling OXPHOS activity. Thus revealing the first snapshots into mitochondrial ribosome assembly in human and yeast. However, numerous assembly factors are known to regulate the assembly pathway of mitoribosomes for which no structural data is available. Therefore, I propose to further employ the recent advances in cryo-EM or cryo-ET to provide fundamental insights into the mitochondrial mitoribosome assembly pathway.

Furthermore, I have obtained high-resolution structures of yeast supercomplexes, determined their architecture and identified novel subunits. For future studies, I suggest studying how respiratory supercomplex factors (RCFs) interact and regulate supercomplexes in both yeast and mammals. These further structural studies will provide answers to existing questions that currently prevail in this research field regarding the role of RCFs in regulating cytochrome c oxidase activity.

33 Moreover, future developments of cryo-ET will enable us to visualise structures of macromolecular complexes in situ and to obtain structures through sub-tomogram averaging. Therefore, techniques that improve cryo-EM sample preparation will open up new horizons in this field.

34 Sammanfattning på svenska

Vi äter för att leva. Ett av huvudskälen till att vi äter är att vi behöver energi för att kunna leva. Hur gör då kroppen för att omvandla föda till energi? Mat består av fetter, kolhydrater och protein, som i våra kroppar, med hjälp av enzymer, bryts ned och omvandlas till fettsyror och enkla sockerarter såsom glukos och aminosyror. Kroppens olika celler använder näringsämnen för att på kemisk väg framställa en mängd olika proteiner, livets byggstenar. För att energi skall kunna utvinnas ur föda behövs mitokondrierna, speciella enheter inuti kroppens celler. Mitokondrierna kallas ofta för cellens kraftstation eftersom de genererar energi lagrad i form av ATP. Det är tack vare denna energigenerering som nödvändiga processer i cellerna fungerar och liv kan upprätthållas.

De flesta proteiner bildar stora komplex som, likt maskiner tillverkade av människan, har en mängd olika funktioner. Till exempel bygger de upp och lagar olika typer av vävnader. Hur sker då tillverkningen av proteiner i kroppen? Proteiner tillverkas av makromolekylära maskiner, så kallade ribosomer. Mitokondrierna har egna specialiserade ribosomer, ”mitoribosom- er”, som behövs för att bilda proteiner till fyra proteinkomplex nödvändiga för andningskedjan. Energin från maten vi äter och syret från luften vi andas används av dessa proteinkomplex för att pumpa protoner över membranet. Den protongradient som då uppstår används sedan av ett annat komplex som genom en turbinliknande rörelse genererar ATP. Intressant nog skapar and- ningskomplexen större strukturer, så kallade ”superkomplex”, som fysiskt interagerar med varandra och därigenom sannolikt förbättrar deras kapacitet att oxidera näringsämnen.

Har du någon gång undrat hur dessa osynliga makromolekylära maskiner, mitokondrierna, faktiskt ser ut? Är det ens möjligt att visualisera dessa maskiner som består av mängder av proteiner och genetiskt material? Svaret är ja. Jag har i denna doktorsavhandling använt mig av en strukturbiologisk teknik som har belönats med Nobelpriset: ”kryoelektronmikroskopi”. Tekni- ken har gjort det möjligt att på bild fånga de beskrivna proteinkomplexen uppförstorade tusentals gånger. Med hjälp av beräkningsverktyg kunde jag ta fram tredimensionella bilder av mitoribosomernas strukturella arkitektur i deras naturliga ihopsatta tillstånd och också av superkomplex sammansatta av flera andningskedjor. På så sätt kunde jag förstå funktionen hos dessa synnerligen dynamiska makromolekyler.

Trots att det finns omfattande information tillgänglig om mitoribosomernas roll i proteinsyntesen, är mycket lite känt om hur de blir till inne i männis-

35 kans mitokondrier. För att lära oss mer om denna process tog jag tillsmam- mans med kollegor fram tredimensionella strukturer av mänskliga mitokon- drieprover. Genom att använda kryo-EM-metoden på dessa mitokondrier upptäckte vi förekomsten av nya faktorer som är inblandade i att kontrollera de olika stadierna i mitoribosomernas ihopmontering och följaktligen också av själva proteinsyntesen. Jag använde också kryo-EM-metoden på jäst för att förstå vad som händer med mitoribosomernas monteringsväg om monter- ingsfaktorerna förstörs genetiskt. Dessa studier är särskilt viktiga eftersom fel i mitokondriell translation leder till mitokondriella sjukdomar som i första hand påverkar vävnad med behov av hög metabolism, t.ex. nervsys- temet, muskler, hjärta, syn och det endokrina systemet. Dessutom fastställde jag tillsammans med kollegor, strukturerna hos superkomplex av andningskedjor i bakjäst som gjorde det möjligt för oss att förstå hur de är ordnade i det inre membranet hos mitokondrier. I dessa strukturer kunde vi identifiera nya proteiner och lipider som behövs för att stabilisera superkomplexen. Slutligen, bidrog jag till att implementera en ny metod för preparering av prover för kryo-EM på en rad olika biologiska prover, som t.ex mitokondrierna.

Resultaten i mina doktorandstudier utgör en grund för framtida forskning.

36 Popular science summary in English

We eat to live. One of the main reason we eat food is to obtain energy to sustain life. So how do we get energy from food? Food, which is a complex mixture of fats, carbohydrates and proteins is broken down by our body by many enzymes into fatty acids, simple sugars, such as glucose and amino acids, respectively. Different cells in our body utilise these nutrients to synthesise several proteins, commonly known as the building blocks of life. To extract the energy from food, our body contains a special compartment inside our cells, the mitochondrion. It is also called powerhouse of the cell since it generates energy in the form of ATP. This energy drives essential cellular processes in our body and keeps us alive.

Most proteins form big complexes that, like machines invented by humans, perform a variety functions such as building and repairing various tissues. But how are these proteins made inside the body? These proteins are made by macromolecular machines called ribosomes. But mitochondria also contain their own specialised ribosomes called “mitoribosomes” that are needed to make proteins for four essential respiratory protein complexes, which use the energy extracted from food and oxygen from the air we breath to pump protons across a membrane. The generated proton gradient is then utilised by another complex that uses a turbine-like motion to generate ATP. Interest- ingly the respiratory complexes form bigger structures called “supercomplexes” by physically interacting with each other that likely improves their efficiency of oxidizing nutrients.

But have you ever wondered how these invisible macromolecular machines of the mitochondria look like? And whether it is possible to visualise these machines that are made up of several proteins and genetic material. The answer to these questions is yes. In the studies performed in this thesis I used a Nobel Prize winning structural biology technique called “electron cryo- microscopy” that made it possible to capture photographs of the described protein complexes with magnifications of many thousand times. Computa- tional tools allowed me to determine the 3D architectures of mitoribosomes in native states of assembly and respiratory supercomplexes to understand the function of these extremely dynamic macromolecules.

Despite vast information available about the function of mitoribsomes in synthesizing proteins, very little is known about how these mitoribosomes are formed inside human mitochondria. To get insights into this process I together with colleagues obtained 3D structures of samples prepared from human mitochondria. Using the cryo-EM approach we discovered presence

37 of novel factors associated with human mitoribosomes that are involved in controlling different stages of mitoribosome assembly and hence protein synthesis. I performed cryo-EM studies to understand the effect that genetic deletion of assembly factors has on the mitoribosome assembly pathway in yeast. These studies are particularly crucial, because defects in mitochondrial translation result in mitochondrial diseases that primarily affect tissues with a high metabolic demand, such as the nervous, muscular, cardiac, ocular and endocrine systems. In addition, I together with colleagues determined the structures of respiratory supercomplexes from bakers’ yeast that enabled us to understand their organisation in the inner mitochondrial membrane. In those structures, we identified new proteins and lipid molecules that are required to stabilise supercomplexes. And finally, I contributed to the successful implementation of a new sample preparation method for cryoEM on a range of biological specimens such as mitochondria.

The findings in my PhD study lay a foundation for future scientific research.

38 Acknowledgement

I would like to express my gratitude to my supervisor, Martin Ott, for providing me a nice lab environment to continue my PhD studies. I am grateful to Stefan Nordlund, Pia Ädelroth, Dan Daley and Martin Högbom for their support.

My sincere thanks goes to Michal Minzcuk lab at MRC-Mitochondrial Biol- ogy Unit, Cambridge, where I got an opportunity to carry out important biochemical studies during my EMBO fellowship period. I also thank all the wonderful researchers I met and worked with at the MBU and LMB.

I would like to extend my thanks to Venki Ramakrishnan lab, Antonio Bar- rientos, Peter Brzezinski, Hongyi Xu from Xiaodong Zou lab with whom I collaborated during the course of my PhD studies. I am truly thankful to the Swedish cryo-EM facility staff for cryo-EM training, IT and software support. I sincerely thank all my fellow labmates at DBB and Scilifelab for eve- ry experience shared, helpful discussions, good times and lab fikas.

I am truly grateful to Sian James for illustrating the cover page of this thesis and bringing out the research theme so nicely. A special thanks to Karin Häggbom Sandberg for her kind words, motivation and for doing the Swe- dish translation. I also thank Maria Sallander for helping me with all PhD program related doubts and finances.

I would like to especially acknowledge my dear friends and former colleagues at Daiichi Sankyo: Deepika and Smita.

Finally, I would like to thank my family: mum, papa, didi for their support, love and blessings. Mum, I would especially like to thank you for the spir- itual guidance, wise words and your countless blessings that motivated me to keep going during hard times. I cannot thank you enough.

I am also deeply grateful to my mum-in-law, dad-in-law, Marc and Oma for all wishes and encouragement.

Last but not the least, I would like to thank my best friend and husband, Julian for standing by me at all times and for your countless prayers. Thank you for all the scientific discussions, cryo-EM mentoring and collaborations. It is really a delight to work with you.

39 References

1. Jang, S. & Javadov, S. Current Challenges in Elucidating Respiratory Supercomplexes in Mitochondria: Methodological Obstacles. Front Physiol 9, 238 (2018). 2. Milenkovic, D., Blaza, J.N., Larsson, N.G. & Hirst, J. The Enigma of the Respiratory Chain Supercomplex. Cell Metab 25, 765-776 (2017). 3. Schagger, H. & Pfeiffer, K. Supercomplexes in the respiratory chains of yeast and mammalian mitochondria. EMBO J 19, 1777-83 (2000). 4. Eubel, H., Jansch, L. & Braun, H.P. New insights into the respiratory chain of plant mitochondria. Supercomplexes and a unique composition of complex II. Plant Physiol 133, 274-86 (2003). 5. Davies, K.M., Blum, T.B. & Kuhlbrandt, W. Conserved in situ arrangement of complex I and III2 in mitochondrial respiratory chain supercomplexes of mammals, yeast, and plants. Proc Natl Acad Sci U S A 115, 3024-3029 (2018). 6. van der Sluis, E.O. et al. Parallel Structural Evolution of Mitochondrial Ribosomes and OXPHOS Complexes. Genome Biol Evol 7, 1235-51 (2015). 7. Ott, M., Amunts, A. & Brown, A. Organization and Regulation of Mitochondrial Protein Synthesis. Annu Rev Biochem 85, 77-101 (2016). 8. Englmeier, R., Pfeffer, S. & Forster, F. Structure of the Human Mitochondrial Ribosome Studied In Situ by Cryoelectron Tomography. Structure 25, 1574-1581 e2 (2017). 9. Pfeffer, S., Woellhaf, M.W., Herrmann, J.M. & Forster, F. Organization of the mitochondrial translation machinery studied in situ by cryoelectron tomography. Nat Commun 6, 6019 (2015). 10. Desai, N., Brown, A., Amunts, A. & Ramakrishnan, V. The structure of the yeast mitochondrial ribosome. Science 355, 528-531 (2017). 11. Amunts, A., Brown, A., Toots, J., Scheres, S.H.W. & Ramakrishnan, V. Ribosome. The structure of the human mitochondrial ribosome. Science 348, 95-98 (2015). 12. Greber, B.J. & Ban, N. Structure and Function of the Mitochondrial Ribosome. Annu Rev Biochem 85, 103-32 (2016). 13. Brown, A. et al. Structure of the large ribosomal subunit from human mitochondria. Science 346, 718-722 (2014). 14. Amunts, A. et al. Structure of the yeast mitochondrial large ribosomal subunit. Science 343, 1485-1489 (2014).

40 15. Yusupov, M.M. et al. Crystal structure of the ribosome at 5.5 A resolution. Science 292, 883-96 (2001). 16. Smirnov, A., Entelis, N., Martin, R.P. & Tarassov, I. Biological significance of 5S rRNA import into human mitochondria: role of ribosomal protein MRP-L18. Genes Dev 25, 1289-305 (2011). 17. Greber, B.J. et al. Ribosome. The complete structure of the 55S mammalian mitochondrial ribosome. Science 348, 303-8 (2015). 18. Arenz, S. et al. Structures of the orthosomycin antibiotics avilamycin and evernimicin in complex with the bacterial 70S ribosome. Proc Natl Acad Sci U S A 113, 7527-32 (2016). 19. Chandrasekaran, V. et al. Mechanism of ribosome stalling during translation of a poly(A) tail. Nat Struct Mol Biol 26, 1132-1140 (2019). 20. Herrmann, J.M., Woellhaf, M.W. & Bonnefoy, N. Control of protein synthesis in yeast mitochondria: the concept of translational activators. Biochim Biophys Acta 1833, 286-94 (2013). 21. Montoya, J., Ojala, D. & Attardi, G. Distinctive features of the 5'- terminal sequences of the human mitochondrial mRNAs. Nature 290, 465-70 (1981). 22. Derbikova, K.S., Levitsky, S.A., Chicherin, I.V., Vinogradova, E.N. & Kamenski, P.A. Activation of Yeast Mitochondrial Translation: Who Is in Charge? Biochemistry (Mosc) 83, 87-97 (2018). 23. Weraarpachai, W. et al. Mutation in TACO1, encoding a translational activator of COX I, results in cytochrome c oxidase deficiency and late-onset Leigh syndrome. Nat Genet 41, 833-7 (2009). 24. Galaz-Montoya, J.G. et al. Alignment algorithms and per-particle CTF correction for single particle cryo-electron tomography. J Struct Biol 194, 383-94 (2016). 25. Marykwas, D.L. & Fox, T.D. Control of the Saccharomyces cerevisiae regulatory gene PET494: transcriptional repression by glucose and translational induction by oxygen. Mol Cell Biol 9, 484- 91 (1989). 26. Costanzo, M.C. & Fox, T.D. Control of mitochondrial gene expression in Saccharomyces cerevisiae. Annu Rev Genet 24, 91- 113 (1990). 27. McMullin, T.W., Haffter, P. & Fox, T.D. A novel small-subunit ribosomal protein of yeast mitochondria that interacts functionally with an mRNA-specific translational activator. Mol Cell Biol 10, 4590-5 (1990). 28. Jones, J.L. et al. Yeast mitochondrial protein Pet111p binds directly to two distinct targets in COX2 mRNA, suggesting a mechanism of translational activation. J Biol Chem 294, 7528-7536 (2019).

41 29. Mayorga, J.P. et al. A Novel Function of Pet54 in Regulation of Cox1 Synthesis in Saccharomyces cerevisiae Mitochondria. J Biol Chem 291, 9343-55 (2016). 30. Yona, A.H. et al. tRNA genes rapidly change in evolution to meet novel translational demands. Elife 2, e01339 (2013). 31. Hallberg, B.M. & Larsson, N.G. Making proteins in the powerhouse. Cell Metab 20, 226-40 (2014). 32. Bauerschmitt, H. et al. Ribosome-binding proteins Mdm38 and Mba1 display overlapping functions for regulation of mitochondrial translation. Mol Biol Cell 21, 1937-44 (2010). 33. Thompson, K. et al. OXA1L mutations cause mitochondrial encephalopathy and a combined oxidative phosphorylation defect. EMBO Mol Med 10(2018). 34. Nikolay, R. et al. Structural Visualization of the Formation and Activation of the 50S Ribosomal Subunit during In Vitro Reconstitution. Mol Cell 70, 881-893 e3 (2018). 35. Lopez-Alonso, J.P. et al. RsgA couples the maturation state of the 30S ribosomal decoding center to activation of its GTPase pocket. Nucleic Acids Res 45, 6945-6959 (2017). 36. Jomaa, A. et al. Cryo-electron microscopy structure of the 30S subunit in complex with the YjeQ biogenesis factor. RNA 17, 2026- 38 (2011). 37. Held, W.A., Mizushima, S. & Nomura, M. Reconstitution of Escherichia coli 30 S ribosomal subunits from purified molecular components. J Biol Chem 248, 5720-30 (1973). 38. De Silva, D., Tu, Y.T., Amunts, A., Fontanesi, F. & Barrientos, A. Mitochondrial ribosome assembly in health and disease. Cell Cycle 14, 2226-50 (2015). 39. Zeng, R., Smith, E. & Barrientos, A. Yeast Mitoribosome Large Subunit Assembly Proceeds by Hierarchical Incorporation of Protein Clusters and Modules on the Inner Membrane. Cell Metab 27, 645- 656 e7 (2018). 40. Kummer, E. et al. Unique features of mammalian mitochondrial translation initiation revealed by cryo-EM. Nature 560, 263-267 (2018). 41. D'Souza, A.R. & Minczuk, M. Mitochondrial transcription and translation: overview. Essays Biochem 62, 309-320 (2018). 42. Bogenhagen, D.F., Martin, D.W. & Koller, A. Initial steps in RNA processing and ribosome assembly occur at mitochondrial DNA nucleoids. Cell Metab 19, 618-29 (2014). 43. Rorbach, J. et al. MRM2 and MRM3 are involved in biogenesis of the large subunit of the mitochondrial ribosome. Mol Biol Cell 25, 2542-55 (2014).

42 44. Rorbach, J., Gammage, P.A. & Minczuk, M. C7orf30 is necessary for biogenesis of the large subunit of the mitochondrial ribosome. Nucleic Acids Res 40, 4097-109 (2012). 45. Spahr, H., Habermann, B., Gustafsson, C.M., Larsson, N.G. & Hallberg, B.M. Structure of the human MTERF4-NSUN4 protein complex that regulates mitochondrial ribosome biogenesis. Proc Natl Acad Sci U S A 109, 15253-8 (2012). 46. Wredenberg, A. et al. MTERF3 regulates mitochondrial ribosome biogenesis in invertebrates and mammals. PLoS Genet 9, e1003178 (2013). 47. Antonicka, H. & Shoubridge, E.A. Mitochondrial RNA Granules Are Centers for Posttranscriptional RNA Processing and Ribosome Biogenesis. Cell Rep 10, 920-932 (2015). 48. Maiti, P., Kim, H.J., Tu, Y.T. & Barrientos, A. Human GTPBP10 is required for mitoribosome maturation. Nucleic Acids Res 46, 11423- 11437 (2018). 49. He, J. et al. Human C4orf14 interacts with the mitochondrial nucleoid and is involved in the biogenesis of the small mitochondrial ribosomal subunit. Nucleic Acids Res 40, 6097-108 (2012). 50. Dennerlein, S., Rozanska, A., Wydro, M., Chrzanowska- Lightowlers, Z.M. & Lightowlers, R.N. Human ERAL1 is a mitochondrial RNA chaperone involved in the assembly of the 28S small mitochondrial ribosomal subunit. Biochem J 430, 551-8 (2010). 51. Uchiumi, T. et al. ERAL1 is associated with mitochondrial ribosome and elimination of ERAL1 leads to mitochondrial dysfunction and growth retardation. Nucleic Acids Res 38, 5554-68 (2010). 52. Arroyo, J.D. et al. A Genome-wide CRISPR Death Screen Identifies Genes Essential for Oxidative Phosphorylation. Cell Metab 24, 875- 885 (2016). 53. Kaur, J. & Stuart, R.A. Truncation of the Mrp20 protein reveals new ribosome-assembly subcomplex in mitochondria. EMBO Rep 12, 950-5 (2011). 54. De Silva, D., Fontanesi, F. & Barrientos, A. The DEAD box protein Mrh4 functions in the assembly of the mitochondrial large ribosomal subunit. Cell Metab 18, 712-25 (2013). 55. Kim, H.J. & Barrientos, A. MTG1 couples mitoribosome large subunit assembly with intersubunit bridge formation. Nucleic Acids Res 46, 8435-8453 (2018). 56. Shoji, S., Dambacher, C.M., Shajani, Z., Williamson, J.R. & Schultz, P.G. Systematic chromosomal deletion of bacterial ribosomal protein genes. J Mol Biol 413, 751-61 (2011). 57. Greber, B.J. et al. Insertion of the Biogenesis Factor Rei1 Probes the Ribosomal Tunnel during 60S Maturation. Cell 164, 91-102 (2016).

43 58. Sosnowski, P. et al. The CryoEM structure of the Saccharomyces cerevisiae ribosome maturation factor Rea1. Elife 7(2018). 59. Kater, L. et al. Visualizing the Assembly Pathway of Nucleolar Pre- 60S Ribosomes. Cell 171, 1599-1610 e14 (2017). 60. Sanghai, Z.A. et al. Modular assembly of the nucleolar pre-60S ribosomal subunit. Nature 556, 126-129 (2018). 61. Goto, S., Muto, A. & Himeno, H. GTPases involved in bacterial ribosome maturation. J Biochem 153, 403-14 (2013). 62. Karbstein, K. Role of GTPases in ribosome assembly. Biopolymers 87, 1-11 (2007). 63. Linder, P. & Jankowsky, E. From unwinding to clamping - the DEAD box RNA helicase family. Nat Rev Mol Cell Biol 12, 505-16 (2011). 64. Tu, Y.T. & Barrientos, A. The Human Mitochondrial DEAD-Box Protein DDX28 Resides in RNA Granules and Functions in Mitoribosome Assembly. Cell Rep 10, 854-864 (2015). 65. Lee, K.W. & Bogenhagen, D.F. Assignment of 2'-O- methyltransferases to modification sites on the mammalian mitochondrial large subunit 16 S ribosomal RNA (rRNA). J Biol Chem 289, 24936-42 (2014). 66. Pintard, L., Bujnicki, J.M., Lapeyre, B. & Bonnerot, C. MRM2 encodes a novel yeast mitochondrial 21S rRNA methyltransferase. EMBO J 21, 1139-47 (2002). 67. Fung, S., Nishimura, T., Sasarman, F. & Shoubridge, E.A. The conserved interaction of C7orf30 with MRPL14 promotes biogenesis of the mitochondrial large ribosomal subunit and mitochondrial translation. Mol Biol Cell 24, 184-93 (2013). 68. Hauser, R. et al. RsfA (YbeB) proteins are conserved ribosomal silencing factors. PLoS Genet 8, e1002815 (2012). 69. Roberti, M. et al. MTERF factors: a multifunction protein family. Biomol Concepts 1, 215-24 (2010). 70. Rugarli, E.I. & Langer, T. Translating m-AAA protease function in mitochondria to hereditary spastic paraplegia. Trends Mol Med 12, 262-9 (2006). 71. Shutt, T.E. & Shadel, G.S. A compendium of human mitochondrial gene expression machinery with links to disease. Environ Mol Mutagen 51, 360-79 (2010). 72. Pagliarini, D.J. et al. A mitochondrial protein compendium elucidates complex I disease biology. Cell 134, 112-23 (2008). 73. Galmiche, L. et al. Exome sequencing identifies MRPL3 mutation in mitochondrial cardiomyopathy. Hum Mutat 32, 1225-31 (2011). 74. Seidel-Rogol, B.L., McCulloch, V. & Shadel, G.S. Human mitochondrial transcription factor B1 methylates ribosomal RNA at a conserved stem-loop. Nat Genet 33, 23-4 (2003).

44 75. Carroll, C.J. et al. Whole-exome sequencing identifies a mutation in the mitochondrial ribosome protein MRPL44 to underlie mitochondrial infantile cardiomyopathy. J Med Genet 50, 151-9 (2013). 76. Smits, P. et al. Mutation in mitochondrial ribosomal protein MRPS22 leads to Cornelia de Lange-like phenotype, brain abnormalities and hypertrophic cardiomyopathy. Eur J Hum Genet 19, 394-9 (2011). 77. Miller, C. et al. Defective mitochondrial translation caused by a ribosomal protein (MRPS16) mutation. Ann Neurol 56, 734-8 (2004). 78. Menezes, M.J. et al. Mutation in mitochondrial ribosomal protein S7 (MRPS7) causes congenital sensorineural deafness, progressive hepatic and renal failure and lactic acidemia. Hum Mol Genet 24, 2297-307 (2015). 79. Kim, H.J., Maiti, P. & Barrientos, A. Mitochondrial ribosomes in cancer. Semin Cancer Biol 47, 67-81 (2017). 80. Park, C.B. et al. MTERF3 is a negative regulator of mammalian mtDNA transcription. Cell 130, 273-85 (2007). 81. Rotig, A. Human diseases with impaired mitochondrial protein synthesis. Biochim Biophys Acta 1807, 1198-205 (2011). 82. Skrtic, M. et al. Inhibition of mitochondrial translation as a therapeutic strategy for human acute myeloid leukemia. Cancer Cell 20, 674-88 (2011). 83. Viale, A. et al. Oncogene ablation-resistant pancreatic cancer cells depend on mitochondrial function. Nature 514, 628-32 (2014). 84. Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646-74 (2011). 85. Lasserre, J.P. et al. Yeast as a system for modeling mitochondrial disease mechanisms and discovering therapies. Dis Model Mech 8, 509-26 (2015). 86. Barrientos, A. Yeast models of human mitochondrial diseases. IUBMB Life 55, 83-95 (2003). 87. Horvath, S.E. & Daum, G. Lipids of mitochondria. Prog Lipid Res 52, 590-614 (2013). 88. Lill, R. et al. The role of mitochondria in cellular iron-sulfur protein biogenesis and iron metabolism. Biochim Biophys Acta 1823, 1491- 508 (2012). 89. Agip, A.A., Blaza, J.N., Fedor, J.G. & Hirst, J. Mammalian Respiratory Complex I Through the Lens of Cryo-EM. Annu Rev Biophys 48, 165-184 (2019). 90. Fiedorczuk, K. & Sazanov, L.A. Mammalian Mitochondrial Complex I Structure and Disease-Causing Mutations. Trends Cell Biol 28, 835-867 (2018).

45 91. Zhu, J., Vinothkumar, K.R. & Hirst, J. Structure of mammalian respiratory complex I. Nature 536, 354-358 (2016). 92. Hirst, J. Mitochondrial complex I. Annu Rev Biochem 82, 551-75 (2013). 93. Matus-Ortega, M.G. et al. New complexes containing the internal alternative NADH dehydrogenase (Ndi1) in mitochondria of Saccharomyces cerevisiae. Yeast 32, 629-41 (2015). 94. Sun, F. et al. Crystal structure of mitochondrial respiratory membrane protein complex II. Cell 121, 1043-57 (2005). 95. Bezawork-Geleta, A., Rohlena, J., Dong, L., Pacak, K. & Neuzil, J. Mitochondrial Complex II: At the Crossroads. Trends Biochem Sci 42, 312-325 (2017). 96. Letts, J.A. & Sazanov, L.A. Clarifying the supercomplex: the higher-order organization of the mitochondrial electron transport chain. Nat Struct Mol Biol 24, 800-808 (2017). 97. Lemire, B.D. & Oyedotun, K.S. The Saccharomyces cerevisiae mitochondrial succinate:ubiquinone oxidoreductase. Biochim Biophys Acta 1553, 102-16 (2002). 98. Solmaz, S.R. & Hunte, C. Structure of complex III with bound cytochrome c in reduced state and definition of a minimal core interface for electron transfer. J Biol Chem 283, 17542-9 (2008). 99. Rathore, S. et al. Cryo-EM structure of the yeast respiratory supercomplex. Nat Struct Mol Biol 26, 50-57 (2019). 100. Hartley, A.M. et al. Structure of yeast cytochrome c oxidase in a supercomplex with cytochrome bc1. Nat Struct Mol Biol 26, 78-83 (2019). 101. Yano, N. et al. X-ray structure of cyanide-bound bovine heart cytochrome c oxidase in the fully oxidized state at 2.0 A resolution. Acta Crystallogr F Struct Biol Commun 71, 726-30 (2015). 102. Tsukihara, T. et al. The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. Science 272, 1136-44 (1996). 103. Marechal, A., Meunier, B., Lee, D., Orengo, C. & Rich, P.R. Yeast cytochrome c oxidase: a model system to study mitochondrial forms of the haem-copper oxidase superfamily. Biochim Biophys Acta 1817, 620-8 (2012). 104. Wikstrom, M., Sharma, V., Kaila, V.R., Hosler, J.P. & Hummer, G. New perspectives on proton pumping in cellular respiration. Chem Rev 115, 2196-221 (2015). 105. Guo, R., Zong, S., Wu, M., Gu, J. & Yang, M. Architecture of Human Mitochondrial Respiratory Megacomplex I2III2IV2. Cell 170, 1247-1257 e12 (2017). 106. Zong, S. et al. Structure of the intact 14-subunit human cytochrome c oxidase. Cell Res 28, 1026-1034 (2018). 107. Mirkin, N., Jaconcic, J., Stojanoff, V. & Moreno, A. High resolution X-ray crystallographic structure of bovine heart cytochrome c and

46 its application to the design of an electron transfer biosensor. Proteins 70, 83-92 (2008). 108. Zhou, A. et al. Structure and conformational states of the bovine mitochondrial ATP synthase by cryo-EM. Elife 4, e10180 (2015). 109. Sousa, J.S., Mills, D.J., Vonck, J. & Kuhlbrandt, W. Functional asymmetry and electron flow in the bovine respirasome. Elife 5(2016). 110. Gu, J. et al. The architecture of the mammalian respirasome. Nature 537, 639-43 (2016). 111. Letts, J.A., Fiedorczuk, K. & Sazanov, L.A. The architecture of respiratory supercomplexes. Nature 537, 644-648 (2016). 112. Dudkina, N.V., Kudryashev, M., Stahlberg, H. & Boekema, E.J. Interaction of complexes I, III, and IV within the bovine respirasome by single particle cryoelectron tomography. Proc Natl Acad Sci U S A 108, 15196-200 (2011). 113. Mileykovskaya, E. et al. Arrangement of the respiratory chain complexes in Saccharomyces cerevisiae supercomplex III2IV2 revealed by single particle cryo-electron microscopy. J Biol Chem 287, 23095-103 (2012). 114. Mourier, A., Matic, S., Ruzzenente, B., Larsson, N.G. & Milenkovic, D. The respiratory chain supercomplex organization is independent of COX7a2l isoforms. Cell Metab 20, 1069-75 (2014). 115. Lapuente-Brun, E. et al. Supercomplex assembly determines electron flux in the mitochondrial electron transport chain. Science 340, 1567-70 (2013). 116. Letts, J.A., Fiedorczuk, K., Degliesposti, G., Skehel, M. & Sazanov, L.A. Structures of Respiratory Supercomplex I+III2 Reveal Functional and Conformational Crosstalk. Mol Cell 75, 1131-1146 e6 (2019). 117. Wiseman, B. et al. Structure of a functional obligate complex III2IV2 respiratory supercomplex from Mycobacterium smegmatis. Nat Struct Mol Biol 25, 1128-1136 (2018). 118. Fleischer, S., Klouwen, H. & Brierley, G. Studies of the electron transfer system. 38. Lipid composition of purified enzyme preparations derived from beef heart mitochondria. J Biol Chem 236, 2936-41 (1961). 119. Mileykovskaya, E. & Dowhan, W. Cardiolipin-dependent formation of mitochondrial respiratory supercomplexes. Chem Phys Lipids 179, 42-8 (2014). 120. Zhang, M., Mileykovskaya, E. & Dowhan, W. Cardiolipin is essential for organization of complexes III and IV into a supercomplex in intact yeast mitochondria. J Biol Chem 280, 29403- 8 (2005).

47 121. Bottinger, L. et al. Phosphatidylethanolamine and cardiolipin differentially affect the stability of mitochondrial respiratory chain supercomplexes. J Mol Biol 423, 677-86 (2012). 122. Martensson, C.U., Doan, K.N. & Becker, T. Effects of lipids on mitochondrial functions. Biochim Biophys Acta Mol Cell Biol Lipids 1862, 102-113 (2017). 123. Wu, M., Gu, J., Guo, R., Huang, Y. & Yang, M. Structure of Mammalian Respiratory Supercomplex I1III2IV1. Cell 167, 1598- 1609 e10 (2016). 124. Hackenbrock, C.R., Chazotte, B. & Gupte, S.S. The random collision model and a critical assessment of diffusion and collision in mitochondrial electron transport. J Bioenerg Biomembr 18, 331- 68 (1986). 125. Althoff, T., Mills, D.J., Popot, J.L. & Kuhlbrandt, W. Arrangement of electron transport chain components in bovine mitochondrial supercomplex I1III2IV1. EMBO J 30, 4652-64 (2011). 126. Lobo-Jarne, T. & Ugalde, C. Respiratory chain supercomplexes: Structures, function and biogenesis. Semin Cell Dev Biol 76, 179- 190 (2018). 127. Maranzana, E., Barbero, G., Falasca, A.I., Lenaz, G. & Genova, M.L. Mitochondrial respiratory supercomplex association limits production of reactive oxygen species from complex I. Antioxid Redox Signal 19, 1469-80 (2013). 128. Sun, J. & Trumpower, B.L. Superoxide anion generation by the cytochrome bc1 complex. Arch Biochem Biophys 419, 198-206 (2003). 129. Blaza, J.N., Serreli, R., Jones, A.J., Mohammed, K. & Hirst, J. Kinetic evidence against partitioning of the ubiquinone pool and the catalytic relevance of respiratory-chain supercomplexes. Proc Natl Acad Sci U S A 111, 15735-40 (2014). 130. Dubochet, J. & McDowall, A. Vitrification of pure water for electron microscopy. Journal of Microscopy 124, 3-4 (1981). 131. Drulyte, I. et al. Approaches to altering particle distributions in cryo- electron microscopy sample preparation. Acta Crystallogr D Struct Biol 74, 560-571 (2018). 132. Glaeser, R.M. Proteins, interfaces, and cryo-EM grids. Current opinion in colloid & interface science 34, 1-8 (2018). 133. Noble, A.J. et al. Routine single particle CryoEM sample and grid characterization by tomography. Elife 7(2018). 134. Dandey, V.P. et al. Spotiton: New features and applications. J Struct Biol 202, 161-169 (2018). 135. Callaway, E. The revolution will not be crystallized: a new method sweeps through structural biology. Nature 525, 172-4 (2015). 136. Passmore, L.A. & Russo, C.J. Specimen Preparation for High- Resolution Cryo-EM. Methods Enzymol 579, 51-86 (2016).

48 137. Herzik, M.A., Jr., Wu, M. & Lander, G.C. High-resolution structure determination of sub-100 kDa complexes using conventional cryo- EM. Nat Commun 10, 1032 (2019). 138. Kato, T. et al. CryoTEM with a Cold Field Emission Gun That Moves Structural Biology into a New Stage. Microscopy and Microanalysis 25, 998-999 (2019). 139. Fernandez-Leiro, R., Conrad, J., Scheres, S.H. & Lamers, M.H. cryo-EM structures of the E. coli replicative DNA polymerase reveal its dynamic interactions with the DNA sliding clamp, exonuclease and tau. Elife 4(2015). 140. Saurer, M. et al. Mitoribosomal small subunit biogenesis in trypanosomes involves an extensive assembly machinery. Science 365, 1144-1149 (2019). 141. Amat, F. et al. Alignment of cryo-electron tomography datasets. Methods Enzymol 482, 343-67 (2010). 142. Kisielowski, C. et al. Detection of single atoms and buried defects in three dimensions by aberration-corrected electron microscope with 0.5-A information limit. Microsc Microanal 14, 469-77 (2008). 143. Bai, X.C., McMullan, G. & Scheres, S.H. How cryo-EM is revolutionizing structural biology. Trends Biochem Sci 40, 49-57 (2015). 144. De Broglie, L. The wave nature of the electron. Nobel lecture 12, 244-256 (1929). 145. Ernst, R. The development of the electron micronscope and of electron micronscopy. (Berlin: Nobel Lecture, 1986). 146. Williams, D.B. & Carter, C.B. The transmission electron microscope. in Transmission electron microscopy 3-17 (Springer, 1996). 147. Downing, K.H. & Glaeser, R.M. Restoration of weak phase-contrast images recorded with a high degree of defocus: the "twin image" problem associated with CTF correction. Ultramicroscopy 108, 921- 8 (2008). 148. Baker, L.A. & Rubinstein, J.L. Radiation damage in electron cryomicroscopy. Methods Enzymol 481, 371-88 (2010). 149. Yonekura, K., Braunfeld, M.B., Maki-Yonekura, S. & Agard, D.A. Electron energy filtering significantly improves amplitude contrast of frozen-hydrated protein at 300kV. J Struct Biol 156, 524-36 (2006). 150. Tan, Y.Z., Cheng, A., Potter, C.S. & Carragher, B. Automated data collection in single particle electron microscopy. Microscopy (Oxf) 65, 43-56 (2016). 151. Wu, C., Huang, X., Cheng, J., Zhu, D. & Zhang, X. High-quality, high-throughput cryo-electron microscopy data collection via beam tilt and astigmatism-free beam-image shift. J Struct Biol 208, 107396 (2019).

49 152. Turchetta, R. et al. A monolithic active pixel sensor for charged particle tracking and imaging using standard VLSI CMOS technology. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 458, 677-689 (2001). 153. McMullan, G., Faruqi, A.R., Clare, D. & Henderson, R. Comparison of optimal performance at 300keV of three direct electron detectors for use in low dose electron microscopy. Ultramicroscopy 147, 156- 63 (2014). 154. Scheres, S.H. Beam-induced motion correction for sub-megadalton cryo-EM particles. Elife 3, e03665 (2014). 155. Nogales, E. & Scheres, S.H. Cryo-EM: A Unique Tool for the Visualization of Macromolecular Complexity. Mol Cell 58, 677-89 (2015). 156. Scheres, S.H. A Bayesian view on cryo-EM structure determination. J Mol Biol 415, 406-18 (2012). 157. Scheres, S.H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J Struct Biol 180, 519-30 (2012). 158. Scheres, S.H. Processing of Structurally Heterogeneous Cryo-EM Data in RELION. Methods Enzymol 579, 125-57 (2016). 159. Nakane, T., Kimanius, D., Lindahl, E. & Scheres, S.H. Characterisation of molecular motions in cryo-EM single-particle data by multi-body refinement in RELION. Elife 7(2018). 160. Kremer, J.R., Mastronarde, D.N. & McIntosh, J.R. Computer visualization of three-dimensional image data using IMOD. J Struct Biol 116, 71-6 (1996).