uu.diva-portal.org › smash › get › diva2:1307157 › FULLTEXT01… · ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2019 Digital Comprehensive Summar ies of Uppsala Disser tations

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1814

Accuracy of protein synthesis andits tuning by mRNA modifications

GABRIELE INDRISIUNAITE

ISSN 1651-6214ISBN 978-91-513-0667-4urn:nbn:se:uu:diva-382490

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





Dissertation presented at Uppsala University to be publicly examined in A1:111a, BMC, Husargatan 3, Uppsala, Tuesday, 4 June 2019 at 13:15 for the degree of Doctor of Philosophy. The examination will be conducted in English. Faculty examiner: Professor Jaanus Remme (University of Tartu, Institute of Molecular and Cell Biology).

AbstractIndrisiunaite, G. 2019. Accuracy of protein synthesis and its tuning by mRNA modifications. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1814. 47 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0667-4.The ribosome is a large macromolecular complex that synthesizes all proteins in the cellin all kingdoms of life. Proteins perform many vital functions, ranging from catalysis ofbiochemical reactions to muscle movement. It is essential for cells and organisms that proteinsare synthesized rapidly and accurately.

This thesis addresses two questions regarding the accuracy of protein synthesis. Howdo bacterial and eukaryotic release factors ensure accurate termination? How do mRNAmodifications affect the accuracy of bacterial protein synthesis?

Bacterial release factors 1 (RF1) and 2 (RF2) are proteins that recognize the stop codonsof mRNA and catalyze the release of a synthesized protein chain from the ribosome. It hasbeen proposed that RFs ensure accurate termination by binding to the ribosome in an inactive,compact conformation and acquire a catalytically active, extended conformation only afterrecognizing a correct stop codon. However, the native compact conformation was too short-lived to be captured by conventional structural methods. We have developed a fast-kineticsapproach for determining when the RFs are in a compact conformation on the ribosome andthen used time-resolved cryogenic electron microscopy to capture the compact conformationsof native RF1 and RF2 bound to a stop codon. We have also measured the effect of eukaryoticrelease factor 3 (eRF3) on the rate and accuracy of peptide release by eukaryotic release factor1 (eRF1) in a yeast (Saccharomyces cerevisiae) in vitro translation system.

Modifications of mRNA nucleotides are post-transcriptional regulators of gene expression,but little is known about their role in protein synthesis. We have studied the effect on accuracyof protein synthesis by two of these modifications: 2’-O-methylation and N6-methylation ofadenosine. 2’-O-methylation greatly reduced the maximal rate (kcat) and efficiency (kcat/Km) ofcognate (correct) codon reading by decreasing the initial GTPase activity in elongation factor Tuand enhancing proofreading losses of cognate aminoacyl-tRNAs. Remarkably, N6-methylationreduced the efficiency of codon reading by cognate aminoacyl-tRNAs and release factors,leaving the efficiency of the corresponding non-cognate reactions much less affected.

Keywords: Ribosome, Protein synthesis, Translation, Accuracy, Release factor, Termination,mRNA modifications

Gabriele Indrisiunaite, Department of Cell and Molecular Biology, Molecular Biology, Box596, Uppsala University, SE-751 24 Uppsala, Sweden.

© Gabriele Indrisiunaite 2019

ISSN 1651-6214ISBN 978-91-513-0667-4urn:nbn:se:uu:diva-382490 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-382490)




























List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I. Indrisiunaite, G., Pavlov, M.Y., Heurgué-Hammard, V., Ehrenberg, M. (2015). On the pH dependence of class-1 RF-dependent termina-tion of mRNA translation. Journal of Molecular Biology, 427, 1848-1860.

II. Fu, Z.*, Indrisiunaite, G.*, Kaledhonkar, S.*, Shah, B., Sun, M.,Chen, B., Grassucci, R.A., Ehrenberg, M. The structural basis for re-lease factor activation during translation termination revealed bytime-resolved cryogenic electron microscopy. (Submitted manu-script)

III. Choi, J., Indrisiunaite, G., DeMirci, H., Ieong, K-W., Wang, J., Pe-trov A., Prabhakar A., Rechavi G., Dominissini D., He C., Ehren-berg M., Puglisi J.D. (2018). 2’-O-methylation in mRNA disruptstRNA decoding during translation elongation. Nature Structural &Molecular Biology, 25, 208–16.

IV. Ieong, K-W.*, Indrisiunaite, G.*, Ehrenberg, M. N6-methyladenosines in mRNA have profound effects on the accuracyof codon reading by tRNAs and peptide release factors. (Manu-script)

*First authors with equal contribution

Reprints were made with permission from the respective publishers.

List of Papers








List of Papers








List of Papers



II. Fu, Z.*, Indrisiunaite, G.*, Kaledhonkar, S.*, Shah, B., Sun, M., Chen, B., Grassucci, R.A., Ehrenberg, M., Frank J. The structural basis for release factor activation during translation termination revealed by time-resolved cryogenic electron microscopy.Nature Communications. In press.





Contents

Introduction..................................................................................................... 9The ribosome............................................................................................ 10Protein synthesis ....................................................................................... 10

Elongation............................................................................................ 11Termination ......................................................................................... 12

The present work........................................................................................... 15Methods .................................................................................................... 15

Escherichia coli in vitro translation system ......................................... 15Saccharomyces cerevisiae in vitro translation system ......................... 16Quench-flow technique........................................................................ 17Stopped-flow technique....................................................................... 17Accuracy of protein synthesis .............................................................. 18

Accuracy of termination in bacteria and yeast.............................................. 20Conformational change in bacterial release factors .................................. 20

Accuracy of bacterial termination ....................................................... 20Structure of bacterial release factors 1 and 2 ....................................... 21Biochemical evidence of conformational change in RFs (Paper I) ..... 21Structural evidence of conformational change in RFs (Paper II) ........ 24

The accuracy of yeast termination............................................................ 28The role of eRF3 in eukaryotic termination......................................... 28The effect of eRF3 on the accuracy of termination by eRF1............... 28

The effect of mRNA modifications on elongation and termination phasesof protein synthesis ....................................................................................... 33

2’-O-methylation of mRNA in bacterial elongation (Paper III) .......... 33N6-methylation of mRNA in bacterial termination (Paper IV) ............ 35

Conclusions and future outlook .................................................................... 38

Sammanfattning på Svenska ......................................................................... 39

Acknowledgements ....................................................................................... 40

References..................................................................................................... 41

Contents













References..................................................................................................... 41

Contents













References..................................................................................................... 41

Contents

Introduction ..................................................................................................... 9 The ribosome ............................................................................................ 10 Protein synthesis ....................................................................................... 10

Elongation ............................................................................................ 11 Termination ......................................................................................... 12

The present work........................................................................................... 15 Methods .................................................................................................... 15

Escherichia coli in vitro translation system ......................................... 15 Saccharomyces cerevisiae in vitro translation system ......................... 16 Quench-flow technique ........................................................................ 17 Stopped-flow technique ....................................................................... 17 Accuracy of protein synthesis .............................................................. 18

Accuracy of termination in bacteria and yeast .............................................. 20 Conformational change in bacterial release factors .................................. 20

Accuracy of bacterial termination ....................................................... 20 Structure of bacterial release factors 1 and 2 ....................................... 21 Biochemical evidence of conformational change in RFs (Paper I) ..... 21 Structural evidence of conformational change in RFs (Paper II) ........ 24

The accuracy of yeast termination ............................................................ 28 The role of eRF3 in eukaryotic termination......................................... 28 The effect of eRF3 on the accuracy of termination by eRF1 ............... 28

The effect of mRNA modifications on elongation and termination phases of protein synthesis ....................................................................................... 33

2’-O-methylation of mRNA in bacterial elongation (Paper III) .......... 33 N6-methylation of mRNA in bacterial termination (Paper IV) ............ 35




References ..................................................................................................... 41

Abbreviations

3H tritiumA accuracy A site aminoacyl-tRNA siteaa-tRNA aminoacyl-tRNA ABCE1 ATP-binding cassette sub-family E member 1 ATP adenosine 5’-triphosphate cryo-EM cryogenic electron microscopyDbp5 DEAD-box protein 5 DNA deoxyribonucleic acid E site exit site E. coli Escherichia coli eEF eukaryotic elongation factor EF elongation factor eRF eukaryotic release factorF proofreading factorfMet formylmethionine FRET Förster resonance energy transferGDP guanosine diphosphate GTP guanosine 5’-triphosphate GTS glycine-threonine-serine motifHis-tag histidine tagHPLC high-pressure liquid chromatography I accuracy of initial selectionm6A N6-methyladenosinemRNA messenger RNA NIKS asparagine-isoleucine-lysine-serine motif P site peptidyl-tRNA siteP(A/V)T proline-(adenine/valine)-threonine motif PABP poly(adenine) binding protein Phe phenylalaninePrmC release factor glutamine methyltransferaseRC ribosomal release complex RF release factorRli1 RNase L inhibitor 1

Abbreviations


Abbreviations


Abbreviations

3H tritium A accuracy A site aminoacyl-tRNA site aa-tRNA aminoacyl-tRNA ABCE1 ATP-binding cassette sub-family E member 1 ATP adenosine 5’-triphosphate cryo-EM cryogenic electron microscopy Dbp5 DEAD-box protein 5 DNA deoxyribonucleic acid E site exit site E. coli Escherichia coli eEF eukaryotic elongation factor EF elongation factor eRF eukaryotic release factor F proofreading factor fMet formylmethionine FRET Förster resonance energy transfer GDP guanosine diphosphate GTP guanosine 5’-triphosphate GTS glycine-threonine-serine motif His-tag histidine tag HPLC high-pressure liquid chromatography I accuracy of initial selection m6A N6-methyladenosine mRNA messenger RNA NIKS asparagine-isoleucine-lysine-serine motif P site peptidyl-tRNA site P(A/V)T proline-(adenine/valine)-threonine motif PABP poly(adenine) binding protein Phe phenylalanine PrmC release factor glutamine methyltransferase RC ribosomal release complex RF release factor Rli1 RNase L inhibitor 1

RNA ribonucleic acid S. cerevisiae Saccharomyces cerevisiaeSPF serine–proline–phenylalanine motif T3 ternary complex tRNA transfer RNA Tyr tyrosine



RNA ribonucleic acid S. cerevisiae Saccharomyces cerevisiae SPF serine–proline–phenylalanine motif T3 ternary complex tRNA transfer RNA Tyr tyrosine

9

Introduction

Each cell and organism store the information about how to be a cell or an organism in their chromosomes, made of deoxyribonucleic acid (DNA). DNA is used for storing genetic information and for passing it on to the nextgeneration. Before each cell division its DNA is duplicated by DNA poly-merases in a process called replication. After cell division, each of the daughter cells inherits a copy of their parent cell’s DNA. The DNA is used as a template for transcription into messenger ribonucleic acid (mRNA). Each nucleotide triplet in mRNA encodes one amino acid. During protein synthesis (translation) the mRNA sequence is used as a template to synthe-size a chain of amino acids linked together by peptide bonds to form a poly-peptide. A functional protein can be made of one or more polypeptides. The genetic code that defines which codon corresponds to which amino acid is universal to all living beings. The flow of information from DNA through RNA to protein is called the central dogma of molecular biology (Figure 1), formulated by Francis Crick (Crick, 1970) and occurs in all kingdoms of life. Also, there are some special cases of information transfer, usually employed by viruses, when RNA is synthesized from RNA or DNA from RNA.

Figure 1. The central dogma of molecular biology describes the flow of infor-mation in biological systems. Solid lines – flows general to all organisms, dotted lines – special cases.

It is essential for cells and organisms that proteins are synthesized rapidly and accurately. Proteins perform many vital functions, ranging from cataly-sis of biochemical reactions to muscle movement. They work as regulators of gene expression, molecule transporters, receptors and structural compo-nents, such as keratin in skin and hair. Proteins also participate in making other proteins. Together with ribosomal RNA (rRNA) they form ribosomes that synthesize all proteins in the cell.

9

Introduction




9

Introduction




9

Introduction

Each cell and organism store the information about how to be a cell or an organism in their chromosomes, made of deoxyribonucleic acid (DNA). DNA is used for storing genetic information and for passing it on to the next generation. Before each cell division its DNA is duplicated by DNA poly-merases in a process called replication. After cell division, each of the daughter cells inherits a copy of their parent cell’s DNA. The DNA is used as a template for transcription into messenger ribonucleic acid (mRNA). Each nucleotide triplet in mRNA encodes one amino acid. During protein synthesis (translation) the mRNA sequence is used as a template to synthe-size a chain of amino acids linked together by peptide bonds to form a poly-peptide. A functional protein can be made of one or more polypeptides. The genetic code that defines which codon corresponds to which amino acid is universal to all living beings. The flow of information from DNA through RNA to protein is called the central dogma of molecular biology (Figure 1), formulated by Francis Crick (Crick, 1970) and occurs in all kingdoms of life. Also, there are some special cases of information transfer, usually employed by viruses, when RNA is synthesized from RNA or DNA from RNA.



10

The ribosome The ribosome is a large macromolecular complex made of proteins andrRNA. Essentially, it is an enzyme that converts its substrate (aminoacyl-tRNAs) to a product (polypeptide). The sequence of a polypeptide is deter-mined by mRNA template. Amino acids are delivered to the ribosome by transfer RNAs (tRNAs). Each aminoacyl-tRNA has an amino acid covalent-ly attached to its acceptor stem and a trinucleotide anticodon, complemen-tary to the mRNA codon of that amino acid. The ribosome recognizes cor-rect tRNAs by checking the geometry of codon-anticodon interaction. Ami-noacyl-tRNAs are charged with appropriate amino acids by specific en-zymes, aminoacyl-tRNA synthetases.

There are two types of mRNA codons – sense codons that encode amino acids and are recognized by their respective tRNAs and stop codons that signal the end of a polypeptide chain and are recognized by release factors (RF). The RFs then release the finished peptide from the P-site tRNA. The codon-anticodon interaction can be either correct (cognate) or incorrect(non-cognate). If non-cognate interaction is mistakenly recognized as cog-nate, wrong amino acid is incorporated or peptide is released on a sense co-don. Such translation errors can produce inactive or even toxic proteins. A special case of non-cognate interaction is near-cognate: when a codon differs from the correct codon by only one nucleotide.

The ribosome is made of two subunits: the small and the large. Ribo-somes and their subunits are named after their sedimentation coefficients in Svedberg units (S). For example, bacterial 70S ribosome is made of 50S and 30S subunits, while a larger eukaryotic 80S ribosome is made of 60S and 40S subunits. The size and composition of the subunits slightly vary betweendifferent kingdoms of life, but all ribosomes have a conserved common core that performs the key processes of protein synthesis (reviewed in(Yusupovaand Yusupov, 2014). The small subunit contains the decoding center (DC),where the codon-anticodon interaction is monitored. The large subunit con-tains the peptidyl transferase center (PTC), where the chemical catalysis ofpeptide bond formation or peptide release from the P-site tRNA occurs. Al-so, the ribosome has three binding sites that span both subunits. The A site isthe binding site for the incoming aminoacyl-tRNA, the P site harbors the peptidyl-tRNA and the E site is where deacylated tRNA binds before exiting the ribosome.

Protein synthesis Protein synthesis can be divided into four phases: initiation, elongation, ter-mination and recycling. In each phase the ribosome is assisted by a differentset of protein factors. In short, during initiation the ribosome is assembled on

10





10





10

The ribosome The ribosome is a large macromolecular complex made of proteins and rRNA. Essentially, it is an enzyme that converts its substrate (aminoacyl-tRNAs) to a product (polypeptide). The sequence of a polypeptide is deter-mined by mRNA template. Amino acids are delivered to the ribosome by transfer RNAs (tRNAs). Each aminoacyl-tRNA has an amino acid covalent-ly attached to its acceptor stem and a trinucleotide anticodon, complemen-tary to the mRNA codon of that amino acid. The ribosome recognizes cor-rect tRNAs by checking the geometry of codon-anticodon interaction. Ami-noacyl-tRNAs are charged with appropriate amino acids by specific en-zymes, aminoacyl-tRNA synthetases.

There are two types of mRNA codons – sense codons that encode amino acids and are recognized by their respective tRNAs and stop codons that signal the end of a polypeptide chain and are recognized by release factors (RF). The RFs then release the finished peptide from the P-site tRNA. The codon-anticodon interaction can be either correct (cognate) or incorrect (non-cognate). If non-cognate interaction is mistakenly recognized as cog-nate, wrong amino acid is incorporated or peptide is released on a sense co-don. Such translation errors can produce inactive or even toxic proteins. A special case of non-cognate interaction is near-cognate: when a codon differs from the correct codon by only one nucleotide.

The ribosome is made of two subunits: the small and the large. Ribo-somes and their subunits are named after their sedimentation coefficients in Svedberg units (S). For example, bacterial 70S ribosome is made of 50S and 30S subunits, while a larger eukaryotic 80S ribosome is made of 60S and 40S subunits. The size and composition of the subunits slightly vary between different kingdoms of life, but all ribosomes have a conserved common core that performs the key processes of protein synthesis (reviewed in(Yusupova and Yusupov, 2014). The small subunit contains the decoding center (DC), where the codon-anticodon interaction is monitored. The large subunit con-tains the peptidyl transferase center (PTC), where the chemical catalysis of peptide bond formation or peptide release from the P-site tRNA occurs. Al-so, the ribosome has three binding sites that span both subunits. The A site is the binding site for the incoming aminoacyl-tRNA, the P site harbors the peptidyl-tRNA and the E site is where deacylated tRNA binds before exiting the ribosome.

Protein synthesis Protein synthesis can be divided into four phases: initiation, elongation, ter-mination and recycling. In each phase the ribosome is assisted by a different set of protein factors. In short, during initiation the ribosome is assembled on

11

mRNA from its small and large subunits, a start codon (AUG) on mRNA is located and a specialized initiator tRNA, charged with amino acid methio-nine, is delivered to the P site. During elongation, the ribosome moves along mRNA, decoding each codon and incorporating a corresponding amino acid into a growing polypeptide chain. When the ribosome encounters one of the stop codons on mRNA (UAA, UAG or UGA), a protein release factor binds to the A site and cleaves the finished polypeptide from the P-site tRNA. Dur-ing recycling phase the ribosome is split by recycling factors and the subu-nits can be used for the next round of initiation. The phases of protein syn-thesis most relevant to the work presented in this thesis (elongation and ter-mination) are discussed in more detail in the following chapters.

Elongation Bacterial elongation starts with a 70S initiation complex, formed during initiation phase: initiator fMet-tRNAfMet is bound to the start codon (AUG) in the P site of the ribosome and the second codon is displayed in the A site. Elongation is a cyclic process consisting of two main events. Incorporationof each amino acid into a growing peptide chain is followed by transloca-tion: movement of the ribosome along the mRNA by one codon. The steps of one amino acid incorporation are shown in Figure 2. Each codon is read by a corresponding tRNA that enters the A site as a part of a ternary complex (T3) consisting of aminoacyl-tRNA (aa-tRNA), elongation factor Tu (EF-Tu) and guanosine triphosphate (GTP). At this stage the aa-tRNA is bound to theribosome in A/T conformation: its anticodon is base-paired with mRNAcodon in the A site, but the acceptor stem is bound to EF-Tu and cannot in-teract with the PTC (Schmeing et al., 2009; Valle et al., 2003). Codon-anticodon interaction is checked by monitoring bases: A1492, A1493 and G530 of 16S rRNA (Ogle et al., 2001). They sense the geometry of the firsttwo base pairs of the codon-anticodon helix. If base pairing is incorrect,there is a high probability for this aa-tRNA to dissociate from the ribosome. If base pairing is correct, they form stable interactions with the minor grooveof the codon-anticodon helix (Ogle et al., 2001). This sends a signal for EF-Tu to hydrolyze GTP. After GTP hydrolysis, EF-Tu dissociates from the ribosome and the aa-tRNA can accommodate in the A site. Its acceptor stem moves to interact with the PTC. When the aa-tRNA is accommodated, the α-amino group of its amino acid performs a nucleophilic attack on the carbonyl carbon of the ester bond connecting the nascent polypeptide to the P-sitetRNA. Thus the P-site tRNA is deacylated, the peptide is lengthened by one amino acid and transferred to the A-site tRNA. Deacylation of the P-site tRNA causes the ribosome to enter a ratcheted state, meaning that its subu-nits are rotated in relation to each other (Frank and Agrawal, 2000; Valle et al., 2003). This relative rotation brings the tRNAs into a hybrid state (A/P orP/E). The anticodon stem of the A-site tRNA remains in the A site on the

11



11



11


Elongation Bacterial elongation starts with a 70S initiation complex, formed during initiation phase: initiator fMet-tRNAfMet is bound to the start codon (AUG) in the P site of the ribosome and the second codon is displayed in the A site. Elongation is a cyclic process consisting of two main events. Incorporation of each amino acid into a growing peptide chain is followed by transloca-tion: movement of the ribosome along the mRNA by one codon. The steps of one amino acid incorporation are shown in Figure 2. Each codon is read by a corresponding tRNA that enters the A site as a part of a ternary complex (T3) consisting of aminoacyl-tRNA (aa-tRNA), elongation factor Tu (EF-Tu) and guanosine triphosphate (GTP). At this stage the aa-tRNA is bound to the ribosome in A/T conformation: its anticodon is base-paired with mRNA codon in the A site, but the acceptor stem is bound to EF-Tu and cannot in-teract with the PTC (Schmeing et al., 2009; Valle et al., 2003). Codon-anticodon interaction is checked by monitoring bases: A1492, A1493 and G530 of 16S rRNA (Ogle et al., 2001). They sense the geometry of the first two base pairs of the codon-anticodon helix. If base pairing is incorrect, there is a high probability for this aa-tRNA to dissociate from the ribosome. If base pairing is correct, they form stable interactions with the minor groove of the codon-anticodon helix (Ogle et al., 2001). This sends a signal for EF-Tu to hydrolyze GTP. After GTP hydrolysis, EF-Tu dissociates from the ribosome and the aa-tRNA can accommodate in the A site. Its acceptor stem moves to interact with the PTC. When the aa-tRNA is accommodated, the α-amino group of its amino acid performs a nucleophilic attack on the carbonyl carbon of the ester bond connecting the nascent polypeptide to the P-site tRNA. Thus the P-site tRNA is deacylated, the peptide is lengthened by one amino acid and transferred to the A-site tRNA. Deacylation of the P-site tRNA causes the ribosome to enter a ratcheted state, meaning that its subu-nits are rotated in relation to each other (Frank and Agrawal, 2000; Valle et al., 2003). This relative rotation brings the tRNAs into a hybrid state (A/P or P/E). The anticodon stem of the A-site tRNA remains in the A site on the

12

30S subunit, but the acceptor stem moves to the P site of the 50S subunit (A/P state). The anticodon stem of the P-site tRNA remains in the P site onthe 30S subunit, but the acceptor stem moves to the E site of the 50S subunit (P/E state). The ratcheted ribosome is bound by elongation factor G (EF-G)and GTP. GTP hydrolysis by EF-G facilitates the movement of the mRNAby one codon in relation to the ribosome, the tRNAs enter P/P and E/E states and the ribosome returns to a non-ratcheted state. EF-G-GDP dissociates, leaving the ribosome ready for the next ternary complex. A more detailedsequence of events prior to aa-tRNA accommodation in the A site is shown in Figure 12 in relation to results discussed in Paper III.

Elongation is the most conserved phase of protein synthesis and proceeds essentially in the same way in eukaryotes, as described above for bacteria.Eukaryotic elongation factor 1A (eEF1A) and elongation factor 2 (eEF2) arestructurally and functionally similar to their bacterial counterparts EF-Tu and EF-G, respectively (reviewed in (Dever et al., 2016). Fungi have an addi-tional elongation factor 3 (eEF3) (Skogerson and Wakatama, 1976). Its func-tion is not completely clear, but eEF3 seems to facilitate the release of deac-ylated tRNA from the E site (Triana-Alonso et al., 1995).

Figure 2. An overview of elongation in bacteria.

Termination Termination of protein synthesis occurs when a translating ribosome encoun-ters one of the three stop codons (UAA, UAG, UGA) on mRNA. These co-dons do not have corresponding tRNAs, but are recognized by protein re-lease factors (RFs). Prokaryotes have two release factors with overlappingstop codon specificity: release factor 1 (RF1) recognizes UAA and UAG codons, while release factor 2 (RF2) recognizes UAA and UGA (Scolnick et al., 1968). Unlike bacteria, eukaryotes have one class I release factor, eRF1,which recognizes all three stop codons (Konecki et al., 1977). RF1 and RF2 share high sequence (Shin et al., 2004) and structural (Petry et al., 2005)similarity, but have different stop codon recognition motifs. The main

12





12





12

30S subunit, but the acceptor stem moves to the P site of the 50S subunit (A/P state). The anticodon stem of the P-site tRNA remains in the P site on the 30S subunit, but the acceptor stem moves to the E site of the 50S subunit (P/E state). The ratcheted ribosome is bound by elongation factor G (EF-G) and GTP. GTP hydrolysis by EF-G facilitates the movement of the mRNA by one codon in relation to the ribosome, the tRNAs enter P/P and E/E states and the ribosome returns to a non-ratcheted state. EF-G-GDP dissociates, leaving the ribosome ready for the next ternary complex. A more detailed sequence of events prior to aa-tRNA accommodation in the A site is shown in Figure 11 in relation to results discussed in Paper III.

Elongation is the most conserved phase of protein synthesis and proceeds essentially in the same way in eukaryotes, as described above for bacteria. Eukaryotic elongation factor 1A (eEF1A) and elongation factor 2 (eEF2) are structurally and functionally similar to their bacterial counterparts EF-Tu and EF-G, respectively (reviewed in (Dever et al., 2016). Fungi have an addi-tional elongation factor 3 (eEF3) (Skogerson and Wakatama, 1976). Its func-tion is not completely clear, but eEF3 seems to facilitate the release of deac-ylated tRNA from the E site (Triana-Alonso et al., 1995).


Termination Termination of protein synthesis occurs when a translating ribosome encoun-ters one of the three stop codons (UAA, UAG, UGA) on mRNA. These co-dons do not have corresponding tRNAs, but are recognized by protein re-lease factors (RFs). Prokaryotes have two release factors with overlapping stop codon specificity: release factor 1 (RF1) recognizes UAA and UAG codons, while release factor 2 (RF2) recognizes UAA and UGA (Scolnick et al., 1968). Unlike bacteria, eukaryotes have one class I release factor, eRF1, which recognizes all three stop codons (Konecki et al., 1977). RF1 and RF2 share high sequence (Shin et al., 2004) and structural (Petry et al., 2005) similarity, but have different stop codon recognition motifs. The main

13

recognition motifs are P(A/V)T (proline-adenine/valine-threonine) in RF1 and SPF (serine-proline-phenylalanine) in RF2 (Ito et al., 2000). These mo-tifs interact with the stop codon in the decoding center of the 30S subunit of the ribosome. Eukaryotic eRF1 has a different structure from prokaryotic RFs (Song et al., 2000) with N-terminal domain harboring several conserved codon recognition motifs: NIKS (asparagine-isoleucine-lysine-serine), YxCxxxF (tyrosine, cysteine, phenylalanine; x - any amino acid) and GTS (glycine-threonine-serine) (Bulygin et al., 2010; Chavatte et al., 2002; Frolova et al., 2002). The stop codon adopts a compact U-turn conformation within a pocket formed by eRF1 and the ribosome which also pulls the 4th nucleotide into the A site (Brown et al., 2015; Matheisl et al., 2015).

Remarkably, class I release factors in all kingdoms of life share a con-served catalytic glycine-glycine-glutamine (GGQ) motif (Frolova et al., 1999). During peptide release the GGQ motif is placed in the peptidyl trans-ferase center of the large ribosomal subunit (Petry et al., 2005; Rawat et al., 2003). The glutamine residue of the GGQ motif coordinates a nucleophile that attacks the ester bond connecting the nascent peptide to the P-site tRNA (Song et al., 2000). The nature of the nucleophile is not completely clarified. A water molecule (Song et al., 2000) or OH- ion (Kuhlenkoetter et al., 2011) has been suggested to perform this function. The glutamine residue of the GGQ is N5-methylated in bacterial (Dincbas-Renqvist et al., 2000) and eu-karyotic (Heurgue-Hamard et al., 2005) class I RFs. In bacteria it is methyl-ated by methyltransferase prmC (Heurgue-Hamard et al., 2002). This modi-fication moderately (around 2-fold) increases termination rate (Dincbas-Renqvist et al., 2000).

An overview of bacterial and eukaryotic termination is presented in Fig-ure 3. Class I RF binds to the stop codon in the A site and catalyzes peptide release from the P-site tRNA. After peptide release, bacterial RF1 and RF2 are recycled by a class II release factor, the GTP-ase RF3 (Freistroffer et al., 1997; Goldstein and Caskey, 1970). RF3 accelerates the dissociation of RF1 and RF2 from the ribosome in a GTP-dependent manner, but has no effect on the rate of peptide release (Freistroffer et al., 2000; Freistroffer et al., 1997). RF3-GTP causes conformational changes in the ribosome that induce dissociation of RF1/2 (Gao et al., 2007; Peske et al., 2014). RF3 then hydro-lyses GTP and also dissociates. In contrast, eukaryotic class II release factor 3 (eRF3) significantly increases the rate of peptide release by eRF1 (Alkalaeva et al., 2006; Eyler et al., 2013). The eRF1 enters the A site as a part of the eRF1·eRF3·GTP complex (Alkalaeva et al., 2006). GTP hydroly-sis by eRF3 induces a conformational change to catalytically active confor-mation in eRF1. This change is followed by A-site binding of eukaryotic ribosome recycling factor ABCE1 and probably dissociation of eRF3 (re-viewed in (Hellen, 2018). ABCE1 then stimulates the catalytic step of pep-tide release (Shoemaker and Green, 2011) and subsequently splits the ribo-some into subunits (Pisarev et al., 2010; Shoemaker and Green, 2011). RF3

13




13




13




14

is not required for cell viability (Grentzmann et al., 1994), while eRF3 is an essential protein (Chao et al., 2003; Ter-Avanesyan et al., 1993). Both bacte-rial and eukaryotic termination will be discussed in more detail in following chapters.

Figure 3. An overview of termination stage of protein synthesis in (a) bacte-ria and (b) eukaryotes.

14



14



14



15

The present work

The aim of this thesis is to answer questions regarding the accuracy of pro-tein synthesis with main focus on bacterial termination. We discuss the hy-pothesis that a conformational change in RF1 and RF2 increases the accura-cy of termination in bacteria. We introduce a rapid kinetics method to deter-mine the rate of such a conformational change (Paper I). This method is combined with time-resolved cryogenic electron microscopy (time-resolved cryo-EM) to obtain the real-time distribution of native structures of release factors on the ribosome, prior and post to the stop-codon dependent confor-mational change (Paper II). We also discuss the effect of eRF3 on the rate and accuracy of termination by eRF1. In addition, we describe the effect of different mRNA modifications on the rate of elongation (Paper III) and on the accuracy of elongation and termination (Paper IV) in bacteria.

Methods Escherichia coli in vitro translation system For all experiments on bacterial protein synthesis we used an in vitro transla-tion system with purified components from the translation machinery of Escherichia coli (E. coli). Since every component is individually purified, each can be added at specific concentrations or omitted altogether, thus tai-loring the translation system for solving each experimental problem. All in vitro experiments were performed in polymix buffer, which has been chosen for high translation speed and optimal accuracy, as first demonstrated in poly-phenylalanine peptide synthesis (Jelenc and Kurland, 1979) and later refined for a more varied set of peptides (Zhang et al., 2016). For the meas-urements of GTP hydrolysis, the polymix was supplemented with an energy regeneration system consisting of ATP, GTP, phosphoenolpyruvate, py-ruvate kinase and myokinase. This system increases the rate and accuracy of translation, likely by shifting the steady-state concentrations towards ATP and GTP and reducing the concentrations of their hydrolytic products to near-undetectable levels (Jelenc and Kurland, 1979).

Ribosomal release complexes (RCs), used for termination experiments (Papers I, II and IV), were assembled in vitro and purified from other com-

15

The present work




15

The present work




15

The present work




16

ponents of the translation system by ultracentrifugation on sucrose cushions. Because of their high molecular weight, the ribosomes are pelleted at the bottom of centrifugation tubes. Unbound or loosely bound components with smaller molecular weight, such as protein factors, tRNAs and mRNAs, re-main suspended and are removed together with the cushion. Peptidyl-tRNAs remained bound to the P site of the majority of ribosomes, as confirmed by cryo-EM experiments (Paper II).

Saccharomyces cerevisiae in vitro translation system 80S ribosomes were purified by ultracentrifugation on sucrose cushions from Saccharomyces cerevisiae (S. cerevisiae) strain PY116, deficient in several vacuolar proteinases and a mitochondrial RNAse NUC1 (Gerik et al., 1998). The yeast cells were harvested in an exponential phase of growth. Since we do not study initiation, this stage was simplified by using leaderless mRNA that can initiate on non-split 80S ribosomes and without initiation factors (Andreev et al., 2006). S. cerevisiae ribosomes can accurately initiate on leaderless mRNAs in vivo, although leaderless mRNAs are translated less efficiently than mRNAs with full-length 5’ leader sequences (Maicas et al., 1990). The only heterogeneous component used in this study was E. coli initiator tRNA fMet-tRNAfMet labelled with tritium (3H) on the methionine. This initiator tRNA was originally chosen to be able to use initiated E. coli ribosomes as a control for yeast ribosome activity. Wild-type elongation factor eEF1A was purified from yeast, His-tagged eEF2 was overexpressed and purified from yeast, eEF1Aβ and eEF3 were His-tagged and overex-pressed in E. coli. S. cerevisiae bulk tRNA and a yeast amino acid synthetase mix were used for peptide synthesis. Release factors eRF1 and eRF3 were His-tagged and overexpressed in E. coli. eRF3 lacked its prion domain (first 253 N-terminal amino acids), which is not essential for termination and in-teraction with eRF1 (Alkalaeva et al., 2006). It has been shown that hydrox-ylation of lysine (K) in the conserved NIKS motif increases termination effi-ciency (kcat/Km) of eRF1 (Feng et al., 2014). Our eRF1 lacks this modifica-tion. However, the kcat/Km increase is moderate (2-fold) and observed only for the first nucleotide of the stop codon (Feng et al., 2014), which is always the cognate U in our system.

Yeast 80S release complexes were purified essentially as described above for E. coli. The products were analyzed by high pressure liquid chromatog-raphy (HPLC) and around 50% of the synthesized peptide was the expected fMet-Phe-Tyr, with lower amounts of fMet-Phe and fMet. No other peptides were detected. The presence of formyl group on the methionine did not af-fect peptide synthesis.

16




16




16




17

Quench-flow technique The quench-flow technique was used to study pre-steady state kinetics of elongation and termination reactions that were too fast to perform by hand. The time resolution of the quench-flow quench-instrument is 2 ms. Equal volumes of two reactants are rapidly mixed together and allowed to react for a pre-selected time while being pushed along a reaction loop (this is the ‘flow’ part of the process). The reaction time depends on the length of the loop and the rate at which the reaction mixture is pushed through it. After a set amount of time the reaction mixture reaches the end of the reaction loop and the reaction is stopped by mixing with a quenching solution (this is the ‘quench’ part of the process). Subsequently, the reaction products at differ-ent reaction times are analyzed and quantified by other methods. In the ex-periments presented here, the quenching solution was formic acid which, in addition to stopping the reaction, also precipitates large RNA molecules. In order to quantify the reaction products, initiator tRNA, fMet-tRNAfMet, used for these experiments was labelled with tritium (3H) on the methionine. For termination studies (Papers I, II and IV), precipitated peptidyl-tRNA was separated from soluble released peptide by centrifugation and the amounts of released and tRNA-bound peptides were determined by liquid scintillation counting of 3H radiation. Scintillation liquid absorbs the energy emitted by radioisotopes and re-emits it as a detectable light signal. Through this exper-imental setup, the average time for all the steps from ternary complex bind-ing to the ribosome up to and including hydrolysis of the ester bond connect-ing peptide with P-site tRNA can be estimated at high precision. At the same time, the rate of released peptide dissociation from the ribosome is not measured. The method allows for validation of those stopped-flow results, to be described below, in which the rate of peptide dissociation can be rate limiting. For protein elongation studies (Paper III), the reaction products were analyzed by HPLC, equipped with an on line radiation detector, that can separate different peptides with single amino acid resolution. For the measurements of GTP hydrolysis (Paper III), the amounts of 3H-GTP and 3H-GDP were also determined by HPLC.

Stopped-flow technique The initial phase of a stopped-flow experiment is similar to that of a quench-flow experiment – equal volumes of two reactants are rapidly mixed, but instead of quenching the reaction at each time point separately, the reaction is continuously monitored in real time in an optical cell. Depending on the setup of the instrument and the reactants used, a change in fluorescence, light absorption or light scattering intensity can be recorded. Since there is no need to take each time point separately, much more points per one reac-tion curve can be monitored to improve precision. Also, since there is no

17



17



17



18

washing step between time points, these experiments are performed much faster than the quench-flow experiments, which is advantageous when using easily degradable components or reaction conditions that might cause the components to lose activity (such as high pH). Since most reactions with native components do not cause directly observable changes, reaction com-ponents for stopped-flow experiments are often labelled with fluorescent molecules. These molecules are often quite bulky, making it important to validate that the presence of a fluorescent label does not significantly affect the kinetics of a reaction. In the experiments presented here stopped-flow was used for measuring the rate of coumarin-labelled peptide release at dif-ferent pH values (Paper I) and validated by comparison with quench-flow data.

Accuracy of protein synthesis The accuracy (A) of an enzymatic reaction is defined as the efficiency (kcat/Km) by which an enzyme forms a reaction product using a correct (cog-nate) substrate divided by the efficiency by which it forms an incorrect product from an incorrect (non-cognate) substrate. Accordingly, the termina-tion accuracy and the total accuracy of elongation are defined as a ratio of the efficiencies of peptide bond formation or peptide release between cog-nate (c) and non-cognate (nc) reactions:

𝐴𝐴 =(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝑐𝑐

(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝑛𝑛𝑐𝑐

The total accuracy of elongation can be further divided into the accuracy of initial selection (I) and proofreading (F):

𝐴𝐴 = 𝐼𝐼 × 𝐹𝐹

These two steps are separated by an irreversible step of GTP hydrolysis on EF-Tu. I is the accuracy of steps leading to GTP hydrolysis and F is the accuracy amplification in steps following GTP hydrolysis where non-cognate aminoacyl-tRNA dissociates with much higher probability than cognate aminoacyl-tRNA. Initial selection is defined as the ratio of efficien-cies of GTP hydrolysis on EF-Tu, (kcat/Km)GTP , during incorporation of cog-nate and non-cognate amino acids:

𝐼𝐼 =(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝐺𝐺𝐺𝐺𝐺𝐺𝑐𝑐

(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝐺𝐺𝐺𝐺𝐺𝐺𝑛𝑛𝑐𝑐

Then the proofreading factor F can be written as:

𝐹𝐹 =𝐴𝐴𝐼𝐼

=(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝐺𝐺𝐺𝐺𝐺𝐺𝑛𝑛𝑐𝑐

(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝑑𝑑𝑑𝑑𝑑𝑑𝑛𝑛𝑐𝑐 /(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝐺𝐺𝐺𝐺𝐺𝐺𝑐𝑐

(𝑘𝑘𝑐𝑐𝑐𝑐𝑐𝑐 /𝐾𝐾𝑀𝑀)𝑑𝑑𝑑𝑑𝑑𝑑𝑐𝑐

18















18















18















19

The concept of kinetic proofreading was first introduced by Hopfield and Ninio (Hopfield, 1974; Ninio, 1975) and later shown to occur in protein elongation (Ruusala et al., 1982; Thompson and Stone, 1977). It was noticed that the accuracy that results only from the energy difference between cog-nate and non-cognate codon-anticodon duplexes was too low to account for the observed elongation accuracy. When using kinetic proofreading, the ribosome can employ the same energy difference in several steps, provided that they are separated by an irreversible step, such as GTP hydrolysis. Then the total accuracy of the whole process follows as the product of the accura-cies in each step. In the case of elongation, the codon-anticodon interaction is first monitored during initial selection. There is a probability close to one that a cognate aa-tRNA in ternary complex with EF-Tu and GTP passes the initial selection step leading to the triggering of GTP hydrolysis on EF-Tu. At the same time the probability that a non-cognate ternary complex passes the initial selection step to GTP hydrolysis is very low. Furthermore, it was early recognized that the number of GTPs hydrolyzed per non-cognate pep-tide bond is much larger than the number of GTPs hydrolyzed per cognate peptide bond, leading to the conclusion that non-cognate tRNAs are discard-ed from the ribosome after GTP hydrolysis with high probability (Ruusala et al., 1982; Thompson and Stone, 1977). Then the accuracy amplification by proofreading, or the proofreading factor F, is the probability that a non-cognate aa-tRNA is discarded from the ribosome after GTP hydrolysis on EF-Tu divided by the corresponding probability for a cognate aa-tRNA.

19


19


19


20

Accuracy of termination in bacteria and yeast

Termination of protein synthesis has to be fast and accurate for optimal cell growth. Here, in Papers I and II, we present biochemical and structural evi-dence that bacterial RF1 and RF2 ensure high termination accuracy by un-dergoing a conformational change from catalytically inactive to active con-formation after recognizing a stop codon. We have also looked into eukary-otic termination and show here that eRF3 increases the accuracy of stop co-don recognition by eRF1.

Conformational change in bacterial release factors Accuracy of bacterial termination Termination of protein synthesis has to be very accurate, because frequenttermination on sense codons would cause accumulation of non-functional or even toxic truncated proteins, harmful to a cell. Unsurprisingly, both in vivo(Jorgensen et al., 1993) and in vitro (Freistroffer et al., 2000) experiments have confirmed that the termination error frequency is indeed very low: around 1 error per 105 termination events. However, determining the mecha-nism used to achieve such high accuracy has not been as straightforward. Identification of RF3, a GTP-ase participating in termination, lead to thesuggestion that the ribosome might increase termination accuracy by em-ploying kinetic proofreading, a mechanism already confirmed for recogni-tion of sense codons by aminoacyl-tRNAs (Ruusala et al., 1982; Thompson and Stone, 1977). However, subsequent experiments demonstrated that this hypothesis is not correct – the presence of RF3 not only does not increase termination accuracy, but even slightly reduces it (Freistroffer et al., 2000). According to molecular dynamics simulations performed by Sund and coworkers, RF-mRNA interactions themselves provide sufficient discrimina-tory power to ensure high termination accuracy (Sund et al., 2010). Accord-ing to another hypothesis, described in the following chapters, RF1 and RF2 undergo a stop codon-induced conformational change as a mechanism to ensure high termination accuracy (Rawat et al., 2003).

20




20




20



Conformational change in bacterial release factors Accuracy of bacterial termination Termination of protein synthesis has to be very accurate, because frequent termination on sense codons would cause accumulation of non-functional or even toxic truncated proteins, harmful to a cell. Unsurprisingly, both in vivo (Jorgensen et al., 1993) and in vitro (Freistroffer et al., 2000) experiments have confirmed that the termination error frequency is indeed very low: around 1 error per 105 termination events. However, determining the mecha-nism used to achieve such high accuracy has not been as straightforward. Identification of RF3, a GTP-ase participating in termination, lead to the suggestion that the ribosome might increase termination accuracy by em-ploying kinetic proofreading, a mechanism already confirmed for recogni-tion of sense codons by aminoacyl-tRNAs (Ruusala et al., 1982; Thompson and Stone, 1977). However, subsequent experiments demonstrated that this hypothesis is not correct – the presence of RF3 not only does not increase termination accuracy, but even slightly reduces it (Freistroffer et al., 2000). According to molecular dynamics simulations performed by Sund and coworkers, RF-mRNA interactions themselves provide sufficient discrimina-tory power to ensure high termination accuracy (Sund et al., 2010). Accord-ing to another hypothesis, described in the following chapters, RF1 and RF2 undergo a stop codon-induced conformational change as a mechanism to ensure high termination accuracy (Rawat et al., 2003).

21

Structure of bacterial release factors 1 and 2 RF1 and RF2 have similar sequences (Shin et al., 2004) and structures (Petryet al., 2005). Each one consists of four domains, of which domains II and III are most important for peptide release. Domain II contains a recognitionmotif (SPF in RF1 and P(A/V)T in RF2, see above) which interacts with thestop codon on mRNA, placed in the decoding center (DC) of the 30S subu-nit. The tip of domain III contains the catalytic GGQ motif. In order to re-lease the peptide from the P-site tRNA, the GGQ motif has to reach into the peptidyl-transfer center (PTC) in the 50S subunit of the ribosome.

The first X-ray structures of RF1 (Shin et al., 2004) and RF2 (Vestergaard et al., 2001) showed the conformations of these factors free in solution. The RFs were in a compact conformation, with the stop codon recognition motif and the GGQ motif just 25 Å apart. However, in early cryo-EM structures ofRF1 (Petry et al., 2005) and RF2 (Klaholz et al., 2003; Rawat et al., 2003) on the ribosome both factors were in an extended conformation, with the stopcodon recognition and the GGQ motifs 75 Å apart and thus spanning thedistance between the DC in the 30S subunit and the PTC in the 50S subunit (Klaholz et al., 2003; Petry et al., 2005). A small-angle X-ray scattering(SAXS) study of E. coli RF1 in solution concluded that the RF was most likely in an extended conformation (Vestergaard et al., 2005), same as on theribosome and contrary to the earlier X-ray structures. It was unclear, whether the compact form was physiologically relevant or just an artefact caused bynon-native crystallization conditions. However, an X-ray structure of a com-pact RF in complex with methyltransferase prmC demonstrated that the compact form had at least one physiological function (Graille et al., 2005). The Graille et al. study was followed by another SAXS study showing thatThermus thermophilus RF2 was predominantly compact, similar to the X-ray structures in crystal (Zoldak et al., 2007).

Back in 2003, Rawat and coworkers offered a solution to the apparently different structures of free and ribosome-bound RFs that could also be rele-vant to high accuracy of bacterial termination. They proposed that RF1 andRF2 enter their ribosome in compact conformation and then, in response to a stop codon in the A site, undergo a conformational change to an extended,catalytically active conformation (Rawat et al., 2003). Strict validation ofthis hypothesis lingered, because at the time there was no experimental method for capturing such short-lived states with sufficient resolution.

Biochemical evidence of conformational change in RFs (Paper I) The first evidence from rapid kinetics experiments that there is a conforma-tional change in class I RFs is described in Paper I. We used stopped-flow technique to measure the effect of pH on the maximal rate of termination(kcat) in an E. coli in vitro translation system. Ribosomes containing fMet-

21





21





21

Structure of bacterial release factors 1 and 2 RF1 and RF2 have similar sequences (Shin et al., 2004) and structures (Petry et al., 2005). Each one consists of four domains, of which domains II and III are most important for peptide release. Domain II contains a recognition motif (SPF in RF1 and P(A/V)T in RF2, see above) which interacts with the stop codon on mRNA, placed in the decoding center (DC) of the 30S subu-nit. The tip of domain III contains the catalytic GGQ motif. In order to re-lease the peptide from the P-site tRNA, the GGQ motif has to reach into the peptidyl-transfer center (PTC) in the 50S subunit of the ribosome.

The first X-ray structures of RF1 (Shin et al., 2004) and RF2 (Vestergaard et al., 2001) showed the conformations of these factors free in solution. The RFs were in a compact conformation, with the stop codon recognition motif and the GGQ motif just 25 Å apart. However, in early cryo-EM structures of RF1 (Petry et al., 2005) and RF2 (Klaholz et al., 2003; Rawat et al., 2003) on the ribosome both factors were in an extended conformation, with the stop codon recognition and the GGQ motifs 75 Å apart and thus spanning the distance between the DC in the 30S subunit and the PTC in the 50S subunit (Klaholz et al., 2003; Petry et al., 2005). A small-angle X-ray scattering (SAXS) study of E. coli RF1 in solution concluded that the RF was most likely in an extended conformation (Vestergaard et al., 2005), same as on the ribosome and contrary to the earlier X-ray structures. It was unclear, whether the compact form was physiologically relevant or just an artefact caused by non-native crystallization conditions. However, an X-ray structure of a com-pact RF in complex with methyltransferase prmC demonstrated that the compact form had at least one physiological function (Graille et al., 2005). The Graille et al. study was followed by another SAXS study showing that Thermus thermophilus RF2 was predominantly compact, similar to the X-ray structures in crystal (Zoldak et al., 2007).

Back in 2003, Rawat and coworkers offered a solution to the apparently different structures of free and ribosome-bound RFs that could also be rele-vant to high accuracy of bacterial termination. They proposed that RF1 and RF2 enter their ribosome in compact conformation and then, in response to a stop codon in the A site, undergo a conformational change to an extended, catalytically active conformation (Rawat et al., 2003). Strict validation of this hypothesis lingered, because at the time there was no experimental method for capturing such short-lived states with sufficient resolution.

Biochemical evidence of conformational change in RFs (Paper I) The first evidence from rapid kinetics experiments that there is a conforma-tional change in class I RFs is described in Paper I. We used stopped-flow technique to measure the effect of pH on the maximal rate of termination (kcat) in an E. coli in vitro translation system. Ribosomes containing fMet-

22

Phe-Phe-tRNAPhe in the P site and UAA stop codon in the A site (RC0 in Fig. 4a) were rapidly mixed with saturating amounts of release factors RF1 orRF2. The methionine in peptidyl-tRNA (Figure 4a) was labelled with a fluo-rescing coumarin derivative and the amount of released peptide at differentOH- ion concentrations was monitored as a change of fluorescence intensity over time.

At rate-saturating RF concentration, the association time of RF and the ri-bosome, 1/(ka[RF]), is very short and thus negligible to the total reaction time. Then the total reaction time is 1/kcat, which is the sum of times for the conformational change of the factor, 1/kconf, hydrolysis of the ester bond in peptidyl-tRNA, 1/khydr, and the time for peptide dissociation from the ribo-some, 1/kdiss, where khydr is strongly pH-dependent (Kuhlenkoetter et al., 2011; Shaw et al., 2012).

Figure 4. Termination in bacteria as described in Paper I. (a) A compact class I RF binds the A site of the ribosomal release complex RC0, forming complex RC·RFcompact with association rate constant ka and compounded rate constant ka·[RF]. RF undergoes conformational change from compact to extended form withpH-independent rate constant kconf, transforming complex RC·RFcompact to RC·RFextended, where the pH-dependent ester bond hydrolysis in peptidyl-tRNA is activated by movement of the universal GGQ motif into the PTC. Hydrolysis occurs with pH-dependent rate constant khydr, leading to complex RC·RFrel. Finally, thereleased peptide dissociates with rate constant kdiss. It leads to post-termination com-plex RC·RFpost. The label L is 3H in quench-flow or a fluorescent coumarin deriva-tive in stopped-flow experiments. (b) Maximal rate constant (kcat) (y-axis) as func-tion of OH- concentration (x-axis), estimated with 0.04 µM RC0 reacting to rate-saturating concentrations of unmethylated RF1 at 37⁰C, as measured in stopped-flow. (c) Quench-flow measurement of the same reaction.

22




22




22

Phe-Phe-tRNAPhe in the P site and UAA stop codon in the A site (RC0 in Fig. 4a) were rapidly mixed with saturating amounts of release factors RF1 or RF2. The methionine in peptidyl-tRNA (Figure 4a) was labelled with a fluo-rescing coumarin derivative and the amount of released peptide at different OH- ion concentrations was monitored as a change of fluorescence intensity over time.


Figure 4. Termination in bacteria as described in Paper I. (a) A compact class I RF binds the A site of the ribosomal release complex RC0, forming complex RC·RFcompact with association rate constant ka and compounded rate constant ka·[RF]. RF undergoes conformational change from compact to extended form with pH-independent rate constant kconf, transforming complex RC·RFcompact to RC·RFextended, where the pH-dependent ester bond hydrolysis in peptidyl-tRNA is activated by movement of the universal GGQ motif into the PTC. Hydrolysis occurs with pH-dependent rate constant khydr, leading to complex RC·RFrel. Finally, the released peptide dissociates with rate constant kdiss. It leads to post-termination com-plex RC·RFpost. The label L is 3H in quench-flow or a fluorescent coumarin deriva-tive in stopped-flow experiments. (b) Maximal rate constant (kcat) (y-axis) as func-tion of OH- concentration (x-axis), estimated with 0.04 µM RC0 reacting to rate-saturating concentrations of unmethylated RF1 at 37⁰C, as measured in stopped-flow. (c) Quench-flow measurement of the same reaction.

23

At low pH, kcat increased with OH- concentration, as observed before (Kuhlenkoetter et al., 2011; Shaw et al., 2012). The linear dependence can beexplained either by titration of a catalytically essential group, or by the OH-

ion itself acting as a nucleophile (Kuhlenkoetter et al., 2011). These twocases cannot be distinguished in our experimental conditions (Paper I). Athigh OH- concentrations, the kcat value became saturated and no longer in-creased with pH (Figure 4b). This saturation was not observed before (Kuh-lenkoetter et al., 2011; Shaw et al., 2012) and we took it to identify the reac-tion step other than the ester bond hydrolysis becoming rate-limiting. We could exclude that ka ·[RF] is rate-limiting and tested the possibility of a ratelimiting kdiss at high OH- concentrations. Since the main change of fluores-cence in stopped-flow experiments occurs when the peptide dissociates from the ribosome into solution, we also applied the quench-flow method to the termination reaction. This method is blind for peptide dissociation from the ribosome after ester bond hydrolysis. Since the saturated kcat value at high OH- concentration was the same (around 60s-1; Figure 4c), we concluded that (i) at low pH the rate-limiting step is the ester bond hydrolysis and (ii) at high pH the reaction is limited by a previously unobserved pH-independentstep. In principle, another possible explanation could have been titration of a catalytically essential group in the PTC, but no such groups have been iden-tified. Therefore we suggested that the rate-limiting step at high pH (kconf) is the previously proposed conformational change in class I release factors upon stop codon recognition (Rawat et al., 2003). Indirect evidence that rate constant kconf corresponded to a reaction step, not occurring in the PTC,came from comparing pH dependence of methylated and unmethylated re-lease factors (Paper I). It had previously been shown that methylation ofglutamine in the GGQ motif increases the catalytic rate of ester bond hy-drolysis around two-fold (Dincbas-Renqvist et al., 2000). In our experiments the kcat value was increased by RF-methylation at pH 7.5 (Table 2 in Paper I), but not at high pH (Table 1 in Paper I). This indicated that the saturated rate corresponded to a process not occurring in the PTC.

Further indirect support for this model came from Förster resonance ener-gy transfer (FRET) experiments that indicated compact form for RF1 in so-lution and a switch to extended form on the ribosome in the presence of a stop codon (Trappl and Joseph, 2016). Domains II and III of RF1 were la-belled with a fluorophore and a quencher pair, so that a change in confor-mation would be represented by a change in fluorescent signal. However, the maximum observable change in these experiments was only 20 Å, around halfway from the compact to extended conformation. These results, eventhough strongly indicating that the conformational change occurs, were notsufficient to unambiguously prove its existence.

23




23




23

At low pH, kcat increased with OH- concentration, as observed before (Kuhlenkoetter et al., 2011; Shaw et al., 2012). The linear dependence can be explained either by titration of a catalytically essential group, or by the OH- ion itself acting as a nucleophile (Kuhlenkoetter et al., 2011). These two cases cannot be distinguished in our experimental conditions (Paper I). At high OH- concentrations, the kcat value became saturated and no longer in-creased with pH (Figure 4b). This saturation was not observed before (Kuh-lenkoetter et al., 2011; Shaw et al., 2012) and we took it to identify the reac-tion step other than the ester bond hydrolysis becoming rate-limiting. We could exclude that ka ·[RF] is rate-limiting and tested the possibility of a rate limiting kdiss at high OH- concentrations. Since the main change of fluores-cence in stopped-flow experiments occurs when the peptide dissociates from the ribosome into solution, we also applied the quench-flow method to the termination reaction. This method is blind for peptide dissociation from the ribosome after ester bond hydrolysis. Since the saturated kcat value at high OH- concentration was the same (around 60s-1; Figure 4c), we concluded that (i) at low pH the rate-limiting step is the ester bond hydrolysis and (ii) at high pH the reaction is limited by a previously unobserved pH-independent step. In principle, another possible explanation could have been titration of a catalytically essential group in the PTC, but no such groups have been iden-tified. Therefore we suggested that the rate-limiting step at high pH (kconf) is the previously proposed conformational change in class I release factors upon stop codon recognition (Rawat et al., 2003). Indirect evidence that rate constant kconf corresponded to a reaction step, not occurring in the PTC, came from comparing pH dependence of methylated and unmethylated re-lease factors (Paper I). It had previously been shown that methylation of glutamine in the GGQ motif increases the catalytic rate of ester bond hy-drolysis around two-fold (Dincbas-Renqvist et al., 2000). In our experiments the kcat value was increased by RF-methylation at pH 7.5 (Table 2 in Paper I), but not at high pH (Table 1 in Paper I). This indicated that the saturated rate corresponded to a process not occurring in the PTC.

Further indirect support for this model came from Förster resonance ener-gy transfer (FRET) experiments that indicated compact form for RF1 in so-lution and a switch to extended form on the ribosome in the presence of a stop codon (Trappl and Joseph, 2016). Domains II and III of RF1 were la-belled with a fluorophore and a quencher pair, so that a change in confor-mation would be represented by a change in fluorescent signal. However, the maximum observable change in these experiments was only 20 Å, around halfway from the compact to extended conformation. These results, even though strongly indicating that the conformational change occurs, were not sufficient to unambiguously prove its existence.

24

Structural evidence of conformational change in RFs (Paper II) With these results in mind, we set out to capture the structures of compact release factors on the ribosome. The field of structural biology has been rap-idly developing since the idea of this conformational change was first pro-posed and a method capable of capturing the short-lived compact state of ribosome-bound release factors at high resolution has become available. We used time-resolved cryogenic electron microscopy (time-resolved cryo-EM) developed by our coworkers in prof. Joachim Frank’s laboratory (Colombia University, New York, USA).

Time-resolved cryo-EM combines kinetic and structural studies by cap-turing biological complexes at different time points of a reaction. Reactants are mixed and sprayed on a cryo-EM grid which is then rapidly plunged into a freezing solution. The captured biological complexes are visualized by transmission electron microscopy, followed by three-dimensional recon-struction (Chen and Frank, 2016). This technique allows the capturing of states as short-lived as 20 ms.

Firstly, using the same method as in Paper I, we determined the rate of the hypothetical conformational change of RF1 and RF2 in the same reaction conditions as would be used for time-resolved cryo-EM: polymix-HEPES buffer and 25⁰C (Figure 5). We used the estimated rate constants to model the steps of the peptide release reaction (Figure 6a) as consecutive first-order reaction steps (Kahley and Novak, 1996). From the model (Figure 6 b-c) we

Figure 5. The maximal rate (kcat) of peptide release by RF2 at different OH- concen-trations in 25⁰C. (a) Extent of RF2-dependent fMet-Phe-Phe peptide release from P-site bound fMet-Phe-Phe-tRNA (y axis) versus time (x axis) at different OH-

concentrations. 0.02 µM release complexes were reacted with saturating (0.8 µM) RF2 concentrations. (b) Maximal rate (kcat) of peptide release by RF2 (y-axis) in-creased to a plateau value (kconf = 11 ± 1 s-1) with increasing [OH-] (x-axis).

24






24






24






25

could determine how the fractions of different termination complexes would change over time after mixing the release complexes with the RFs in a time-resolved cryo-EM instrument. According to these simulations, after 20 msfrom reaction start, majority of both RF1 and RF2 should be bound to theribosome in compact conformation. After 60 ms from reaction start, majority of RFs should be bound to the ribosome in extended forms (Figure 6 b-c). These predictions were confirmed by the time-resolved cryo-EM results, even though the fractions of compact RFs were smaller than anticipated fromthe modelling (Figure 6 d-e).

Figure 6. Time evolution of ribosome ensembles in termination of translation (a) Visualization of the pathway from RF-free release complex (RC0) to peptide release. Compact release factor (RF) binds to RC0 and forms the RC·RFcompact complex with compounded rate constant ka·[RFfree]. Stop codon recognition induces conformation-al change in the RF which brings the complex RC·RFcompact to RC·RFextended with rate constant kconf. The ester bond between the peptide and the P-site tRNA is hydrolyzed with rate constant khydr (b-c) Predicted fractions of ribosomes in RC0, RC·RFcompactand RC·RFextended forms (y-axis) as functions of time (x-axis). (d-e) Populations of RC·RFcompact (red) and RC·RFextended (blue) at 20 ms, 60 ms and long incubation timepoints as obtained by time-resolved cryo-EM.

25



25



25

could determine how the fractions of different termination complexes would change over time after mixing the release complexes with the RFs in a time-resolved cryo-EM instrument. According to these simulations, after 20 ms from reaction start, majority of both RF1 and RF2 should be bound to the ribosome in compact conformation. After 60 ms from reaction start, majority of RFs should be bound to the ribosome in extended forms (Figure 6 b-c). These predictions were confirmed by the time-resolved cryo-EM results, even though the fractions of compact RFs were smaller than anticipated from the modelling (Figure 6 d-e).

Figure 6. Time evolution of ribosome ensembles in termination of translation (a) Visualization of the pathway from RF-free release complex (RC0) to peptide release. Compact release factor (RF) binds to RC0 and forms the RC·RFcompact complex with compounded rate constant ka·[RFfree]. Stop codon recognition induces conformation-al change in the RF which brings the complex RC·RFcompact to RC·RFextended with rate constant kconf. The ester bond between the peptide and the P-site tRNA is hydrolyzed with rate constant khydr (b-c) Predicted fractions of ribosomes in RC0, RC·RFcompact and RC·RFextended forms (y-axis) as functions of time (x-axis). (d-e) Populations of RC·RFcompact (red) and RC·RFextended (blue) at 20 ms, 60 ms and long incubation time points as obtained by time-resolved cryo-EM.

26

Compact forms of RF1 and RF2 on the ribosome were captured with 3.5 - 4 Å resolution (Paper II). In the compact structure, the GGQ motif was around 60 Å away from the PTC and positioned next to domain II of the RFand the anticodon stem of the P-site tRNA (Figure 7 a-b). The overall posi-tion of domain III is similar to X-ray structures of RFs in solution.

Figure 7. Time-resolved cryo-EM structures of E.coli 70S ribosome with RF1. (a-b) RF1 bound to the ribosome in a compact conformation. (d-e) RF1 bound to the ribo-some in an extended conformation. Light blue: 50S large subunit; light gold: 30S small subunit; green: tripeptide; orange: P-site tRNA; pink: mRNA; red: compact RF1; dark blue: extended RF1. Conformations of the switch loop region in the com-pact (c) and extended (f) RF1 conformation on the ribosome. Gold: A1492 and A1493 rRNA nucleotides; and orange: S12 protein. Structures by Zi-ao Jack Fu.

At 60 ms, the GGQ motif is placed in the PTC (Figure 7 d-e). The position of domain III in relation to domain II is changed by a rearrangement of a switch loop region, connecting the domains II and III (Korostelev et al., 2008; Laurberg et al., 2008). In the extended structure the switch loop is stabilized by interactions within a pocket formed by ribosomal proteinS12, the loop of 16S rRNA, the β-sheet of domain II, and A1493 and A1913 rRNA residues (Figure 7f). No interaction with S12 was observed for the compact structures (Figure 7c).

26




26




26

Compact forms of RF1 and RF2 on the ribosome were captured with 3.5 - 4 Å resolution (Paper II). In the compact structure, the GGQ motif was around 60 Å away from the PTC and positioned next to domain II of the RF and the anticodon stem of the P-site tRNA (Figure 7 a-b). The overall posi-tion of domain III is similar to X-ray structures of RFs in solution.


At 60 ms, the GGQ motif is placed in the PTC (Figure 7 d-e). The position of domain III in relation to domain II is changed by a rearrangement of a switch loop region, connecting the domains II and III (Korostelev et al., 2008; Laurberg et al., 2008). In the extended structure the switch loop is stabilized by interactions within a pocket formed by ribosomal protein S12, the loop of 16S rRNA, the β-sheet of domain II, and A1493 and A1913 rRNA residues (Figure 7f). No interaction with S12 was observed for the compact structures (Figure 7c).

27

While these results were being prepared for publication, an X-ray struc-ture of a compact RF1 bound to a stop codon was published (Svidritskiy and Korostelev, 2018). Svidritskiy and Korostelev had mutated the switch loop region to slow down the conformational change and used the antibiotic blas-ticidin S to block the GGQ motif from docking into the PTC. In their struc-ture the decoding region is in a conformation more similar to that of extend-ed RFs in our structures, but the domain III is in a position similar to that in the compact RFs (Supplementary Figure 4b in Paper II). It is possible, that their structure could resemble an intermediate on the pathway from the com-pact to extended RF (Paper II).

27


27


27


28

The accuracy of yeast termination

The role of eRF3 in eukaryotic termination Termination of eukaryotic protein synthesis is more complex than the corre-sponding process in bacteria and employs more protein factors. The role of each factor and the sequence of termination events in eukaryotes have beenextensively studied in recent years (reviewed in (Hellen, 2018; Jackson et al., 2012), but major questions remain unanswered. One of them is how eRF1 recognizes all three stop codons (UAA, UAG, UGA), but effectively discriminates against the tryptophan-encoding UGG codon.

It has been proposed that stop codon recognition and peptide release by eRF1 is coupled by GTP-ase activity of eRF3 (Salas-Marco and Bedwell,2004). eRF3 forms a complex with eRF1 in solution (Stansfield et al., 1995;Zhouravleva et al., 1995) and has been proposed to enter the ribosomal A site as a part of the eRF1·eRF3·GTP complex (Alkalaeva et al., 2006). How-ever, in their recent paper Beissel and coworkers suggest that helicase Dbp5, and not eRF3, delivers eRF1 to the already eRF3-bound ribosome (Beissel et al., 2019). eRF3 accelerates peptide release by eRF1 (Alkalaeva et al., 2006; Shoemaker and Green, 2011; Zhouravleva et al., 1995), especially when eRF1 is in sub-stoichiometric concentrations to the ribosomes (Eyler et al., 2013).

When bound on the ribosome in a pre-termination complex with eRF3 and non-hydrolysable GTP analog, eRF1 is in catalytically inactive confor-mation with the GGQ motif away from the PTC (Preis et al., 2014). Howev-er, eRF1 is in catalytically active conformation when bound on the ribosomein a post-termination complex with the eukaryotic ribosome recycling factorABCE1 (Brown et al., 2015; Preis et al., 2014). These findings are compati-ble with a model where GTP hydrolysis by eRF3 induces a conformational change to catalytically active conformation in eRF1. This change is followed by A-site binding of ABCE1 (Rli1 in yeast) and probably dissociation ofeRF3 (reviewed in (Hellen, 2018). ABCE1 then stimulates the catalytic stepof peptide release (Shoemaker and Green, 2011) and after peptide releasesplits the ribosome into subunits (Pisarev et al., 2010; Shoemaker and Green, 2011).

The effect of eRF3 on the accuracy of termination by eRF1 We have measured the accuracy of termination by eRF1 as a ratio of effi-ciencies (kcat/Km) on stop and sense codons in a yeast (S. cerevisiae) in vitro translation system (see Methods). Our preliminary results show that eRF1 alone could not efficiently discriminate between stop codon UAA and sense

28






28






28


The role of eRF3 in eukaryotic termination Termination of eukaryotic protein synthesis is more complex than the corre-sponding process in bacteria and employs more protein factors. The role of each factor and the sequence of termination events in eukaryotes have been extensively studied in recent years (reviewed in (Hellen, 2018; Jackson et al., 2012), but major questions remain unanswered. One of them is how eRF1 recognizes all three stop codons (UAA, UAG, UGA), but effectively discriminates against the tryptophan-encoding UGG codon.

It has been proposed that stop codon recognition and peptide release by eRF1 is coupled by GTP-ase activity of eRF3 (Salas-Marco and Bedwell, 2004). eRF3 forms a complex with eRF1 in solution (Stansfield et al., 1995; Zhouravleva et al., 1995) and has been proposed to enter the ribosomal A site as a part of the eRF1·eRF3·GTP complex (Alkalaeva et al., 2006). How-ever, in their recent paper Beissel and coworkers suggest that helicase Dbp5, and not eRF3, delivers eRF1 to the already eRF3-bound ribosome (Beissel et al., 2019). eRF3 accelerates peptide release by eRF1 (Alkalaeva et al., 2006; Shoemaker and Green, 2011; Zhouravleva et al., 1995), especially when eRF1 is in sub-stoichiometric concentrations to the ribosomes (Eyler et al., 2013).

When bound on the ribosome in a pre-termination complex with eRF3 and non-hydrolysable GTP analog, eRF1 is in catalytically inactive confor-mation with the GGQ motif away from the PTC (Preis et al., 2014). Howev-er, eRF1 is in catalytically active conformation when bound on the ribosome in a post-termination complex with the eukaryotic ribosome recycling factor ABCE1 (Brown et al., 2015; Preis et al., 2014). These findings are compati-ble with a model where GTP hydrolysis by eRF3 induces a conformational change to catalytically active conformation in eRF1. This change is followed by A-site binding of ABCE1 (Rli1 in yeast) and probably dissociation of eRF3 (reviewed in (Hellen, 2018). ABCE1 then stimulates the catalytic step of peptide release (Shoemaker and Green, 2011) and after peptide release splits the ribosome into subunits (Pisarev et al., 2010; Shoemaker and Green, 2011).


29

codon UGG, but terminated with very high accuracy in the presence of eRF3.

0.04 µM of purified 80S release complexes, containing fMet-Phe-Tyr-tRNATyr in the P site and UAA or UGG codon in the A site were reacted to increasing amounts or eRF1, either in presence or absence of saturating amounts of eRF3 (10 µM). The reactions were performed in polymix-HEPES buffer, supplemented with energy regeneration system at 30⁰C. The experiments were performed either in quench-flow or by hand and stopped at different time points by quenching with formic acid. Precipitated peptidyl-tRNA was separated from soluble released peptide by centrifugation and the amount of released and non-released peptide determined by scintillation counting of 3H radiation.

The plots of released peptide over time were biphasic with the fast phase, corresponding to fMet-Phe-Tyr peptide release, comprising around 50% of total amplitude (a-b in Figures 8 and 9) in our preliminary analysis presented here. The slow phase most likely corresponded to the release of shorter pep-tides.

The results showed a strong effect of eRF3 on the accuracy of termination by eRF1. The accuracy was expressed as the ratio of efficiencies (kcat/Km) of peptide release on UAA and UGG codons. On its own, eRF1 discriminated between UAA and UGG with the accuracy of only 4 (Table 1; Figure 8). It means that, assuming equal concentrations of ribosomes containing these two codons, eRF1 would be only 4 times less likely to terminate on UGG than on UAA codon. However, with eRF3 and GTP present, eRF1 discrimi-nated against UGG with the accuracy of 10,400 (Table 1; Figure 9).

Table 1. Accuracy of termination by eRF1 with and without eRF3.

Protein UAA UGG Accuracy kcat/Km (s-1µM-1) kcat (s-1) kcat/Km (s-1 µM-1) eRF1 0.04 ± 0.008 0.35 ± 0.04 0.01 ± 0.003 4

± 1.4 eRF1·eRF3 15.6 ± 4 0.77 ± 0.05 0.0015 ± 0.0001 10400

± 2750 Fold increase (↑) or decrease (↓) 390↑ 2↑ 7↓ 2600↑

29







± 1.4 eRF1·eRF3 15.6 ± 4 0.77 ± 0.05 0.0015 ± 0.0001 10400


29







± 1.4 eRF1·eRF3 15.6 ± 4 0.77 ± 0.05 0.0015 ± 0.0001 10400


29







± 1.4 eRF1·eRF3 15.6 ± 4 0.77 ± 0.05 0.0015 ± 0.0001 10400


30

Figure 8. Peptide release by eRF1 on UAA and UGG codons. (a-b) Percentage of released fMet-Phe-Tyr peptide (y axis) over time (x axis, log10 display) at different eRF1 concentrations. (c-d) The rate constant of peptide release (krel) (y axis) over eRF1 concentration (x axis) on UAA and UGG codons. All reactions were per-formed with 0.04 µM release complexes containing fMet-Phe-Tyr-tRNATyr in the P site and indicated codon in the A site in Polymix-HEPES buffer with 2.2 mM GTP and energy regeneration system at 30⁰C.

The accuracy of termination on UGG versus UAA with eRF3 present (10,400) was comparable to that measured in an in vitro system with E. coli components: accuracy of 13,000 by RF1 and 2400 by RF2 (Freistroffer et al., 2000). This effect was mainly caused by almost 400-fold increase of kcat/Km on UAA, but also by moderate reduction of kcat/Km on UGG (Table 1). This finding is in line with molecular dynamics simulations performed by Lind and coworkers that predict the energetic penalty of reading UGG com-pared to reading UAA to be 6.6 kcal/mol, or a reduction of termination effi-ciency by around 3 orders of magnitude (Lind et al., 2017).

eRF3 also increased the kcat of termination 2-fold: from 0.35 s-1 to 0.77 s-1 (Table 1), comparable to 4-fold kcat increase reported by Eyler and cowork-ers for Met-Lys peptide release (Eyler et al., 2013). One would expect the in

30




30




30




31

vivo kcat values to be higher than observed in our experiments, because our in vitro translation system did not include Rli1 and eIF5A, both reported to increase the kcat of peptide release (Schuller et al., 2017; Shoemaker andGreen, 2011).

Figure 9. Peptide release by eRF1 in complex with eRF3 and GTP. (a-b) Percentage of released fMet-Phe-Tyr peptide (y axis) over time (x axis, log10 display) at differ-ent eRF1 concentrations. (c-d) The rate constant of peptide release (krel) (y-axis) over eRF1 concentration (x-axis) on UAA and UGG codons. All reactions were performed with saturating amounts (10 µM) eRF3, 0.04 µM release complexes con-taining fMet-Phe-Tyr-tRNATyr in the P site and indicated codon in the A site in Polymix-HEPES buffer with 2.2 mM GTP and energy regeneration system at 30⁰C.

The observed large increase of efficiency and accuracy of termination by eRF1 caused by the presence of eRF3 is comparable to the effect of EF-Tu on tRNA selection in bacterial elongation. The low intrinsic selectivity of eRF1 interaction with a stop codon or tRNA-only interaction with a sense codon (Ieong et al., 2016) is strongly enhanced by the presence of respec-

31




31




31

vivo kcat values to be higher than observed in our experiments, because our in vitro translation system did not include Rli1 and eIF5A, both reported to increase the kcat of peptide release (Schuller et al., 2017; Shoemaker and Green, 2011).



32

tively, eRF3 of EF-Tu. The large increase in kcat/Km might indicate that eRF3 is required for efficient binding of eRF1 to the ribosome. However, this model is at odds with the recent suggestion that eRF1 is delivered to the ribosome by another protein, helicase Dbp5 (Beissel et al., 2019). More ex-periments and further improvements to our yeast in vitro translation system are needed to precisely define the role of eRF3 in eukaryotic termination.

32


32


32


33

The effect of mRNA modifications on elongation and termination phases of protein synthesis

Modified RNA nucleotides are abundant and functionally important in trans-fer RNA and ribosomal RNA (reviewed in(Ontiveros et al., 2019). Recently, some of these modifications have also been identified in messenger RNA (Table 1 in (Nachtergaele and He, 2018) and there is a growing body of evi-dence for their importance in regulation of gene expression. Less is known about the effect these modifications have on protein synthesis. In this thesis work, we have tested the effect of 2’-O-methylation on elongation (Paper III) and the influence of N6-methylation on termination (Paper IV) of protein synthesis. The modifications are shown in Figure 11.

Figure 10. Modifications on mRNA. (a) 2’-O-methylation, (b) N6-methylation on adenosine. Added methyl groups are shown in red.

2’-O-methylation of mRNA in bacterial elongation (Paper III) Recently, many 2’-O-methylation sites have been identified in human mRNA: around 2000 sites in HeLa and around 700 in HEK293 cell lines(Dai et al., 2017). 2’-O-methylation sites are the most common in 5’ untrans-lated and protein coding regions, while the most frequently modified nucleo-tide is uracil: around 60% of all 2’-O-methylated sites (Dai et al., 2017). Thefunction of this modification is not completely clarified, but it has been

33





33





33


Modified RNA nucleotides are abundant and functionally important in trans-fer RNA and ribosomal RNA (reviewed in(Ontiveros et al., 2019). Recently, some of these modifications have also been identified in messenger RNA (Table 1 in (Nachtergaele and He, 2018) and there is a growing body of evi-dence for their importance in regulation of gene expression. Less is known about the effect these modifications have on protein synthesis. In this thesis work, we have tested the effect of 2’-O-methylation on elongation (Paper III) and the influence of N6-methylation on termination (Paper IV) of proteinsynthesis. The modifications are shown in Figure 10.


2’-O-methylation of mRNA in bacterial elongation (Paper III) Recently, many 2’-O-methylation sites have been identified in human mRNA: around 2000 sites in HeLa and around 700 in HEK293 cell lines (Dai et al., 2017). 2’-O-methylation sites are the most common in 5’ untrans-lated and protein coding regions, while the most frequently modified nucleo-tide is uracil: around 60% of all 2’-O-methylated sites (Dai et al., 2017). The function of this modification is not completely clarified, but it has been

34

shown to help eukaryotic cells to discriminate between their own and viralmRNA (Daffis et al., 2010; Zust et al., 2011) and to mark correctly capped mRNAs, thus protecting them from degradation (Picard-Jean et al., 2018). 2’-O-methylation reduces peptide yield and induces ribosome stalling when the modification is in the second position of the codon (Hoernes et al., 2016)through an unknown mechanism.

In order to obtain a clear picture of how 2’-O-methylation affects differ-ent elongation steps, we have used E. coli in vitro translation system for measuring initial selection and proofreading values (see Methods) on modi-fied and unmodified codons. Initial selection and proofreading stages ofamino acid incorporation are separated by irreversible GTP hydrolysis onEF-Tu. Most non-cognate tRNAs are rejected at the initial selection stage. During the proofreading stage, codon-anticodon interaction is monitored again, so that only cognate aa-tRNAs proceed with high probability to A-site accommodation and peptide bond formation, while the small fraction ofremaining non-cognate tRNAs is further culled by dissociation through theproofreading pathway. The steps from T3 binding to tRNA accommodation and the effect of 2’-O-methylation, discussed below, are summarized in Fig-ure 11.

We used ribosomes programmed with 3H-labelled initiator fMet-tRNAfMet

in the P-site and lysine-encoding AAA codon in the A site. This codon was either unmodified (AAA) or 2’-O-methylated at the second position(AAmA). The ribosomes were reacted to a ternary complex (T3), consisting of Lys-tRNALys, EF-Tu and GTP. The amounts of peptides formed and GTPshydrolyzed over time were determined by HPLC.

The efficiency (kcat/Km) of initial codon selection was reduced 300-fold on AAmA codon, compared to unmodified AAA (Figure 4 in Paper III). The proofreading factor was increased 5-fold, compared to that of the unmodified codon (Figure 5 in Paper III). The effect on proofreading was reduced by increasing free Mg2+ concentration (Figure 5 in Paper III), which is in linewith the modification mainly affecting the first, Mg2+ dependent, proofread-ing step (Ieong et al., 2016). These results show that with the AAmA codon in the A site, the cognate Lys-tRNALys was 300 times more likely to be re-jected in the initial selection stage and 5 times more likely to be rejected in the proofreading stage. These numbers in combination estimate a total modi-fication-induced efficiency loss in AAA codon reading by its cognate tRNA to a factor of 1500.

Since the modified codon could not be efficiently read by its cognate tRNA, its presence most likely would cause the ribosome to stall. Indeed, single-molecule FRET measurements performed by our co-workers from thePuglisi group (Stanford University, Palo Alto, USA) showed long stalling times on modified codons and multiple unproductive Lys-tRNALys binding events that did not result in peptide bond formation (Paper III). This effectwas most likely caused by the modification disrupting the interaction of 16S

34







34







34

shown to help eukaryotic cells to discriminate between their own and viral mRNA (Daffis et al., 2010; Zust et al., 2011) and to mark correctly capped mRNAs, thus protecting them from degradation (Picard-Jean et al., 2018). 2’-O-methylation reduces peptide yield and induces ribosome stalling when the modification is in the second position of the codon (Hoernes et al., 2016) through an unknown mechanism.

In order to obtain a clear picture of how 2’-O-methylation affects differ-ent elongation steps, we have used E. coli in vitro translation system for measuring initial selection and proofreading values (see Methods) on modi-fied and unmodified codons. Initial selection and proofreading stages of amino acid incorporation are separated by irreversible GTP hydrolysis on EF-Tu. Most non-cognate tRNAs are rejected at the initial selection stage. During the proofreading stage, codon-anticodon interaction is monitored again, so that only cognate aa-tRNAs proceed with high probability to A-site accommodation and peptide bond formation, while the small fraction of remaining non-cognate tRNAs is further culled by dissociation through the proofreading pathway. The steps from T3 binding to tRNA accommodation and the effect of 2’-O-methylation, discussed below, are summarized in Fig-ure 11.

We used ribosomes programmed with 3H-labelled initiator fMet-tRNAfMet in the P-site and lysine-encoding AAA codon in the A site. This codon was either unmodified (AAA) or 2’-O-methylated at the second position (AAmA). The ribosomes were reacted to a ternary complex (T3), consisting of Lys-tRNALys, EF-Tu and GTP. The amounts of peptides formed and GTPs hydrolyzed over time were determined by HPLC.

The efficiency (kcat/Km) of initial codon selection was reduced 300-fold on AAmA codon, compared to unmodified AAA (Figure 4 in Paper III). The proofreading factor was increased 5-fold, compared to that of the unmodified codon (Figure 5 in Paper III). The effect on proofreading was reduced by increasing free Mg2+ concentration (Figure 5 in Paper III), which is in line with the modification mainly affecting the first, Mg2+ dependent, proofread-ing step (Ieong et al., 2016). These results show that with the AAmA codon in the A site, the cognate Lys-tRNALys was 300 times more likely to be re-jected in the initial selection stage and 5 times more likely to be rejected in the proofreading stage. These numbers in combination estimate a total modi-fication-induced efficiency loss in AAA codon reading by its cognate tRNA to a factor of 1500.

Since the modified codon could not be efficiently read by its cognate tRNA, its presence most likely would cause the ribosome to stall. Indeed, single-molecule FRET measurements performed by our co-workers from the Puglisi group (Stanford University, Palo Alto, USA) showed long stalling times on modified codons and multiple unproductive Lys-tRNALys binding events that did not result in peptide bond formation (Paper III). This effect was most likely caused by the modification disrupting the interaction of 16S

35

rRNA monitoring bases (A1942, A1943 and G530) with the codon-anticodon helix in the A site (Ogle et al., 2001). Both bulk kinetics and sin-gle molecule FRET results showed that the effect of 2’-O-methylation can bereverted by the introduction of error-inducing antibiotics like paromomycin and neomycin (Figure 6 in Paper III). These antibiotics induce translation errors by hyper-activating the monitoring bases (Ogle and Ramakrishnan,2005).

Figure 11. Mechanism of elongation stall, induced by 2’-O- methylation of mRNA (Paper III).

2’-O-methylated sites present in the untranslated region near the 5’capserve as markers indicating that mRNA is correctly capped and native to thecell and thus should not be degraded or induce an immune response (Daffis et al., 2010; Picard-Jean et al., 2018; Zust et al., 2011). Since this region is not translated by the ribosome, the profound reduction of translation rate caused by 2’-O-methylation is of no consequence in this case. It is rather more puzzling that such a modification would also be present in the coding sequences, as has been recently shown (Dai et al., 2017). It is possible, that 2’-O-methylation sites are used to modulate the translation rate, for example by inducing a pause that would allow a nascent peptide to fold correctly.Alternatively, ribosome stalling at 2’-O-methylated sites might recruit spe-cialized protein factors that enhance decoding of the modified codons, simi-lar to translation of polyproline stretches in bacteria by elongation factor P (EF-P) (Doerfel et al., 2013; Ude et al., 2013).

N6-methylation of mRNA in bacterial termination (Paper IV) N6-methyladenosine (m6A) is the most abundant mRNA modification, pre-sent in all three kingdoms of life (reviewed in (Roignant and Soller, 2017). It is of a particular interest, because abnormal changes in N6-methylation pat-terns were shown to play part in the development of human diseases, such asleukemia (Li et al., 2017) and human immunodeficiency virus infection

35





35





35

rRNA monitoring bases (A1942, A1943 and G530) with the codon-anticodon helix in the A site (Ogle et al., 2001). Both bulk kinetics and sin-gle molecule FRET results showed that the effect of 2’-O-methylation can be reverted by the introduction of error-inducing antibiotics like paromomycin and neomycin (Figure 6 in Paper III). These antibiotics induce translation errors by hyper-activating the monitoring bases (Ogle and Ramakrishnan, 2005).


2’-O-methylated sites present in the untranslated region near the 5’cap serve as markers indicating that mRNA is correctly capped and native to the cell and thus should not be degraded or induce an immune response (Daffis et al., 2010; Picard-Jean et al., 2018; Zust et al., 2011). Since this region is not translated by the ribosome, the profound reduction of translation rate caused by 2’-O-methylation is of no consequence in this case. It is rather more puzzling that such a modification would also be present in the coding sequences, as has been recently shown (Dai et al., 2017). It is possible, that 2’-O-methylation sites are used to modulate the translation rate, for example by inducing a pause that would allow a nascent peptide to fold correctly. Alternatively, ribosome stalling at 2’-O-methylated sites might recruit spe-cialized protein factors that enhance decoding of the modified codons, simi-lar to translation of polyproline stretches in bacteria by elongation factor P (EF-P) (Doerfel et al., 2013; Ude et al., 2013).

N6-methylation of mRNA in bacterial termination (Paper IV) N6-methyladenosine (m6A) is the most abundant mRNA modification, pre-sent in all three kingdoms of life (reviewed in (Roignant and Soller, 2017). It is of a particular interest, because abnormal changes in N6-methylation pat-terns were shown to play part in the development of human diseases, such as leukemia (Li et al., 2017) and human immunodeficiency virus infection

36

(Kennedy et al., 2016; Lichinchi et al., 2016; Tirumuru et al., 2016). The main function of N6-methylation in mRNA seems to be the fine-tuning of gene expression by affecting mRNA stability, alternative splicing and trans-lation in response to changing environment or developmental status (re-viewed in (Nachtergaele and He, 2018; Roignant and Soller, 2017).

The m6A modification induces cognate tRNA rejection during initial se-lection and proofreading stages of elongation (Choi et al., 2016), but to a much lesser degree than 2’-O-methylation described above. The presence of m6A perturbs the canonical A-U base pairing in the codon-anticodon duplex (Choi et al., 2016), an earlier step than the interaction of monitoring bases, disrupted by 2’-O-methylation (Choi et al., 2018). Our decision to look intothe effect of m6A on termination was influenced by studies, showing thatm6A is enriched around stop codons (Meyer et al., 2012). Here, we used a bacterial in vitro translation system to test the effect of m6A on termination by RF2 (Paper IV). Purified ribosomal release complexes containing fMet-Phe-Tyr-tRNATyr in the P site were reacted with RF2 manually or in a quench-flow instrument, in order to determine the termination rate constants on modified and unmodified codons.

RF2, which normally terminates at UAA and UGA codons, strongly se-lected against termination at UAG, as shown before (Freistroffer et al.,2000). Indeed, RF2 preferred termination at its cognate UAA codon with an accuracy of 14,000 (Table 2). When the m6A modification was introduced into the 2nd position of the UAA codon (Um6AA), there was no significanteffect on the catalytic rate constant kcat value of peptide release, but the effi-ciency kcat/Km was reduced almost 7-fold (Table 2). These results show thatthe modification has no effect on the rate constant for hydrolysis of the esterbond, connecting the peptide to the P-site tRNA, but most likely affects theefficiency of RF2 binding to the A site and decoding the stop codon. When the m6A modification was introduced into the 2nd position of the non-cognate UAG codon (Um6AG), both kcat and kcat/Km values were reduced about 2-fold (Table 2).

Table 2. Kinetic parameters of peptide release by RF2 on N6-methylated codons.

Rate constants of peptide release Accuracy (A)

Codon kcat (s-1) kcat/Km (µM-1 s-1) Codon pair Accuracy

UAA 2.8 ± 0.18 70 ± 13 UAA / UAG 14,000 ± 3256

Um6AA 3.0 ± 0.06 11 ± 0.5 Um6AA / UAG 2200 ± 324

UAG 0.06 ± 0.006 0.005 ± 0.0007 UAA / Um6AA 6 ± 1

Um6AG 0.03 ± 0.013 0.002 ± 0.0002 UAG / Um6AG 2.5 ± 0.4

36







UAA 2.8 ± 0.18 70 ± 13 UAA / UAG 14,000 ± 3256

Um6AA 3.0 ± 0.06 11 ± 0.5 Um6AA / UAG 2200 ± 324

UAG 0.06 ± 0.006 0.005 ± 0.0007 UAA / Um6AA 6 ± 1

Um6AG 0.03 ± 0.013 0.002 ± 0.0002 UAG / Um6AG 2.5 ± 0.4

36







UAA 2.8 ± 0.18 70 ± 13 UAA / UAG 14,000 ± 3256

Um6AA 3.0 ± 0.06 11 ± 0.5 Um6AA / UAG 2200 ± 324

UAG 0.06 ± 0.006 0.005 ± 0.0007 UAA / Um6AA 6 ± 1

Um6AG 0.03 ± 0.013 0.002 ± 0.0002 UAG / Um6AG 2.5 ± 0.4

36


The m6A modification induces cognate tRNA rejection during initial se-lection and proofreading stages of elongation (Choi et al., 2016), but to a much lesser degree than 2’-O-methylation described above. The presence of m6A perturbs the canonical A-U base pairing in the codon-anticodon duplex (Choi et al., 2016), an earlier step than the interaction of monitoring bases, disrupted by 2’-O-methylation (Choi et al., 2018). Our decision to look into the effect of m6A on termination was influenced by studies, showing that m6A is enriched around stop codons (Meyer et al., 2012). Here, we used a bacterial in vitro translation system to test the effect of m6A on termination by RF2 (Paper IV). Purified ribosomal release complexes containing fMet-Phe-Tyr-tRNATyr in the P site were reacted with RF2 manually or in a quench-flow instrument, in order to determine the termination rate constants on modified and unmodified codons.

RF2, which normally terminates at UAA and UGA codons, strongly se-lected against termination at UAG, as shown before (Freistroffer et al., 2000). Indeed, RF2 preferred termination at its cognate UAA codon with an accuracy of 14,000 (Table 2). When the m6A modification was introduced into the 2nd position of the UAA codon (Um6AA), there was no significant effect on the catalytic rate constant kcat value of peptide release, but the effi-ciency kcat/Km was reduced almost 7-fold (Table 2). These results show that the modification has no effect on the rate constant for hydrolysis of the ester bond, connecting the peptide to the P-site tRNA, but most likely affects the efficiency of RF2 binding to the A site and decoding the stop codon. When the m6A modification was introduced into the 2nd position of the non-cognate UAG codon (Um6AG), both kcat and kcat/Km values were reduced about 2-fold (Table 2).




UAA 2.8 ± 0.18 70 ± 13 UAA / UAG 14,000 ± 3256

Um6AA 3.0 ± 0.06 11 ± 0.5 Um6AA / UAG 2200 ± 324

UAG 0.06 ± 0.006 0.005 ± 0.0007 UAA / Um6AA 6 ± 1

Um6AG 0.03 ± 0.013 0.002 ± 0.0002 UAG / Um6AG 2.5 ± 0.4

37

These results are in agreement with the effect of m6A on the accuracy oftranslation elongation. N6-methylation reduced the kcat/Km value for cognate peptidyl transfer reaction, leaving the kcat/Km of non-cognate reaction virtual-ly unaltered (Paper IV). These results suggest that m6A affects elongation and termination stages through a similar mechanism, presumably by disrupt-ing the interaction between the mRNA codon and either tRNA anticodon(Choi et al., 2016) or recognition motif of a release factor.

37


37


37

These results are in agreement with the effect of m6A on the accuracy of translation elongation. N6-methylation reduced the kcat/Km value for cognate peptidyl transfer reaction, leaving the kcat/Km of non-cognate reaction virtual-ly unaltered (Paper IV). These results suggest that m6A affects elongation and termination stages through a similar mechanism, presumably by disrupt-ing the interaction between the mRNA codon and either tRNA anticodon (Choi et al., 2016) or recognition motif of a release factor.

38

Conclusions and future outlook

The aim of the work presented in this thesis has been to answer several ques-tions regarding the accuracy of protein synthesis and its tuning by mRNA modifications. The main conclusions from this work are:

1. Bacterial class I release factors RF1 and RF2 undergo a conformational change upon binding to the stop programmed ribosome, presumably as a mechanism to ensure accurate termination.

2. Yeast class II release factor eRF3 strongly increases the accuracy of stopcodon recognition by eRF1.

3. 2’-O-methylation of sense codons in mRNA induces strong rejection ofcorrect aminoacyl-tRNAs mainly in the initial selection and, to a lesser extent, in the proofreading stages of the elongation phase of protein syn-thesis.

4. N6-methylation of adenosine in mRNA reduces the efficiency, kcat/Km, of cognate termination by RF2, but has no effect on the catalytic rate con-stant, kcat, for peptide release and only minor effect on RF2-dependent termination at non-cognate codons.

The time-resolved cryo-EM structures of ribosomal complexes with RF1 and RF2 in contact with a stop codon in a compact conformation likely pro-vide the definitive proof that these class I RFs undergo a conformationaltransition from compact to open form on the native pathway from free factorand ribosome to factor-induced ester bond hydrolysis on the terminating ribosome. In contrast, there are still many unanswered questions regardingthe exact order of events in eukaryotic termination and individual contribu-tions of participating protein factors.

The data on mRNA modifications have been rapidly accumulating in therecent years and it is very likely that more modifications will be identified in the near future. Their function both in translation and in the general regula-tion of gene expression in the cell have become a new exciting field of re-search. In due time, it might offer therapeutic solutions to diseases caused byaberrant mRNA modifications.

38









38









38



1. Bacterial class I release factors RF1 and RF2 undergo a conformationalchange upon binding to the stop programmed ribosome, presumably as amechanism to ensure accurate termination.


3. 2’-O-methylation of sense codons in mRNA induces strong rejection of correct aminoacyl-tRNAs mainly in the initial selection and, to a lesser extent, in the proofreading stages of the elongation phase of protein syn-thesis.

4. N6-methylation of adenosine in mRNA reduces the efficiency, kcat/Km, ofcognate termination by RF2, but has no effect on the catalytic rate con-stant, kcat, for peptide release and only minor effect on RF2-dependenttermination at non-cognate codons.

The time-resolved cryo-EM structures of ribosomal complexes with RF1 and RF2 in contact with a stop codon in a compact conformation likely pro-vide the definitive proof that these class I RFs undergo a conformational transition from compact to open form on the native pathway from free factor and ribosome to factor-induced ester bond hydrolysis on the terminating ribosome. In contrast, there are still many unanswered questions regarding the exact order of events in eukaryotic termination and individual contribu-tions of participating protein factors.

The data on mRNA modifications have been rapidly accumulating in the recent years and it is very likely that more modifications will be identified in the near future. Their function both in translation and in the general regula-tion of gene expression in the cell have become a new exciting field of re-search. In due time, it might offer therapeutic solutions to diseases caused by aberrant mRNA modifications.

39

Sammanfattning på Svenska

Ribosomen är ett stort makromolekylärt komplex som syntetiserar alla proteiner i alla levande varelsers celler. Proteinerna utför merparten av cellernas livsuppehållande funktioner från katalysering av biokemiska reaktioner till muskelrörelser. En förutsättning för allt liv på jorden är att proteinerna sammanfogas snabbt och noggrannt från aminosyror.

Denna avhandling fokuserar på två frågor angående proteinsyntesens noggrannhet. Hur kan ribosomens termineringsfaktorer i bakterier (RF1 och RF2: RF1/2) och eukaryoter (eRF1) avsluta proteinsyntesen med hög precision? Hur påverkar kemiska modifieringar av budbärar-RNA (mRNA) hastighet och noggrannhet vid översättnng av kodord för aminosyror av tRNA och stoppsignaler av termineringsfaktorer?

RF1/2 känner igen stopp-kodord i mRNA och frigör syntetiseradeproteiner som peptid-kedjor. RF1/2 har antagits får hög noggrannhet genom att i inaktiv, kompakt form binda till ribosomen och sedan övergå i en aktiv, utsträckt form när termingsfaktorn läser ett stopp-kodon. I naturligaribsosomkomplex termineringsfaktorernas kompakta form kortlivad och hartidigare bara observats i långlivat tillstånd i närvaro av ribosom-mutationer och inhibitorer.

Vi har utvecklat en snabb-kinetikmetod för att estimera tiden RF1/2förblir kompakta efter ribosom-bindning i närvaro av stopp-kodon och använde sedan tidsupplöst kryo-elektronmikroskopi för att visualisera termineringsfaktorernas kompakta former . Vi har också undersökt hur eukaryot termineringsfaktor 3 (eRF3) påverkar hastighet och noggrannhet vid peptid-frigörelse från ribosomer inducerad av eukaryot termineringsfaktor 1 i ett in vitro system för proteinsyntes med komponenterfrån jäst (Sacharomyces cerevisiae).

Vi har studerat effekterna av två kemiska modifieringar av mRNA: 2’-O-metylering och N6-metylering av adenosin. 2’-O-metylering minskar kraftigt den maximala hastigheten (kcat) och effektiviteten (kcat/Km) vid kognat (korrekt) kodon-läsning genom drastisk reduktion av den GTPas- aktivitet som är första delen av kodon-läsningen ochgenom en kraftig ökning av förlusten av kognata tRNA genom överdriven korrekturläsning. Anmärkningsvärt nog ökar N6-metylering felet vid kodon-läsning genom att minska effektiviteten för kognat kodon-läsning med termneringsfaktorer ochtRNA, medan de motsvarande icke-kognata reaktionerna påverkas mycket mindre, vilket är mycket ovanligt.

39







39







39




RF1/2 känner igen stopp-kodord i mRNA och frigör syntetiserade proteiner som peptid-kedjor. RF1/2 har antagits får hög noggrannhet genom att i inaktiv, kompakt form binda till ribosomen och sedan övergå i en aktiv, utsträckt form när termingsfaktorn läser ett stopp-kodon. I naturliga ribsosomkomplex termineringsfaktorernas kompakta form kortlivad och har tidigare bara observats i långlivat tillstånd i närvaro av ribosom-mutationer och inhibitorer.

Vi har utvecklat en snabb-kinetikmetod för att estimera tiden RF1/2 förblir kompakta efter ribosom-bindning i närvaro av stopp-kodon och använde sedan tidsupplöst kryo-elektronmikroskopi för att visualisera termineringsfaktorernas kompakta former . Vi har också undersökt hur eukaryot termineringsfaktor 3 (eRF3) påverkar hastighet och noggrannhet vid peptid-frigörelse från ribosomer inducerad av eukaryot termineringsfaktor 1 i ett in vitro system för proteinsyntes med komponenter från jäst (Sacharomyces cerevisiae).

Vi har studerat effekterna av två kemiska modifieringar av mRNA: 2’-O-metylering och N6-metylering av adenosin. 2’-O-metylering minskar kraftigt den maximala hastigheten (kcat) och effektiviteten (kcat/Km) vid kognat (korrekt) kodon-läsning genom drastisk reduktion av den GTPas- aktivitet som är första delen av kodon-läsningen ochgenom en kraftig ökning av förlusten av kognata tRNA genom överdriven korrekturläsning. Anmärkningsvärt nog ökar N6-metylering felet vid kodon-läsning genom att minska effektiviteten för kognat kodon-läsning med termneringsfaktorer och tRNA, medan de motsvarande icke-kognata reaktionerna påverkas mycket mindre, vilket är mycket ovanligt.

40

Acknowledgements

My supervisor, Måns, thank you for your time and your science.

My co-supervisor, Suparna, for helping with things big and small.

Zi-ao Jack Fu, Sandip Kaledhonkar and prof. Joachim Frank for collab-oration on the RF conformation project, letting me work in their lab and sharing their knowledge about time-resolved cryo-EM.

Galina Bartish, for starting the work on yeast in vitro translation system.

Valérie Heurgué-Hamard and Emmeline Huvelle, for letting me work in their lab and sharing their knowledge about eukaryotic proteins.

Prof. Erik Johansson and Pia Osterman, for their help with large scale yeast cultures.

Junhong Choi and prof. Jody Puglisi, for collaboration on the 2’-O-methylation project.

Ka-Weng: for joint projects and answering my numerous questions.

Anneli, Chandu, Jingji, Mikael, for small things that add up to big things.

Ray, for supplying bacterial components.

Mamai, tėčiui, Simonui, Rimai, Aurelijai, Giedrei, Rūtai ir Simonai, ačiū už viską!

40

Acknowledgements












40

Acknowledgements












40

Acknowledgements








Ka-Weng, for joint projects and answering my numerous questions.




41

References

Alkalaeva, E.Z., Pisarev, A.V., Frolova, L.Y., Kisselev, L.L., and Pestova, T.V. (2006). In vitro reconstitution of eukaryotic translation reveals cooperativity between release factors eRF1 and eRF3. Cell 125, 1125-1136.

Andreev, D.E., Terenin, I.M., Dunaevsky, Y.E., Dmitriev, S.E., and Shatsky, I.N. (2006). A leaderless mRNA can bind to mammalian 80S ribosomes and direct polypeptide synthesis in the absence of translation initiation factors. Molecular and cellular biology 26, 3164-3169.

Beissel, C., Neumann, B., Uhse, S., Hampe, I., Karki, P., and Krebber, H. (2019). Translation termination depends on the sequential ribosomal entry of eRF1 and eRF3. Nucleic acids research.

Brown, A., Shao, S., Murray, J., Hegde, R.S., and Ramakrishnan, V. (2015). Structural basis for stop codon recognition in eukaryotes. Nature 524, 493-496.

Bulygin, K.N., Khairulina, Y.S., Kolosov, P.M., Ven'yaminova, A.G., Graifer, D.M., Vorobjev, Y.N., Frolova, L.Y., Kisselev, L.L., and Karpova, G.G. (2010). Three distinct peptides from the N domain of translation termination factor eRF1 surround stop codon in the ribosome. Rna 16, 1902-1914.

Chao, A.T., Dierick, H.A., Addy, T.M., and Bejsovec, A. (2003). Mutations in eukaryotic release factors 1 and 3 act as general nonsense suppressors in Drosophila. Genetics 165, 601-612.

Chavatte, L., Seit-Nebi, A., Dubovaya, V., and Favre, A. (2002). The invariant uridine of stop codons contacts the conserved NIKSR loop of human eRF1 in the ribosome. The EMBO journal 21, 5302-5311.

Chen, B., and Frank, J. (2016). Two promising future developments of cryo-EM: capturing short-lived states and mapping a continuum of states of a macromolecule. Microscopy (Oxf) 65, 69-79.

Choi, J., Indrisiunaite, G., DeMirci, H., Ieong, K.W., Wang, J., Petrov, A., Prabhakar, A., Rechavi, G., Dominissini, D., He, C., et al. (2018). 2'-O-methylation in mRNA disrupts tRNA decoding during translation elongation. Nature structural & molecular biology 25, 208-216.

Crick, F. (1970). Central dogma of molecular biology. Nature 227, 561-563. Dever, T.E., Kinzy, T.G., and Pavitt, G.D. (2016). Mechanism and Regulation of

Protein Synthesis in Saccharomyces cerevisiae. Genetics 203, 65-107. Dincbas-Renqvist, V., Engstrom, A., Mora, L., Heurgue-Hamard, V., Buckingham,

R., and Ehrenberg, M. (2000). A post-translational modification in the GGQ motif of RF2 from Escherichia coli stimulates termination of translation. The EMBO journal 19, 6900-6907.

Eyler, D.E., Wehner, K.A., and Green, R. (2013). Eukaryotic release factor 3 is required for multiple turnovers of peptide release catalysis by eukaryotic release factor 1. The Journal of biological chemistry 288, 29530-29538.

Feng, T., Yamamoto, A., Wilkins, S.E., Sokolova, E., Yates, L.A., Munzel, M., Singh, P., Hopkinson, R.J., Fischer, R., Cockman, M.E., et al. (2014). Optimal

41

References















41

References















41

References















42

translational termination requires C4 lysyl hydroxylation of eRF1. Molecular cell 53, 645-654.

Frank, J., and Agrawal, R.K. (2000). A ratchet-like inter-subunit reorganization of the ribosome during translocation. Nature 406, 318-322.

Freistroffer, D.V., Kwiatkowski, M., Buckingham, R.H., and Ehrenberg, M. (2000). The accuracy of codon recognition by polypeptide release factors. Proceedings of the National Academy of Sciences of the United States of America 97, 2046-2051.

Freistroffer, D.V., Pavlov, M.Y., MacDougall, J., Buckingham, R.H., and Ehrenberg, M. (1997). Release factor RF3 in E.coli accelerates the dissociation of release factors RF1 and RF2 from the ribosome in a GTP-dependent manner. The EMBO journal 16, 4126-4133.

Frolova, L., Seit-Nebi, A., and Kisselev, L. (2002). Highly conserved NIKS tetrapeptide is functionally essential in eukaryotic translation termination factor eRF1. Rna 8, 129-136.

Frolova, L.Y., Tsivkovskii, R.Y., Sivolobova, G.F., Oparina, N.Y., Serpinsky, O.I., Blinov, V.M., Tatkov, S.I., and Kisselev, L.L. (1999). Mutations in the highly conserved GGQ motif of class 1 polypeptide release factors abolish ability of human eRF1 to trigger peptidyl-tRNA hydrolysis. Rna 5, 1014-1020.

Gao, H., Zhou, Z., Rawat, U., Huang, C., Bouakaz, L., Wang, C., Cheng, Z., Liu, Y., Zavialov, A., Gursky, R., et al. (2007). RF3 induces ribosomal conformational changes responsible for dissociation of class I release factors. Cell 129, 929-941.

Gerik, K.J., Li, X., Pautz, A., and Burgers, P.M. (1998). Characterization of the two small subunits of Saccharomyces cerevisiae DNA polymerase delta. The Journal of biological chemistry 273, 19747-19755.

Goldstein, J.L., and Caskey, C.T. (1970). Peptide chain termination: effect of protein S on ribosomal binding of release factors. Proceedings of the National Academy of Sciences of the United States of America 67, 537-543.

Graille, M., Heurgue-Hamard, V., Champ, S., Mora, L., Scrima, N., Ulryck, N., van Tilbeurgh, H., and Buckingham, R.H. (2005). Molecular basis for bacterial class I release factor methylation by PrmC. Molecular cell 20, 917-927.

Grentzmann, G., Brechemier-Baey, D., Heurgue, V., Mora, L., and Buckingham, R.H. (1994). Localization and characterization of the gene encoding release factor RF3 in Escherichia coli. Proceedings of the National Academy of Sciences of the United States of America 91, 5848-5852.

Hellen, C.U.T. (2018). Translation Termination and Ribosome Recycling in Eukaryotes. Cold Spring Harb Perspect Biol 10.

Heurgue-Hamard, V., Champ, S., Engstrom, A., Ehrenberg, M., and Buckingham, R.H. (2002). The hemK gene in Escherichia coli encodes the N(5)-glutamine methyltransferase that modifies peptide release factors. The EMBO journal 21, 769-778.

Heurgue-Hamard, V., Champ, S., Mora, L., Merkulova-Rainon, T., Kisselev, L.L., and Buckingham, R.H. (2005). The glutamine residue of the conserved GGQ motif in Saccharomyces cerevisiae release factor eRF1 is methylated by the product of the YDR140w gene. The Journal of biological chemistry 280, 2439-2445.

Hopfield, J.J. (1974). Kinetic proofreading: a new mechanism for reducing errors in biosynthetic processes requiring high specificity. Proceedings of the National Academy of Sciences of the United States of America 71, 4135-4139.

Ieong, K.W., Uzun, U., Selmer, M., and Ehrenberg, M. (2016). Two proofreading steps amplify the accuracy of genetic code translation. Proceedings of the

42

















42

















42

















43

National Academy of Sciences of the United States of America 113, 13744-13749.

Ito, K., Uno, M., and Nakamura, Y. (2000). A tripeptide 'anticodon' deciphers stop codons in messenger RNA. Nature 403, 680-684.

Jackson, R.J., Hellen, C.U., and Pestova, T.V. (2012). Termination and post-termination events in eukaryotic translation. Advances in protein chemistry andstructural biology 86, 45-93.

Jelenc, P.C., and Kurland, C.G. (1979). Nucleoside triphosphate regeneration decreases the frequency of translation errors. Proceedings of the NationalAcademy of Sciences of the United States of America 76, 3174-3178.

Jorgensen, F., Adamski, F.M., Tate, W.P., and Kurland, C.G. (1993). Release factor-dependent false stops are infrequent in Escherichia coli. Journal of molecular biology 230, 41-50.

Kahley, M.J., and Novak, M. (1996). A Practical Procedure for Determining RateConstants in Consecutive First-Order Systems. Journal of Chemical Education 73, 359.

Klaholz, B.P., Pape, T., Zavialov, A.V., Myasnikov, A.G., Orlova, E.V., Vestergaard, B., Ehrenberg, M., and van Heel, M. (2003). Structure of the Escherichia coli ribosomal termination complex with release factor 2. Nature 421, 90-94.

Konecki, D.S., Aune, K.C., Tate, W., and Caskey, C.T. (1977). Characterization of reticulocyte release factor. The Journal of biological chemistry 252, 4514-4520.

Korostelev, A., Asahara, H., Lancaster, L., Laurberg, M., Hirschi, A., Zhu, J., Trakhanov, S., Scott, W.G., and Noller, H.F. (2008). Crystal structure of a translation termination complex formed with release factor RF2. Proceedings of the National Academy of Sciences of the United States of America 105, 19684-19689.

Kuhlenkoetter, S., Wintermeyer, W., and Rodnina, M.V. (2011). Different substrate-dependent transition states in the active site of the ribosome. Nature 476, 351-354.

Laurberg, M., Asahara, H., Korostelev, A., Zhu, J., Trakhanov, S., and Noller, H.F. (2008). Structural basis for translation termination on the 70S ribosome. Nature 454, 852-857.

Maicas, E., Shago, M., and Friesen, J.D. (1990). Translation of the Saccharomyces cerevisiae tcm1 gene in the absence of a 5'-untranslated leader. Nucleic acids research 18, 5823-5828.

Matheisl, S., Berninghausen, O., Becker, T., and Beckmann, R. (2015). Structure of a human translation termination complex. Nucleic acids research 43, 8615-8626.

Ninio, J. (1975). Kinetic amplification of enzyme discrimination. Biochimie 57, 587-595.

Ogle, J.M., Brodersen, D.E., Clemons, W.M., Jr., Tarry, M.J., Carter, A.P., and Ramakrishnan, V. (2001). Recognition of cognate transfer RNA by the 30S ribosomal subunit. Science 292, 897-902.

Ontiveros, R.J., Stoute, J., and Liu, K.F. (2019). The chemical diversity of RNA modifications. The Biochemical journal 476, 1227-1245.

Peske, F., Kuhlenkoetter, S., Rodnina, M.V., and Wintermeyer, W. (2014). Timingof GTP binding and hydrolysis by translation termination factor RF3. Nucleic acids research 42, 1812-1820.

Petry, S., Brodersen, D.E., Murphy, F.V.t., Dunham, C.M., Selmer, M., Tarry, M.J., Kelley, A.C., and Ramakrishnan, V. (2005). Crystal structures of the ribosome in complex with release factors RF1 and RF2 bound to a cognate stop codon. Cell 123, 1255-1266.

43



















43



















43



Jackson, R.J., Hellen, C.U., and Pestova, T.V. (2012). Termination and post-termination events in eukaryotic translation. Advances in protein chemistry and structural biology 86, 45-93.

Jelenc, P.C., and Kurland, C.G. (1979). Nucleoside triphosphate regeneration decreases the frequency of translation errors. Proceedings of the National Academy of Sciences of the United States of America 76, 3174-3178.


Kahley, M.J., and Novak, M. (1996). A Practical Procedure for Determining Rate Constants in Consecutive First-Order Systems. Journal of Chemical Education 73, 359.











Peske, F., Kuhlenkoetter, S., Rodnina, M.V., and Wintermeyer, W. (2014). Timing of GTP binding and hydrolysis by translation termination factor RF3. Nucleic acids research 42, 1812-1820.


44

Pisarev, A.V., Skabkin, M.A., Pisareva, V.P., Skabkina, O.V., Rakotondrafara, A.M., Hentze, M.W., Hellen, C.U., and Pestova, T.V. (2010). The role of ABCE1 in eukaryotic posttermination ribosomal recycling. Molecular cell 37, 196-210.

Preis, A., Heuer, A., Barrio-Garcia, C., Hauser, A., Eyler, D.E., Berninghausen, O.,Green, R., Becker, T., and Beckmann, R. (2014). Cryoelectron microscopic structures of eukaryotic translation termination complexes containing eRF1-eRF3 or eRF1-ABCE1. Cell reports 8, 59-65.

Rawat, U.B., Zavialov, A.V., Sengupta, J., Valle, M., Grassucci, R.A., Linde, J., Vestergaard, B., Ehrenberg, M., and Frank, J. (2003). A cryo-electron microscopic study of ribosome-bound termination factor RF2. Nature 421, 87-90.

Ruusala, T., Ehrenberg, M., and Kurland, C.G. (1982). Is there proofreading during polypeptide synthesis? The EMBO journal 1, 741-745.

Salas-Marco, J., and Bedwell, D.M. (2004). GTP hydrolysis by eRF3 facilitates stop codon decoding during eukaryotic translation termination. Molecular and cellular biology 24, 7769-7778.

Schmeing, T.M., Voorhees, R.M., Kelley, A.C., Gao, Y.G., Murphy, F.V.t., Weir, J.R., and Ramakrishnan, V. (2009). The crystal structure of the ribosome bound to EF-Tu and aminoacyl-tRNA. Science 326, 688-694.

Scolnick, E., Tompkins, R., Caskey, T., and Nirenberg, M. (1968). Release factors differing in specificity for terminator codons. Proceedings of the National Academy of Sciences of the United States of America 61, 768-774.

Shin, D.H., Brandsen, J., Jancarik, J., Yokota, H., Kim, R., and Kim, S.H. (2004).Structural analyses of peptide release factor 1 from Thermotoga maritima revealdomain flexibility required for its interaction with the ribosome. Journal of molecular biology 341, 227-239.

Shoemaker, C.J., and Green, R. (2011). Kinetic analysis reveals the ordered coupling of translation termination and ribosome recycling in yeast. Proceedings of the National Academy of Sciences of the United States of America 108, E1392-1398.

Skogerson, L., and Wakatama, E. (1976). A ribosome-dependent GTPase from yeast distinct from elongation factor 2. Proceedings of the National Academy of Sciences of the United States of America 73, 73-76.

Song, H., Mugnier, P., Das, A.K., Webb, H.M., Evans, D.R., Tuite, M.F., Hemmings, B.A., and Barford, D. (2000). The crystal structure of human eukaryotic release factor eRF1--mechanism of stop codon recognition and peptidyl-tRNA hydrolysis. Cell 100, 311-321.

Stansfield, I., Jones, K.M., Kushnirov, V.V., Dagkesamanskaya, A.R., Poznyakovski, A.I., Paushkin, S.V., Nierras, C.R., Cox, B.S., Ter-Avanesyan, M.D., and Tuite, M.F. (1995). The products of the SUP45 (eRF1) and SUP35 genes interact to mediate translation termination in Saccharomyces cerevisiae. The EMBO journal 14, 4365-4373.

Sund, J., Ander, M., and Aqvist, J. (2010). Principles of stop-codon reading on the ribosome. Nature 465, 947-950.

Ter-Avanesyan, M.D., Kushnirov, V.V., Dagkesamanskaya, A.R., Didichenko, S.A., Chernoff, Y.O., Inge-Vechtomov, S.G., and Smirnov, V.N. (1993). Deletion analysis of the SUP35 gene of the yeast Saccharomyces cerevisiae reveals twonon-overlapping functional regions in the encoded protein. Molecular microbiology 7, 683-692.

44















44















44


Preis, A., Heuer, A., Barrio-Garcia, C., Hauser, A., Eyler, D.E., Berninghausen, O., Green, R., Becker, T., and Beckmann, R. (2014). Cryoelectron microscopic structures of eukaryotic translation termination complexes containing eRF1-eRF3 or eRF1-ABCE1. Cell reports 8, 59-65.






Shin, D.H., Brandsen, J., Jancarik, J., Yokota, H., Kim, R., and Kim, S.H. (2004). Structural analyses of peptide release factor 1 from Thermotoga maritima reveal domain flexibility required for its interaction with the ribosome. Journal of molecular biology 341, 227-239.






Ter-Avanesyan, M.D., Kushnirov, V.V., Dagkesamanskaya, A.R., Didichenko, S.A., Chernoff, Y.O., Inge-Vechtomov, S.G., and Smirnov, V.N. (1993). Deletion analysis of the SUP35 gene of the yeast Saccharomyces cerevisiae reveals two non-overlapping functional regions in the encoded protein. Molecular microbiology 7, 683-692.

45

Thompson, R.C., and Stone, P.J. (1977). Proofreading of the codon-anticodoninteraction on ribosomes. Proceedings of the National Academy of Sciences of the United States of America 74, 198-202.

Triana-Alonso, F.J., Chakraburtty, K., and Nierhaus, K.H. (1995). The elongation factor 3 unique in higher fungi and essential for protein biosynthesis is an E site factor. The Journal of biological chemistry 270, 20473-20478.

Valle, M., Zavialov, A., Li, W., Stagg, S.M., Sengupta, J., Nielsen, R.C., Nissen, P., Harvey, S.C., Ehrenberg, M., and Frank, J. (2003). Incorporation of aminoacyl-tRNA into the ribosome as seen by cryo-electron microscopy. Nat Struct Biol 10, 899-906.

Vestergaard, B., Sanyal, S., Roessle, M., Mora, L., Buckingham, R.H., Kastrup, J.S.,Gajhede, M., Svergun, D.I., and Ehrenberg, M. (2005). The SAXS solution structure of RF1 differs from its crystal structure and is similar to its ribosomebound cryo-EM structure. Molecular cell 20, 929-938.

Vestergaard, B., Van, L.B., Andersen, G.R., Nyborg, J., Buckingham, R.H., and Kjeldgaard, M. (2001). Bacterial polypeptide release factor RF2 is structurally distinct from eukaryotic eRF1. Molecular cell 8, 1375-1382.

Yusupova, G., and Yusupov, M. (2014). High-resolution structure of the eukaryotic 80S ribosome. Annu Rev Biochem 83, 467-486.

Zhang, J., Ieong, K.W., Mellenius, H., and Ehrenberg, M. (2016). Proofreading neutralizes potential error hotspots in genetic code translation by transfer RNAs.Rna 22, 896-904.

Zhouravleva, G., Frolova, L., Le Goff, X., Le Guellec, R., Inge-Vechtomov, S., Kisselev, L., and Philippe, M. (1995). Termination of translation in eukaryotes is governed by two interacting polypeptide chain release factors, eRF1 andeRF3. The EMBO journal 14, 4065-4072.

Zoldak, G., Redecke, L., Svergun, D.I., Konarev, P.V., Voertler, C.S., Dobbek, H.,Sedlak, E., and Sprinzl, M. (2007). Release factors 2 from Escherichia coli andThermus thermophilus: structural, spectroscopic and microcalorimetric studies. Nucleic acids research 35, 1343-1353.

45










45










45

Thompson, R.C., and Stone, P.J. (1977). Proofreading of the codon-anticodon interaction on ribosomes. Proceedings of the National Academy of Sciences of the United States of America 74, 198-202.



Vestergaard, B., Sanyal, S., Roessle, M., Mora, L., Buckingham, R.H., Kastrup, J.S., Gajhede, M., Svergun, D.I., and Ehrenberg, M. (2005). The SAXS solution structure of RF1 differs from its crystal structure and is similar to its ribosome bound cryo-EM structure. Molecular cell 20, 929-938.



Zhang, J., Ieong, K.W., Mellenius, H., and Ehrenberg, M. (2016). Proofreading neutralizes potential error hotspots in genetic code translation by transfer RNAs. Rna 22, 896-904.

Zhouravleva, G., Frolova, L., Le Goff, X., Le Guellec, R., Inge-Vechtomov, S., Kisselev, L., and Philippe, M. (1995). Termination of translation in eukaryotes is governed by two interacting polypeptide chain release factors, eRF1 and eRF3. The EMBO journal 14, 4065-4072.

Zoldak, G., Redecke, L., Svergun, D.I., Konarev, P.V., Voertler, C.S., Dobbek, H., Sedlak, E., and Sprinzl, M. (2007). Release factors 2 from Escherichia coli and Thermus thermophilus: structural, spectroscopic and microcalorimetric studies. Nucleic acids research 35, 1343-1353.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1814

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-382490

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019





ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Documents

uu.diva-portal.org › smash › get › diva2:1307157 › FULLTEXT01… · ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2019 Digital Comprehensive Summar ies of Uppsala Disser tations