Removal and Replacement of Ribosomal Proteins

Home , 16S ribosomal RNA, 23S ribosomal RNA, Bacterial translation, Elongation factor P, Peptidyl transferase, Ribosome

The Road Not Taken - Robert Frost

Two roads diverged in a yellow wood, And sorry I could not travel both And be one traveler, long I stood And looked down one as far as I could To where it bent in the undergrowth;

Then took the other, as just as fair, And having perhaps the better claim, Because it was grassy and wanted wear; Though as for that the passing there Had worn them really about the same,

And both that morning equally lay In leaves no step had trodden black. Oh, I kept the first for another day! Yet knowing how way leads on to way, I doubted if I should ever come back.

I shall be telling this with a sigh Somewhere ages and ages hence: Two roads diverged in a wood, and I— I took the one less traveled by, And that has made all the difference.

List of Papers

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

I Tobin, C., Mandava, CS., Ehrenberg, M., Andersson, DI., San- yal, S. (2010) Ribosomes lacking protein S20 are defective in mRNA binding and subunit association. Journal of Molecular Biology, 397(3):767–76 II Tobin, C., Nord, S., Wikström, PM., Näsvall, J., Sanyal, S., Andersson, DI. (2011) Ribosomal protein L1 is required for balanced subunit formation. (Manuscript) III Lind, PA., Tobin, C., Berg, OG., Kurland, CG., Andersson, DI. (2010) Compensatory gene amplification restores fitness after inter-species gene replacements. Molecular Microbiology, 75(5):1078-89

Reprints were made with permission from the respective publishers.

Cover illustration made in PyMol (www.pymol.org) using the coordinates of the X- ray crystal structure of the T. thermophilus 70S ribosome (PDB I.D. 2J00 & 2J01) (Selmer et al. 2006).

Contents

Preface ...... 11 The Magnificence of the Machine ...... 11 Background...... 13 The ribosome is a ribozyme but what about the proteins?...... 13 The necessary tools for understanding the ribosome and translation ...... 14 The nuts and bolts of the translation machinery ...... 14 Messenger RNA (mRNA) ...... 16 Transfer RNA (tRNA)...... 17 Translation factors...... 19 The ribosome at work ...... 19 Initiation ...... 20 Elongation ...... 22 Termination ...... 26 Ribosome recycling...... 28 Ribosome biogenesis and assembly...... 29 rRNA synthesis...... 29 R-protein synthesis ...... 30 Processing and modification of rRNA and r-proteins ...... 31 Subunit assembly...... 33 Degradation and turnover of ribosomal components ...... 37 Ribosomal Proteins ...... 38 General features...... 38 Roles in and outside the ribosome...... 39 Conservation...... 41 Non-essential r-proteins ...... 43 R-proteins of particular interest...... 44 Ribosomal complexity – an evolutionary riddle ...... 47 Concept of fitness...... 49 Adaptive and non-adaptive forces drive evolution ...... 49 Compensatory evolution and adaptation...... 50 Ribosomal complexity – a consequence of compensation? ...... 51 Horizontal Gene Transfer (HGT)...... 52 Constraints on HGT...... 52

Present Investigations ...... 54 The model organism...... 55 Construction and characterization of r-protein mutants...... 55 Recombineering...... 55 In vitro system ...... 57 In vivo system...... 57 Fitness costs of r-protein removal and replacement...... 59 Removal of r-proteins confers substantial fitness costs ...... 59 Fitness costs of replacement are generally weaker than removal ...... 60 Small fitness costs restrict HGT of r-proteins ...... 60 Phenotypes of S20 and L1 mutants...... 61 Removal of S20 perturbs mRNA binding and subunit association..... 61 S20 is required for correct structural positioning of h44...... 62 Removal of L1 impairs ribosome biogenesis ...... 63 Mutations in 50S r-proteins mitigate fitness costs of L1 removal ...... 64 High functional conservation of r-proteins ...... 66 Increased dosage rescues sub-optimal r-proteins...... 67 Future Perspectives ...... 68 Paper I...... 68 Paper II ...... 68 Paper III...... 70 Concluding remarks...... 72 Acknowledgements...... 74 References...... 77

Abbreviations

DNA Deoxyribonucleic acid RNA Ribonucleic acid rRNA ribosomal RNA mRNA messenger RNA tRNA transfer RNA aa-tRNA amino-acyl tRNA r-protein ribosomal protein MDa Mega Dalton Å Ångström bps base pairs nt nucleotide ss single-stranded ds double-stranded CP Central Protuberance PTC Peptidyl-Transferase Centre SD Shine-Dalgarno sequence aSD anti-Shine-Dalgarno sequence TIR Translation Initiation Region IC Initiation Complex ASL Anticodon stem loop AARSs Amino-acyl tRNA synthetases A-site Amino-acyl site P-site Peptidyl site E-site Exit site IF Initiation Factor EF Elongation Factor RF Release Factor RRF Ribosome recycling factor GTP Guanosine triphosphate GDP Guanosine diphosphate ATP Adenosine triphosphate ppGpp Guanosine 5’-diphosphate, 3’- diphosphate pppGpp Guanosine 5’-triphosphate, 3’- diphosphate RBS Ribosome Binding Site

ITS Internal Transcribed Spacer IVS Intervening sequence Cryo-EM Cryogenic Electron Microscopy FRET Fluorescence Resonance Energy Transfer Met Methionine fMet formylated Methionine GEF Guanine nucleotide Exchange Factor ram ribosomal ambiguity E. coli Escherichia coli S. typhimurium Salmonella enterica var Typhi- murium LT2 T. thermophilus Thermus thermophilus H. marismortui Haloarcula marismortui HIV Human Immunodeficiency Virus HGT Horizontal Gene Transfer H-NS Histone-like Nucleoid Structuring Protein SRP Signal Recognition Particle mitoribosome Mitochondrial ribosome Fe Iron ORF Open reading frame NG N-methyl-N'-nitro-N- nitrosoguanidine IPTG Isopropyl -D-1- thiogalactopyranoside CAI Codon Adaptation Index

Preface

The Magnificence of the Machine Of all the molecules present in the living organism, proteins are the most abundant and arguably the most fundamental. Often referred to as the biological workhorses of the cell, proteins form the cornerstone of the most elementary cellular processes such as enzyme catalysis, structural support and immune defence, to name but a few. As such, efficient and orderly production of proteins is vital for cell fitness and survival. This fact is reflected in the recognition that an expanding catalogue of human diseases is caused by modifications of various components that function in protein synthesis. For example, mutated variants of mammalian ribosomal protein S19 are the cause of approximately 25% of cases of the rare erythropoietic disorder Diamond-Blackfan anaemia (DBA) (Scheper et al. 2007). In every living organism, conversion of the genetic code into functional proteins is performed on the ribosome. This dynamic, macromolecular machine acts as a translator, producing biologically active protein molecules from a nucleic acid message. Despite the enormity of its task, the ribosome manages to manufacture proteins at a high rate and with few errors. In Es- cherichia coli (E.coli), ribosomes account for up to 50% of the dry mass of the cell (Nierhaus 2006) and during rapid cell division ribosome production consumes up to 40% of the cell’s total energy (Nierhaus 1991). Indeed, considering the enormous energy expense, the process has evolved to become highly efficient and tightly regulated. The highly significant nature of the task of the ribosome is reflected in the sophistication of its structure, where the modern ribosome constitutes one of the most complex cellular machines. Despite substantial and continual progress in understanding the mechanism of protein synthesis, many aspects of this elaborate process remain to be elucidated. The work presented in this thesis is an attempt to resolve ribosomal complexity to some degree by means of studies designed to reveal the significance and distinct functional roles of certain ribosomal proteins (r-proteins) in translation. R-proteins were either completely removed or replaced with orthologous counterparts and the associated phenotypes were examined using in vivo and in vitro techniques. In addition, compensatory evolution experiments were performed which serve to expand our knowledge and appre- ciation of the highly co-operative nature of the translation apparatus.

11 12 Background

The ribosome is a ribozyme but what about the proteins? At approximately 2.4 MDa in mass and over 200 Å in width, ribosomes are one of the largest molecular particles in the bacterial cell (Williamson 2009). They are composed of ribosomal RNA (rRNA) and ribosomal proteins (r- proteins), with rRNA constituting approximately two thirds of the total mass. Not only is rRNA predominant in terms of structural composition but it also forms the core of ribosomal function. Initially, questions surrounding the catalytic activity of the ribosome were posed from a purely proteocentric perspective, befitting the widely held belief at that time that all enzymes were protein in nature. This viewpoint started to lose integrity when experiments by Noller indicated that 50S subunits from Thermus aquaticus retained peptidyl transferase activity when stripped of almost all of their ribosomal proteins (Noller et al. 1992). The idea that proteins could be responsible for ribosomal enzymatic activity was completely crushed upon visualization of the first high-resolution crystal structure of the large ribosomal subunit (Nissen et al. 2000). The structure explicitly showed that the peptidyl transferase centre was completely devoid of protein. This finding firmly established that the ribosome was indeed a ribozyme and secured a starring role for RNA in translation. While RNA assumed centre stage, the r-proteins largely evaded intense scrutiny and were ascribed a more general architectural role in stabilizing the rRNA to maintain the overall morphology of the ribosome (Moore & Steitz 2002). In spite of this generalization, certain proteins are known to enrich the ribosome beyond architectural support such as S12 of the small subunit and its role in mRNA decoding. However, with an impressive array of 54 individual proteins, a comprehensive understanding of their distinct functional roles and their overall contribution to the spectacular process of translation is an area where research is scarce.

13 The necessary tools for understanding the ribosome and translation A complete understanding of how the ribosome works relies on well- developed genetic, biochemical and structural techniques. In the past, structural elucidation of the ribosome at the atomic level proved difficult due to general problems with crystallographic techniques combined with the in- credible size of the ribosome. During this period of struggle, the properties of neutron scattering and electron microscopy were exploited in an effort to tease out the structural mysteries (Capel et al. 1987; Frank et al. 1995). Al- though informative, such studies failed to capture high-resolution images. Beginning in the year 2000, crystallographic images of the ribosomal subunits from Thermus thermophilus (Wimberly et al. 2000) and Haloarcula marismortui (Ban et al. 2000) at the atomic level appeared and today we have detailed images of the full ribosome with ligands bound (Carter et al. 2001; Yusupova et al. 2001; Selmer et al. 2006; Jin et al. 2010). These landmarks in ribosome research have complimented the increasingly sophisticated genetic and biochemical tools in dissecting the mechanism of translation as well as providing many new detailed descriptions of ribosomal dynamics.

The nuts and bolts of the translation machinery Translation refers to the process in which the genetic code, contained in nucleic acid, is converted into its corresponding sequence of amino acids and is performed by the ribosome. The procedure relies on the availability of template that is supplied in the form of mRNA in addition to amino acid substrates that are delivered to the ribosome bound to tRNA molecules. In order to translate efficiently, the ribosome is assisted by a number of protein factors although all steps of the translation cycle can be performed in the absence of such factors. Thus the ribosome can be regarded as the central component of translation and takes responsibility for orchestrating the entire process.

Bacterial ribosomes In all cells the ribosome is composed of two subunits of unequal size. The prokaryotic ribosome consists of the small 30S subunit and the larger 50S subunit that together constitute the 70S ribosome. Both subunits exist in the cell cytoplasm as free, separated molecules when inactive but associate upon the initiation of translation. The small subunit is composed of one strand of 16S rRNA (1,542 nt) and 21 r-proteins, while the large subunit contains two strands of rRNA, the 23S (2,904 nt) and 5S (120 nt) rRNAs and an additional 33 r-proteins (Noller 1996). The two subunits differ widely in their

14 overall shape. The 50S subunit is a more compact, consolidated structure consisting of a rounded, hemispherical base with three protuberances designated the L1 stalk, the central protuberance (CP) and the L7/L12 stalk (Fig- ure 1B). The six secondary structures of the 23S rRNA are tightly inter- twined forming a dense structure. The 30S on the other hand is not only smaller, but morphologically it is also less compact. Unlike the 50S subunit, the three-dimensional structure is easily correlated to its secondary structure domains. It is organized into a head region, which is separated from the body of the subunit by a narrow neck and below the head, the platform and shoulder domains are found (Figure 1A). The subunits not only possess unique structures but they also contribute unique functions to protein synthesis. The 30S subunit is allocated the major tasks of initial mRNA binding and decoding, whilst the 50S catalyzes peptide bond formation. Upon the commencement of translation, the two subunits associate via interactions along their interface sides and cooperate during all subsequent phases to efficiently and faithfully transform the RNA message into a functional polypeptide. (A) Head

Platform

Shoulder

Decoding site CP

Body

L1 stalk L7/L12 stalk Spur

PTC

(B)

Figure 1: Tertiary structures of the bacterial 30S (A) and 50S (B) subunits viewed from the subunit interface. Ribosomal RNA is depicted as green ribbons and r- proteins as orange ribbons. Abbreviations are as follows: CP, central protuberance; PTC, peptidyl-transferase centre. The image was created in PyMol (www.pymol.org) using the coordinates of the X-ray crystal structure of the T. thermophilus 70S ribosome (PDB ID 2J00 and 2J01) (Selmer et al. 2006).

15 Messenger RNA (mRNA) The information that dictates the sequence of amino acids in a protein is presented to the ribosome in the form of an mRNA molecule. The nucleotide bases are read sequentially by the ribosome in combinations of three (known as codons), with each codon specifying a particular amino acid. Due to the degeneracy of the genetic code, more than one codon can specify the same amino acid. In addition the start and stop of the mRNA coding sequence must be defined to set the correct reading frame on the mRNA and thereby support the fidelity of gene expression. The codon AUG (encoding methionine) indicates the translation start site for most eubacterial genes and is a perfect match for the anticodon CAU of the initiator tRNA, fMet-tRNAfMet. Translation initiation has also been observed on GUG and UUG at frequencies of 5% and 1% respectively in E. coli (Schneider et al. 1986). The infC gene, encoding initiation factor IF3, is unusual in its use of AUU as the start codon (La Teana et al. 1993). Three alternative stop signals, (UAA, UAG and UGA), are specifically recognized by release factors to terminate translation of the message. The existence of an mRNA channel on the 30S subunit was first proposed by Frank (Frank et al. 1995) and later demonstrated by X-ray crystallography (Yusupova et al. 2001). This channel allows the mRNA to wrap around the neck of the 30S subunit, entering between the head and shoulder and exiting on the opposite side of the subunit through the opening between the head and platform. To solve the problem of mRNA secondary structure, which would hinder passage of the mRNA through the narrow channel, the ribosome possesses helicase activity to unwind stable helices in the incoming mRNA during elongation. Analysis of this activity in vitro revealed that proteins S3, S4 and S5 that enclose the mRNA at its entry point are required for the melting of stable secondary structures (Takyar et al. 2005). The small and large ribosomal subunits assemble on the mRNA at a site known as the translation initiation region (TIR). Exact positioning of the mRNA on the 30S subunit is critical to ensure selection of the correct start codon during initiation. In prokaryotes, this is primarily achieved by complementary base-pairing of a short stretch of purine rich nucleotides on the mRNA that precedes the start codon, known as the Shine-Dalgarno (SD) sequence, to the anti-SD sequence (aSD) located in the 3’ tail of 16S rRNA (Shine & Dalgarno 1974). Other mRNA sequence elements that influence formation of the 30S initiation complex (IC) include the type of start codon, the distance between the SD motif and the start codon, the nucleotide composition as well as secondary structure of the TIR region (Simonetti et al. 2009). Not all mRNAs carry a strong SD sequence and it has been suggested that initial recognition and tethering of such mRNAs to the 30S occurs via interactions between a region upstream of the SD sequence and the solvent- exposed S1 protein (Sengupta et al. 2001).

16 Considering that all mRNAs are structured to some degree, this creates a dilemma for the firm attachment of the 30S subunit due to possible occlusion of the SD sequence. Cryo-electron microscopy (cryo-EM) images suggest that recognition of these structured mRNAs occurs in three stages: first tran- sient docking of the mRNA on the platform, followed by unfolding of secondary structures before final accommodation in the mRNA channel (Marzi et al. 2007). Stalling of initiation in this manner is promoted by ligands that stabilize the structured form of the mRNA and it is suggested that this translational mechanism may be prevalent in the control of gene expression (Marzi et al. 2007). The involvement of mRNA parameters, other than the classical SD-aSD interaction, in determining the efficiency of translation initiation is also a feature of the “stand-by model”. This model posits that the attachment of the 30S subunit to the mRNA occurs in two stages: initially rapid non-specific binding on a single stranded site before shifting into the correct position and inducing productive, stable binding via the SD-aSD and codon-anticodon interactions (de Smit & van Duin 2003). This mRNA unfolding and adaptation process has also been monitored by fluorescence resonance energy transfer (FRET) which has shown that stable binding and final adjustment of the mRNA relies on the SD sequence and start codon, initiator tRNA and IF2 in complex with GTP (Studer & Joseph 2006).

Transfer RNA (tRNA) In 1955 Crick proposed the requirement for "adaptor" molecules to link amino acids to their corresponding mRNA codons. These were later identified as soluble RNAs and in time adopted the term transfer RNAs (tRNA) (Woese 2001). These small molecules are pivotal in protein synthesis as they provide the link connecting the worlds of nucleic acid and protein and are thus responsible for ensuring correct interpretation of the genetic code as amino acids. In prokaryotes, tRNAs normally contain approximately 75 nucleotides and adopt the classical cloverleaf structure due to base pairing between the 5’ and 3’ termini and at three other sites creating three “leaves” or loops (Figure 2B). The middle loop (blue) is known as the anticodon stem loop (ASL) and as the name implies, contains the anticodon (grey) that forms Watson-Crick base pairs with the mRNA codon, thus determining the specificity of each tRNA molecule. The amino-acyl stem (violet) is situated opposite the ASL and contains the universally conserved CCA sequence (yellow), which serves as the attachment site for the specific amino acid residue. All tRNAs adopt a so-called L-shape three dimensional structure (Figure 2A), which separates the anticodon and amino-acyl acceptor by approximately 75 Å and was one of the first indications that decoding and peptidyl transfer occurred at distant sites on the ribosome, consistent with the partitioning of these two processes on the 30S and 50S subunit respectively

17 (Liljas 2004). This shape also endows the tRNA with a high degree of flexibility that is required for conformational rearrangements during the translation cycle (Korostelev & Noller 2007).

(A) (B)

Figure 2: Structure and domains of tRNA. (A) Crystallographic three-dimensional structure of tRNAPhe from yeast. (B) Schematic illustration of cloverleaf secondary structure. Image created by Yikrazuul (own work) [CC-BY-SA-3.0 (www.creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons. The overall fidelity of translation is highly dependent on the correct charging of tRNAs with their corresponding amino acids. This is a two-step reaction involving ATP and is catalyzed by a group of enzymes known as amino-acyl tRNA synthetases (AARSs) (Ibba & Soll 2000). These enzymes must specifically recognize both the amino acid and the correct tRNA with high accuracy since errors in charging will not be detected in subsequent steps of translation. To ensure accurate discrimination of the correct tRNA and amino acid from a large cellular pool of competing substrates, these enzymes make use of a series of contacts with the amino acid and tRNA at the active site and some AARSs complement the accuracy of selection using proofreading and editing mechanisms (Ibba & Soll 2000). Upon binding to the ribosome, the charged tRNA molecules span the intersubunit face positioning the acceptor stem in the 50S subunit and the anticodon stem in the 30S subunit. All tRNAs, with the exception of the initiator tRNA, enter the ribosome via the A-site where decoding occurs. The tRNA will either be accommodated or rejected based on complementarity of the codon-anticodon interaction. The P-site holds the peptidyl-tRNA which carries the growing nascent chain of amino acids and the E-site is the exit site for the deacylated tRNA. All tRNAs (excluding the initiator tRNA), pass

18 sequentially from the A-site, through the P-site and are subsequently ejected from the ribosome at the E-site.

Translation factors During translation the ribosome is assisted by a number of protein factors that accelerate various steps of the process. These include initiation factors (IFs), of which there are three in eubacteria, known as IF1, IF2, and IF3. Upon completion of initiation complex (IC) formation, the ribosome progresses to the elongation phase which involves elongation factors (EFs) EF- Tu, EF-Ts, EF-G, EF-P and EF-4. Termination of the elongation phase is modulated by release factors (RF1, 2 and 3), followed by ribosome recycling involving RRF (ribosome recycling factor), EF-G and IF3. Several of these factors are GTPases and act as molecular switches to in- terconvert from active (GTP-bound) to inactive (GDP-bound) conformations, which in turn modulates their affinity for substrates and their activity during translation (Rodnina et al. 2000). To translate rapidly and progres- sively, the bacterial ribosome requires four G-proteins IF2, EF-Tu, EF-G and RF3. EF-4 is also a G-protein, but experiments indicate that this factor is only required during environmental stress conditions (Pech et al. 2011). All of these factors are highly conserved among bacteria with the exception of RF3 which is absent in some species (Sprinzl et al. 2000). The intrinsic GTPase activity of these factors is low but is stimulated upon interaction with the ribosome (Rodnina et al. 2000).

The ribosome at work The rapid and accurate conversion of an mRNA into a polypeptide chain places high demands on the ribosome. Efficient translation requires both local and global conformational rearrangements of the ribosome, its substrates and co-factors. In addition to biochemical approaches, aimed at understanding the mechanistic details of each translational step, cryo-EM, X- ray crystallography and FRET studies have been integral in describing the nature of the structural dynamics of the ribosome at work.

Translation can be conceptually subdivided into four major phases and these are illustrated in Figure 3. Each major step is propelled forward via a number of sub-steps, the details of which are outlined below.

19 Initiation

50S GDP GTP GTP IF2 Pi IF2 IF2 fMet-tRNAfMet IF3 IF1 IF1 IF1 mRNA IF3 APE IF3 APE APE E AP 30S mRNA & IF binding initiator tRNA binding 50S subunit docking 70S initiation complex to 30S & release of IFs

Elongation GDP EF-G Pi GTP GTP EF-TuGTP EF-G EF-G

E AP E AP E AP E P A

Pi EF-TuGDP aa-tRNA delivery GTP hydrolysis peptidyl-transfer & EF-G GTP hydrolysis, and accomodation hybrid tRNA formation tRNA translocation Termination & E-site tRNA release GDP RF3 Pi GDP GTP GDP RF3 RF3 RF 1/2 RF 1/2 RF3 GTP

E P A E P A E P A E P A

RF 1/2 release factor 1/2 binding release factor 3 exchange of GDP for GTP GTP hydrolysis & peptide release binding on RF3 & release of RF1/2 & release of RF3 Ribosome recycling

GDP EF-G Pi GTP EF-G RRF RRF IF3 IF3 IF3 E P A E P A E P A E P A

RRF & EF-G binding 50S dissociation rapid binding of IF3 30S subunit & release of P-site tRNA ready to re-initiate + mRNA

Figure 3: Schematic overview of the four phases of translation: initiation, elongation, termination and ribosome recycling.

Initiation The initiation phase of protein synthesis has been described as the most complex, most regulated and most divergent phase in translation among the different kingdoms of life (Marintchev & Wagner 2004; Laursen et al. 2005; Simonetti et al. 2009). Due to the coupling of transcription and translation, ribosomes can initiate before mRNA transcription is complete on the nascent mRNA. In the overall scheme of translation, initiation is rate-limiting with reported rates of 0.3 per second, in contrast to the rapid incorporation of up to twenty amino acids per second during elongation in E. coli (Kennell & Riezman 1977).

20 This multi-phase step incorporates assembly of the components needed for synthesis of a new peptide and is intimately coupled to subunit dissociation following the previous translation cycle. After subunit splitting, the 30S is rapidly bound by IF3, which stimulates dissociation of the uncharged tRNA and mRNA and prevents premature association of the 50S subunit to an empty 30S (Karimi et al. 1999; Peske et al. 2005). Initiation contributes heavily to the fidelity of translation by defining the mRNA reading frame and selecting the initiator tRNA from a large pool of competing tRNA species. Formation of the translationally active 70S initiation complex (IC) is kinetically and selectively tuned via the interplay of three initiation factors IF1, IF2 and IF3 (Antoun et al. 2006a; Antoun et al. 2006b). The GTPase IF2 is the principal factor involved in selection and binding of the initiator tRNA to the 30S subunit. In all organisms, the initiator tRNA is charged with methionine and in almost all bacteria, as well as eukaryotic organelles such as the mitochondria and chloroplast, this methionine is formylated (fMet-tRNAfMet). The formyl moiety, along with other distinctive structural features, differentiates the initiator tRNA from elongator Met- tRNAMet and facilitates its recognition by IF2 (Varshney et al. 1993). Two possible pathways have been proposed for the recruitment of fMet-tRNAfMet to the 30S subunit. Experiments by Wu et al. suggest that initiator tRNA is escorted to the 30S subunit in a ternary complex with IF2 and GTP (Wu & RajBhandary 1997). However recent data demonstrate that IF2 complexed with GTP binds the 30S subunit first and promotes the subsequent binding of fMet-tRNAfMet to the 30S P-site (Milon et al. 2010). Although IF2 is the major determinant of initiator tRNA binding, both IF1 and IF3 modulate this event. In the presence of all three IFs, the rate and accuracy of fMet- tRNAfMet association to 30S subunits in vitro is maximized (Antoun et al. 2006a; Antoun et al. 2006b). The effect of IF3 alone is interesting as the factor increases the rate of dissociation of initiator tRNA more than its association rate resulting in an overall loss in the affinity of the tRNA for the 30S subunit which is amplified in the presence of IF1 (Antoun et al. 2006b). IF1 has also been shown to stimulate the affinity-enhancing effect of IF2 on initiator tRNA binding (Antoun et al. 2006b). Rather than direct interactions with initiator tRNA, structural studies suggest that IF1 and IF3 play indirect roles in tRNA binding via conformational changes of the 30S IC (Carter et al. 2001; Dallas & Noller 2001). Progression to the elongation phase of translation requires association of the 50S subunit to form an elongation competent 70S IC. IF2•GTP and the initiator tRNA have been allocated a role during this step in stimulating recruitment of the large subunit (Antoun et al. 2003). Due to its anti- association properties, release of IF3 from the 30S IC is required for 50S docking (Dallas & Noller 2001; Antoun et al. 2006b). Association of the large subunit prompts the hydrolysis of GTP causing release of IF2•GDP and Pi. A brief outline of eubacterial initiation is shown in Figure 3.

21 Elongation The outcome of each cycle of elongation is an extension of the nascent polypeptide chain by one amino acid. The end of the initiation phase is heralded upon formation of the 70S IC that contains an initiator tRNA at the P-site, while the A-site is vacant exposing the next mRNA codon for interaction with the incoming aa-tRNA (amino-acyl tRNA). Selection and accommodation of the correct (cognate) aa-tRNA is rapidly followed by peptidyl transfer which leaves the P-site tRNA deacylated and the A-site tRNA attached to the nascent polypeptide. Concerted movement of the tRNAs from the A- and P-sites to the P- and E-sites is termed translocation and occurs in two steps. This results in movement of the mRNA by exactly one codon and exposure of the ensuing codon at the A-site while the deacylated tRNA is ejected from the E-site. This cyclic process is iterated for each sense codon until a stop codon is admitted to the A-site and signals the end of elongation. As the polypeptide grows, it is conducted through a polypeptide exit tunnel that opens at the PTC and travels through the centre of the 50S subunit where it exits at the other side after a distance of 80Å (Steitz 2008). It has been estimated that the tunnel can hold up to 30-50 amino acid residues and that it may be involved in initial folding of the polypeptide (Gilbert et al. 2004). The main events that occur during elongation are illustrated in Figure 3.

Decoding at the A-site With the exception of initiation, all incoming aa-tRNAs enter the ribosome via the A-site and so it is at this site that the ribosome must distinguish the correct aa-tRNA (cognate tRNA) from a large pool of competing substrates. This process of discrimination is termed decoding and is crucial for the maintenance of high fidelity. However, the ribosome faces a tough challenge in consideration of the sheer number of different aa-tRNAs in the cell (41 different tRNA anticodons exist in E. coli and even more in most eukaryotes) (Wilson & Nierhaus 2003). As many as five or six different aa-tRNAs possess an anticodon similar to the cognate tRNA (termed near-cognate), while the remaining tRNAs possess a dissimilar anticodon (termed non- cognate). To circumvent these potential obstacles the ribosome has evolved an intricate decoding system to ensure a high level of accuracy. In the cell, most aa-tRNAs exist bound to EF-Tu•GTP. Elongation begins by binding of an aa-tRNA to the A-site as a ternary complex with EF- Tu•GTP. Selectivity of the aa-tRNA is ultimately based on complementary base pairing between the mRNA codon and the aa-tRNA anticodon. How- ever, binding energy differences between cognate and near-cognate aa- tRNAs are too small to account for the observed high-level accuracy in vivo (misreading error rates estimated at 10-3 to 10-4 per codon (Kramer & Farabaugh 2007)). To account for the high level of accuracy, both biochemical and structural approaches have yielded data to describe how aa-tRNA

22 discrimination proceeds following delivery of the ternary complex at the A- site.

Recognition of the cognate aa-tRNA Initial binding of the ternary complex is rapid and results in the formation of a kinetically labile intermediate. This is followed by a codon recognition step wherein non-cognate tRNAs are likely to be rejected due to their lower affinity for the mRNA codon and thus dissociate before stimulating the hydrolysis of GTP on EF-Tu. Those cognate and near-cognate complexes that manage to proceed are screened again by means of a proofreading strategy after the irreversible hydrolysis of GTP (Rodnina & Wintermeyer 2001). This discriminatory mechanism promotes initial rejection based on the affinity of aa-tRNA for the codon without requiring the expenditure of GTP. However, if this fails the ribosome is afforded a second, more costly opportunity to reject the aa-tRNA based on an induced fit mechanism. Kinetic proofreading essentially enhances aa-tRNA selectivity as the overall accuracy is the product of the accuracies of two separate selection steps. From a structural perspective, base pairing at the A site is geometrically screened by means of conformational rearrangements that occur upon cognate aa-tRNA binding. These rearrangements involve universally conserved bases A1492 and A1493 located within helix 44 (h44) and G530 from the shoulder domain (Ogle et al. 2003). Together these three bases monitor the shape of the codon-anticodon helix and stabilize Watson-Crick base pairs in the first and second positions upon cognate tRNA binding. Along with G530 r-protein S12 plays a small role in recognition by coordinating a magnesium ion at the interface of bases at the third position. However, the third or “wobble” position is not stringently monitored in agreement with the observation that non-Watson-Crick base pairs are accepted here (Ogle et al. 2003). Due to the distorted conformation of a near-cognate codon-anticodon helix, the interactions described above are not induced prompting efficient rejection of the aa-tRNA. In addition to these local conformational rearrangements, the cognate reaction is accompanied by global changes in the architecture of the 30S subunit. Of particular functional interest, is the movement of the shoulder and head domains relative to the rest of the body resulting in a “closed” conformation (Ogle et al. 2002). The induced conformational changes that occur upon the cognate reaction lead to the rapid hydrolysis of GTP on EF-Tu and it is released from the ribosome as EF-Tu•GDP and Pi. The amino-acyl end of the tRNA is liberated to swing into position at the A-site of the 50S subunit (accommodation) for immediate peptidyl-transfer. This chain of events is disrupted if mis- matches occur in the codon-anticodon complex and the tRNA dissociates with high probability. Regeneration of EF-Tu•GTP in the cell is accom- plished by the action of EF-Ts, that acts as a guanine nucleotide exchange factor (GEF) to replace GDP with GTP, thereby recycling EF-Tu in its GTP

23 form for successive rounds of ternary complex formation (Rodnina et al. 2000).

Peptidyl transfer The peptidyl transferase reaction is the core enzymatic activity of the large 50S subunit and the residues involved belong to domain V of the 23S rRNA, an area almost completely devoid of protein. One of the last surviving protein candidates, considered to play a potential role in catalysis, was r-protein L27. The 70S crystal structure by Selmer et al. captured the tail of L27 in the vicinity of the PTC in close proximity to the P-site tRNA (Selmer et al. 2006). However, rather than a direct role in peptidyl-transfer it seems more likely that this protein contributes mildly to catalysis via stabilization of the amino-acyl end of the P-site tRNA. This is supported by a study in which peptidyl-transferase activity was reduced using versions of L27 truncated at the N-terminus (Maguire et al. 2005) but its role is not essential as a strain lacking the protein is viable (Wower et al. 1998). Furthermore, the lack of universal conservation of L27 supports the idea that this protein does not form part of an evolutionary conserved mechanism. The ribosome catalyzes protein synthesis by providing an environment that favours optimal alignment and positioning of the substrates (A- and P- site tRNAs) for spontaneous product formation. Prior to accommodation of the A-site tRNA, the rRNA of the PTC shields the carbonyl-carbon of the P- site tRNA to prevent a nucleophilic attack by water molecules. Accommoda- tion of the A-site tRNA induces a conformational change that re-orients the rRNA surrounding the P-site tRNA to expose the carbonyl-carbon that is subsequently attacked by the -amino group of the A-site tRNA (Steitz 2008). Thus the catalytic ability of the ribosome is the result of correct orientation of its substrates, an event that accelerates the peptidyl-transferase reaction. The product is a pre-translocation complex consisting of a new peptidyl tRNA, extended by one amino acid in the A-site and a deacylated tRNA in the P-site. The highly conserved protein factor EF-P is an essential translation factor in bacteria and is present in the cell at about 0.1 to 0.2 copies per ribosome (Aoki et al. 1997). This factor was first detected based on its ability to enhance peptide bond formation between fMet-tRNAfMet and puromycin. Thus it was proposed that EF-P had an analogous function to eukaryotic eIF-5a in promoting synthesis of the first peptide bond. This was supported by the results of chemical footprinting studies that mapped its binding site to domain V of the 23S rRNA, in proximity of the PTC (Aoki et al. 2008). Inter- estingly, the tertiary structure of EF-P mimics the “L-shape” of tRNAs (Hanawa-Suetsugu et al. 2004), although in the X-ray crystal structure of EF-P bound to the 70S ribosome from T. thermophilus, the protein does not bind to a canonical tRNA-binding site. It is rather positioned between the P- and E-sites, spanning both ribosomal subunits and interacting with the L1

24 stalk and the backbone of the P-site tRNA (Blaha et al. 2009). Rather than inducing any conformational changes in the tRNA or PTC, the crystal structure suggests an indirect role for EF-P in stimulating peptide bond formation by stabilizing the orientation of the P-site tRNA for rapid peptidyl-transfer.

Translocation Following peptidyl transfer, the ribosome is configured in a pre-translocation state with the new peptidyl tRNA in the A-site and a deacylated tRNA in the P-site. For the progressive addition of amino acids to the nascent chain, the tRNAs and mRNA must be re-positioned by a distance of exactly one codon. This synchronous movement, termed translocation, is required for presenta- tion of the next codon at the A-site for further rounds of decoding and peptidyl transfer. This occurs in a stepwise manner and is catalyzed by the hydrolysis of GTP on EF-G.

Hybrid states model The two-step or hybrid model of translocation is now widely accepted and refers to the uncoupling of tRNA movement on the small and large subunits. Experimental evidence for the formation of tRNA hybrid states was first demonstrated by chemical probing studies designed to track the pathway of tRNA movement after peptidyl transfer (Moazed & Noller 1989). This landmark paper showed that A- and P-site tRNAs move spontaneously after peptide bond formation with respect to the 50S subunit. In other words, the relative positions of each tRNA differ with respect to each subunit. The A- site tRNA occupies the A/P hybrid state (anticodon stem remains at the A- site on the 30S with the amino-acyl stem positioned at the P-site on the 50S). Likewise the P-site tRNA occupies the P/E hybrid state (anticodon stem remains at the P-site on the 30S with the amino-acyl stem positioned at the E-site on the 50S). This sampling of intermediate tRNA binding sites on the ribosome has since been visualized by cryo-EM (Agirrezabala et al. 2008) and is accompanied by a ratcheting of the small subunit relative to the large subunit. Single molecule fluorescence resonance energy transfer (smFRET) experiments utilizing fluorescently labeled A- and P-site tRNAs have established that in the absence of EF-G, the tRNAs fluctuate between the classical and hybrid states (Blanchard et al. 2004) and upon EF-G binding the hybrid configuration is stabilized (Spiegel et al. 2007). In the second step the tRNAs move with respect to the 30S subunit to oc- cupy classical E/E and P/P states which is coupled to the movement of mRNA by one codon. This configuration represents the post-translocation complex and requires the hydrolysis of GTP on EF-G. This protein binds the 70S ribosome at the intersubunit space on the A-site side and according to the Wintermeyer group it binds initially in its GTP form (Wilden et al. 2006). This is followed by rapid hydrolysis of GTP where the released energy acts as a motor to “unlock” the ribosome and drive formation of the

25 post-translocated state (Rodnina et al. 1997). Subsequent Pi release results in further conformational changes prompting dissociation of EF-G•GDP from the ribosome. The numerous events that occur during translocation immediately invites the expectation that the ribosome itself undergoes dynamic rearrangements. Indeed the process requires many large-scale re-organizations that have been illuminated by techniques such as cryo-EM and smFRET. One particularly striking alteration is known as ratcheting and involves a rotation of the small subunit relative to the large subunit (Frank & Agrawal 2000) as well as swiveling of the head of the small subunit towards the E-site (Ratje et al. 2010). These rearrangements are accompanied by movement of the L1 stalk on the 50S subunit towards the E-site. The specific details of the dynamics of L1 stalk movement and its coupling to tRNA translocation and subunit rotation is currently an unresolved matter. However, this region of the ribosome, which consists of helices 76-78 of the 23S rRNA and r-protein L1, appears to play an active role in formation of the hybrid state, tRNA translocation and release of deacylated tRNAs from the E-site (Trabuco et al. 2010). A recent publication by Fischer et al. takes advantage of time resolved cryo-EM to demonstrate that a surfeit of ribosomal configurations are sampled during translocation implying that the array of rearrangements is larger than formerly anticipated (Fischer et al. 2010).

Termination In eubacteria arrest of the elongation phase is signaled upon exposure of one of three stop codons (UAA, UAG or UGA) at the ribosomal A-site. As opposed to sense codons, which are recognized by tRNAs, release factor proteins that have overlapping specificity recognize these codons. The class I release factor RF1 responds to the stop codons UAA and UAG, while RF2 responds to UAA and UGA (Kisselev & Buckingham 2000). The scenario in eukaryotes and archaea differs as a single release factor recognizes all three stop codons. The class II release factor RF3 is a GTPase and replenishes the cellular pool of free RF1 and RF2 by stimulating the dissociation of RF1/2 from the ribosome after release of the completed polypeptide (Zavialov et al. 2001). RF1 and RF2 are highly conserved among eubacteria whereas some species lack RF3 (Sprinzl et al. 2000). Until recently, the detailed mechanism of stop codon recognition and polypeptide release has been poorly understood. The molecular basis of stop codon discrimination by RF1 and RF2 was deciphered by Ito et al., where tripeptide motifs in domain 2 of each factor were shown to be responsible for direct recognition of the codons exposed at the A-site (Ito et al. 2000). By means of amino acid substitution studies, the authors showed that the tripeptide Pro-X-Thr of RF1 and Ser-Pro-Phe of RF2 were analogous to the anticodon triplet of tRNAs in determining specificity of the release factor. The

26 authors could show that exchange of these motifs between RF1 and RF2 also exchanged their stop codon specificity. Recent crystal structures provide some support for this model although they also reveal that these motifs alone are only partially responsible for determining specificity (Korostelev et al. 2008a; Laurberg et al. 2008; Weixlbaumer et al. 2008). Class I RFs have dual activity in recognizing the stop codon and trigger- ing release of the newly synthesized protein from the P-site tRNA. In both prokaryotes and eukaryotes, domain 3 of RF1 and RF2 contains a conserved Gly-Gly-Gln (GGQ) motif that is positioned at the PTC in all three high- resolution crystal structures of the post-termination complex (Korostelev et al. 2008a; Laurberg et al. 2008; Weixlbaumer et al. 2008). The Gln residue of this motif is methylated by the methyltransferase PrmC in E.coli and the crystal structure demonstrates that RF1 adopts a compact structure when complexed with this enzyme (Graille et al. 2005). In contrast the release factors must adopt an extended conformation on the ribosome to enable simultaneous interactions at the 30S A-site and PTC of the 50S subunit. The structures predict that docking of the GGQ motif at the PTC induces a conformational rearrangement of the 23S rRNA to expose the ester bond that tethers the peptide to the P-site tRNA for a nucleophilic attack by a water molecule. Furthermore, the methylated Gln residue which stimulates the efficiency of termination, is proposed to coordinate the water molecule for the hydrolytic reaction as predicted by molecular-dynamics simulations (Trobro & Aqvist 2007). Stop codon recognition and hydrolysis are strictly coupled which raises questions regarding how the two events are coordinated. Although this question has no clear answer, Noller suggests that specific conformational rearrangements in the rRNA upon stop codon recognition directs and stabilizes the positioning of the GGQ motif in the PTC for hydrolysis to occur (Korostelev et al. 2008a). Following release of the newly synthesized polypeptide, RF3 recycles the class I RFs by catalyzing their dissociation from the post-termination complex. The mechanism governing this reaction has been dissected biochemically. According to experiments performed by Zavialov et al., RF3 initially binds to the ribosome after peptide release in its GDP form. Exchange of GDP for GTP on RF3 occurs on the ribosome and in this new conformation, dissociation of the class I RFs is stimulated. This is followed by GTP hydrolysis on RF3 and release of the factor (Zavialov et al. 2001). In its GTP- bound state, RF3 was also shown to induce ratcheting of the ribosome providing a structural basis for the rapid release of RF1/2 from the ribosome (Gao et al. 2007).

27 Ribosome recycling

At the end of the termination phase, the 70S ribosome is still bound to the mRNA template and carries a deacylated tRNA in the P-site. During the last phase, these ligands are removed and the 70S is split to allow the subunits to participate in further rounds of translation. This event requires the interplay of three translation factors: ribosome recycling factor (RRF), EF-G and IF3. The sequence of events that leads to disassembly of post-termination complexes has been examined biochemically by Karimi et al. By means of ribosome sedimentation analysis, it was shown that post-termination 70S complexes require RRF and EF-G in its GTP state for dissociation into subunits. Furthermore, the release of deacylated tRNA was prevented if GTP was replaced by a non-cleavable analogue implying that GTP hydrolysis is required for splitting. The authors propose that RRF inhibits the release of EF-G•GDP, which has low affinity for the ribosome, to channel the energy of the conformational change in EF-G into ribosome splitting. The product of this event is a free 50S and a 30S subunit bound to mRNA with a deacylated tRNA in the partial P-site. This 30S complex is the substrate for IF3 that binds and ejects the tRNA. The authors propose that this IF3-bound 30S subunit slides along the mRNA to allow either re-initiation or mRNA release (Karimi et al. 1999). The structural similarities between RRF and tRNA advocated the idea that EF-G promoted translocation of RRF from the A-site to the P-site during recycling and this action promoted tRNA ejection (Hirokawa et al. 2002). However a recent publication by Peske et al., re-examined this hypothesis and found that the position of mRNA relative to the ribosome was unchanged during recycling providing direct evidence that a translocation- like event does not occur (Peske et al. 2005). Using a FRET assay, their experiments validated the conclusions of Karimi et al., and showed that RRF and EF-G•GTP were necessary and sufficient for 70S splitting. Since subunit dissociation was ten times faster than tRNA release, the data suggested that splitting occurs first and tRNA ejection follows as a second step, in agreement with the sequence of events proposed previously (Karimi et al. 1999). Structural studies of RRF bound to the ribosome have yielded contradictory interpretations of its role in subunit dissociation. Cryo-EM images suggest that RRF could be capable of disrupting the intersubunit bridges B2a and B3 (Gao et al. 2005), while a high-resolution crystal structure predicts that such changes in intersubunit contact upon RRF binding appear unlikely (Weixlbaumer et al. 2007).

28 Ribosome biogenesis and assembly The complexity of the ribosome is adequately reflected in the elaborate cascade of overlapping events that ultimately lead to construction of the macromolecular particle. Ribosome biogenesis entails synthesis, folding, processing and assembly of 3 large strands of rRNA as well as the addition of more than 50 proteins. The large number of components not only makes the process inherently complicated but it also demands a highly organized system of synthesis and assembly. Considering the fact that cell fitness and optimal growth are tightly coupled to efficient protein synthesis, the production of functional ribosomes is of crucial importance (Bremer 1996). With the identification of a growing number of processing enzymes and auxiliary factors as well as significant advancements in the techniques used to monitor assembly in vitro, the biogenesis field has been enjoying a long- awaited boost over the past few years. What has emerged is the realisation that rather than a series of discrete, ordered steps, biogenesis consists of a network of pathways that occur sequentially and in parallel to secure coordinated construction of the subunits (Talkington et al. 2005; Adilakshmi et al. 2008; Mulder et al. 2010). Although the in vitro studies have yielded a wealth of important information, how biogenesis proceeds in vivo persists as one of the daunting problems in the field. Two key issues are exclusive to the in vivo process: (1) the existence of a large number of ancillary assembly factors and (2) the likely co-transcriptional coupling of assembly in the cell. However a recent study reports the identification of an ensemble of heterogeneous 30S and 50S assembly intermediates in vivo (Sykes et al. 2010) implying that multiple parallel assembly trajectories also exist during cellular assembly. Although an exact scheme of assembly events has yet to be established, the apparent similarities of intermediate states in vivo and in vitro implies that the experimental approaches are moving in the right direction. rRNA synthesis In both prokaryotes and eukaryotes, rRNA represents the major product of all cellular transcription even though the rRNA operons (rrn) encompass only 0.5% of the E. coli genome (Bremer 1996). Considering that r-protein synthesis is regulated by the availability of free rRNA in the cell, the transcription of rRNA is rate-limiting for the biogenesis of ribosomes (Paul et al. 2004). In E. coli and S. typhimurium, there are seven rrn operons designated rrnA, rrnB, rrnC, rrnD, rrnE, rrnG and rrnH. Initially, the rRNAs are transcribed as a precursor primary transcript containing a single copy of 16S, 23S and 5S rRNA as well as tRNAs, the identity of which depends on the particular operon. The operons contain two strong promoters, P1 and P2, which are separated by 120 bp. The P1 promoter is the most active during

29 moderate and fast growth (Paul et al. 2004) whereas P2 appears to be used during periods of slow growth (Kaczanowska & Ryden-Aulin 2007). The rate of rRNA synthesis is tailored to the nutritional conditions of the cell to allow fastest growth when nutrients are plentiful and curtail metabolic effort and energy costs during starvation. The number of ribosomes in a rapidly dividing E. coli cell has been estimated to reach as many as 70,000 and during slower growth this number can drop approximately 10-fold (Bremer 1996). Thus, several mechanisms are in place to control rRNA synthesis and cellular ribosome content to ensure that most of those synthesized are engaged in translation and are required for growth. Briefly, these include the stringent response that is induced during amino acid starvation and mediated by up-regulation of the alarmones ppGpp and its precursor pppGpp. The physiological effects of these molecules are complex and lead to alterations in the gene expression profile of cells. Although the exact mechanism of (p)ppGpp mediated regulation of ribosome synthesis is unresolved, it has been proposed that these two molecules are negative regulators of transcription from the P1 promoter and thus reduce the wasteful synthesis of rRNA when the cell is starved for charged tRNA molecules (Gourse et al. 1996). Ribosome production increases with the square of the growth rate to meet changing demands for protein synthesis (Bremer 1996), a phenomenon known as growth-rate dependent control. Whether this mechanism is mediated by ppGpp or nucleoside triphosphate (NTP) concentration is currently a topic of debate in the literature (Nomura 1999). A phenomenon known as feedback control has been discussed within the context of this regulatory mechanism, and refers to changes in rRNA transcription according to the translational capacity of the cell. This mechanism was illuminated by studies where either a decrease or increase in rrn gene dosage was compensated via the up- or down-regulation of rrn expression from the available operons to maintain rRNA synthesis at a level appropriate for steady-state growth (Jinks-Robertson et al. 1983; Condon et al. 1993).

R-protein synthesis In the majority of bacteria the genes encoding all of the r-proteins are present as a single copy, and are mostly clustered together in large polycistronic operons. Their rate of synthesis is dictated by the extent of rRNA transcription and is often autogenously regulated at the translational level. This means that one of the products of the operon recognizes and binds to a site on the polycistronic mRNA and inhibits translation of all genes encoded by the operon. This can occur via two alternative mechanisms known as displacement and entrapment (Schlax & Worhunsky 2003). The term displacement refers to the competitive binding of the repressor and 30S subunit to a common binding site on the mRNA. The repressor acts to “displace” or physically block the 30S subunit from accessing the ribosome binding site

30 (RBS). The entrapment model is more complicated in that the repressor regulates the equilibrium of active and inactive/misfolded mRNA conformations. The crucial difference between this mechanism and displacement is the fact that binding of the repressor and 30S subunit is not mutually exclusive as they do not share a common target site. The repressor binds to and stabilizes the misfolded mRNA conformation and traps the 30S subunit in an inactive ternary complex. This mode of autogenous regulation is mediated by the r-proteins S15 and S4 (Benard et al. 1996; Schlax & Worhunsky 2003). Dual RNA binding activity prompted Nomura to suggest that the binding sites of autoregulatory r-proteins on rRNA and mRNA mimic each other structurally and constitute competitive binding sites (Nomura et al. 1980). This model is attractive in the sense that it provides an elegant mechanism that couples r-protein expression to the translational demands of the cell. As long as the preferential rRNA binding sites sequester the regulatory r- proteins, the mRNAs that encode them are liberated from this feedback regulation. The target sites of small subunit protein S8 provide one good example of this type of molecular mimicry. Merianos et al. have shown that the crystal structure of S8, bound to its mRNA target site, is more or less identical to that of S8 bound to helix 21 of the 16S rRNA (Merianos et al. 2004). With relevance to the r-proteins under investigation here, both L1 and S20 act as translational repressors. L1 regulates synthesis of both products of the rplK-rplA operon, L1 and L11. By means of site-directed mutagenesis, the binding site of L1 was mapped to a stem structure on the bicistronic mRNA that overlapped the L11 translation initiation site (Baughman & Nomura 1984), suggestive of a displacement mechanism of inhibition. This stem loop resembles the L1 target site on the 23S rRNA. In the presence of 23S rRNA, L1 suppression of r-protein synthesis was eradicated, implying that L1 binds preferentially to its rRNA target (Yates & Nomura 1981). The rpsT gene that encodes S20 sits alone in an operon and feedback regulates expression of itself (Parsons & Mackie 1983). Experiments by Parsons et al. demonstrated that the S20 target site on its own mRNA included its unusual UUG start codon and possibly adjacent regions in the 5’ leader sequence (Parsons et al. 1988). The authors also raise the possibility that efficient repression may require the mRNA to be bound to the ribosome. The gene encoding r-protein L17 is part of the alpha-operon that is autogenously regulated by small subunit protein S4 via the entrapment mechanism.

Processing and modification of rRNA and r-proteins In wild-type E. coli, immature rRNA represents only 1-2% of the total rRNA indicating that processing to mature products occurs rapidly (Srivastava & Schlessinger 1990). Maturation is a multistep process consisting of two main events (1) cleavage by RNAses to produce rRNA strands of the correct

31 length and (2) chemical modification of the rRNA nucleotides at a number of positions. In E. coli and S. typhimurium, rRNA operons have the following basic order: 16S – internal transcribed spacer I (ITS-1) – tRNA – ITS-2 – 23S – ITS- 3 – 5S (Figure 4). Complementary regions that flank the 16S, 23S and 5S sequences are base paired in the primary transcript to form so-called “processing stems”, the substrates for endoribonuclease cleavage. The first enzyme to act is RNase III which separates the rRNAs and tRNAs from each other to generate pre-16S rRNA (17S), pre-23S rRNA, pre-5S rRNA (9S) and depending on the operon, a few tRNA precursors (Deutscher 2009). In an RNase III deficient E. coli strain, mature 16S is formed but normal maturation of the 23S rRNA is perturbed although it is still functional (King et al. 1984). The final trimming of the precursor rRNAs is performed by exonucleases in the context of pre-ribosomal particles (Deutscher 2009). The RNases E and G are required to generate the mature 5’ end of 16S rRNA and an unknown enzyme is responsible for final maturation of the 3’ end. RNase T in E. coli performs final maturation of the 3’ end of 23S rRNA. In Salmonella and some other proteobacteria processing of 23S rRNA is complicated due to the presence of intervening sequences (IVSs) that are removed by RNase III resulting in a 23S rRNA that is fragmented but still functional (Burgin et al. 1990). Interestingly, ribosomes from a Salmonella mutant lacking RNase III retain the IVSs and remain functional suggesting that fragmentation is not required for the activity of Salmonella ribosomes (Mattatall & Sanderson 1998). Both termini of the 9S rRNA species are processed by RNase E and final maturation is performed by RNase T at the 3’ end and an unknown enzyme at the 5’ end generating 5S rRNA (Deutscher 2009). These rRNA processing steps are believed to initiate co- transcriptionally before transcription of the entire operon is complete.

16S 23S 5S G ? ? T ? T E tRNA III E III III III E III III 5’ P1 P2 3’ T1 T2 Figure 4: Schematic drawing depicting the organization of an rrn operon. Comple- mentary base pairing results in the formation of processing stems enclosing each rRNA species. The positions of RNase cleavage sites are indicated with arrows. The letters “P” and “T” refer to promoter and terminator sequences respectively. Adapted from (Kaczanowska & Ryden-Aulin 2007).

32 Modification of rRNA is common to all organisms although the exact position and number of these modifications differs. The two main types are methylations and pseudo-uridinylations that are mostly found at functionally important, conserved regions of the ribosome (Decatur & Fournier 2002). The functional significance of these modifications is unclear due to the apparent non-essential nature of some of them (Green & Noller 1996; O'Connor & Gregory 2011). Since they expand the chemical diversity of rRNA nucleosides, it has been proposed that they may alter the folding and interactions of rRNA (Kaczanowska & Ryden-Aulin 2007). Recently, underlying functional roles in ribosome biogenesis have been established for some modification enzymes. One example is provided by the universally conserved KsgA enzyme that dimethylates two adjacent adenosines in the 16S rRNA. The viability of kasugamycin resistant strains, that lack modification at these sites, demonstrates that its methylation function is non-essential. The enzyme interacts with the same site of the 30S subunit as IF3 and 50S subunits (Xu et al. 2008). Thus its release is required to allow formation of the 70S initiation complex. Using a mutant allele of ksgA, incapable of methylation, Connolly et al. cleverly showed that its methyltransferase activity is a pre-requisite for dissociation of the enzyme suggesting that methylation per se was secondary to its role in regulating 30S biogenesis (Connolly et al. 2008). Many of the r-proteins are also subject to chemical modification, the most well known being the acetylated form of L12, referred to as L7. Small subunit proteins S5, S6, S11, S12 and S18 along with large subunit proteins L3, L11, L16 and L33 also undergo modification, some carrying more than one modification. The most common r-protein modification is methylation and it is observed in all three kingdoms (Polevoda & Sherman 2007). Similar to rRNA, the significance of these modifications remains unclear, although some of the methylated proteins are functionally important during various steps in translation. For example, L7/L12 is involved in the recruitment and GTPase activation of translation factors (Savelsbergh et al. 2000) and is required for rapid subunit association in the presence of IF2 (Huang et al. 2010).

Subunit assembly Construction of the ribosomal subunits from their raw materials has been a subject of intense interest ever since the discovery of the ribosome more than 50 years ago. As substantiated from the work of in vitro reconstitution studies, it can be described as a self-assembly process as all of the information required to manufacture each subunit is contained within the primary sequences of the rRNA and r-proteins. However given the non-physiological requirements of in vitro assembly, an array of accessory factors that stream- line and accelerate the process are involved in vivo (Wilson & Nierhaus

33 2007). The existence of a so-called “assembly gradient” has the added advantage of directionality due to the coupling of assembly to rRNA transcription in the cell. This simplifies the process and makes it more energetically favourable due to the involvement of fewer initial components (Nierhaus 1991). Much of what we know is based on the 30S subunit, which due to its smaller size, represents the simplest ribosomal complex.

30S subunit assembly In the early 1970s, Nomura and colleagues pioneered the development of an in vitro assembly map for the 30S subunit by reconstituting mature 16S rRNA and a mixture of small subunit proteins (Mizushima & Nomura 1970; Held et al. 1974). These experiments were the first indication that the addition of r-proteins to rRNA was a hierarchical and highly cooperative process. As such the proteins were classified as primary, secondary or tertiary binders depending on whether they were capable of binding naked rRNA or required an RNA-protein complex for binding. Proteins S4 and S7 were later identified as the assembly initiator proteins, befitting the observation that they directly contact the rRNA (Nowotny & Nierhaus 1988). The seminal experiments of Nomura provided a springboard that has lead to the application of a host of other techniques aimed at dissecting the process further. One successful approach made use of time-resolved hydroxyl radical footprinting, a technique developed in the Woodson lab that measures the solvent accessibility of the rRNA backbone (Woodson 2008). Using the 5’ domain of 16S rRNA, Adilakshmi et al. were able to demonstrate that, in the presence of high Mg2+, almost all of the predicted tertiary interactions were made in the absence of proteins (Adilakshmi et al. 2005). This clearly indicated that the folding instructions are an inherent feature of the rRNA primary sequence. It was also noted that the proteins were required to stabilize the structure at physiological conditions and since up to half of the rRNA population was trapped in misfolded conformations, it seemed likely that the proteins were also needed to facilitate re-folding of non-productive intermediate states. In a follow-up study, this time using the entire 16S rRNA sequence and 30S proteins, they found that nucleotides from different domains had similar rates of protection which would imply simultaneous and independent folding of the 16S rRNA domains and r-protein binding (Adilakshmi et al. 2008). Their data supports the existence of multiple parallel assembly intermediates containing various subsets of proteins. Another informative approach utilized to monitor the kinetics of assembly makes use of pulse-chase labeling monitored by quantitative mass spec- trometry. Talkington et al. developed this effective method to measure the rates of r-protein binding to assembling 30S complexes (Talkington et al. 2005). The results demonstrate that the order of protein binding matches that predicted by Nomura’s equilibrium assembly map (Figure 5) and rather than a single trajectory, multiple routes are available to achieve the same final

34 conformation. This notion is also supported by an in vivo study, in which an ensemble of heterogeneous 30S and 50S intermediates was observed when cells were treated with the assembly inhibitor neomycin (Sykes et al. 2010). Such an assembly landscape would have the advantage of circumventing complete breakdown of the biogenesis cascade in the event of mutations or deficiencies. This is exemplified by the viability of an E. coli mutant lacking r-protein S15, a protein initially believed to be essential for assembly of the platform of 30S subunits (Bubunenko et al. 2006). Time-resolved electron microscopy has also facilitated visualization of structural intermediates during 30S assembly, and this method provides further evidence of the existence of parallel assembly pathways (Mulder et al. 2010).

16S ribosomal RNA 5’ 5’ domain Central domain 3’ domain 3’

S17 S20 S4 S8 S11 S15 S7

S16 S6 S18 S9S13 S19

S12 S5 S21 S10 S14

S3 S2

Figure 5: Order and dependencies of protein binding during in vitro assembly of the 30S subunit based on the binding kinetics of r-proteins. Proteins are coloured according to their binding rates; red are the first to bind, followed by orange, green, and blue. R-proteins coloured grey are the last to bind. Adapted from (Talkington et al. 2005).

50S subunit assembly Much less attention has been devoted to assembly of the large subunit owing to the fact that the process is compounded by twice as much rRNA and almost twice as many proteins compared to the 30S subunit. The lack of corre- spondence between the six secondary domains of the 23S rRNA and its tertiary domains provided the first clue that the assembly process would be more complex than that of the 30S subunit. Nevertheless, an assembly map for the 50S subunit was developed using a similar reconstitution technique as that employed by Nomura (Rohl & Nierhaus 1982) (Figure 6). Similar to the

35 30S, two proteins were identified as assembly initiator proteins, L3 and L24 (Nowotny & Nierhaus 1982) and L20 was later shown to partially fulfill the role of L24 when this protein was absent during reconstitution (Franceschi & Nierhaus 1988). Proteins L5, L18 and L25 mediate interactions between the 23S and 5S rRNAs and together with the 5S rRNA constitute the central protuberance of the large subunit (Korepanov et al. 2007). Given the large repertoire of 50S intermediates generated upon treatment with neomycin in vivo (Sykes et al. 2010), it seems evident that the complexity of 50S assembly shares similarities with that of the 30S subunit although we are much further away from a detailed scheme of events.

23S ribosomal RNA 5‘ 13S 8S 12S 3‘

L20 L4

L13 L22 L17 L24 L9 L1 L19 L3 L29 L23 L21 L2

L11 L5 L7/L12 L10 5S L18 L14 L16 L15

L27 L6 L25 Figure 6: In vitro assembly map of the 50S subunit. Thick, thin and dashed arrows indicate strong, weak, and very weak binding interdependencies. The line enclosing proteins L5, L18 and L15 indicates that these three proteins mediate interactions with the 5S rRNA. Adpated from (Rohl & Nierhaus 1982). Structural studies have indicated that L25 replaces L15 as one of the principal mediators of 5S rRNA interactions (Lu & Steitz 2000).

Ancillary factors In addition to the various enzymes that process and modify the ribosomal components, a seemingly ever-expanding number of protein factors appear to be directly involved in assembly. These include GTPases, helicases, chaperones, ribosome maturation factors and a diverse array of factors of unknown function (Connolly & Culver 2009). In all likelihood such factors influence the course of assembly in vivo despite the fact that some are non- essential. Indeed, many of these assembly factors were indentified by means of deletion studies in which ribosomal precursor particles and immature

36 rRNA species were detected (Charollais et al. 2003; Peil et al. 2008; Nord et al. 2009). Delineating the temporal order of their binding and departure on the assembling ribosome, their precise targets and their potential interactions with other factors will undoubtedly help in piecing together the assembly puzzle. The high level of conservation among bacteria as well as the essential nature of some of these extra-ribosomal factors, for example Era (Inoue et al. 2003) and Der (Hwang & Inouye 2006), has generated much interest as they could potentially be exploited as targets for emerging antibacterial compounds (Comartin & Brown 2006).

Degradation and turnover of ribosomal components No biological process is entirely error-free and given the inherent complexity of ribosome synthesis and assembly, departures from the productive pathways can and do occur, even in wild-type cells. Misassembled ribosomes can pose a threat to the cell by, for example, sequestering translational factors and r-proteins or impeding the activity of their productive counterparts. Under normal growth conditions rRNA and tRNA are very stable, however direct evidence to support a quality control system targeted at rRNA degradation has recently become available. Cheng et al. discovered that a double knockout of the exoribonucleases RNase R (rnr gene) and PNPase (pnp gene) is lethal in E. coli by combining an rnr deletion with a temperature-sensitive mutation in pnp (Cheng & Deutscher 2003). Shifting the growth temperature from permissive to non-permissive resulted in the accumulation of a large amount of fragmented rRNA accompanied with a decrease in 16S and 23S rRNA leading to eventual cell death. Although the manner of fragment formation was not established, the data suggest that they are by-products of normal rRNA metabolism and are rapidly cleared in wild- type cells due to the action of RNase R and PNPase. Regarding r-protein turnover, Pulk et al., have shown that upon chemical damage a sub-set of r-proteins can be exchanged on the ribosome in vitro to restore activity (Pulk et al. 2010). By using N-ethylmaleimide to modify cysteine residues the authors could show that this treatment leads to defects in ribosome activity. However incubation of these damaged ribosomes with an extract of 70S proteins resulted in a two-fold improvement in activity, suggesting that the modified r-proteins were replaced with undamaged counterparts. Unlike ribosome reconstitution from isolated components, this r- protein replacement was temperature independent indicating that no major conformational changes were required. In line with this observation, it was found that surface-exposed, loosely bound proteins were responsible for reactivation. In addition, the authors provided evidence that such replacement occurred in vivo during stationary phase when proteins are most prone to chemical damage as a result of oxidative stress (Nystrom 2004). As opposed

37 to degradation of rRNA, the data here suggests that a fraction of defective r- proteins are replaceable which is ultimately more beneficial for the cell in terms of energy consumption. Further evidence for the recycling of r- proteins is provided by a study in which the origin of r-proteins in precursor ribosomal particles was detected using pulse-labeling (Sykes et al. 2010). In this study a mixture of labeled (newly synthesized) and unlabeled (re-used) r-proteins were detected in the immature 30S subunit peak following sedimentation analysis. Whether these precursor 30S subunits were new, matur- ing particles or old, degraded particles could not be distinguished although fragmented rRNA was not detected suggesting that the former scenario is more feasible.

Ribosomal Proteins The majority of research in the ribosome field is dedicated towards an understanding of the detailed mechanisms governing translation, which has lead to particular fixation on rRNA. This is not surprising considering that rRNA predominates at the main functional sites and is responsible for decoding and peptidyl-transfer, the fundamental tasks of the ribosome. Generally speaking the r-proteins are considered integral to endowing an otherwise flaccid RNA structure with crucial structural support to tighten rRNA domain interactions thereby favouring a firm and robust structure (Cech 2000). The exact number of proteins tends to vary according to species but in E. coli and S. typhimurium, 21 proteins are associated with the small subunit (S proteins) and 33 with the large subunit (L proteins) (Wilson & Nierhaus 2005). Eukaryotic ribosomes contain approximately 20-30 additional r- proteins (Doudna & Rath 2002).

General features Bacterial r-proteins come in a range of sizes, the smallest encompassing 46 amino acids (L34), and the largest 557 amino acids (S1). All r-proteins are present as a single copy per ribosome, the only exception being L7/L12 which is present as a dimer of dimers, referred to as the L7/L12 stalk (Ramakrishnan & White 1998). Atomic resolution structures of the ribosomal subunits revealed important insights into the structure and spatial organization of the r-proteins and their interactions with rRNA. The majority of the proteins are unevenly distributed at the periphery and back of both subunits, remote from the major functional sites of the ribosome. Many of the proteins are in direct contact with rRNA and consist of a globular, surface exposed domain with extended projections that insert into the mesh of rRNA helices to reach deep into the ribosome interior (Maguire & Zimmermann 2001). The amino acid composition of these long tails appears

38 to be suited to extensive rRNA contact as they contain a high density of basic amino acids such as lysine and arginine, possibly to neutralize the negative charges of the phosphate residues in the rRNA backbone (Wilson & Nierhaus 2005). This feature is likely to be important for promoting correct folding of the rRNA. They also contain a high proportion of glycine residues that are proposed to facilitate threading of the tails through the narrow crev- ices surrounding the RNA helices. As opposed to the globular domains, the structures of these extensions are often disordered in the isolated protein probably because they only assume an ordered structure upon interaction with rRNA (Steitz & Moore 2003). Unlike other nucleic acid binding proteins, the r-proteins seem to rely on rRNA shape rather than sequence for their binding specificities as the majority of the interactions occur via the rRNA backbone rather than base-specific contacts (Brodersen et al. 2002; Klein et al. 2004). As one would suspect rRNA-protein contact is extensive, especially in the large subunit, where 23 of the 50S proteins contact at least two rRNA domains creating an interwoven web of contacts. One extreme example is L22 which interacts with all six domains of the 23S rRNA (Steitz & Moore 2003). Given the comprehensive array of interactions, it is not surprising that the proteins are ascribed a general function in maintaining the overall morphology and architecture of the subunits.

Roles in and outside the ribosome The high level of co-operativity within the ribosome hinders the assignment of defined functions to particular proteins. Yet, research has revealed that certain r-proteins do in fact enrich the ribosome with more than merely structural support and are involved in a variety of processes both within and outside of translation.

Functions within translation The requirement of certain r-proteins for proper ribosome biogenesis and assembly is well documented (Kaczanowska & Ryden-Aulin 2007, and refernces therein), and with the results of paper II large subunit protein L1 can now be added to this list. Other proteins with designated functions include the largest r-protein S1. It is located on the small subunit at the junc- tion of the head, platform and body and interacts with the 5’ region of mRNA to tether it to the 30S during initiation (Sengupta et al. 2001). How- ever the SD-aSD interaction promotes the stronger and more specific attachment of mRNA. Mutants with altered versions of small subunit protein S12 probably represent the most studied r-protein mutants due to the involvement of this protein in decoding. The antibiotic streptomycin is known to alter translational accuracy by inducing an “error-prone” phenotype, the basis of which has been demonstrated structurally. Streptomycin facilitates closure of the 30S

39 subunit by making contacts between the shoulder domain and central part of the subunit. Such interactions resemble those induced upon formation of a cognate codon-anticodon duplex and thus the antibiotic stabilizes domain closure even in the presence of incorrect aa-tRNA substrates (Ogle et al. 2003). This activity results in an increase in the erroneous incorporation of near- and non-cognate tRNAs which in turn reduces the fidelity of translation (Carter et al. 2000). Certain mutations in S12, at residues that contact helices 27 and 44 of the 16S rRNA upon domain closure, confer resistance to streptomycin. These mutated versions of S12 counteract this ram (ribosomal ambiguity) phenotype possibly by destabilizing the closed 30S conformation to introduce an error-restrictive conformation to the A-site while reducing the rate of translation (Ogle et al. 2002). In certain cases the S12 mutants are dependent on streptomycin for growth as the hyper-accurate phenotype is so extreme that it requires destabilization by the antibiotic to restore acceptable levels of error restriction and rescue translation (Carter et al. 2000). In addition mutations in S4 and S5 promote formation of the ram state and are found to compensate the slow-growth phenotype of error- restrictive S12 mutants (Kurland 1996; Maisnier-Patin et al. 2002). Together with S3, r-proteins S4 and S5 have an additional role in unwinding stable secondary structures in the mRNA to facilitate its passage through the entry pore during translation (Takyar et al. 2005). In terms of proteins on the large subunit, the L1 stalk is proposed to be involved in tRNA translocation and deacylated tRNA release (see section on L1). Protein L27 as well as S9 and S13 on the small subunit appear to be important in the positioning of tRNAs at the A- and P-sites for peptidyl transfer, although their mutation or deletion does not completely abolish peptidyl transferase activity implying more of a supportive rather than an essential role (Hoang et al. 2004; Maguire et al. 2005). It is reported that the highly mobile L7/L12 stalk binds directly to and recruits the G-protein factors to the ribosome (Helgstrand et al. 2007) and may play a role in GTP hydrolysis by stabilizing the activated GTPase confirmation of these factors (Diaconu et al. 2005). Extended regions of proteins L4 and L22 protrude into the interior of the polypeptide exit tunnel at a site that is proposed to be involved in interactions with the nascent polypeptide that can in turn regulate translation (Wilson & Nierhaus 2005). Furthermore, the proteins L22, L23, L24 and L29 that are suggested to function as anchors for chaperones, encir- cle the exit site of the tunnel. L23 was shown to be the docking site of the signal recognition particle (SRP) that targets newly synthesized proteins to the cytoplasmic membrane (Gu et al. 2003) and it also binds the chaperone Trigger Factor required to assist the folding of nascent peptides (Kramer et al. 2002).

40 Extraribosomal functions The most well known function of the r-proteins outside of protein synthesis is their ability to feedback regulate their own expression at the translational level as described earlier. Furthermore, one third of all large subunit proteins in E. coli possess chaperone activity and in the case of L1 this is conserved across the three phylogenetic domains (Semrad et al. 2004; Ameres et al. 2007). This observation suggests that these proteins may be involved in chaperoning ribosome assembly but may also function in rescuing or pre- venting RNA misfolding outside of the context of the ribosome. Small subunit protein S10 (NusE) is involved in transcriptional antitermination and recently a direct connection between the RNA polymerase (via NusG) and the ribosome (via NusE) was detected (Burmann et al. 2010). This interaction is likely to facilitate coupling of transcription and translation by keeping the ribosome in close proximity to the nascent mRNA transcript during transcription elongation. In addition, this tethering of the ribosome to the polymerase may also help in the prevention of RNA polymerase back- tracking, suggestive of a role for the ribosome in controlling the rate of transcription (Proshkin et al. 2010).

Conservation In the majority of bacteria, the r-protein genes exist as single copies and are, for the most part, clustered in a few operons on the chromosome. In eukaryotes many r-protein genes exist as paralogous pairs that are commonly scat- tered across the chromosomes (Lecompte et al. 2002). Both Archaea and Eukarya contain more r-proteins than Bacteria and a higher degree of conservation exists between the r-protein complement of Archaea and Eukarya (Lecompte et al. 2002). This would suggest that the archaeal ribosome is a small-scale version of the eukaryotic ribosome. In addition the lack of a shared r-protein pool between the Bacteria and Archaea/Eukarya lends support to the view that the bacterial lineage diverged from the common ances- tor before separation of the Archaea and Eukarya from each other (Figure 7). All three kingdoms share a common pool of 34 r-protein families (Fig- ure 7). The conservation of rRNA and r-proteins follows a similar pattern, in that those proteins that lack universal conservation tend to reside at the periphery of the subunits and are remote from the functional core (Mears et al. 2002). Twenty-three bacterial specific r-protein families exist, indicative of specialization of the bacterial ribosome at an early point in evolution. Among the bacterial lineage, a stable pool of approximately fifty proteins exists (19 in the small subunit and 31 in the large subunit). Proteins S1, S21, L25, and the universally conserved protein L30 show disparate distributions (Lecompte et al. 2002; Mears et al. 2002). In the case of S1 and S21, the pattern of loss shows inconsistencies, as their absence occurs in only a few

41 widely dispersed organisms. This would indicate independent gene losses. It has been suggested that L25 is a late-comer to the bacterial ribosome which may explain its absence in certain lineages (Lu & Steitz 2000). Loss of L30, on the other hand, is a little more puzzling, since it is absent in some bacterial lineages and is conserved in the Archaea and Eukarya (Lecompte et al. 2002).

Eukarya: 78 Archaea: 68 (32:46) (28;40) E: 11 AE: 33 A: 1 (13;20) (4;7) (0;1) BAE: 34 (15;19) BE: 0 BA: 0

B: 23 (8;15)

Bacteria: 57 (23;34) Figure 7: General distribution of r-protein families among the three kingdoms: Eukarya (E), Archaea (A) and Bacteria (B). The numbers indicate the average number of proteins present in each domain or shared between domains, numbers en- closed in parentheses refer to proteins found in the small;large subunits respectively. Adapted from (Lecompte et al. 2002).

As for the role of the extra r-proteins present in eukaryotes, cryo-EM studies have demonstrated that those proteins lacking bacterial homologues form multiple contacts with the extra rRNA sequences (expansion segments) present in the eukaryotic structure (Wool et al. 1995). It is likely that these extra proteins assist in folding and assembly of the larger rRNA species. Other potential roles include interactions with eukaryotic specific translation factors or they may play important extra-ribosomal roles (Dresios et al. 2006). Taking organellar conservation into consideration, a complete rever- sal of the RNA:protein ratio is observed. Mammalian mitochondrial ribosomes are protein-rich and about half of their proteins are homologous to bacterial r-proteins, while the remainder are unique to the mitoribosome (O'Brien 2002). These extra proteins tend to be larger on average than bacterial r-proteins and are found at regions containing truncated rRNA, implying that they fulfill certain functional roles, for example the intersubunit bridges of the mitoribosome consist largely of protein (Sharma et al. 2003).

42 Non-essential r-proteins Despite quite a high level of conservation, many ribosomal proteins are non- essential in bacteria (Dabbs 1991; Ryden-Aulin et al. 1993; Wower et al. 1998; Bubunenko et al. 2007; Korepanov et al. 2007). Loss of these proteins is not fatal, although the ribosome suffers various degrees of impairment. Thus, it is certain that the interplay of proteins and rRNA is required for optimal ribosomal activity. The ribosome is indeed a ribozyme but as Cech pointed out "admittedly one dependent on structural support from protein components" (Cech 2000). The first described mutant lacking a ribosomal protein was identified as a suppressor of a temperature sensitive valyl-tRNA synthetase mutation in E. coli (Wittmann & Stoffler 1975). One of the strains that conferred partial suppression of the phenotype failed to demonstrate r-protein S20 when tested by two-dimensional (2D) gel electrophoresis and using antiserum against S20. A few years later, Dabbs and colleagues set up a selection system to hunt for alterations in r-proteins in E. coli with the aim of assessing ribosomal function and to aid mapping of the r-protein genes on the chromosome (Dabbs & Wittman 1976). Dabbs used a streptomycin-dependent strain of E. coli K12 that had the unique property of spontaneously generating streptomycin-independent mutants. Several of the isolated mutants showed alterations in a total of 28 different r-proteins (as determined by 2D-gel electrophoresis). These spontaneous streptomycin-independent mutants arose at low frequency (approximately 10-10). In an attempt to broaden the variety and frequency of altered proteins, the mutagen N-methyl-N'-nitro-N- nitrosoguanidine (NG) was used to initially mutagenize the streptomycin dependent strain which gave rise to antibiotic-independent revertants at much higher frequencies and alterations in almost all r-proteins (Dabbs 1978). To take it a step further and isolate mutants that lacked an r-protein, antibiotic sensitive cells were mutagenized in another study. These cells were subsequently selected for dependence on erythromycin and the mutants isolated were then selected for reversion to erythromycin independence (Dabbs 1979). The revertants obtained appeared to lack one of six individual r-proteins. In a series of other selection studies, using various chemical mutagens and antibiotic selection screens, strains were isolated lacking one or more of 16 individual r-proteins (S1, S6, S9, S13, S17, S20, L1, L11, L15, L19, L24, L27, L28, L29, L30 and L33) (Dabbs 1991). In all of these studies, absence of each protein was only verified on the protein level by means of 2D-gel and immunological methods using antibodies specific for each missing protein. The mutants exhibited a conditional lethal phenotype and were varied in terms of growth properties. These studies were certainly seminal in terms of showing the non-essential nature of certain r-proteins but follow-up studies designed to determine the nature of the ribosomal defects and growth impairments were hampered by the chemical mutagenesis and

43 selection methods used for the isolation of mutants. These strains were subject to mutagenesis on a genome-wide scale and were therefore not fully defined genetically, which creates difficulty in specifically linking the phenotype to absence of the protein. Since then many advances have been made in the techniques used for bacterial genetic engineering. In recent years, these new methods have been exploited for the generation of E. coli mutants in which the gene of interest is precisely replaced with a selectable antibiotic resistance marker (Datsenko & Wanner 2000; Sharan et al. 2009). Using this approach a number of r- protein null mutants have now been isolated and represent important tools for the investigation of the phenotypes associated with r-protein removal (Wower et al. 1998; Cukras & Green 2005; Bubunenko et al. 2006; Korepanov et al. 2007).

R-proteins of particular interest A brief summary of the features, proposed functions and published data regarding the three r-proteins that are the major focus of the studies in this thesis follows below.

S20 Ribosomal protein S20 is encoded by the rpsT gene, which, in contrast to most other r-protein genes, exists singly in its own operon. Although it is a designated small subunit protein, it has occasionally been co-purified with the large subunit and mistakenly identified as large subunit protein L26 (Wilson & Nierhaus 2005). It is specific to bacterial ribosomes as no homo- logue exists in either the Eukarya or Archaea. The S20 protein, encompassing 87 amino acids, exhibits mRNA binding activity and modulates its own expression at the level of translation as described earlier. S20 represents one of the primary rRNA binders and interacts directly with the 16S rRNA during assembly of the 30S subunit (Figure 5). The primary binders are considered important in initiating proper folding of the major structural domains and facilitating the addition of later binding proteins. In Nomura’s original assembly map, binding of S13 was dependent on prior attachment of S20 (Mizushima & Nomura 1970). Since then, S13 has been repositioned on the assembly map (Grondek & Culver 2004) and the only protein that appears to require S20 during 30S assembly is S16, which also relies on S4 for attachment (Figure 5). High-resolution structural studies have placed S20 at the base of the body of the 30S subunit. It is one of the few small subunit proteins that interacts with two 16S rRNA domains where it contacts several helices in the 5’ domain and helix 44 (h44) of the 3’ minor domain (Brodersen et al. 2002). Hydroxyl radicals generated from Fe(II)-tethered S20 were used to determine its interactions with the 5’ and 3’ minor rRNA domains during 16S

44 folding and 30S assembly (Dutca & Culver 2008). The results of this study suggest that the organization of both domains, relative to S20, changes depending on the assembly stage such that interactions with the 5’ domain predominate early and strong interactions with h44 occur later in the presence of additional r-proteins. It has been suggested that since S20 is the only protein present at the base of the 30S body and the main protein interacting with h44, it is probably crucial for maintaining and stabilizing the correct conformational arrangements in this region in mature 30S particles (Brodersen et al. 2002). The most well studied S20 mutants were isolated by Dabbs as suppressors of streptomycin and erythromycin dependence (Dabbs 1978; Dabbs 1979). The doubling time of these strains was estimated to be at least 75 minutes versus 22 minutes for the wild-type (Dabbs 1979). Follow-up studies aimed at understanding the ribosomal defect caused by loss of the protein indicated poor association of the mutated 30S with wild-type 50S subunits based on sucrose gradient ultracentrifugation experiments and protection of P-site tRNA from hydrolysis by peptidyl tRNA hydrolase (Gotz et al. 1989). In a second article, partial reactions exclusively using the mutant 30S subunits revealed defects in translation initiation (Gotz et al. 1990). More specifically, no defect in the binding of poly(U) or poly(AUG) was detected, however, the authors propose that the slow growth rate was due to slow initiator tRNA binding and a defective codon-anticodon reaction at the P-site as well as instability of initiation complexes. An S20 mutant, obtained as a spontaneous erythromycin-dependent suppressor, was also isolated in a later study (Wild 1988). Ribosome sedimentation profiles revealed that this strain produced precursor 30S particles and fewer 70S ribosomes. In a separate study, an E. coli rpsT mutant was isolated as a suppressor of nonsense codons (Ryden-Aulin et al. 1993). This slow-growing mutant had a deletion cover- ing the S20 ORF and grew poorly at 42°C. In addition, subunit association was defective (determined by sucrose gradient ultracentrifugation) and alterations in the methylation pattern of 16S rRNA were observed. The authors 5 6 speculate that two particular 16S rRNA modifications (m C and m2 A), which were under-represented in the association-defective 30S subunits, were required for the formation of S20-depleted 70S complexes. More recently, the recombineering approach has been used to isolate in-frame deletions of rpsT but no data to describe the phenotypic effects of these deletions is currently available (Baba et al. 2006; Bubunenko et al. 2007).

L1 Ribosomal protein L1 is encoded by the rplA gene and is one of the largest r- proteins comprising 234 amino acids. It occupies an operon with the structural gene for r-protein L11 and as mentioned earlier, L1 regulates expression of itself and L11 (Baughman & Nomura 1984). The protein is univer-

45 sally conserved and homologues of bacterial L1 exist in the Archaea and Eukarya. It binds directly to the 23S rRNA and is not known to directly influence the binding of any other proteins during 50S assembly (Figure 6). Along with helices 76-78 of the 23S rRNA, the protein constitutes a highly flexible region of the 50S subunit known as the L1 stalk. Consistent with its inherent mobility, this region was disordered in the first high-resolution crystal structure of the H. marismortui 50S subunit (Ban et al. 2000). In the crystal structure of the T. thermophilus 70S with bound tRNAs at the A-, P- and E-sites, the L1 stalk adopts the so-called closed conformation. In this position, L1 makes contact with the elbow of the E-site tRNA and, together with small subunit protein S7, was proposed to act as a gate blocking tRNA release (Yusupov et al. 2001). That same year, the crystal structure of the 50S subunit from D. radiodurans was published in which the L1 stalk was extended further from the main 50S body implying that it could also adopt an open conformation (Harms et al. 2001). Comparison of the crystal structures were the first indications that the stalk might be involved in translocation and alluded to a role in regulating the release of tRNA from the E-site. Subsequent to this a number of experiments have been performed with the aim of delineating stalk dynamics during translocation. A cryo-EM study suggested that the stalk actively participates in pulling the deacylated tRNA from the P- to the E-site during translocation (Valle et al. 2003). In recent years, single-molecule fluorescence resonance energy transfer (smFRET) technology has revolutionized investigations of the dynamics of L1 stalk movement. One study utilised L1-depleted ribosomes and a destabilization of tRNAs in the hybrid state was observed. The authors suggested that rather than direct involvement, the stalk facilitated the structural rearrangements that increased the affinity of the deacylated tRNA for the E-site (Munro et al. 2007). Furthermore, a number of FRET studies have yielded conflicting results regarding the coupling of stalk motions to intersubunit rotation and tRNA movement during translocation (Fei et al. 2008; Cornish et al. 2009; Fei et al. 2009; Munro et al. 2010). Regardless of whether these motions occur in a correlated or stochastic manner, the overall picture that emerges suggests that the stalk is involved in the transit of deacylated tRNA to the E-site followed by its release from the ribosome. The underlying contribution of protein L1 to the movements and participation of the stalk in this process is unclear. Although L1 constitutes a major component of the stalk, its mobility is suggested to be an inherent feature of the flexibility of the rRNA stem by virtue of conserved G-U base pairs in H76 (Korostelev et al. 2008b). Apart from its role in translation, L1 exhibits RNA chaperone activity, a feature shared with many large subunit proteins (Semrad et al. 2004; Ameres et al. 2007). Despite the universal conservation of L1, E. coli strains lacking this protein have been isolated. Two independent L1 mutant strains were isolated as

46 spontaneous revertants of a spectinomycin- and kasugamycin-dependent strain respectively (Dabbs 1977; Dabbs 1980). The antibiotic dependent strains were initially obtained following ethylmethane sulfonate mutagenesis. Absence of L1 was only verified on the protein level by means of 2D-gel electrophoresis and immunological tests (Dabbs et al. 1981). The first publication detailing functional activity of the L1 mutant ribosomes suggested that they were approximately 50% less active in the in vitro production of polypeptide in both poly(U) and phage MS2 RNA directed systems (Subramanian & Dabbs 1980). The growth rate of the mutants was also approximately 50% slower than wild-type and polypeptide synthetic activity was restored to wild-type levels upon reconstitution with purified L1. In a later study, using a poly(U) directed system, it was found that both the binding of N-acetyl-Phe-tRNA to the P-site and stimulation of EF-G dependent GTP hydrolysis in the presence of tRNA and mRNA were significantly compromised with these same ribosomes (Sander 1983). A defect at the translocation step was not detected however. Another L1 deletion strain was isolated as a spontaneous erythromycin-independent revertant from an erythromycin-dependent strain that was created by mutagenesis with NG. This L1 mutant strain showed an excess of free ribosomal subunits and fewer 70S ribosomes when the sedimentation profile of ribosomal particles was examined (Wild 1988). As opposed to the aforementioned L1 lacking strains obtained using antibiotic selection systems, a genetically defined rplA deletion strain has been constructed in E. coli (Baba et al. 2006), but at the time of writing, details of the phenotype of this mutant are unknown.

L17 Large subunit protein L17 is encoded by the rplQ gene and forms part of the alpha operon that also encodes r-proteins S13, S11, S4 and the alpha subunit of RNA polymerase. R-protein S4 acts as an autogenous regulator of this operon. L17 is 127 amino acids in length and is highly conserved among bacteria but is absent in the Eukarya and Archaea. In mature 50S subunits of T. thermophilus, the protein is located on the solvent side on the main body of the subunit (Selmer et al. 2006). The rplQ gene is categorized as an essential gene in E. coli K12 since gene disruption by means of recombineering failed to produce a knockout of this gene (Baba et al. 2006). No precise functional role has been assigned to L17, either within or outside of its ribosomal context.

Ribosomal complexity – an evolutionary riddle To assert that biology is accustomed to complexity is a straightforward notion. We are familiar with the complex arrangements of atoms that make up biomolecules and the complex assemblies of cells and tissues that make up

47 organs and entire organisms. Even in the humble bacterial cell the ribosome represents a biomolecule of vast complexity, however very little is known about how this complexity emerged. We can be certain that evolution has shaped ribosomal complexity but a more interesting question addresses how this might have occurred. In a sense the situation is unfortunate, the more we learn about the ribosome, the more perplexing its evolution appears as the complexity and co-operativity of both its structure and function is magnified even further. Translation, as we know it today, is carried out by a ribonucleoprotein. Whatever its origins, we do know that proteins are an integral component of the current coded system. The most obvious question that arises is how were the first r-proteins synthesized if ribosome function depends on proteins? The realization that RNA is responsible for the major functional tasks of the ribosome has paved a way forward out of this apparent paradox (Moore & Steitz 2010). Furthermore simple RNA molecules selected in vitro have de- monstratable peptidyl-transferase activity, the core enzymatic activity of the ribosome (Zhang & Cech 1997). Thus the idea that the modern ribosome evolved from an all RNA particle sits comfortably with most people. To imagine how the spectacular process of contemporary protein synthesis could have evolved from a primitive RNA world is a much more difficult concept to grasp. The ribosome is often viewed as a stepping stone from a primordial nucleic acid based world to the current protein based world of modern cells (Woese 2001). One would predict that the first peptides were quite simple in nature, possibly of fixed or random sequence. It has been suggested that these primitive peptides might have acted as non-specific chaperones and stabilized RNA structures (Poole et al. 1998). Noller extends this hypothesis further and proposes that translation evolved to expand the repertoire of RNA structural motifs and thus augment its functional capabilities (Noller 2010). It is an attractive hypothesis considering that no positive selection for a protein based world would have existed until the translation apparatus developed the capacity to synthesize active proteins with discrete selectable functions of their own. It is supported by a variety of observations where small peptides and ligands are known to induce the formation of unusual RNA tertiary rearrangements such as the Tat protein of HIV virus that requires only a short stretch of arginine residues to cause complete rearrangement of the TAR RNA that is essential for virulence (Puglisi et al. 1992). In addition the r-proteins themselves are ascribed a major role in folding the rRNA into biologically active conformations (Cech 2000). The complexity of ribosome function is addressed in this thesis in terms of the functional contributions made by specific ribosomal proteins. As expected both removal and replacement of native proteins conferred fitness costs although the costs of removal were much greater than those of replacement suggesting that in the complete absence of a protein stabilizer the

48 rRNA adopts sub-optimal conformations. To probe deeper into the functional roles of the proteins, and to gain some insight into the complex cooperative nature of ribosome function, r-protein mutants were evolved to determine the mechanisms by which ribosome function and fitness were improved.

Concept of fitness The topic of evolution is tightly linked to the concept of fitness. There are a number of alternative definitions of fitness but in essence they all refer to the ability of an organism or genotype to survive and reproduce in a given environment (Orr 2009). Naturally, the environment of an organism can vary over time and this will be hugely influential in determining which genotypes win and which genotypes lose. In any particular environment, the fittest genotype will be more successful at survival and propagation and thereby contribute more genes to subsequent generations. Likewise, genotypes of low fitness contribute less genetic material to successive generations and will eventually become lost from a population. Throughout this body of work the term relative fitness is encountered frequently. This is a common term in population genetics and is often expressed as the fitness of a particular genotype or strain normalized to wild-type fitness (Orr 2009). Relative fitness, as opposed to absolute fitness, is generally a more informative statistic as it describes differences in fitness upon which selection acts to determine the evolutionary success of different genotypes. Bacterial fitness can be measured in the laboratory using a number of techniques. One of the most straightforward methods utilizes generation times measured during exponential growth. One weakness of this measurement is its failure to evaluate growth performance over the entire growth cycle. This can be overcome by means of pair-wise competition experiments in which the test strain and reference strain are differentially labeled with genetic markers and mixed in an equal ratio before the commencement of growth. During the competition experiment representative samples are with- drawn and the relative proportions of each strain are determined. This method is not only more sensitive than exponential growth rates but it also gives a composite fitness measurement as it accounts for all stages of the growth cycle.

Adaptive and non-adaptive forces drive evolution In crude terms we can define evolution as changes in the characteristics of an organism over time (Futuyma 1998). Thus it represents the most vast and comprehensive biological phenomenon in the sense that all organisms and

49 their characteristics are the products of it. At the molecular level, these changes are the result of alterations in allele frequencies which is governed by two fundamental mechanisms, one of which was articulated by Charles Darwin over 150 years ago. Darwin’s theory of natural selection challenged the prevailing worldview at that time and revolutionized our understanding of evolution. The process of natural selection operates by the non-random selection of randomly generated variants in a population. This means that genotypes harboring advantageous mutations increase in frequency and ultimately reach fixation in a population whereas those that carry deleterious mutations become less frequent over time and are eventually eliminated. Thus the process is adaptive and dictated by fitness. Without the existence of genetic variation in a population, and thus individuals of different fitness, natural selection cannot act. The supply of genetic variation is thus provided by mutation and recombination. The second major mechanism that deter- mines the path of evolution is known as genetic drift and opposes natural selection in terms of its stochastic nature and its indifference to fitness. As such it does not necessarily foster adaptation but rather leads to the random loss of genotypes from a population. The fate of an individual in a population relies on the interplay of these two forces.

Compensatory evolution and adaptation Any given bacterial population consists of genotypes of varying fitness and the generation of this variation is primarily governed by the random appear- ance of mutations. A newly arising mutation has the potential to alter fitness, however since the majority of mutations in all organisms are deleterious, the effect is most often detrimental rather than beneficial. Although such mutations confer fitness costs, they can persist in a population especially if the deleterious effect(s) can be relieved or remedied by second-site suppressor mutations, also known as compensatory mutations (Maisnier-Patin & Andersson 2004). These are defined as pairs of mutations that are independently deleterious but neutralize each other’s costs when combined (Kimura 1990). They are often more common than true reversion in reducing the deleterious effect of a mutation due, at least in part to, the larger target size for compensation compared to reversion. Compensatory mutations are frequently discussed within the framework of antibiotic resistance, regarding the threat they pose in allowing resistant organisms to persist even in the absence of drug selective pressure (Maisnier-Patin et al. 2002; Nilsson et al. 2006). However they are also important in the more general context of adaptation and evolution and may represent a powerful strategy for the evolution of complexity in cellular systems (Zuckerkandl 1997; Zuckerkandl 2001). By facilitating access to otherwise prohibited evolutionary pathways, compensatory mutations can gen-

50 erate functional novelty and augment complexity. This theory has been used to describe the evolution of complex regulatory networks, for example to explain the evolution of multi-factorial gene transcription. It is well known that expression of many genes is controlled by the interactions of transcription factors with regulatory elements located upstream of the coding region. The precise timing and balance of expression is often determined by the interaction of multiple factors. It has been suggested that this complex scenario is the result of repeated cycles of deleterious mutation followed by selection of compensatory mutations leading to an accumulation of required factors to restore functional activity (Zuckerkandl 1997). Thus the generation of complexity occurs as an indirect response or a by-product of an event that was initially non-adaptive. This theory, which has been discussed in the context of slightly deleterious mutations, asserts the potential power of non-adaptive processes in the generation of complexity. In this way, weakly deleterious mutations can be viewed as important facilitators for the emergence of novelty and complexity in biological systems.

Ribosomal complexity – a consequence of compensation? The hypothesis formulated by Zuckerkandl could also be used to rationalize the requirement for ribosomal proteins in modern translation. We can propose that the contemporary ribosome descended from a smaller, simpler RNA structure that evolved slowly over time. During its evolution this particle was subject to mutation, mostly deleterious in nature leading to destabilization of the structure and loss of translational efficiency. The r-proteins may have been recruited as a compensatory response to rescue the structure while restoring or even improving ribosome function. Successive deleterious mutations and resultant compensation would lead to an RNA-protein particle of higher complexity. The repetitive cycle of damage followed by repair acts as an intrinsic drive to push the system forward resulting in an end product that is as fit or even fitter than the initial state though structurally more complex. Were the proteins selected as by products of a deleterious process or were they adopted simply because of the selective advantage they endowed? The lack of universal conservation of all r-proteins and the heterogeneity of their structures gives the impression that they were added slowly over time. On the one hand natural selection may have acted alone so that upon the arrival of protein molecules, they were incorporated due to the beneficial impact they endowed. Indeed, even small refinements and improvements impose strong selective advantages. Alternatively, they may have been recruited via an initially non-adaptive process to suppress or rescue poor functioning ribosome precursors. Upon addition to the particle, the proteins themselves would subsequently co-evolve with the ribosome and thereby have the potential to co-adapt and attain discrete functional roles in translation. In an

51 effort to learn more about the functional requirements of r-proteins, mutants lacking an r-protein or harboring a non-native version were evolved in the laboratory in papers II and III to determine the mechanisms of compensatory adaptation.

Horizontal Gene Transfer (HGT) HGT is a collective term that refers to those processes involved in the lateral transfer of genetic information between organisms as opposed to inheritance by vertical descent from a parent cell. In eubacteria, it can occur between closely related as well as phylogenetically distant species via three major mechanisms: transformation, transduction and conjugation. Transformation refers to the uptake of naked DNA from the extracellular environment. In most bacterial species, this mechanism requires the recipient to adjust to a competent physiological state by expressing the genes that encode specialized uptake machinery and is usually a response to a change in environmental conditions (Chen et al. 2005; Thomas & Nielsen 2005). DNA transfer mediated by bacteriophages is known as transduction. Imprecise excision of integrated prophages from the chromosome of donor cells results in the transfer of flanking donor DNA (specialized transduction) or the phage can inadvertently package and transfer exclusively chromosomal DNA from the donor (generalized transduction) (Ochman et al. 2000). The abundance and stability of bacteriophages in the environment makes these agents highly efficient at gene transfer although recipient range is limited to those cells that express compatible phage receptors (Canchaya et al. 2003). In contrast to transformation and transduction, conjugation occurs via physical contact between the donor and recipient cells and requires formation of a conducting channel for DNA transfer. This mechanism is often associated with conjuga- tive plasmids and transposons but theoretically any genetic material can be transferred this way as long as it becomes part of the unit that contains the transfer origin (Thomas & Nielsen 2005). Upon uptake and successful entry, the transferred genetic material must be established in the host chromosome or be capable of autonomous replication to become a functional element of the recipient’s genome.

Constraints on HGT The barriers that curb HGT include both mechanistic and selective constraints. In terms of the mechanistic barriers, restriction endonucleases of the host can target foreign DNA for degradation, conjugal transfer is limited by the ability of the transfer machinery to operate between species and phage- mediated transfer is restricted to hosts that express receptors recognizable by the transducing phage (Chen et al. 2005). Even in the absence of such obsta-

52 cles, many HGT events are likely to cause detrimental effects upon integration due to the high coding density of the chromosome and are therefore selectively deleterious (Thomas & Nielsen 2005). Disturbances in genome organization, DNA supercoiling and gene order are expected to confer substantial costs. Even if the insertion is neutral, expression of the foreign material also confers costs that operate on both the transcriptional and translational levels. In addition, the products of both of these events (mRNA and protein) may in turn be selectively disadvantageous (Stoebel et al. 2008), for example if the foreign protein interferes with the activity of an orthologous counterpart in the host. Base compositional divergence represents another barrier as documented by the transcriptional silencing of AT-rich DNA by the histone-like nucleoid structuring protein (H-NS) in enteric bacteria such as S. typhimurium (Navarre et al. 2006). These barriers must be surmounted for successful integration in the host genome and to allow expression of the foreign genetic material. To ensure long-term maintenance in the population, the transferred gene must also be subject to sufficient positive selection to prevent elimination by genetic drift. Aside from the aforementioned constraints, HGT events are biased towards genes that encode proteins that function independently, belong to simple functional networks or are transferred as complete operons (Rivera et al. 1998). This has been explained in terms of the complexity hypothesis and asserts that operational genes (housekeeping genes whose products require few or no physical interactions) are preferentially transferred as they do not require integration in large cellular complexes for function (Jain et al. 1999). According to this hypothesis informational core genes, such as those involved in replication, transcription or translation, are rarely successfully transferred due to the high degree of physical interactions and mechanistic co-operations required for function. In other words, the fixation probability of informational genes is expected to be significantly lower than that of operational genes that have fewer interacting partners. The relevance of this hypothesis was experimentally tested in paper III, wherein r-protein genes from related and phylogenetically distant species were used to replace their native orthologous counterparts and the costs of such HGT events were measured to determine whether this represents a significant constraint for transfer.

53 Present Investigations

The ribosome is undoubtedly one of the most complex and sophisticated cellular machines ever made. The abundance of RNA in the overall structure and its predominance at the major functional sites, almost certainly reflects a predecessor composed largely or perhaps entirely of RNA. The proteins, despite their exclusion from the spotlight, represent integral components of contemporary ribosomes. They also represent a level of complexity that is not yet fully understood. The flexibility and plasticity of the translation apparatus in terms of its protein components is the major focus of this thesis. Our primary goal was to address how r-protein removal or replacement alters bacterial physiology. We were specifically interested in the potential fitness costs associated with r-protein removal and replacement in terms of the effects on bacterial growth rate and ribosome function. The experimental approach allowed us to ascribe discrete functional roles to particular r-proteins by comparing the phenotype of r-protein depleted ribosomes with those from a congenic wild-type strain. In papers I and II the genes encoding ribosomal proteins S20 and L1 were removed from the chromosome of Salmonella typhimurium by precise genetic replacement and the underlying functional roles of each protein were examined using a combination of in vivo and in vitro techniques. As a further step to deepen our understanding of the significance of the deleted protein and to reveal potentially novel functional interactions, compensatory evolution was performed with the strain lacking L1 in paper II. The identification of compensatory mutations, that mitigate the fitness costs of r- protein removal, provide unique insights into the highly cooperative nature of the translation apparatus. In a related vein, the secondary aspect of this study addresses similar questions but rather than complete removal, the r- proteins of interest (S20, L1 and L17) were replaced by closely related as well as phylogenetically distant orthologues in paper III. Such hybrid mutants address questions regarding translational robustness and the plasticity of the bone fide components. An additional feature of the experimental set- up, facilitated by the use of such hybrid strains, allowed us to examine the selective barriers that constrain inter-species gene replacements. The following section provides a concise summary of the experimental design as well as the major findings and conclusions of the studies included in this thesis.

54 The model organism Appropriate model systems rely on the ability to generalize findings to other organisms and taxa. The ultimate aim of this study was to further our understanding of the contributions made by particular r-proteins to translation and to shed some light on the co-operative nature of ribosome function. The ribosome is universally conserved, and although eukaryotic and archaeal ribosomes possess more RNA and more proteins, the tertiary structure and basic function is conserved among the three phylogenetic domains. By virtue of their short generation times, large population sizes and the availability of a wide range of tools for genetic manipulation, bacterial organisms are ideal for the study of both genetics and experimental evolution. Escherichia coli (E. coli) has long served as a model organism in the ribosome field. However, due to its close genetic relationship to E. coli, Sal- monella enterica serovar Typhimurium (referred to as S. typhimurium) represents an appropriate alternative for studies on ribosome function and is becoming increasingly popular as a model organism in bacterial genetics. Salmonellae are Gram-negative rod-shaped bacteria that belong to the class of -proteobacteria and are a leading cause of gastroenteritis in humans (Ohl & Miller 2001). S. typhimurium strain LT2 was isolated in the 1940s and is the main Salmonella strain used in molecular biology (McClelland et al. 2001). All strains used in this study are derivatives of LT2, which is less virulent than other strains of Salmonella due to decreased expression of the stationary-phase sigma factor s (Swords et al. 1997). Today, with its fully sequenced 4.95 Mb genome (McClelland et al. 2001) and extensively studied metabolic and regulatory networks, this bacterium represents a conven- ient model for the purposes of this study.

Construction and characterization of r-protein mutants Recombineering The use of recombineering for targeted gene disruption in vivo is now a well- established technique. Due to its efficiency, precision and simplicity, the method employed here to construct r-protein knockout and replacement mutants was based on -Red mediated homologous recombination. Linear double-stranded DNA fragments (usually antibiotic resistance cassettes generated by PCR) are used as homologous substrates for targeted replacement, while a recombination-proficient environment is provided by the Red functions encoded by phage (Sharan et al. 2009). Usually only very short ho- mologies of 50bps are required for successful recombination (Sawitzke et al. 2007).

55 Normally linear DNA introduced into bacteria is rapidly degraded by the action of intracellular exonucleases. However induced expression of the Red proteins preserves the incoming substrate DNA and also facilitates recombination at the target site. The Gam protein inhibits the host cell RecBCD and SbcCD nuclease activities to prevent substrate degradation, the Exo protein is a dsDNA-dependent exonuclease that generates 3’ ssDNA overhangs to which the Beta protein binds and promotes annealing of complementary DNA molecules (Sharan et al. 2009). The timing and level of expression of these genes is critical to maintain efficiency of the system. The Red functions can be supplied on a defective prophage in which the repressor is utilized for controlled expression or they may be provided on an inducible plasmid (Sawitzke et al. 2007). In our case the genes are supplied on a temperature sensitive plasmid (pKD46), under the control of an arabinose- inducible promoter (Datsenko & Wanner 2000). This allows easy curing of the plasmid to prevent the occurrence of undesirable events following recombination. For the generation of r-protein gene deletions a kanamycin resistance cassette was PCR amplified using primers with 50 nt 5' tails homologous to the flanking sequence of the target gene (rplA and rpsT). Electroporation of the linear DNA product into the cell resulted in substitution of the r-protein ORF with the selectable kanamycin resistance cassette (Figure 8A). A similar procedure was used for the generation of r-protein replacement mutants however in this case the PCR fragment consisted of the ORF of the orthologous replacement gene and an adjacent kanamycin resistance cassette. Ho- mologous recombination resulted in replacement of the target r-protein ORF (rpsT, rplA and rplQ) with its orthologous ORF and the kanamycin resistance cassette (Figure 8B). Congenic wild-type control strains were also constructed and carried the kanamycin cassette insertion adjacent to the native r-protein ORFs. In the case of both types of constructions the native promoter and terminator sequences were preserved. The kanamycin cassette is flanked by FRT (FLP recognition target) sites to eliminate the marker following construction if desired (Datsenko & Wanner 2000). However in our case the kanamycin cassette was left intact for selective purposes.

50 nt 50 nt 50 nt 50 nt KanR orthologous ORF KanR homologous homologous recombination bacterial recombination r-protein ORF r-protein ORF chromosome

KanR orthologous ORF KanR (A) Gene Removal (B) Gene Replacement

Figure 8: Schematic representation of the method used for construction of S. typhimurium r-protein mutants.

56 In this study, S. typhimurium strains lacking r-proteins S20 ( S20) (paper I) and L1 ( L1) (paper II) and hybrid strains with orthologous replacements of S20, L1 and L17 (paper III) were constructed using the -Red recombination technique. This method of r-protein removal and replacement is advantageous compared to the mutagenesis and antibiotic selection systems used to obtain r-protein mutants in previous studies (Dabbs 1991). Most importantly, the system allows precise genetic replacement in a genetically defined background as opposed to the genome-wide mutagenesis strategy wherein undesired mutations elsewhere in the genome and/or ribosome are an unavoidable concern and constitute a major caveat for phenotypic studies.

In vitro system One major challenge encountered in the characterization of ribosome function in vitro is the extent to which the parameters measured correspond to the in vivo situation. The system designed by Ehrenberg and used for the characterization of ribosomes in papers I and II is optimized to meet in vivo ex- pectations in so far as it allows measurement of in vitro reactions in the mil- lisecond range (Ehrenberg 1989). Ribosomes and all of the required translation components are prepared to high purity according to standard protocols to ensure robustness and optimization of the system. Additionally, all experiments are performed in a buffer (polymix) that mimics the pH and complex ionic conditions of the bacterial cell. All steps of translation can be measured using a variety of biochemical techniques that allow dissection of each event in minute detail. Since each step of translation progresses forward via multiple smaller steps, kinetic measurements detailing the formation and consumption of intermediate states as well as transitions between intermediates allows precise and specific comparison of mutant ribosomes with their wild-type counterparts. Although the system was developed for E. coli, the results of a previous study established that ribosomes from S. typhimurium function well in this system (Tubulekas et al. 1991).

In vivo system Functional characterization of ribosomes lacking r-proteins In papers I and II polypeptide elongation rates of exponentially growing cultures were measured by determining the exact time taken to produce the first detectable -galactosidase molecule following IPTG induction. The amount of o-nitrophenol produced, normalized to the amount of cells in the culture and the incubation time was used to calculate elongation rate according to Miller's protocol (Miller 1992). As Salmonella does not harbor the lac

57 operon, the normal inducible lac operon was supplied on an F’ factor that was transferred to strains by means of conjugation. The synthesis of -galactosidase was also used to determine the frequency of stop codon read-through and frameshifting of an L1-depleted ribosome in paper II. In this instance strains carried an F’ factor in which the lacI and lacZ genes are fused for constitutive expression of -galactosidase. How- ever, the ribosome must read-through a premature stop codon (UGA189 or UAG220) or a +1 or -1 frameshift in lacI to produce an active molecule of the enzyme, the activity of which is detected as described above. A read-through value is calculated based on the amount of enzyme produced by the mutated lacIZ construct relative to that produced in a congenic strain harbouring the non-mutated lacIZ construct (Bjorkman et al. 1999). It has previously been shown that stop codon read-through correlates well with ribosomal misreading in vivo and in vitro (Kurland 1996). A second assay based on luciferase expression was also employed in paper II to determine if L1-depleted ribosomes were prone to general misreading (Kramer & Farabaugh 2007).

Compensatory evolution and whole-genome sequencing Compensation, in genetic terms, refers to the ability of evolutionary processes to create genetic variants of improved fitness from a pool of less fit variants that carry a deleterious mutation. Experimental evolution was employed here to ascertain if and by what mechanisms fitness compensation could be achieved in mutants either lacking r-protein L1 (paper II) or harboring a non-native version of r-proteins S20 and L1 (paper III). The type of compensated mutants selected during evolution depends on a number of factors, including the mutation rate specific to each type of compensatory mutation, the relative fitness of the compensated mutants and the population bottlenecks associated with evolution. Thus, in large populations that are not subject to bottlenecks, the fittest variant is expected to eventually reach fixation. In contrast, if the population is subject to small bottlenecks, then less fit compensated variants are afforded the opportunity to fix if they appear at a faster rate than the rarer high-fitness mutants (Levin et al. 2000; Maisnier-Patin et al. 2002). The experiment is designed to optimize purify- ing selection, which reduces genetic diversity so that the population stabilizes on the most favourable (fittest) genotype. During the evolutionary process compensatory mutations that reduce the fitness cost conferred by loss or replacement of the r-protein become more common in successive generations. To maintain continuous population growth and to ensure that all genetic variants were represented during the entire evolutionary process, large bottlenecks of 106 cells were used. This experimental design maxi- mizes the potential to fix the compensated mutant of highest fitness. The advent of whole-genome sequencing has greatly facilitated the rapid detection and identification of the mutations that arise during the experimental time-frame by comparison to the ancestral strain. Whole-genome sequencing

58 was employed in paper II to determine the compensatory mutations that suppressed the slow-growth phenotype of a mutant lacking L1.

Fitness costs of r-protein removal and replacement In papers I, II and III fitness measurements were performed to determine the costs conferred by r-protein removal and orthologous replacement. In our case we have defined fitness as growth performance and this was evaluated at different temperatures (papers I and II) and using different carbon sources (paper III). Early exponential growth rates are an informative fitness parameter and were the primary method used in our studies. Relative fitness values were calculated as the growth rate of each mutant as a fraction of the growth rate of the wild-type reference strain used in each experiment (set to 1.0). By only examining early exponential growth rate, information regarding growth performance over the entire growth cycle is lost and the method is limited to the detection of fitness differences of about 3% or higher. Therefore, head-to-head competitions between the wild-type and hybrid strains were performed in paper III to permit detection of small fitness costs that might otherwise go unnoticed. Such costs are expected to be highly influential in determining the fixation probability of a HGT event and are therefore crucial to the study of the selective constraints that might limit r-protein transfer in nature.

Removal of r-proteins confers substantial fitness costs The large fitness costs associated with loss of the r-proteins S20 and L1 were immediately evident upon growth on solid media and the S20 strain grew most poorly at high temperature. For more precise determination of the fitness costs, early exponential growth rates were measured for both knockout strains. The experiments revealed that loss of S20 reduced exponential growth rate approximately three-fold relative to the wild-type under standard growth conditions (LB broth at 37°C). This translated to a relative fitness value of 0.33. In addition a reproducible extension of the lag phase was apparent (paper I). In the case of the L1 mutant the generation time was approximately 2-fold slower than wild-type (relative fitness value of 0.52) under standard growth conditions and furthermore the fitness cost was more pronounced during growth at a lower temperature indicative of a cold- sensitivity phenotype (paper II). Restoration of wild-type growth rate occurred upon expression of the deleted gene from an inducible plasmid, specifically identifying removal of each respective protein as the cause of reduced fitness.

59 Fitness costs of replacement are generally weaker than removal Given the high fitness costs associated with complete removal of S20 and L1, and the essential nature of L17, it was expected that replacement proteins would improve fitness even if the alien protein lacked a high degree of homology and functioned sub-optimally in its non-native context. In paper III, exponential growth rates were measured in four different media where the generation time of the wild-type strain varied from 20-120 minutes and was set to a value of 1.0 in each respective medium. Relative to this, all of the hybrid S20 strains displayed higher fitness than the null mutant, irrespec- tive of the growth medium and even when the amino acid identity of the replacement protein dropped to as low as 32%. Similarly, the relative fitness of the L1 hybrid strains was only modestly reduced relative to wild-type in most cases, however, a large fitness reduction was observed when amino acid identity reached as low as 20-30%. Since L17 is an essential r-protein, comparisons to a null mutant were not feasible, however, a substantial fitness reduction relative to the wild-type was only observed when amino acid identity of the orthologue fell below 50%. Higher fitness of certain L17 hybrid mutants relative to wild-type was also evident. A general trend that emerged was a reduction in relative fitness as phylogenetic distance increased, however the costs of replacement for the more homologous proteins were diminutive in comparison to the costs of removal.

Small fitness costs restrict HGT of r-proteins For those hybrid strains that displayed very small fitness costs, pair-wise competition experiments with the wild-type strain were also performed to obtain more sensitive fitness measurements. The competing strains were genetically tagged with fluorescent markers and were mixed in a 1:1 ratio at the commencement of growth. Continuous growth was maintained by serial transfer for a period of up to 12 cycles ( 60 generations) and the ratio of hybrid to wild-type cells was measured after each cycle using a fluorescence-activated cell sorter (FACS). The ability to track large numbers of single cells (105) allowed us to calculate an accurate selection coefficient for each orthologous protein and observe very small differences in fitness. The results revealed that the costs conferred by the non-native proteins were often less than 10% and in many cases were in fact less than 1%. These small fitness costs are contradictory to the idea that the transfer of components that function within tightly coordinated complexes and have many interacting partners should confer large fitness costs as posited by the complexity hypothesis (Jain et al. 1999). However, modeling showed that the fixation probability of such weakly counter-selected orthologues in large bacterial populations is vanishingly small so the calculated fitness costs do constitute an effective constraint in terms of the success of HGT.

60 Phenotypes of S20 and L1 mutants The ability to obtain genetically defined mutants lacking r-proteins S20 and L1 allowed us to investigate the role of these proteins in translation by comparing their in vivo and in vitro phenotypes with a congenic wild-type strain.

Removal of S20 perturbs mRNA binding and subunit association In paper I we established that the rate of polypeptide chain elongation in vivo was unperturbed in the S20 mutant, however, a reproducible reduction in the rate of accumulation of synthesized protein was evident. Considering that S20 is a designated 30S subunit protein and given the results of previous studies implicating the involvement of S20 in initiation (Gotz et al. 1990; Ryden-Aulin et al. 1993), we hypothesized that the reduced rate of protein accumulation was likely the result of impaired initiation. Thus, the activity of S20-depleted 30S subunits were compared to those obtained from the wild-type using a number of in vitro methods designed to measure each distinct step of translation initiation. The rate of template binding was measured using a radioactively labeled mRNA with a strong SD sequence. In both the presence and absence of initiator tRNA and the initiation factors, the rate of mRNA binding to S20 30S subunits was reduced approximately 3.5-fold and maximum binding required 40 minutes compared to only 10 minutes for the wild-type. Since similar rates of mRNA binding were observed with and without the initiation ligands, a primary impairment in mRNA binding could be concluded. Re- constitution with purified S20 restored the wild-type mRNA binding phenotype and established that defective mRNA association was a direct consequence of S20 removal. In addition, titration of mRNA revealed that a substantial fraction ( 25%) of the S20-depleted 30S subunits were completely inactive in the binding of mRNA. Since the rate of fMet-tRNAfMet binding emulated that of mRNA and was therefore used to detect the level of mRNA bound to 30S subunits in this assay, the possibility of unstable initiator tRNA binding could not be entirely excluded as the underlying cause of incomplete S20 30S occu- pancy with mRNA. Thus the stability of initiator tRNA binding was measured by its rate of dissociation from 30S pre-initiation complexes. The results demonstrated that dissociation of fMet-tRNAfMet occurred at the same rate for the mutant and wild-type. Hence, we could conclude that reduced initiator tRNA binding was a secondary, consequential effect of a primary impairment in mRNA binding. The kinetics of 50S association to wild-type and S20-depleted 30S subunits was also measured in a stopped-flow instrument using light scattering. This assay demonstrated that naked S20 30S subunits were severely impaired in association with wild-type 50S subunits. However, a pronounced

61 improvement (from 20% to 40%) in the association capacity of the mutant 30S was observed upon preincubation with mRNA, initiator tRNA and the initiation factors. Similar to the results of mRNA binding, subunit association was further improved upon extension of the S20 30S preincubation time with initiation components to 40 minutes. Defects in the rate and extent of mRNA binding, as well as the poor association capacity of S20-depleted 30S subunits are likely to be accountable for the reduced rate of protein accumulation observed in vivo and the pro- longed generation time of the S20 mutant.

S20 is required for correct structural positioning of h44 The involvement of S20 in promoting mRNA binding and subunit association was puzzling considering its topographical location at the base of the body of the 30S subunit far removed from the mRNA binding channel and distal to the intersubunit face. However S20 has been shown to interact with helix 44 of the 16S rRNA during 30S assembly and in mature 30S particles (Brodersen et al. 2002; Dutca & Culver 2008). High-resolution structural studies have shown that this helix extends from the base of the 30S body to the mRNA binding site where it forms the A- and P-sites (Schluenzen et al. 2000) and it appears to be involved in fixing the position of mRNA on the 30S subunit (Selmer et al. 2006). It also makes extensive contact with the 50S subunit via a number of bridges (Schuwirth et al. 2005; Selmer et al. 2006). Furthermore, the structural studies suggest that S20 may be responsible for anchoring h44 in the correct orientation (Schluenzen et al. 2000). We therefore propose that in the absence of S20, h44 adopts a structural orientation that hinders both mRNA binding and subunit association (see Figure 9). Upon extended incubation with mRNA, initiator tRNA and the initiation factors these defects were concealed to some extent, suggesting that these 30S ligands promote a structural rearrangement of h44 that permits stable mRNA binding and facilitates docking of the 50S subunit.

62 mRNA

B2a

S20 h44 Figure 9: Structure of the 30S subunit highlighting the proposed involvement of S20 in mRNA binding and subunit association. S20 (green ribbon) interacts with and possibly stabilizes h44 (red) in an orientation that is optimized for mRNA (orange; A-site, purple; P-site, cyan) binding and interactions with the 50S subunit (bridges B2a, B3, B5 and B6 depicted in yellow). All other rRNA and protein residues are depicted in grey. The image was created in PyMol (www.pymol.org) from the coordinates of the T. thermophilus 30S subunit (PDB ID 2J00) (Selmer et al. 2006).

Removal of L1 impairs ribosome biogenesis In paper II we examined the effects of rplA deletion with a particular focus on ribosome function in vivo. Although L1 removal had no effect on the rate of polypeptide chain elongation in vivo or on the rate of peptidyl transfer in vitro, a reduction in the rate of protein accumulation in vivo was evident, reminiscent of that observed for the S20 mutant in paper I. Furthermore, loss of L1 increased read-through of the UGA stop codon approximately three-fold indicating that L1 may have a minor role in antagonizing a ribosomal ambiguity (ram) phenotype. However the mutant was not prone to frameshifting and a general misreading phenotype was not detected. The distribution of ribosomal particles in the cell was determined using polysome profiles of total cell extracts. The profiles showed that in the absence of L1 the fraction of free subunits increased with a resultant decrease in the proportion of 70S particles from 0.85 in the wild-type to 0.65 in the L1 mutant. However, the ratio of 70S to polysomes was the same as that observed in wild-type cells implying that those L1-depleted 70S that are

63 formed can translate as efficiently as wild-type ribosomes, consistent with the peptidyl-transfer and elongation rate data. Interestingly, the increase in free subunits was strongly biased towards 50S subunits, indicating that L1 is required to promote balanced formation of the ribosomal subunits. Com- pared to the wild-type ratio of 1:1 for free 30S and 50S subunits, in the L1 mutant this ratio was 1:2. Although L1 has no known role in 30S biogenesis, the results suggest that in its absence the formation of 30S subunits is decreased or they are more prone to degradation. Upon expression of the rplA gene from an inducible plasmid, the distribution of ribosomal particles in the L1 mutant was restored to wild-type particle partitioning verifying that removal of L1 was specifically responsible for the altered polysome profile. Considering that L1 is a component of mature 50S subunits, it is more probable that its absence directly affects proper assembly of 50S subunits. It is thus conceivable that in an L1-depleted environment, biogenesis of the 50S is perturbed leading to the production of misassembled 50S subunits or trapped intermediates. As a consequence, assembly factors that are required for the maturation of both subunits may become hijacked leading to the down-regulation of 30S formation. Alternatively, if the 30S subunits lack properly assembled association partners, final maturation of the 30S may be impaired and/or the 30S subunits may be more susceptible to the action of degradative RNases as previously suggested (Deutscher 2009). Either scenario may explain the observed high 50S-to-30S ratio that was observed in the L1 mutant. In conclusion, removal of large subunit protein L1 causes aberrant ribosome biogenesis and this is supported by the observed cold sensitivity phenotype of the L1 mutant. Defective ribosome biogenesis culminated as an increase in free ribosomal subunits and a resultant decrease in the cellular pool of 70S species as monosomes and polysomes. We expect that this di- minishes the overall translational capacity of the cell and is ultimately responsible for the reduced yield of protein production in vivo and the slow- growth phenotype of the L1 mutant.

Mutations in 50S r-proteins mitigate fitness costs of L1 removal Twelve independent lineages of the slow-growing L1 mutant were evolved for approximately 300 generations to determine if and by what mechanisms compensation could be achieved. Upon isolation of faster-growing mutants, four were chosen at random and subject to whole-genome sequencing to identify candidate compensatory mutations responsible for suppressing the fitness cost associated with loss of L1. Single nucleotide substitutions in the genes rplB and rplN were detected as well as a small deletion in the rplS gene. These mutations resulted in the amino acid substitutions E194K in large subunit protein L2 (encoded by rplB) and P102L in protein L14 (encoded by rplN). The deletion in rplS resulted in loss of the Leu codon at

64 position 100 of L19 (referred to as rplS L100) (Figure 10). Strain recon- structions confirmed that each individual mutation was necessary and sufficient for fitness improvement and all three increased growth rate to a similar extent. In any combination, all three compensatory mutations demonstrated negative epistasis in the L1 background. In the wild-type strain each individual mutation reduced fitness and this became even more pronounced when the mutations were combined. We investigated the effect of each individual mutation (rplB E194K, rplN P102L and rplS L100) on polypeptide elongation rate and UGA suppression in the L1 and wild-type background. In each case, the presence of each individual mutation had no impact on the rate of elongation. In terms of UGA read-through, none of the compensatory mutations altered the mild ram phenotype of the L1 strain but actually increased UGA suppression when separated from the L1 deletion in a wild-type background. However, the E194K amino acid substitution in L2 conferred a small but reproducible reduction in the proportion of free ribosomal subunits and a consequent increase in the fraction of 70S complexes in the L1 mutant. This restoration towards wild-type particle partitioning is a possible explana- tion for the mechanism of fitness improvement. From the polysome profiles it was not immediately clear how the compensatory mutations in rplN and rplS mediated their beneficial effect on fitness. However both L14 and L19 make a number of bridging contacts with the 30S subunit during subunit association (Gao et al. 2003; Selmer et al. 2006). It is thus conceivable that the mutations in these two proteins alter the interactions between an L1- depleted 50S subunit and the 30S subunit to confer additional stability to 70S complexes.

65 L1

L14 P102L L2

ΔL100 E194K L19

Figure 10: Schematic illustration of the 50S subunit showing the topographical locations of the compensatory mutations relative to r-protein L1 (green ribbon). Compensatory mutations were mapped to r-proteins L2 (red ribbon), L14 (purple ribbon) and L19 (orange ribbon). Yellow spheres represent the amino acid residues of each protein that were altered in the compensated strains. All other proteins and the rRNA are coloured in grey. The image was created in PyMol (www.pymol.org) using the coordinates of the T. thermophilus 50S subunit (PDB ID 2J01) (Selmer et al. 2006).

High functional conservation of r-proteins Since the rate of bacterial growth varies according to the quantity and quality of protein synthesis, the fitness assays (which are based on growth performance) also serve as an indirect measurement of the translational capacity of the hybrid ribosomes. In paper III fitness was measured in both rich and poor growth conditions. During exponential growth in rich medium, such as LB, ribosomes become saturated with substrate and are pushed towards operating at or close to maximum efficiency (Ehrenberg & Kurland 1984). Thus, the observed fitness costs tend to be more pronounced during growth in rich medium versus poorer medium. For example, the relative fitness of the L1 hybrid strain, in which L1 was derived from Sulfolobus acidocal- darius, was approximately 0.85 in M9 medium supplemented with acetate but dropped to approximately 0.3 in LB medium.

66 Furthermore, we could be certain that the majority of the non-native proteins were at least semi-functional considering the much larger fitness costs associated with complete loss of function in the null mutants. Despite extremely large differences in both nucleotide and amino acid identity as well as codon adaptation index (CAI) and GC content, in general the non-native proteins function extraordinarily well in their new environment. This finding suggests that there is very high functional conservation of r-proteins and implies that the functional requirement is superior to conservation of the primary sequence of amino acids. For example, L1 from Helicobacter pylori only shares 50% amino acid identity with the native protein, the cost of replacement however is extremely small (approximately 1%).

Increased dosage rescues sub-optimal r-proteins To determine if and how fitness compensation of the hybrid strains could be achieved, an S20 and L1 hybrid strain of low fitness were subject to compensatory evolution. Independent lineages of the S20 hybrid (containing the Haemophilus influenzae orthologue) and the L1 hybrid (containing the Sac- charomyces cerevisiae orthologue) were serially passaged in rich medium for between 40-250 generations of growth. Compensated mutants were re- covered and were examined for increased copy number of the transferred gene. Indeed, a two- to three-fold increase in gene copy number was detected and this was also apparent at the protein level when the S20 compensated mutants were tested. To verify that overproduction of the non-native protein was sufficient to mitigate the fitness cost of replacement, the orthologous proteins were expressed from an inducible high-copy number plasmid. Growth rate measurements confirmed that for the majority of mutants, increased dosage of the alien protein conferred higher fitness. This implies that the observed fitness costs are at least partially due to either in- sufficient expression of the non-native protein (possibly due to unstable mRNA or protein) or a requirement for more of the alien protein to restore proper ribosome assembly and/or function. Imbalances in the stoichiometric amounts of components that function co-operatively in the cell have previously been suggested to have deleterious effects (Papp et al. 2003). Finally, the observation that increased dosage of transferred r-proteins improves fitness suggests that sub-optimal gene expression may represent a general barrier to HGT and gene amplification may be a general compensatory mechanism for initial stabilization of deleterious HGT events.

67 Future Perspectives Paper I Removal of small subunit protein S20 perturbed the rate and extent of mRNA binding, as well as dramatically reducing the ability of the ribosomal subunits to associate. From careful examination of the X-ray crystal structures of the 30S subunit, we proposed that upon loss of S20 h44 adopts a conformational orientation that perturbs the binding of mRNA and hinders the productive association of ribosomal subunits. With the aim of providing experimental support for this conclusion, S20-depleted 30S subunits are currently undergoing cryo-EM to ascertain the orientation of h44 in the presence and absence of the ribosomal ligands mRNA, initiator tRNA and IFs. Based on the results of the in vitro biochemical assays, we would predict that, compared to the wild-type 30S subunit, h44 would be structurally repositioned in naked S20 30S subunits. Furthermore, our in vitro data predicts that upon extended incubation with the initiation components, h44 becomes stabilized in a more productive conformation, akin to that observed in the wild-type.

Paper II Based on our results, we propose that large subunit L1 plays a major role in ribosome biogenesis. The most striking observation from the polysome profile of the L1 mutant was a considerable increase in the fraction of free ribosomal subunits at the expense of functional 70S complexes. At this point we are uncertain why the increase in free subunits is disproportionate but it seems plausible to suspect that the L1-depleted 50S subunits are more prone to misassembly than the wild-type. It is possible that this directly leads to perturbations in the formation of 30S subunits. Alternatively, 30S formation occurs normally but the misassembled 50S subunits cannot form competent association partners, leading to increased degradation of the 30S species and reduced formation of 70S particles. To investigate this phenotype more thor- oughly, it is possible to purify the subunits from the L1 mutant and examine their propensity to associate in a subunit association assay. By independently evaluating the association capacity of each subunit from the mutant with wild-type subunits we have the possibility of determining if lack of L1 affects biogenesis of only the 50S subunit or of both subunits. Regarding the compensatory mutations, we observed that the rplB E194K mutation reduced the proportion of free subunits and increased the formation of 70S ribosomes. Thus it is evident that this mutation may help in the re- structuring of L1-depleted 50S subunits to promote more stable interactions with the 30S subunit. Since the fitness cost was not entirely removed, this mutation alone is not completely sufficient for restoration of optimal subunit

68 association and this is reflected in the fact that the polysome profile was not restored to wild-type particle partitioning. Based on the 70S crystal structure, we would predict that rplN P102L and rplS L100 alter subunit interactions since both proteins make extensive contact with the 30S subunit (Selmer et al. 2006). Although the polysome profiles provided no obvious evidence that this is the case for the rplS mutation, a substantial increase in 70S monosomes was obtained in the presence of the mutation in rplN. It would therefore be interesting to examine whether the rplS and rplN mutations alter the intersubunit bridging contacts of L1-depleted 50S subunits using a subunit association assay.

New candidate compensatory mutations In addition to the three compensatory mutations described in paper II, whole-genome sequencing has since been undertaken to detect potentially new candidate mutations in the remaining evolved lineages. In one lineage a single nucleotide substitution in rpsM (encoding S13) was detected that results in amino acid substitution R92H. This is the first indication that a mutation on the small subunit can compensate for lack of L1. S13 binds to the head region of mature 30S subunits and its extended tail penetrates the P-site (Selmer et al. 2006). E. coli mutants with truncations of the tail demonstrated a cold-sensitivity phenotype and the 30S subunits had reduced affinity for tRNA (Hoang et al. 2004). Another mutant in which the entire rpsM gene was deleted showed defects in initiation, translocation and subunit association (Cukras & Green 2005). The subunit association defect correlates with the position of S13 on the intersubunit face where it acts as a contact point between the head of the 30S and the central protuberance of the 50S subunit via bridges B1a and B1b. Interestingly bridge B1a is formed between helix 38 of the 23S rRNA and a highly conserved region of S13 that localizes to amino acid positions 92-94 of the T. thermophilus protein (Yusupov et al. 2001). This region is rich in arginine residues and is a close match to the position of the candidate compensatory mutation that we detected. Although this mutation has yet to be confirmed and characterized it is tempting to speculate that the R92H substitution alters association of the 30S subunit with L1-depleted 50S subunits and may enhance 70S complex formation. In three other independent lineages single nucleotide substitutions were detected upstream of the essential gene engA that encodes a tandem GTPase known as Der. The first indication that Der functioned in ribosome biogenesis was the finding that its overexpression suppressed the slow growth of a strain lacking modification of a 23S rRNA residue due to absence of the methyltransferase RrmJ (Tan et al. 2002). More recent studies have demonstrated that Der associates with 50S subunits in the presence of GTP (Bharat et al. 2006; Hwang & Inouye 2006). Decreased expression of Der produced aberrant polysome profiles in which 50S precursor particles were observed

69 with sub-stoichiometric amounts of large subunit proteins L9 and L18 in addition to an accumulation of unprocessed 16S and 23S rRNA (Hwang & Inouye 2006). Based on these observations it is evident that Der is involved in the biogenesis and stability of 50S subunits. The engA gene is the second cistron in a transcription unit with yfgL (bamB in E. coli K12). In our compensated mutants the nucleotide substitutions are positioned after the stop codon of yfgL in a region that appears to contain a complex arrangement of transcriptional attenuators. Although we presently lack experimental evidence for how these mutations improve fitness, it seems probable that they de-stabilize the terminator sequence of yfgL and consequently may increase expression of the Der protein. If loss of L1 causes misassembly of 50S subunits, as our data from paper II suggests, then overexpression of this 50S biogenesis factor may facilitate or assist in re-structuring the L1-depleted 50S subunits to increase 70S complex formation. Initially we intend to confirm this hypothesis by overexpressing Der from an inducible plasmid in the L1 mutant and examining its effect on fitness. In order to establish that these new candidate compensatory mutations are responsible for increasing fitness of the L1 mutant, the next step will require reconstruction of each mutation in the L1 background. Once fitness improvements have been confirmed the effect of each mutation on ribosome biogenesis and function can be evaluated using the assays described in paper II. Although any combination of the confirmed compensatory mutations reduced fitness of the L1 mutant it would be interesting to determine if any combination of the whole set of mutations has additive or synergistic effects on fitness. Finally, during the serial transfer experiment mutants of intermediate fitness were isolated and saved at -80°C. These mutants facilitate an analysis of the dynamics of the compensatory process as they represent "fossils" of prior evolutionary events. By examining the mechanisms of compensation in these mutants we may be able to identify the varied routes taken towards the fitness optima in each independent lineage.

Paper III Although the hybrid ribosomes had comparatively small fitness costs when the r-protein was transferred from close relatives, in general larger costs were observed as phylogenetic distance increased. Compensatory evolution revealed that increased dosage of the sub-optimal replacement protein was one mechanism of improving fitness and presumably ribosome function. Since gene duplication occurs much more frequently than rare point mutations that improve function, it is reasonable to propose that gene amplification might represent a general response to initially increase fitness upon transfer of a heterologous protein to a new environment as observed in pa-

70 per III. However, not all evolved lineages carried amplifications of the non- native gene. In one case a mutation in the promoter region was detected but the mechanism of compensation in the remaining lineages is presently unknown. Whole-genome sequencing of these evolved strains should reveal additional mechanisms that mitigate the fitness cost of r-protein replacement with distantly related orthologues. Furthermore, the evolution experiment was only performed for up to 25 cycles, thus it is possible that further fitness gains are attainable if the experiment is continued. It may even be possible to evolve secondary adaptive mutations in the amplified genes since the increased copy number provides a larger target for mutational events to act on. With regard to those hybrid ribosomes that function poorly, it may also be interesting to examine how they function. Are any particular steps in translation defective? Or are there problems in ribosome biogenesis that might be due to poor expression of the heterologous gene? From the experiments performed on the null mutants we are aware of the functional contributions of native S20 and L1, however, L17 is essential and no information on its role in translation is available. Thus a comprehensive analysis of the function of the L17 hybrid ribosomes that confer high fitness costs offers a unique opportunity to delineate the importance of this protein in translation. In terms of the barriers that restrict HGT, it would be informative to use another model gene to examine if the observed constraints are general or specific to r-proteins. In paper III, we observe that functional constraints are highly conserved between species and fitness costs appear to be partly the result of poor expression levels. R-protein genes are special in the sense that they are highly conserved, strongly expressed and have optimized codon usage. In addition, we know that the ribosome is an ancient biomolecule and these proteins have co-evolved with it, which may mean that the evolutionary constraints differ widely from those operating on another, more “average” gene. Furthermore, use of other model genes would allow us to examine the importance and relevance of gene amplification as a general response after HGT.

71 Concluding remarks

The most fundamental aspect of this thesis is directed towards understanding ribosome complexity in terms of the selective contributions made by certain r-proteins. Why does the ribosome contain such an impressive array of proteins if rRNA performs the basic functional tasks? What, if any, are the benefits afforded by the r-proteins? Given the large fitness costs and functional defects upon complete removal of S20 or L1, we are assured of their importance in maintaining optimal activity. However in the case of both deletion mutants, the ribosomal defects could not be specifically ascribed to loss of the protein, but rather loss of rRNA structural integrity. The significance of rRNA-protein interactions is illuminated by various r-protein mutants in which the phenotype is a consequence of alterations in rRNA structure (Allen & Noller 1989). Although certain proteins, such as S12, contribute to distinct steps in protein synthesis, the major role of S20 and L1 appears to be in the support and maintenance of the higher order structure of rRNA. The results of paper I suggest that removal of S20 perturbs the orientation of h44 leading to defective mRNA binding and subunit association. This conclusion is strengthened by the suggestion that S20 anchors this helix in its correct position in mature 30S subunits (Schluenzen et al. 2000). Similarly removal of L1 in paper II alters the distribution of ribosomal particles in the cell, and although we lack conclusive experimental evidence, it is reasonable to suspect that this is due to misassembly of 50S subunits. L1 has previously been shown to act as an RNA chaperone and may be required during ribosome assembly (Ameres et al. 2007). Moreover, compensatory mutations selected as a response to L1 removal were mapped to other proteins dispersed at various sites of the large subunit, suggesting that they might amend the structural perturbations to some degree. Thus, in both cases it seems that loss of the protein indirectly perturbs protein synthesis via alterations in the overall morphology of the ribosome. Perhaps this is not so surprising considering that both proteins are primary rRNA binders and are thus considered particularly important for initiating global folding of the major rRNA domains. In contrast to the large fitness costs of r-protein removal, in general the replacement proteins (paper III) functioned surprisingly well in their new context even when the transferred protein shared as little as 50% amino acid identity with the native protein. This clearly demonstrates the robustness of the ribosome and protein synthesis in general and it also implies that functional

72 constraints are highly conserved between these proteins despite having evolved in species that were, in some cases, only distantly related on the phylogenetic level. This compatibility between highly divergent r-proteins was observed in a previous study in which S18 from rye chloroplast, that shares only 35% amino acid identity with E. coli S18, was incorporated into 30-40% of E. coli ribosomes and polysomes when overexpressed from a plasmid (Weglohner et al. 1997). The identification of a whole host of compensatory mutations that suppress the fitness cost of L1 removal in paper II also highlights the enormous flexibility and co-operativity of ribosome activity. The low fitness costs associated with most of the hybrid strains suggests conservation in the overall shape of the proteins or at least those residues that are required for interaction with the ribosome, prompting the idea that the orthologous proteins may act as structural mimics of the native proteins. Even if the overall fold is not analogous to the native protein, previous observations have suggested that divergent proteins can stabilize similar RNA structures (Klein et al. 2004). Overall the findings of paper III emphasize the superiority of functional requirement over conservation of the primary sequence of amino acids. Hybrid ribosomes have been constructed previously wherein the E. coli proteins of the GTPase-associated centre were replaced with their rat counterparts in vitro (Uchiumi et al. 2002). These hybrid ribosomes required eukaryotic elongation factors to support protein synthesis in vitro reflecting the involvement of these proteins in translation factor recruitment and activation. It thus seems likely that the generally small costs of S20, L1 and L17 replacement reflect their inertness in specialized ribosomal functions. It is suggested that the precursor of the modern ribosome was formed as the result of dimerization of two self-folding RNA chains to form the symmetrical pocket that we see today as the PTC of the large subunit (Davidovich et al. 2009). The transition from relatively simple peptidyl transfer to the elaborate coded process of translation in modern cells probably required a supporting network of RNA and proteins to contour and stabilize the catalytic site allowing it to achieve more efficiency. Since the fundamental structure of the ribosome is an inherent feature of its rRNA, and the r-proteins are considered late additions to the structure, it is likely that rRNA structure constituted the main selective pressure for evolution of the proteins. Hence we observe high functional conservation of the proteins S20, L1 and L17 despite large variations in the primary sequences of orthologues from various species. The fact that the main impairments caused by loss of the proteins L1 and S20 appear to be the result of perturbations in the conformation of rRNA structure, rather than direct defects at particular steps in translation, reinforces the superiority of rRNA over protein in ribosome function. However, impaired ribosome activity upon removal and replacement of r-proteins also highlights the fundamental importance of structural support from the r-proteins, without which, the translation apparatus operates sub-optimally.

73 Acknowledgements

The road travelled as a PhD student has often been a lonely one. Neverthe- less, without the support and guidance of those nearest and dearest, I doubt that I would ever have made it this far. Here are my thanks to all of you.

First of all, my main supervisor Dan Andersson. You never fail to impress me with your wealth of knowledge and ideas. I seriously envy the great scientific mind you have! Thank you for having faith in me, especially during those times when I lacked it myself. I am also very grateful for the opportu- nities you have provided that allowed me to travel so much during my time as a PhD student.

My co-supervisors Suparna Sanyal and Lars Sundström. Suparna, you are one of the most driven people I have ever met and a truly inspirational scien- tist. I am so grateful for the time you have spent with me in planning and analyzing the in vitro experiments. Thank you for your encouraging words along the way, they meant more than you know. Lars, although we may not have had as much contact I really appreciate the interest you showed in my project even though it was outside of your scientific realm.

My co-authors especially Chandu, the ribosome guru, thank you for being so generous with your time and for being such an easy-going guy. I have no idea how I would have coped with all the ribosome preps if you weren’t on stand-by to help, regardless of the time of the day or night! Peter you’re a true master of evolutionary theory and genetics and a real pleasure to discuss science with. Jocke, thank you for sharing your talent as a microbial geneti- cist and for being willing to discuss various scientific problems that were encountered during the projects. Måns, I’m very grateful for your insightful suggestions during the design of the biochemical experiments and your comments on my written work. Stefan and Mikael, thank you for your mar- velous work with the polysome profiles as well as the interesting discussions that ensued!

To the “GODS” that are the present and former members of the DA group. Sanna, from the moment you picked me up at Uppsala train station when I first arrived we became and I’m sure that we’ll always remain good friends. Thank you for always being available to discuss the PhD blues and for tak-

74 ing time to bounce scientific ideas around with. Peter, thank you for listen- ing to my frequent rants about the project and for reassuring me that things eventually work themselves out. Linus, thank you for being that someone that all of us in the lab can count on and for being a great listener. Maria, you bring such positivity to the lab and I’ve really enjoyed the distraction of discussing movies rather than science every now and then. Anna, the one that always made sure that things got done in the lab, thanks for reminding me that work doesn’t always need to be taken so seriously. Song, now the most senior PhD student although I do think you’re over-qualified! You have a great attitude and friendliness that I admire. Jocke, I really appreciate your general interest in problem solving and your willingness to share the occa- sional beer or two (or five!). Hava, thanks for taking the pressure off me as the only native English speaker! Seriously though, I’m delighted to have someone around that’s so easy to talk to and so laid-back. Hervé, you have such a great sense of humour and really make me laugh out loud. You have also brought so much knowledge to the group and I’m very grateful for the time you have taken to discuss the project with me. Ulrika, how did we ever cope without you?! I’m so glad that such a considerate and friendly person joined the group. Amira, I think that we’re bad for each other in the sense that we’re both such worriers! Thanks for the therapeutic lunchtime chats though. Marlen, Lisa, Erik and Marius, you’ve all added fresh enthusiasm to the group and I look forward to being more present with all of you in the lab again soon. Former group members, especially Camilla, my ally during the early days, how I’ve missed having you around. I am also very grateful to Peter and Jocke for proofreading and constructive criticism of this thesis.

Thanks to all of my IMBIM colleagues, many of whom I consider dear friends especially the virology girls Ellenor, Heidi, Monika and Anna. You have always been prepared to offer advice and give your support when times are tough. Lena M., for knowing everything about everything and encouraging me to learn Swedish during the early days. Andrea, Göte, Nizar, Amany and Carolina, the faces I see everyday, you’ve all been great for chats and encouragement, especially now towards the end. Previous members of the department, especially Jens and Katarina - IMBIM has not been the same without you guys!

My sincere gratitude goes to all the IMBIM staff for your efficiency and dedication to the smooth running of the department. Especially Barbro – always reliable, ready to help and a constant source of encouragement!

To my secondary school biology teacher Dr. Mary Kelly for introducing me to the world of science. Your great teaching skills, enthusiasm and passion for science certainly sparked my interest in the subject all those years ago.

75 To the lecturers in the School of Biological Sciences at Dublin Institute of Technology, especially Dr. Patrick McHale for introducing me to the microbial world and convincing me to come to IMBIM on Erasmus.

To all of those wonderful friends that have made my time here in Sweden so enjoyable. Firstly the Irish crew Fiona, Lesley, Nikki and Lucy – to think that we may never have met unless our paths crossed here. Thanks for all the fun times that we shared over the years and for putting up with my frenzied screaming during the rugby games… C’mon Ireland! My dear Swedish friends especially Johan A, Susanna and Micke for the fantastic inter- railing adventures and the various social gatherings that have added so much joy to my life. The ever-splendid Jenny and Kristin, although not so frequent these days, I really enjoy our fikas and chats about everything other than science.

To my friends back home in Ireland, especially Corina, Claire and Brian, you mean more to me than the rare phone calls and emails may indicate.

My Swedish family, the Kåhrströms. Thanks to you all for welcoming me so warmly into your lives and hearts. Especially Ann-Kristin and Per Einar, your support has helped me enormously and I really appreciate the effort that you make in creating a relaxing and homely environment when I need it most.

I would never have made it to this point without the love and support of my own family, especially my mam Carmel and dad Séamus. You have both taught me more about hard work than anyone else. Thanks so much for your constant encouragement and your unfaultering belief in my ability. Thanks for making such a huge effort to help me relax during my visits home. Words are not adequate to express how much I appreciate all that you’ve done for me and the sacrifices you’ve made. To my sister Sharon and brother James, I may be the annoying older sister but you are both experts at putting me firmly in my place!

Oh Nooooo… there’s bubbles!!! Thanks to my wonderful Johan for your company and assistance during those late-night encounters with the ultracen- trifuge, at this point you could probably easily do a ribosome prep on your own! I simply can’t thank you enough for your patience, understanding and support. I am so grateful for your willingness to help and to discuss my nig- gling thoughts about the ribosome. This PhD has been tough on both of us so most of all thank you for not giving up on me and for walking this path by my side. Du är bäst helt enkelt och jag älskar dig!

76 References

Adilakshmi, T., Bellur, D. L. and Woodson, S. A. (2008). "Concurrent nucleation of 16S folding and induced fit in 30S ribosome assembly." Nature 455(7217): 1268-1272. Adilakshmi, T., Ramaswamy, P. and Woodson, S. A. (2005). "Protein- independent folding pathway of the 16S rRNA 5' domain." J Mol Biol 351(3): 508-519. Agirrezabala, X., Lei, J., Brunelle, J. L., Ortiz-Meoz, R. F., Green, R. and Frank, J. (2008). "Visualization of the hybrid state of tRNA binding promoted by spontaneous ratcheting of the ribosome." Mol Cell 32(2): 190-197. Allen, P. N. and Noller, H. F. (1989). "Mutations in ribosomal proteins S4 and S12 influence the higher order structure of 16 S ribosomal RNA." J Mol Biol 208(3): 457-468. Ameres, S. L., Shcherbakov, D., Nikonova, E., Piendl, W., Schroeder, R. and Semrad, K. (2007). "RNA chaperone activity of L1 ribosomal proteins: phylogenetic conservation and splicing inhibition." Nucleic Acids Res 35(11): 3752-3763. Antoun, A., Pavlov, M. Y., Andersson, K., Tenson, T. and Ehrenberg, M. (2003). "The roles of initiation factor 2 and guanosine triphosphate in initiation of protein synthesis." EMBO J 22(20): 5593-5601. Antoun, A., Pavlov, M. Y., Lovmar, M. and Ehrenberg, M. (2006a). "How initiation factors maximize the accuracy of tRNA selection in initiation of bacterial protein synthesis." Mol Cell 23(2): 183-193. Antoun, A., Pavlov, M. Y., Lovmar, M. and Ehrenberg, M. (2006b). "How initiation factors tune the rate of initiation of protein synthesis in bacteria." EMBO J 25(11): 2539-2550. Aoki, H., Dekany, K., Adams, S. L. and Ganoza, M. C. (1997). "The gene encoding the elongation factor P protein is essential for viability and is required for protein synthesis." J Biol Chem 272(51): 32254-32259. Aoki, H., Xu, J., Emili, A., Chosay, J. G., Golshani, A. and Ganoza, M. C. (2008). "Interactions of elongation factor EF-P with the Escherichia coli ribosome." FEBS J 275(4): 671-681. Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K. A., Tomita, M., Wanner, B. L. and Mori, H. (2006). "Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection." Mol Syst Biol 2: 2006 0008.

77 Ban, N., Nissen, P., Hansen, J., Moore, P. B. and Steitz, T. A. (2000). "The complete atomic structure of the large ribosomal subunit at 2.4 A resolution." Science 289(5481): 905-920. Baughman, G. and Nomura, M. (1984). "Translational regulation of the L11 ribosomal protein operon of Escherichia coli: analysis of the mRNA target site using oligonucleotide-directed mutagenesis." Proc Natl Acad Sci U S A 81(17): 5389-5393. Benard, L., Philippe, C., Ehresmann, B., Ehresmann, C. and Portier, C. (1996). "Pseudoknot and translational control in the expression of the S15 ribosomal protein." Biochimie 78(7): 568-576. Bharat, A., Jiang, M., Sullivan, S. M., Maddock, J. R. and Brown, E. D. (2006). "Cooperative and critical roles for both G domains in the GTPase activity and cellular function of ribosome-associated Escherichia coli EngA." J Bacteriol 188(22): 7992-7996. Bjorkman, J., Samuelsson, P., Andersson, D. I. and Hughes, D. (1999). "Novel ribosomal mutations affecting translational accuracy, antibiotic resistance and virulence of Salmonella typhimurium." Mol Microbiol 31(1): 53-58. Blaha, G., Stanley, R. E. and Steitz, T. A. (2009). "Formation of the first peptide bond: the structure of EF-P bound to the 70S ribosome." Science 325(5943): 966-970. Blanchard, S. C., Kim, H. D., Gonzalez, R. L., Jr., Puglisi, J. D. and Chu, S. (2004). "tRNA dynamics on the ribosome during translation." Proc Natl Acad Sci U S A 101(35): 12893-12898. Bremer, H. a. D., P. P. (1996). Modulation of Chemical Composition and Other Parameters of the Cell by Growth Rate. Escherichia coli and Salmonella: Cellular and Molecular Biology. F. C. Neidhart. Washington D.C., ASM Press. 2: 1553-1566. Brodersen, D. E., Clemons, W. M., Jr., Carter, A. P., Wimberly, B. T. and Ramakrishnan, V. (2002). "Crystal structure of the 30 S ribosomal subunit from Thermus thermophilus: structure of the proteins and their interactions with 16 S RNA." J Mol Biol 316(3): 725-768. Bubunenko, M., Baker, T. and Court, D. L. (2007). "Essentiality of ribosomal and transcription antitermination proteins analyzed by systematic gene replacement in Escherichia coli." J Bacteriol 189(7): 2844-2853. Bubunenko, M., Korepanov, A., Court, D. L., Jagannathan, I., Dickinson, D., Chaudhuri, B. R., Garber, M. B. and Culver, G. M. (2006). "30S ribosomal subunits can be assembled in vivo without primary binding ribosomal protein S15." RNA 12(7): 1229-1239. Burgin, A. B., Parodos, K., Lane, D. J. and Pace, N. R. (1990). "The excision of intervening sequences from Salmonella 23S ribosomal RNA." Cell 60(3): 405-414.

78 Burmann, B. M., Schweimer, K., Luo, X., Wahl, M. C., Stitt, B. L., Gottesman, M. E. and Rosch, P. (2010). "A NusE:NusG complex links transcription and translation." Science 328(5977): 501-504. Canchaya, C., Fournous, G., Chibani-Chennoufi, S., Dillmann, M. L. and Brussow, H. (2003). "Phage as agents of lateral gene transfer." Curr Opin Microbiol 6(4): 417-424. Capel, M. S., Engelman, D. M., Freeborn, B. R., Kjeldgaard, M., Langer, J. A., Ramakrishnan, V., Schindler, D. G., Schneider, D. K., Schoenborn, B. P., Sillers, I. Y. and et al. (1987). "A complete mapping of the proteins in the small ribosomal subunit of Escherichia coli." Science 238(4832): 1403-1406. Carter, A. P., Clemons, W. M., Brodersen, D. E., Morgan-Warren, R. J., Wimberly, B. T. and Ramakrishnan, V. (2000). "Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics." Nature 407(6802): 340-348. Carter, A. P., Clemons, W. M., Jr., Brodersen, D. E., Morgan-Warren, R. J., Hartsch, T., Wimberly, B. T. and Ramakrishnan, V. (2001). "Crystal structure of an initiation factor bound to the 30S ribosomal subunit." Science 291(5503): 498-501. Cech, T. R. (2000). "Structural biology. The ribosome is a ribozyme." Science 289(5481): 878-879. Charollais, J., Pflieger, D., Vinh, J., Dreyfus, M. and Iost, I. (2003). "The DEAD-box RNA helicase SrmB is involved in the assembly of 50S ribosomal subunits in Escherichia coli." Mol Microbiol 48(5): 1253- 1265. Chen, I., Christie, P. J. and Dubnau, D. (2005). "The ins and outs of DNA transfer in bacteria." Science 310(5753): 1456-1460. Cheng, Z. F. and Deutscher, M. P. (2003). "Quality control of ribosomal RNA mediated by polynucleotide phosphorylase and RNase R." Proc Natl Acad Sci U S A 100(11): 6388-6393. Comartin, D. J. and Brown, E. D. (2006). "Non-ribosomal factors in ribosome subunit assembly are emerging targets for new antibacterial drugs." Curr Opin Pharmacol 6(5): 453-458. Condon, C., French, S., Squires, C. and Squires, C. L. (1993). "Depletion of functional ribosomal RNA operons in Escherichia coli causes increased expression of the remaining intact copies." EMBO J 12(11): 4305-4315. Connolly, K. and Culver, G. (2009). "Deconstructing ribosome construction." Trends Biochem Sci 34(5): 256-263. Connolly, K., Rife, J. P. and Culver, G. (2008). "Mechanistic insight into the ribosome biogenesis functions of the ancient protein KsgA." Mol Microbiol 70(5): 1062-1075. Cornish, P. V., Ermolenko, D. N., Staple, D. W., Hoang, L., Hickerson, R. P., Noller, H. F. and Ha, T. (2009). "Following movement of the L1 stalk between three functional states in single ribosomes." Proc Natl Acad Sci U S A 106(8): 2571-2576.

79 Cukras, A. R. and Green, R. (2005). "Multiple effects of S13 in modulating the strength of intersubunit interactions in the ribosome during translation." J Mol Biol 349(1): 47-59. Dabbs, E. R. (1977). "A spectinomycin dependent mutant of Escherichia coli." Mol Gen Genet 151(3): 261-267. Dabbs, E. R. (1978). "Mutational alterations in 50 proteins of the Escherichia coli ribosome." Mol Gen Genet 165(1): 73-78. Dabbs, E. R. (1979). "Selection for Escherichia coli mutants with proteins missing from the ribosome." J Bacteriol 140(2): 734-737. Dabbs, E. R. (1980). "The ribosomal components responsible for kasugamycin dependence, and its suppression, in a mutant of Escherichia coli." Mol Gen Genet 177(2): 271-276. Dabbs, E. R. (1991). "Mutants lacking individual ribosomal proteins as a tool to investigate ribosomal properties." Biochimie 73(6): 639-645. Dabbs, E. R., Ehrlich, R., Hasenbank, R., Schroeter, B. H., Stoffler- Meilicke, M. and Stoffler, G. (1981). "Mutants of Escherichia coli lacking ribosomal protein L1." J Mol Biol 149(4): 553-578. Dabbs, E. R. and Wittman, H. G. (1976). "A strain of Escherichia coli which gives rise to mutations in a large number of ribosomal proteins." Mol Gen Genet 149(3): 303-309. Dallas, A. and Noller, H. F. (2001). "Interaction of translation initiation factor 3 with the 30S ribosomal subunit." Mol Cell 8(4): 855-864. Datsenko, K. A. and Wanner, B. L. (2000). "One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products." Proc Natl Acad Sci U S A 97(12): 6640-6645. Davidovich, C., Belousoff, M., Bashan, A. and Yonath, A. (2009). "The evolving ribosome: from non-coded peptide bond formation to sophisticated translation machinery." Res Microbiol 160(7): 487-492. de Smit, M. H. and van Duin, J. (2003). "Translational standby sites: how ribosomes may deal with the rapid folding kinetics of mRNA." J Mol Biol 331(4): 737-743. Decatur, W. A. and Fournier, M. J. (2002). "rRNA modifications and ribosome function." Trends Biochem Sci 27(7): 344-351. Deutscher, M. P. (2009). "Maturation and degradation of ribosomal RNA in bacteria." Prog Mol Biol Transl Sci 85: 369-391. Diaconu, M., Kothe, U., Schlunzen, F., Fischer, N., Harms, J. M., Tonevitsky, A. G., Stark, H., Rodnina, M. V. and Wahl, M. C. (2005). "Structural basis for the function of the ribosomal L7/12 stalk in factor binding and GTPase activation." Cell 121(7): 991-1004. Doudna, J. A. and Rath, V. L. (2002). "Structure and function of the eukaryotic ribosome: the next frontier." Cell 109(2): 153-156. Dresios, J., Panopoulos, P. and Synetos, D. (2006). "Eukaryotic ribosomal proteins lacking a eubacterial counterpart: important players in ribosomal function." Mol Microbiol 59(6): 1651-1663.

80 Dutca, L. M. and Culver, G. M. (2008). "Assembly of the 5' and 3' minor domains of 16S ribosomal RNA as monitored by tethered probing from ribosomal protein S20." J Mol Biol 376(1): 92-108. Ehrenberg, M., Bilgin, N., Kurland, C. G. (1989). Design and use of a fast and accurate in vitro translation system. Ribosomes and protein synthesis: a practical approach. G. Spedding. New York, Oxford University Press: 101-129. Ehrenberg, M. and Kurland, C. G. (1984). "Costs of accuracy determined by a maximal growth rate constraint." Q Rev Biophys 17(1): 45-82. Fei, J., Bronson, J. E., Hofman, J. M., Srinivas, R. L., Wiggins, C. H. and Gonzalez, R. L., Jr. (2009). "Allosteric collaboration between elongation factor G and the ribosomal L1 stalk directs tRNA movements during translation." Proc Natl Acad Sci U S A 106(37): 15702-15707. Fei, J., Kosuri, P., MacDougall, D. D. and Gonzalez, R. L., Jr. (2008). "Coupling of ribosomal L1 stalk and tRNA dynamics during translation elongation." Mol Cell 30(3): 348-359. Fischer, N., Konevega, A. L., Wintermeyer, W., Rodnina, M. V. and Stark, H. (2010). "Ribosome dynamics and tRNA movement by time-resolved electron cryomicroscopy." Nature 466(7304): 329-333. Franceschi, F. J. and Nierhaus, K. H. (1988). "Ribosomal protein L20 can replace the assembly-initiator protein L24 at low temperatures." Biochemistry 27(18): 7056-7059. Frank, J. and Agrawal, R. K. (2000). "A ratchet-like inter-subunit reorganization of the ribosome during translocation." Nature 406(6793): 318-322. Frank, J., Zhu, J., Penczek, P., Li, Y., Srivastava, S., Verschoor, A., Radermacher, M., Grassucci, R., Lata, R. K. and Agrawal, R. K. (1995). "A model of protein synthesis based on cryo-electron microscopy of the E. coli ribosome." Nature 376(6539): 441-444. Futuyma, D. J. (1998). Evolutionary Biology. Sunderland, Massachusetts, Sinauer Associates, Inc. Gao, H., Sengupta, J., Valle, M., Korostelev, A., Eswar, N., Stagg, S. M., Van Roey, P., Agrawal, R. K., Harvey, S. C., Sali, A., Chapman, M. S. and Frank, J. (2003). "Study of the structural dynamics of the E coli 70S ribosome using real-space refinement." Cell 113(6): 789-801. Gao, H., Zhou, Z., Rawat, U., Huang, C., Bouakaz, L., Wang, C., Cheng, Z., Liu, Y., Zavialov, A., Gursky, R., Sanyal, S., Ehrenberg, M., Frank, J. and Song, H. (2007). "RF3 induces ribosomal conformational changes responsible for dissociation of class I release factors." Cell 129(5): 929- 941. Gao, N., Zavialov, A. V., Li, W., Sengupta, J., Valle, M., Gursky, R. P., Ehrenberg, M. and Frank, J. (2005). "Mechanism for the disassembly of the posttermination complex inferred from cryo-EM studies." Mol Cell 18(6): 663-674.

81 Gilbert, R. J., Fucini, P., Connell, S., Fuller, S. D., Nierhaus, K. H., Robinson, C. V., Dobson, C. M. and Stuart, D. I. (2004). "Three- dimensional structures of translating ribosomes by Cryo-EM." Mol Cell 14(1): 57-66. Gotz, F., Dabbs, E. R. and Gualerzi, C. O. (1990). "Escherichia coli 30S mutants lacking protein S20 are defective in translation initiation." Biochim Biophys Acta 1050(1-3): 93-97. Gotz, F., Fleischer, C., Pon, C. L. and Gualerzi, C. O. (1989). "Subunit association defects in Escherichia coli ribosome mutants lacking proteins S20 and L11." Eur J Biochem 183(1): 19-24. Gourse, R. L., Gaal, T., Bartlett, M. S., Appleman, J. A. and Ross, W. (1996). "rRNA transcription and growth rate-dependent regulation of ribosome synthesis in Escherichia coli." Annu Rev Microbiol 50: 645- 677. 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." Mol Cell 20(6): 917-927. Green, R. and Noller, H. F. (1996). "In vitro complementation analysis localizes 23S rRNA posttranscriptional modifications that are required for Escherichia coli 50S ribosomal subunit assembly and function." RNA 2(10): 1011-1021. Grondek, J. F. and Culver, G. M. (2004). "Assembly of the 30S ribosomal subunit: positioning ribosomal protein S13 in the S7 assembly branch." RNA 10(12): 1861-1866. Gu, S. Q., Peske, F., Wieden, H. J., Rodnina, M. V. and Wintermeyer, W. (2003). "The signal recognition particle binds to protein L23 at the peptide exit of the Escherichia coli ribosome." RNA 9(5): 566-573. Hanawa-Suetsugu, K., Sekine, S., Sakai, H., Hori-Takemoto, C., Terada, T., Unzai, S., Tame, J. R., Kuramitsu, S., Shirouzu, M. and Yokoyama, S. (2004). "Crystal structure of elongation factor P from Thermus thermophilus HB8." Proc Natl Acad Sci U S A 101(26): 9595-9600. Harms, J., Schluenzen, F., Zarivach, R., Bashan, A., Gat, S., Agmon, I., Bartels, H., Franceschi, F. and Yonath, A. (2001). "High resolution structure of the large ribosomal subunit from a mesophilic eubacterium." Cell 107(5): 679-688. Held, W. A., Ballou, B., Mizushima, S. and Nomura, M. (1974). "Assembly mapping of 30 S ribosomal proteins from Escherichia coli. Further studies." J Biol Chem 249(10): 3103-3111. Helgstrand, M., Mandava, C. S., Mulder, F. A., Liljas, A., Sanyal, S. and Akke, M. (2007). "The ribosomal stalk binds to translation factors IF2, EF-Tu, EF-G and RF3 via a conserved region of the L12 C-terminal domain." J Mol Biol 365(2): 468-479. Hirokawa, G., Kiel, M. C., Muto, A., Selmer, M., Raj, V. S., Liljas, A., Igarashi, K., Kaji, H. and Kaji, A. (2002). "Post-termination complex

82 disassembly by ribosome recycling factor, a functional tRNA mimic." EMBO J 21(9): 2272-2281. Hoang, L., Fredrick, K. and Noller, H. F. (2004). "Creating ribosomes with an all-RNA 30S subunit P site." Proc Natl Acad Sci U S A 101(34): 12439-12443. Huang, C., Mandava, C. S. and Sanyal, S. (2010). "The ribosomal stalk plays a key role in IF2-mediated association of the ribosomal subunits." J Mol Biol 399(1): 145-153. Hwang, J. and Inouye, M. (2006). "The tandem GTPase, Der, is essential for the biogenesis of 50S ribosomal subunits in Escherichia coli." Mol Microbiol 61(6): 1660-1672. Ibba, M. and Soll, D. (2000). "Aminoacyl-tRNA synthesis." Annu Rev Biochem 69: 617-650. Inoue, K., Alsina, J., Chen, J. and Inouye, M. (2003). "Suppression of defective ribosome assembly in a rbfA deletion mutant by overexpression of Era, an essential GTPase in Escherichia coli." Mol Microbiol 48(4): 1005-1016. Ito, K., Uno, M. and Nakamura, Y. (2000). "A tripeptide 'anticodon' deciphers stop codons in messenger RNA." Nature 403(6770): 680-684. Jain, R., Rivera, M. C. and Lake, J. A. (1999). "Horizontal gene transfer among genomes: the complexity hypothesis." Proc Natl Acad Sci U S A 96(7): 3801-3806. Jin, H., Kelley, A. C., Loakes, D. and Ramakrishnan, V. (2010). "Structure of the 70S ribosome bound to release factor 2 and a substrate analog provides insights into catalysis of peptide release." Proc Natl Acad Sci U S A 107(19): 8593-8598. Jinks-Robertson, S., Gourse, R. L. and Nomura, M. (1983). "Expression of rRNA and tRNA genes in Escherichia coli: evidence for feedback regulation by products of rRNA operons." Cell 33(3): 865-876. Kaczanowska, M. and Ryden-Aulin, M. (2007). "Ribosome biogenesis and the translation process in Escherichia coli." Microbiol Mol Biol Rev 71(3): 477-494. Karimi, R., Pavlov, M. Y., Buckingham, R. H. and Ehrenberg, M. (1999). "Novel roles for classical factors at the interface between translation termination and initiation." Mol Cell 3(5): 601-609. Kennell, D. and Riezman, H. (1977). "Transcription and translation initiation frequencies of the Escherichia coli lac operon." J Mol Biol 114(1): 1-21. Kimura, M. (1990). "Some models of neutral evolution, compensatory evolution, and the shifting balance process." Theor Popul Biol 37(1): 150-158. King, T. C., Sirdeshmukh, R. and Schlessinger, D. (1984). "RNase III cleavage is obligate for maturation but not for function of Escherichia coli pre-23S rRNA." Proc Natl Acad Sci U S A 81(1): 185-188. Kisselev, L. L. and Buckingham, R. H. (2000). "Translational termination comes of age." Trends Biochem Sci 25(11): 561-566.

83 Klein, D. J., Moore, P. B. and Steitz, T. A. (2004). "The roles of ribosomal proteins in the structure assembly, and evolution of the large ribosomal subunit." J Mol Biol 340(1): 141-177. Korepanov, A. P., Gongadze, G. M., Garber, M. B., Court, D. L. and Bubunenko, M. G. (2007). "Importance of the 5 S rRNA-binding ribosomal proteins for cell viability and translation in Escherichia coli." J Mol Biol 366(4): 1199-1208. Korostelev, A., Asahara, H., Lancaster, L., Laurberg, M., Hirschi, A., Zhu, J., Trakhanov, S., Scott, W. G. and Noller, H. F. (2008a). "Crystal structure of a translation termination complex formed with release factor RF2." Proc Natl Acad Sci U S A 105(50): 19684-19689. Korostelev, A., Ermolenko, D. N. and Noller, H. F. (2008b). "Structural dynamics of the ribosome." Curr Opin Chem Biol 12(6): 674-683. Korostelev, A. and Noller, H. F. (2007). "The ribosome in focus: new structures bring new insights." Trends Biochem Sci 32(9): 434-441. Kramer, E. B. and Farabaugh, P. J. (2007). "The frequency of translational misreading errors in E. coli is largely determined by tRNA competition." RNA 13(1): 87-96. Kramer, G., Rauch, T., Rist, W., Vorderwulbecke, S., Patzelt, H., Schulze- Specking, A., Ban, N., Deuerling, E. and Bukau, B. (2002). "L23 protein functions as a chaperone docking site on the ribosome." Nature 419(6903): 171-174. Kurland, C. G., Hughes, D., Ehrenberg, M. (1996). Limitations of Translational Accuracy, ASM, Washington DC. La Teana, A., Pon, C. L. and Gualerzi, C. O. (1993). "Translation of mRNAs with degenerate initiation triplet AUU displays high initiation factor 2 dependence and is subject to initiation factor 3 repression." Proc Natl Acad Sci U S A 90(9): 4161-4165. 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(7206): 852-857. Laursen, B. S., Sorensen, H. P., Mortensen, K. K. and Sperling-Petersen, H. U. (2005). "Initiation of protein synthesis in bacteria." Microbiol Mol Biol Rev 69(1): 101-123. Lecompte, O., Ripp, R., Thierry, J. C., Moras, D. and Poch, O. (2002). "Comparative analysis of ribosomal proteins in complete genomes: an example of reductive evolution at the domain scale." Nucleic Acids Res 30(24): 5382-5390. Levin, B. R., Perrot, V. and Walker, N. (2000). "Compensatory mutations, antibiotic resistance and the population genetics of adaptive evolution in bacteria." Genetics 154(3): 985-997. Liljas, A. (2004). Structural Aspects of Protein Synthesis. Singapore, World Scientific Publishing Co. Pte. Ltd.

84 Lu, M. and Steitz, T. A. (2000). "Structure of Escherichia coli ribosomal protein L25 complexed with a 5S rRNA fragment at 1.8-A resolution." Proc Natl Acad Sci U S A 97(5): 2023-2028. Maguire, B. A., Beniaminov, A. D., Ramu, H., Mankin, A. S. and Zimmermann, R. A. (2005). "A protein component at the heart of an RNA machine: the importance of protein l27 for the function of the bacterial ribosome." Mol Cell 20(3): 427-435. Maguire, B. A. and Zimmermann, R. A. (2001). "The ribosome in focus." Cell 104(6): 813-816. Maisnier-Patin, S. and Andersson, D. I. (2004). "Adaptation to the deleterious effects of antimicrobial drug resistance mutations by compensatory evolution." Res Microbiol 155(5): 360-369. Maisnier-Patin, S., Berg, O. G., Liljas, L. and Andersson, D. I. (2002). "Compensatory adaptation to the deleterious effect of antibiotic resistance in Salmonella typhimurium." Mol Microbiol 46(2): 355-366. Marintchev, A. and Wagner, G. (2004). "Translation initiation: structures, mechanisms and evolution." Q Rev Biophys 37(3-4): 197-284. Marzi, S., Myasnikov, A. G., Serganov, A., Ehresmann, C., Romby, P., Yusupov, M. and Klaholz, B. P. (2007). "Structured mRNAs regulate translation initiation by binding to the platform of the ribosome." Cell 130(6): 1019-1031. Mattatall, N. R. and Sanderson, K. E. (1998). "RNase III deficient Salmonella typhimurium LT2 contains intervening sequences (IVSs) in its 23S rRNA." FEMS Microbiol Lett 159(2): 179-185. McClelland, M., Sanderson, K. E., Spieth, J., Clifton, S. W., Latreille, P., Courtney, L., Porwollik, S., Ali, J., Dante, M., Du, F., Hou, S., Layman, D., Leonard, S., Nguyen, C., Scott, K., Holmes, A., Grewal, N., Mulvaney, E., Ryan, E., Sun, H., Florea, L., Miller, W., Stoneking, T., Nhan, M., Waterston, R. and Wilson, R. K. (2001). "Complete genome sequence of Salmonella enterica serovar Typhimurium LT2." Nature 413(6858): 852-856. Mears, J. A., Cannone, J. J., Stagg, S. M., Gutell, R. R., Agrawal, R. K. and Harvey, S. C. (2002). "Modeling a minimal ribosome based on comparative sequence analysis." J Mol Biol 321(2): 215-234. Merianos, H. J., Wang, J. and Moore, P. B. (2004). "The structure of a ribosomal protein S8/spc operon mRNA complex." RNA 10(6): 954- 964. Miller, J. (1992). A Short Course in Bacterial Genetics. Cold Spring Harbour, NY., Cold Spring Harbour Laboratory Press. Milon, P., Carotti, M., Konevega, A. L., Wintermeyer, W., Rodnina, M. V. and Gualerzi, C. O. (2010). "The ribosome-bound initiation factor 2 recruits initiator tRNA to the 30S initiation complex." EMBO Rep 11(4): 312-316. Mizushima, S. and Nomura, M. (1970). "Assembly mapping of 30S ribosomal proteins from E. coli." Nature 226(5252): 1214.

85 Moazed, D. and Noller, H. F. (1989). "Intermediate states in the movement of transfer RNA in the ribosome." Nature 342(6246): 142-148. Moore, P. B. and Steitz, T. A. (2002). "The involvement of RNA in ribosome function." Nature 418(6894): 229-235. Moore, P. B. and Steitz, T. A. (2010). "The Roles of RNA in the Synthesis of Protein." Cold Spring Harb Perspect Biol. Mulder, A. M., Yoshioka, C., Beck, A. H., Bunner, A. E., Milligan, R. A., Potter, C. S., Carragher, B. and Williamson, J. R. (2010). "Visualizing ribosome biogenesis: parallel assembly pathways for the 30S subunit." Science 330(6004): 673-677. Munro, J. B., Altman, R. B., O'Connor, N. and Blanchard, S. C. (2007). "Identification of two distinct hybrid state intermediates on the ribosome." Mol Cell 25(4): 505-517. Munro, J. B., Altman, R. B., Tung, C. S., Cate, J. H., Sanbonmatsu, K. Y. and Blanchard, S. C. (2010). "Spontaneous formation of the unlocked state of the ribosome is a multistep process." Proc Natl Acad Sci U S A 107(2): 709-714. Navarre, W. W., Porwollik, S., Wang, Y., McClelland, M., Rosen, H., Libby, S. J. and Fang, F. C. (2006). "Selective silencing of foreign DNA with low GC content by the H-NS protein in Salmonella." Science 313(5784): 236-238. Nierhaus, K. H. (1991). "The assembly of prokaryotic ribosomes." Biochimie 73(6): 739-755. Nierhaus, K. H. (2006). "Bacterial Ribosomes: Assembly." Encyclopedia of Life Sciences. Nilsson, A. I., Zorzet, A., Kanth, A., Dahlstrom, S., Berg, O. G. and Andersson, D. I. (2006). "Reducing the fitness cost of antibiotic resistance by amplification of initiator tRNA genes." Proc Natl Acad Sci U S A 103(18): 6976-6981. Nissen, P., Hansen, J., Ban, N., Moore, P. B. and Steitz, T. A. (2000). "The structural basis of ribosome activity in peptide bond synthesis." Science 289(5481): 920-930. Noller, H. F. (2010). "Evolution of Protein Synthesis from an RNA World." Cold Spring Harb Perspect Biol. Noller, H. F., Hoffarth, V. and Zimniak, L. (1992). "Unusual resistance of peptidyl transferase to protein extraction procedures." Science 256(5062): 1416-1419. Noller, H. F. a. N., M. (1996). Ribosomes. Escherichia coli and Salmonella: Cellular and Molecular Biology. F. C. Neidhart. Washington D.C., ASM Press. 2. Nomura, M. (1999). "Regulation of ribosome biosynthesis in Escherichia coli and Saccharomyces cerevisiae: diversity and common principles." J Bacteriol 181(22): 6857-6864. Nomura, M., Yates, J. L., Dean, D. and Post, L. E. (1980). "Feedback regulation of ribosomal protein gene expression in Escherichia coli:

86 structural homology of ribosomal RNA and ribosomal protein MRNA." Proc Natl Acad Sci U S A 77(12): 7084-7088. Nord, S., Bylund, G. O., Lovgren, J. M. and Wikstrom, P. M. (2009). "The RimP protein is important for maturation of the 30S ribosomal subunit." J Mol Biol 386(3): 742-753. Nowotny, V. and Nierhaus, K. H. (1982). "Initiator proteins for the assembly of the 50S subunit from Escherichia coli ribosomes." Proc Natl Acad Sci U S A 79(23): 7238-7242. Nowotny, V. and Nierhaus, K. H. (1988). "Assembly of the 30S subunit from Escherichia coli ribosomes occurs via two assembly domains which are initiated by S4 and S7." Biochemistry 27(18): 7051-7055. Nystrom, T. (2004). "Stationary-phase physiology." Annu Rev Microbiol 58: 161-181. O'Brien, T. W. (2002). "Evolution of a protein-rich mitochondrial ribosome: implications for human genetic disease." Gene 286(1): 73-79. O'Connor, M. and Gregory, S. T. (2011). "Inactivation of the RluD pseudouridine synthase has minimal effects on growth and ribosome function in wild-type Escherichia coli and Salmonella enterica." J Bacteriol 193(1): 154-162. Ochman, H., Lawrence, J. G. and Groisman, E. A. (2000). "Lateral gene transfer and the nature of bacterial innovation." Nature 405(6784): 299- 304. Ogle, J. M., Carter, A. P. and Ramakrishnan, V. (2003). "Insights into the decoding mechanism from recent ribosome structures." Trends Biochem Sci 28(5): 259-266. Ogle, J. M., Murphy, F. V., Tarry, M. J. and Ramakrishnan, V. (2002). "Selection of tRNA by the ribosome requires a transition from an open to a closed form." Cell 111(5): 721-732. Ohl, M. E. and Miller, S. I. (2001). "Salmonella: a model for bacterial pathogenesis." Annu Rev Med 52: 259-274. Orr, H. A. (2009). "Fitness and its role in evolutionary genetics." Nat Rev Genet 10(8): 531-539. Papp, B., Pal, C. and Hurst, L. D. (2003). "Dosage sensitivity and the evolution of gene families in yeast." Nature 424(6945): 194-197. Parsons, G. D., Donly, B. C. and Mackie, G. A. (1988). "Mutations in the leader sequence and initiation codon of the gene for ribosomal protein S20 (rpsT) affect both translational efficiency and autoregulation." J Bacteriol 170(6): 2485-2492. Parsons, G. D. and Mackie, G. A. (1983). "Expression of the gene for ribosomal protein S20: effects of gene dosage." J Bacteriol 154(1): 152- 160. Paul, B. J., Ross, W., Gaal, T. and Gourse, R. L. (2004). "rRNA transcription in Escherichia coli." Annu Rev Genet 38: 749-770. Pech, M., Karim, Z., Yamamoto, H., Kitakawa, M., Qin, Y. and Nierhaus, K. H. (2011). "Elongation factor 4 (EF4/LepA) accelerates protein

87 synthesis at increased Mg2+ concentrations." Proc Natl Acad Sci U S A 108(8): 3199-3203. Peil, L., Virumae, K. and Remme, J. (2008). "Ribosome assembly in Escherichia coli strains lacking the RNA helicase DeaD/CsdA or DbpA." FEBS J 275(15): 3772-3782. Peske, F., Rodnina, M. V. and Wintermeyer, W. (2005). "Sequence of steps in ribosome recycling as defined by kinetic analysis." Mol Cell 18(4): 403-412. Polevoda, B. and Sherman, F. (2007). "Methylation of proteins involved in translation." Mol Microbiol 65(3): 590-606. Poole, A. M., Jeffares, D. C. and Penny, D. (1998). "The path from the RNA world." J Mol Evol 46(1): 1-17. Proshkin, S., Rahmouni, A. R., Mironov, A. and Nudler, E. (2010). "Cooperation between translating ribosomes and RNA polymerase in transcription elongation." Science 328(5977): 504-508. Puglisi, J. D., Tan, R., Calnan, B. J., Frankel, A. D. and Williamson, J. R. (1992). "Conformation of the TAR RNA-arginine complex by NMR spectroscopy." Science 257(5066): 76-80. Pulk, A., Liiv, A., Peil, L., Maivali, U., Nierhaus, K. and Remme, J. (2010). "Ribosome reactivation by replacement of damaged proteins." Mol Microbiol 75(4): 801-814. Ramakrishnan, V. and White, S. W. (1998). "Ribosomal protein structures: insights into the architecture, machinery and evolution of the ribosome." Trends Biochem Sci 23(6): 208-212. Ratje, A. H., Loerke, J., Mikolajka, A., Brunner, M., Hildebrand, P. W., Starosta, A. L., Donhofer, A., Connell, S. R., Fucini, P., Mielke, T., Whitford, P. C., Onuchic, J. N., Yu, Y., Sanbonmatsu, K. Y., Hartmann, R. K., Penczek, P. A., Wilson, D. N. and Spahn, C. M. (2010). "Head swivel on the ribosome facilitates translocation by means of intra- subunit tRNA hybrid sites." Nature 468(7324): 713-716. Rivera, M. C., Jain, R., Moore, J. E. and Lake, J. A. (1998). "Genomic evidence for two functionally distinct gene classes." Proc Natl Acad Sci U S A 95(11): 6239-6244. Rodnina, M. V., Savelsbergh, A., Katunin, V. I. and Wintermeyer, W. (1997). "Hydrolysis of GTP by elongation factor G drives tRNA movement on the ribosome." Nature 385(6611): 37-41. Rodnina, M. V., Stark, H., Savelsbergh, A., Wieden, H. J., Mohr, D., Matassova, N. B., Peske, F., Daviter, T., Gualerzi, C. O. and Wintermeyer, W. (2000). "GTPases mechanisms and functions of translation factors on the ribosome." Biol Chem 381(5-6): 377-387. Rodnina, M. V. and Wintermeyer, W. (2001). "Fidelity of aminoacyl-tRNA selection on the ribosome: kinetic and structural mechanisms." Annu Rev Biochem 70: 415-435.

88 Rohl, R. and Nierhaus, K. H. (1982). "Assembly map of the large subunit (50S) of Escherichia coli ribosomes." Proc Natl Acad Sci U S A 79(3): 729-733. Ryden-Aulin, M., Shaoping, Z., Kylsten, P. and Isaksson, L. A. (1993). "Ribosome activity and modification of 16S RNA are influenced by deletion of ribosomal protein S20." Mol Microbiol 7(6): 983-992. Sander, G. (1983). "Ribosomal protein L1 from Escherichia coli. Its role in the binding of tRNA to the ribosome and in elongation factor g- dependent gtp hydrolysis." J Biol Chem 258(16): 10098-10103. Savelsbergh, A., Mohr, D., Wilden, B., Wintermeyer, W. and Rodnina, M. V. (2000). "Stimulation of the GTPase activity of translation elongation factor G by ribosomal protein L7/12." J Biol Chem 275(2): 890-894. Sawitzke, J. A., Thomason, L. C., Costantino, N., Bubunenko, M., Datta, S. and Court, D. L. (2007). "Recombineering: in vivo genetic engineering in E. coli, S. enterica, and beyond." Methods Enzymol 421: 171-199. Scheper, G. C., van der Knaap, M. S. and Proud, C. G. (2007). "Translation matters: protein synthesis defects in inherited disease." Nat Rev Genet 8(9): 711-723. Schlax, P. J. and Worhunsky, D. J. (2003). "Translational repression mechanisms in prokaryotes." Mol Microbiol 48(5): 1157-1169. Schluenzen, F., Tocilj, A., Zarivach, R., Harms, J., Gluehmann, M., Janell, D., Bashan, A., Bartels, H., Agmon, I., Franceschi, F. and Yonath, A. (2000). "Structure of functionally activated small ribosomal subunit at 3.3 angstroms resolution." Cell 102(5): 615-623. Schneider, T. D., Stormo, G. D., Gold, L. and Ehrenfeucht, A. (1986). "Information content of binding sites on nucleotide sequences." J Mol Biol 188(3): 415-431. Schuwirth, B. S., Borovinskaya, M. A., Hau, C. W., Zhang, W., Vila- Sanjurjo, A., Holton, J. M. and Cate, J. H. (2005). "Structures of the bacterial ribosome at 3.5 A resolution." Science 310(5749): 827-834. Selmer, M., Dunham, C. M., Murphy, F. V. t., Weixlbaumer, A., Petry, S., Kelley, A. C., Weir, J. R. and Ramakrishnan, V. (2006). "Structure of the 70S ribosome complexed with mRNA and tRNA." Science 313(5795): 1935-1942. Semrad, K., Green, R. and Schroeder, R. (2004). "RNA chaperone activity of large ribosomal subunit proteins from Escherichia coli." RNA 10(12): 1855-1860. Sengupta, J., Agrawal, R. K. and Frank, J. (2001). "Visualization of protein S1 within the 30S ribosomal subunit and its interaction with messenger RNA." Proc Natl Acad Sci U S A 98(21): 11991-11996. Sharan, S. K., Thomason, L. C., Kuznetsov, S. G. and Court, D. L. (2009). "Recombineering: a homologous recombination-based method of genetic engineering." Nat Protoc 4(2): 206-223. Sharma, M. R., Koc, E. C., Datta, P. P., Booth, T. M., Spremulli, L. L. and Agrawal, R. K. (2003). "Structure of the mammalian mitochondrial

89 ribosome reveals an expanded functional role for its component proteins." Cell 115(1): 97-108. Shine, J. and Dalgarno, L. (1974). "The 3'-terminal sequence of Escherichia coli 16S ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites." Proc Natl Acad Sci U S A 71(4): 1342-1346. Simonetti, A., Marzi, S., Jenner, L., Myasnikov, A., Romby, P., Yusupova, G., Klaholz, B. P. and Yusupov, M. (2009). "A structural view of translation initiation in bacteria." Cell Mol Life Sci 66(3): 423-436. Spiegel, P. C., Ermolenko, D. N. and Noller, H. F. (2007). "Elongation factor G stabilizes the hybrid-state conformation of the 70S ribosome." RNA 13(9): 1473-1482. Sprinzl, M., Brock, S., Huang, Y., Milovnik, P., Nanninga, M., Nesper- Brock, M., Rutthard, H. and Szkaradkiewicz, K. (2000). "Regulation of GTPases in the bacterial translation machinery." Biol Chem 381(5-6): 367-375. Srivastava, A. K. and Schlessinger, D. (1990). "Mechanism and regulation of bacterial ribosomal RNA processing." Annu Rev Microbiol 44: 105- 129. Steitz, T. A. (2008). "A structural understanding of the dynamic ribosome machine." Nat Rev Mol Cell Biol 9(3): 242-253. Steitz, T. A. and Moore, P. B. (2003). "RNA, the first macromolecular catalyst: the ribosome is a ribozyme." Trends Biochem Sci 28(8): 411- 418. Stoebel, D. M., Dean, A. M. and Dykhuizen, D. E. (2008). "The cost of expression of Escherichia coli lac operon proteins is in the process, not in the products." Genetics 178(3): 1653-1660. Studer, S. M. and Joseph, S. (2006). "Unfolding of mRNA secondary structure by the bacterial translation initiation complex." Mol Cell 22(1): 105-115. Subramanian, A. R. and Dabbs, E. R. (1980). "Functional studies on ribosomes lacking protein L1 from mutant Escherichia coli." Eur J Biochem 112(2): 425-430. Swords, W. E., Cannon, B. M. and Benjamin, W. H., Jr. (1997). "Avirulence of LT2 strains of Salmonella typhimurium results from a defective rpoS gene." Infect Immun 65(6): 2451-2453. Sykes, M. T., Shajani, Z., Sperling, E., Beck, A. H. and Williamson, J. R. (2010). "Quantitative proteomic analysis of ribosome assembly and turnover in vivo." J Mol Biol 403(3): 331-345. Takyar, S., Hickerson, R. P. and Noller, H. F. (2005). "mRNA helicase activity of the ribosome." Cell 120(1): 49-58. Talkington, M. W., Siuzdak, G. and Williamson, J. R. (2005). "An assembly landscape for the 30S ribosomal subunit." Nature 438(7068): 628-632. Tan, J., Jakob, U. and Bardwell, J. C. (2002). "Overexpression of two different GTPases rescues a null mutation in a heat-induced rRNA methyltransferase." J Bacteriol 184(10): 2692-2698.

90 Thomas, C. M. and Nielsen, K. M. (2005). "Mechanisms of, and barriers to, horizontal gene transfer between bacteria." Nat Rev Microbiol 3(9): 711-721. Trabuco, L. G., Schreiner, E., Eargle, J., Cornish, P., Ha, T., Luthey- Schulten, Z. and Schulten, K. (2010). "The role of L1 stalk-tRNA interaction in the ribosome elongation cycle." J Mol Biol 402(4): 741- 760. Trobro, S. and Aqvist, J. (2007). "A model for how ribosomal release factors induce peptidyl-tRNA cleavage in termination of protein synthesis." Mol Cell 27(5): 758-766. Tubulekas, I., Buckingham, R. H. and Hughes, D. (1991). "Mutant ribosomes can generate dominant kirromycin resistance." J Bacteriol 173(12): 3635-3643. Uchiumi, T., Honma, S., Nomura, T., Dabbs, E. R. and Hachimori, A. (2002). "Translation elongation by a hybrid ribosome in which proteins at the GTPase center of the Escherichia coli ribosome are replaced with rat counterparts." J Biol Chem 277(6): 3857-3862. Valle, M., Zavialov, A., Sengupta, J., Rawat, U., Ehrenberg, M. and Frank, J. (2003). "Locking and unlocking of ribosomal motions." Cell 114(1): 123-134. Varshney, U., Lee, C. P. and RajBhandary, U. L. (1993). "From elongator tRNA to initiator tRNA." Proc Natl Acad Sci U S A 90(6): 2305-2309. Weglohner, W., Junemann, R., von Knoblauch, K. and Subramanian, A. R. (1997). "Different consequences of incorporating chloroplast ribosomal proteins L12 and S18 into the bacterial ribosomes of Escherichia coli." Eur J Biochem 249(2): 383-392. Weixlbaumer, A., Jin, H., Neubauer, C., Voorhees, R. M., Petry, S., Kelley, A. C. and Ramakrishnan, V. (2008). "Insights into translational termination from the structure of RF2 bound to the ribosome." Science 322(5903): 953-956. Weixlbaumer, A., Petry, S., Dunham, C. M., Selmer, M., Kelley, A. C. and Ramakrishnan, V. (2007). "Crystal structure of the ribosome recycling factor bound to the ribosome." Nat Struct Mol Biol 14(8): 733-737. Wild, D. G. (1988). "Reversion from erythromycin dependence in Escherichia coli: strains altered in ribosomal sub-unit association and ribosome assembly." J Gen Microbiol 134(5): 1251-1263. Wilden, B., Savelsbergh, A., Rodnina, M. V. and Wintermeyer, W. (2006). "Role and timing of GTP binding and hydrolysis during EF-G- dependent tRNA translocation on the ribosome." Proc Natl Acad Sci U S A 103(37): 13670-13675. Williamson, J. R. (2009). "The ribosome at atomic resolution." Cell 139(6): 1041-1043. Wilson, D. N. and Nierhaus, K. H. (2003). "The ribosome through the looking glass." Angew Chem Int Ed Engl 42(30): 3464-3486.

91 Wilson, D. N. and Nierhaus, K. H. (2005). "Ribosomal proteins in the spotlight." Crit Rev Biochem Mol Biol 40(5): 243-267. Wilson, D. N. and Nierhaus, K. H. (2007). "The weird and wonderful world of bacterial ribosome regulation." Crit Rev Biochem Mol Biol 42(3): 187-219. Wimberly, B. T., Brodersen, D. E., Clemons, W. M., Jr., Morgan-Warren, R. J., Carter, A. P., Vonrhein, C., Hartsch, T. and Ramakrishnan, V. (2000). "Structure of the 30S ribosomal subunit." Nature 407(6802): 327-339. Wittmann, H. G. and Stoffler, G. (1975). "Alteration of ribosomal proteins in revertants of a valyl-tRNA synthetase mutant of Escherichia coli." Mol Gen Genet 141(4): 317-329. Woese, C. R. (2001). "Translation: in retrospect and prospect." RNA 7(8): 1055-1067. Woodson, S. A. (2008). "RNA folding and ribosome assembly." Curr Opin Chem Biol 12(6): 667-673. Wool, I. G., Chan, Y. L. and Gluck, A. (1995). "Structure and evolution of mammalian ribosomal proteins." Biochem Cell Biol 73(11-12): 933- 947. Wower, I. K., Wower, J. and Zimmermann, R. A. (1998). "Ribosomal protein L27 participates in both 50 S subunit assembly and the peptidyl transferase reaction." J Biol Chem 273(31): 19847-19852. Wu, X. Q. and RajBhandary, U. L. (1997). "Effect of the amino acid attached to Escherichia coli initiator tRNA on its affinity for the initiation factor IF2 and on the IF2 dependence of its binding to the ribosome." J Biol Chem 272(3): 1891-1895. Xu, Z., O'Farrell, H. C., Rife, J. P. and Culver, G. M. (2008). "A conserved rRNA methyltransferase regulates ribosome biogenesis." Nat Struct Mol Biol 15(5): 534-536. Yates, J. L. and Nomura, M. (1981). "Feedback regulation of ribosomal protein synthesis in E. coli: localization of the mRNA target sites for repressor action of ribosomal protein L1." Cell 24(1): 243-249. Yusupov, M. M., Yusupova, G. Z., Baucom, A., Lieberman, K., Earnest, T. N., Cate, J. H. and Noller, H. F. (2001). "Crystal structure of the ribosome at 5.5 A resolution." Science 292(5518): 883-896. Yusupova, G. Z., Yusupov, M. M., Cate, J. H. and Noller, H. F. (2001). "The path of messenger RNA through the ribosome." Cell 106(2): 233-241. Zavialov, A. V., Buckingham, R. H. and Ehrenberg, M. (2001). "A posttermination ribosomal complex is the guanine nucleotide exchange factor for peptide release factor RF3." Cell 107(1): 115-124. Zhang, B. and Cech, T. R. (1997). "Peptide bond formation by in vitro selected ribozymes." Nature 390(6655): 96-100. Zuckerkandl, E. (1997). "Neutral and nonneutral mutations: the creative mix--evolution of complexity in gene interaction systems." J Mol Evol 44 Suppl 1: S2-8.

92 Zuckerkandl, E. (2001). "Intrinsically driven changes in gene interaction complexity. I. Growth of regulatory complexes and increase in number of genes." J Mol Evol 53(4-5): 539-554.