ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1933

Antibiotic Resistance and theCellular Currency S-adenosyl-methionine

Modification of aminoglycosides and nucleic acids

SANDESH KANCHUGAL P

ISSN 1651-6214ISBN 978-91-513-0944-6urn:nbn:se:uu:diva-408674

Dissertation presented at Uppsala University to be publicly examined in A1:111a, Biomedicalcentrum (BMC), Husargatan 3, Uppsala, Thursday, 4 June 2020 at 09:15 for the degree ofDoctor of Philosophy. The examination will be conducted in English. Faculty examiner:Associate Professor Ditlev Brodersen (Department of Molecular Biology and Genetics -Structural Biology, Aarhus University).

AbstractKanchugal P, S. 2020. Antibiotic Resistance and the Cellular Currency S-adenosyl-methionine. Modification of aminoglycosides and nucleic acids. Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty of Science and Technology 1933. 61 pp.Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0944-6.

Streptomycin and spectinomycin are antibiotics that bind to ribosomes and inhibit proteinsynthesis. Common resistance mechanisms involve enzymatic modification of the two drugsby aminoglycoside nucleotidyltransferases (ANTs). The first part of this thesis covers thestructural mechanism of two ANT enzymes. The first is the dual-specificity AadA, belongingto the ANT(3")(9) family, which modifies the 3" position of streptomycin and position 9 ofspectinomycin. The second is ANT(9) that only modifies spectinomycin at position 9.

We solved crystal structures of both enzymes, AadA in complex with ATP and streptomycinand ANT(9) with ATP and spectinomycin. The two enzymes show overall structural similarityand both consist of an N-terminal nucleotidyltransferase domain and a C-terminal helicaldomain. The binding of ATP between the two domains induces a conformational change thatallows the drug to bind. The modified hydroxyl groups of both drugs align at similar positionsin the active site, even though the drugs are chemically distinct. Comparison of the ANT(9)and AadA structures shows that spectinomycin specificity is explained by the straight α5 helixfollowed by a short loop in ANT(9) that would clash with larger drug streptomycin. Thesefindings allowed us to explain the substrate recognition of these enzymes and propose a catalyticmechanism.

In the second and third parts of this thesis, I studied two enzymes that use S-adenosyl-methionine (SAM), RlmF in site-specific methylation of ribosomal RNA (rRNA) and Svi3-3in SAM degradation. SAM is an essential molecule for normal cellular function in all-livingcells and termed as a ‘cellular currency’. Knowledge is lacking about the substrate recognitionof rRNA methyltransferases and the role of the modifications that they add during ribosomeassembly. Here, we identify the residues of RlmF that are critical for binding of the cofactorSAM and the lithium chloride core particle substrate that mimics a 50S ribosome assemblyintermediate.In the third part, I present structural and ligand-binding studies of a newlydiscovered SAM degrading enzyme Svi3-3 from bacteriophage.

Keywords: Antibiotic resistance, streptomycin, spectinomycin, X-ray crystallography,methyltransferase

Sandesh Kanchugal P, Department of Cell and Molecular Biology, Structural Biology, Box596, Uppsala University, SE-751 24 Uppsala, Sweden.

© Sandesh Kanchugal P 2020

ISSN 1651-6214ISBN 978-91-513-0944-6urn:nbn:se:uu:diva-408674 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-408674)

To my parents Amma and Appa

List of Papers

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

I Ana Laura Stern, Sander Egbert Van der Verren, Sandesh Kanchu-gal P, Joakim Näsvall,Hugo Gutiérrez-de-Terán, and Maria Selmer. (2018) Structural mechanism of AadA, a dual-specificity aminogly-coside adenylyltransferase from Salmonella enterica. Journal of Bio-logical Chemistry. 293(29): 11481–11490

II Sandesh Kanchugal P, Maria Selmer. (2020) Structural Recogni-tion of Spectinomycin by Resistance enzyme ANT(9) from Entero-coccus faecalis. Antimicrobial Agents and Chemotherapy. 64(6):(In press)

III Sandesh Kanchugal P, Daniel Larsson, Avinash S. Punekar, Kristi-na Bäckbro, Maria Selmer. (2020) Substrate recognition by 23S RNA methyltransferase RlmF. (Manuscript)

IV Xiaohu Guo, Annika Söderholm, Sandesh Kanchugal P, Geir Isaksen, Omar Warsi, Ulrich Eckhard, Silvia Trigüis, Adolf Gogoll, Jon Jerlström-Hultqvist, Johan Åqvist, Dan I. Andersson & Maria Selmer. (2020) Structure of a phage-encoded SAM degrading en-zyme. (Manuscript)

Reprints were made with permission from the respective publishers.

Contributions Paper I. I refined and analysed one structure, performed mutagenesis, puri-fied proteins, performed the thermal shift assays and actively participated in writing the revised manuscript Paper II. I performed all experiments and took main responsibility in writing the manuscript. Paper III. I performed homology modelling, designed and prepared mutants, purified proteins and performed assays. I actively participated in writing the manuscript. Paper IV. I collected crystallography and SAXS data, solved one structure and built and refined two structures. I performed thermal shift assays and wrote parts of the manuscript. Additional Publication

I Claudia Durall, Sandesh Kanchugal P, Maria Selmer and Peter Lindblad (2020) Oligomerization and characteristics of phos-phoenolpyruvate carboxylase in Synechococcus PCC 7002. Sci-entific Reports. 10: 3607

Contents

Introduction ................................................................................................... 11

Bacterial ribosomes and translation .............................................................. 13Bacterial translation .................................................................................. 14

Ribosome biogenesis .................................................................................... 16Modifications of rRNA by methyltransferases ........................................ 17

Post-transcriptional modification of RNA ........................................... 17Ribosomal RNA modifications ........................................................... 18Presence of rRNA modifications at functional sites ............................ 18Ribosomal RNA methyltransferases ................................................... 19Rossmann-fold methyltransferases and conserved sequence motifs ... 20

S-adenosyl methionine: a cellular currency .................................................. 22S-adenosyl methionine in methyltransferases .......................................... 22Restriction modification system in bacteria and methyltransferases ....... 23

Defense against the restriction modification systems .......................... 23Discovery of novel bacteriophage SAM hydrolase enzyme ............... 24

Antibiotic resistance ...................................................................................... 25Ribosomes and antibiotics ................................................................... 25

Aminoglycosides ...................................................................................... 26Ribosomal binding sites of streptomycin and spectinomycin ............. 27

Resistance to aminoglycosides ................................................................. 28Efflux-mediated resistance .................................................................. 28Target mutations and modifications .................................................... 28Enzymatic modification of aminoglycosides or aminoglycoside-modifying enzymes (AMEs) ............................................................... 29

Mechanism of adenyltransferases ................................................................. 33

Summary of papers ....................................................................................... 35Paper I: Structural studies of a dual-specificity adenylyltransferase AadA .................................................................................................................. 35

Structure of AadA with ATP or ATP and streptomycin ..................... 35Structure-based sequence alignment ................................................... 36Basis for streptomycin recognition ...................................................... 37Docking of spectinomycin to AadA .................................................... 37

Paper II: Structural recognition of spectinomycin by resistance enzyme ANT(9) ..................................................................................................... 38

Overall structure of apo-ANT(9) ......................................................... 38The first structures of ANT(9) in complex with spectinomycin or ATP and spectinomycin ............................................................................... 38How different is the active site of ANT(9) from that of the dual-specific AadA? .................................................................................... 39Can spectinomycin analogues bind to ANT(9) enzymes? ................... 39

Paper III: Substrate recognition by RlmF, a 23S rRNA methyltransferase .................................................................................................................. 41

Homology model and design of RlmF mutants ................................... 41In vitro methylation assay and cofactor SAM binding studies ............ 42

Paper IV: Structure of SAM lyase ............................................................ 43Structure determination and SAXS studies ......................................... 43Thermal shift binding assays ............................................................... 43

Conclusions ................................................................................................... 45

Future perspectives ....................................................................................... 47

Populärvetenskaplig sammanfattning ........................................................... 48

Acknowledgements ....................................................................................... 50

References ..................................................................................................... 54

Abbreviations

rRNA ribosomal RNA mRNA messenger RNA tRNA transfer RNA ATP Adenosine-tri-phosphate WHO World Health Organisation ADP Adenosine-di-phosphate AMP Adenosine-mono-phosphate SAM S-adenosyl-methionine SAH S-adenosyl-homocysteine MTA 5′-methylthioadenosine PDB Protein Data Bank ITC Isothermal titration calorimetry SAXS Small angle X-ray scattering MD Molecular dynamics GTP Guanosine-tri-phosphate PTM Post-transcriptional modification

11

Introduction

The widespread use of antibiotics in the past century to treat bacterial infec-tions has helped us to prevent the spread of several deadly diseases. Howev-er, this has also led to the emergence of a new threat to humanity. The acqui-sition of mechanisms to counteract the antibiotics in several pathogenic strains has led to the WHO declaring antibiotic resistance as a global prob-lem needing highly organized response (WHO, 2014). A large number of antibiotics act by targeting different components of the cellular machinery that translates the genetic code into proteins, the ribosomes. The smaller subunit of the ribosome is the target of a class of antibiotics referred to as aminoglycosides. One of the mechanisms employed to acquire aminoglyco-side resistance is through enzymatic modification of aminoglycosides. Stud-ying how these enzymes recognize their antibiotic substrate and modify them would not only enable a deeper understanding of resistance mechanism but also aid in the development of next generation antimicrobial agents. In the first half of my thesis (Paper I and II), I focus on two adenyltransferases, AadA from Salmonella enterica and ANT(9) from Enterococcus faecalis. AadA is a dual-specificity enzyme that modifies both streptomycin and spec-tinomycin, while ANT(9) acts specifically on spectinomycin. Through struc-tural and biochemical studies, I explore answers to the following questions: How does AadA recognize and modify two chemically dissimilar drugs? What are the structural differences and similarities between ANT(9) and AadA? The second half of my thesis involves investigations into a cofactor that is important for various biochemical processes, S-adenosyl-methionine (SAM). The ubiquitous use of SAM in a myriad of processes has resulted in its clas-sification along with adenosine-tri-phosphate (ATP), as a ‘cellular currency’. Currencies in an economy display several properties such as limited availa-bility, durability, divisibility, uniformity, limited supply and universal ac-ceptability. Similar to currencies in an economy, SAM is a molecule whose quantities are tightly regulated, it is used by many classes of enzymes, evolu-tionarily conserved and the products derived from its metabolism are utilized as a precursor for the other reactions in the cell. In papers III and IV, I have studied two enzymes, which employ SAM to perform their function, the rRNA methyltransfease RlmF and a SAM degrading enzyme from bacterio-phage. RlmF is an enzyme involved in ribosome assembly that catalyzes the

12

methylation of the A1618 base of Escherichia coli 23S rRNA using SAM as a methyl donor. However, the role of this modification and how methyltrans-ferases recognize their substrate rRNA residues is currently unknown. In this project, we show the key residues of RlmF that are involved in the binding of cofactor SAM and substrate. We also determined the structure of LiCl core particles of 50S ribosomal subunit, the substrate of RlmF. In paper IV, we characterized a SAM degrading enzyme that was recently discovered from an unidentified bacteriophage isolated from a local pond Svandammen in Uppsala. This type of enzyme has been reported to target the anti-restriction mechanism of bacteria. Here, I was involved in structure determi-nation and ligand binding studies of this enzyme.

13

Bacterial ribosomes and translation

The ribosome is the protein synthesis machinery of all living cells. This large ribonucleoprotein particle—along with aminoacyl-transfer RNA using in-formation encoded in messenger RNA—produces the proteins in the cell. The bacterial ribosome consists of two subunits: the large 50S and the small 30S subunit. The large subunit consists of 23S rRNA (~2900 nucleotides), 5S rRNA (~120 nucleotides), and approximately 30 different proteins. The small subunit consists of 16S rRNA (~1500 nucleotides) and nearly 20 dif-ferent proteins. In addition, several translation factors interact with the ribo-some during the various stages of translation (Schmeing and Ramakrishnan, 2009). Ribosome crystallography was challenging during the early 1990s and it pushed the boundaries of the technique forever (Pennisi, 1999). The first high-resolution structures of the bacterial ribosomal subunits were published in 2000, the large subunit (Ban et al., 2000) and the small subunit (Wimberly et al., 2000). The following year, the complete Thermus thermophilus 70S ribosome structure with mRNA and tRNA at 5.5 Å resolution was published (Yusupov et al., 2001). Many other ribosome structures were solved; these structures reveal detailed interactions between tRNAs, mRNA, and riboso-mal subunits, together with translation factors during various stages of trans-lation, for example (Selmer et al., 2006; Voorhees et al., 2009). The most striking impact of the ribosome structures was that they allowed visualiza-tion of many antibiotics bound to the ribosome, helping to explain their mechanisms of action (Carter et al., 2000; Schlünzen et al., 2001; Tu et al., 2005). Notably, in 2009, the Nobel Prize in Chemistry was awarded to Ven-katraman Ramakrishnan, Tomas Steitz, and Ada Yonath for their crystallo-graphic structure and function studies of the ribosome. The basic architecture of the ribosome is shown in Figure 1. It contains three sites for tRNA binding: A, P and E. The incoming aminoacyl tRNA (aa-tRNA) binds to the A-site; the peptidyl tRNA with nascent peptide to the P-site; and deacylated tRNA to the E-site. The mRNA sits on the 30S subunit. The decoding center in 30S ribosomal subunit monitors the correct base pairing between the codon of the mRNA and anticodon of the aa-tRNA, ensuring the correct aa-tRNA is selected. The peptidyl transferase center (PTC) in the 50S ribosomal subunit catalyzes the peptide bond formation

14

between the nascent polypeptide chain on the P-site tRNA and the amino acid moiety attached to the A-site tRNA (reviewed in Schmeing and Ramakrishnan, 2009).

Figure 1. The 70S structure of ribosome in complex with mRNA and A-, P- and E- site tRNAs. The 30S ribosomal RNA is shown in sky blue and 30S ribosomal pro-teins are shown in deep blue. The 50S RNA is shown in magenta and 50S ribosomal proteins in deep salmon. The A-, P- and E- site tRNAs are shown in surface repre-sentation with green, yellow and red respectively. The mRNA is shown in orange surface representation. PDB code: 4V5D (Voorhees et al., 2009).

Bacterial translation The bacterial translation cycle can be divided into four stages namely initia-tion, elongation, termination, and ribosome recycling. The Figure 2 below shows the translation cycle in bacteria.

15

Figure 2. The translation cycle. Each cycle consists of four major stages: Initiation, elongation, release or termination and recycling. The role of individual protein trans-lation factors and tRNAs that act on ribosome during the translation cycle are (re-viewed in Schmeing and Ramakrishnan, 2009). Two antibiotics used in my study streptomycin and spectinomycin (highlighted in green) act at two different steps during elongation stage. Figure adapted from (Schmeing and Ramakrishnan, 2009)

16

Ribosome biogenesis

The production and maintenance of ribosomes is an energy- and resource-demanding process. During the exponential growth phase of bacteria, the mature ribosomes constitute nearly 30 % of the cell mass (Wilson and Nierhaus, 2008). The production of a single ribosome in a bacteria takes approximately 2 minutes (Chen et al., 2012). Thus, the bacterial transcrip-tion and translation systems must be robust to generate the ribosomes in such a short time. Moreover, the fidelity of these newly synthesized ribosomes ensures error-free protein synthesis. In order to meet these stringent require-ments, the ribosome biogenesis represents a tightly regulated process (Davis and Williamson, 2017). Initial steps of ribosome biogenesis involve the synthesis of two compo-nents: rRNA and ribosomal proteins (r-proteins). The production of rRNA involves its transcription, processing by RNases, and post-transcriptional modifications. These modifications involve the action of methyltransferases, helicases, chaperones and specific GTPases. In parallel, the r-proteins are transcribed, translated and post-translationally modified. At the next step, the rRNA and r-proteins, fold and assemble to their 50S and 30S subunits (Shajani, Sykes and Williamson, 2011; Davis and Williamson, 2017). The stoichiometry of individual components is tightly regulated by an auto-regulatory feedback system so that no one component is produced in excess (Zengel and Lindahl, 1994). Despite decades of research about ribosome biogenesis, many structural details of each protein factor and the various assembly stages are still not well established. The recent advances in cryo-electron microscopy (Cryo-EM) helped us to structurally visualize the assembly intermediates of each subunit during the process under perturbed conditions like in bL17-depletion strains (Davis and Williamson, 2017). Biophysical methods that include pulse labeling—combined with quantitative mass spectrometry—have also contributed to a better understanding of ribosome biogenesis in vivo (McCarthy, Britten and Roberts, 1962; Chen et al., 2012). The post-transcriptional modification of rRNA is the scope of my research. Therefore, I will discuss this process in more detail and focus on corresponding en-zymes involved in rRNA modification.

17

Modifications of rRNA by methyltransferases Post-transcriptional modification of RNA RNA in all living cells undergoes post-transcriptional modification (PTM), which changes the structural and functional properties of the RNA. There are more than 172 different types of PTMs of RNA according to the Modomics database (Boccaletto et al., 2018) (accessed in Feb 2020). The most diverse set of modifications is found in the tRNAs, a few dozens in the rRNA, and the most dynamic modifications are found in the mRNA (Sergiev et al., 2018). Early studies of PTMs in RNA focused on tRNA because of its abun-dance and it was easy to isolate. Changes in tRNA modifications were stud-ied upon altering growth temperature or applying other environmental stresses in bacteria (Agris, Koh and Söll, 1973). The rRNA modifications, on the other hand, were mainly studied in the context of antibiotic resistance. They were discovered in an E. coli strain that was resistant to the antibiotic kasugamycin and lacking specific modifications in the bases A1518 and A1519 of the 16S rRNA (Helser, Davies and Dahlberg, 1971). This led re-searchers to study the enzymes involved in these rRNA modifications, name-ly methyltransferases (MTases), and their role in methylation during ribo-some biogenesis (Thammana and Held, 1974; Baldridge and Contreras, 2014). All the rRNA modifications and their methyltransferases are known for E. coli, with RlmJ methyltransferase being the last one revealed (Golovina et al., 2012). Advances in next-generation sequencing methods enabled identification and characterization of modifications in different types of RNA, such as mRNA and small nuclear RNA. The enzymes involved in modifications of mRNA are so called ‘writers’ like example METTL3-METTL14-WTAP. The en-zymes recognizing these modifications are ‘readers’ like YTH and eIF3 and the enzymes those remove the modification are ‘erasers’ like ALKBH5 and FTO. Hence the mRNA modifications are dynamic, when any of these are missing, it would lead to a disease (reviewed in Meyer and Jaffrey, 2017). Meanwhile, recent cryo-EM and X-ray crystal structures have enabled visu-alization of all the rRNA modifications on the 70S ribosomes from E. coli and T. thermophilus (Noeske et al., 2015; Polikanov et al., 2015; Stojković et al., 2020) . Eukaryotic and mitochondrial ribosome modifications are also mapped, and structural comparison of all these modifications show some are conserved across all kingdoms of life (reviewed in Sergiev et al., 2018). Here I will only focus on rRNA modifications in the context of bacteria and the enzymes involved in modification.

18

Ribosomal RNA modifications There are three kinds of rRNA modifications common in bacteria. i) Formation of Pseudouridines ii) Base methylation iii) 2’O ribose methylation There are 36 modifications of rRNA identified in E. coli. They occur both in 16S and 23S rRNA. Site-specific modifications are performed by 30 differ-ent enzymes that include 23 methyltransferases and 7 pseudouridine syn-thases (Sergiev et al., 2011a; Popova and Williamson, 2014). In contrast, eukaryotic rRNA has more than 100 modifications and archaea has 88. Eu-karyotes and archaea use the RNA-guided rRNA modification system. The RNA-protein complex (RNP) consists of a guide RNA, called small nucleo-lar RNA (snoRNA) that base pairs with the substrate rRNA and a set of pro-teins; one of these protein is the rRNA modification enzyme. The RNP com-plex catalyzes the modification. The advantage of this mechanism is that it uses only a limited number of protein factors along with specific snoRNAs to modify the rRNAs (Decatur and Fournier, 2003). In contrast, bacteria use site-specific enzymes, where each enzyme modifies one or two nucleotides during the process of ribosome biogenesis (Sergiev et al., 2011a). In vivo experiments with E.coli show that rRNA modification events during ribosome biogenesis can be divided into three groups: early, intermediate and late assembly specific modifications. Seven out of 11 modifications in 16S rRNA occur at the late steps of assembly. In contrast, in 23S rRNA as-sembly, 16 out of 25 modifications occur at early steps (Siibak and Remme, 2010). These in vivo results are consistent with other study using isotope labeling combined with mass spectrometry (Popova and Williamson, 2014).

Presence of rRNA modifications at functional sites Each individual rRNA modifications may not have a direct functional role, but many modifications together could have large role (Decatur and Fournier, 2002a). Some rRNA modifications have been linked to specific functions like checkpoint during ribosome assembly or antibiotic resistance (Connolly, Rife and Culver, 2008). The modified nucleotides in rRNA are not randomly distributed. Rather, they are all concentrated in functional re-gions in the ribosomes. Examples of such regions include around the mRNA, A- and P- sites, peptidyl transferase center, peptide exit tunnel, and inter-subunit bridges (Figure 3) (Decatur and Fournier, 2002b; Sergiev et al., 2011b). A broader understanding of rRNA modifications is still lacking. In a recent study in E. coli, the knockout strains of rRNA modification enzymes shows slight growth defects and only few of them affected the ribosome

19

assembly. They also show that many knockout stains are unable to cope exogenous protein overexpression. Hence, the study concludes that rRNA modifications may have a great influence on translation and translation regu-lation (Pletnev et al., 2020).

Figure 3. The distribution of modified nucleotides in the rRNA of E. coli ribosome. The 23S rRNA (PDB code 6PJ6 (Stojković et al., 2020)) is shown as grey ribbon, the N6-methyladenosine modified nucleotides of the rRNA are shown as red spheres and other modifications in purple. The 16S rRNA (PDB code 4YBB (Noeske et al., 2015)) is shown as ribbon with head (left part of the molecule) and body (right part of the molecule) regions. The modified nucleotides are represented same as in 23S rRNA. Blue triangle (close to the A-site) and green square represent the binding sites of streptomycin and spectinomycin respectively, the two antibiotics that I focus later in this thesis.

Ribosomal RNA methyltransferases Of the three different types of rRNA modifications mentioned above, base methylation is the most common type found in E. coli. N6-methyladenosine (m6A) is the most abundant base methylation of all PTMs in RNA across all kingdoms (Sergiev et al., 2016). The m6A methylation in rRNA is carried out by site-specific methyltransferases. In E. coli, four m6A-methylated ba-ses (2 each from 16S and 23S rRNA) are modified by three different methyl-transferases. KsgA (RsmA) dimethylates the bases A1518 and A1519 in 16S rRNA. In 23S rRNA, RlmF and RlmJ mono-methylate at positions A1618 (close to peptide exit tunnel) and A2030 (close to PTC), respectively (Figure 3) (Sergiev et al., 2016).

20

RlmJ was the last methyltransferase (MTase) to be discovered in E. coli (Golovina et al., 2012). The crystal structure of the enzyme shows it consists of a Rossmann-like methyltransferase fold along with an inserted helical subdomain. RlmJ uses S-adenosyl-methionine (SAM) as a methyl donor to modify the substrate RNA. Upon binding to SAM and substrate, the N-terminal motif-X tail region of the enzyme undergoes a conformational change to enclose SAM and the substrate, which brings the substrate and active site residues closer together for catalysis. Importantly, RlmJ can mod-ify the in-vitro-transcribed 23S rRNA, as well as its smaller substrate (helix 72 of 23S rRNA). The helical subdomain is involved in substrate recogni-tion, along with the Rossmann-like MTase domain (Punekar et al., 2013). In paper III of this thesis, I will investigate the structure and substrate recogni-tion of the other 23S rRNA m6A MTase, RlmF.

Rossmann-fold methyltransferases and conserved sequence motifs Most MTases catalyze the SAM-dependent modification of several sub-strates, which includes rRNA mentioned above. The majority of these en-zymes adopts a Rossmann-fold and is grouped into the Class-1 MTase fami-ly. The Rossmann-fold MTases have an α/β architecture; they contain seven stranded β-sheets at the core and are flanked with a variable number of α-helices on both sides of the core. The fold gives rise to a characteristic αβα sandwich with the seven β-strands ordered as 3-2-1-4-5-7-6 (Figure 4), where the seventh β-strand is anti-parallel to the rest of the others. The structure-guided sequence alignment of Rossmann-fold MTases reveals nine conserved sequence motifs (motif I-VIII & X) (Malone, Blumenthal and Cheng, 1995). In this section I will discuss the four motifs (motif I-IV) crucial for SAM binding and catalysis. Motif I consists of a consensus GxGxG/A sequence, found mostly at the N-terminus on the loop after strand β1, and it is conserved in most Rossmann-fold SAM-MTases. At least one or two glycines interact with SAM. Motif II consists of strand β2 and the ad-joining turn. In Motif II, partially conserved acidic residues are common. Motif III encompasses strand β2 and a conserved acidic acid residue. Motif IV contains a sequence of conserved residues (D/N/S)-P-P- (F/Y). This motif is located in the loop after the β4-strand and is involved in SAM binding and catalysis (Figure 4) (Kozbial and Mushegian, 2005; Gana et al., 2013).

21

Figure 4. Topoplogy diagram of the Rossmann fold SAM-Mtases. The β-strands are shown as triangles and α-helices as circles, approximate location of motifs shown as M-I to IV and SAM: S-adenosyl-methionine. Figure adapted from (Schubert, Blumenthal and Cheng, 2003).

22

S-adenosyl methionine: a cellular currency

S-adenosyl methionine (SAM)—first discovered in 1952—is one of the es-sential cofactors in all cells (Cantoni, 1952). After ATP, this molecule is probably the most used cofactor (Lu, 2000). Hence it functions as a ‘cellular currency’. SAM is involved in many key biological processes, of which I mention three here. First, many enzymes uses this molecule as methyl donor in methylation of DNA, RNA, lipids, proteins, and small molecules. Second, SAM is involved in the synthesis of cysteine and glutathione through the trans-sulfuration pathway. Third, SAM is a precursor for polyamine biosyn-thesis (Mato et al., 1997). In this section, I will describe the role of SAM in MTases and in a newly discovered bacteriophage enzyme.

S-adenosyl methionine in methyltransferases SAM is a ubiquitous methyl donor that most of the MTases use to modify substrates. The SAM-dependent MTases use the methyl group attached to the positively charged sulfur atom and transfers it to either the carbon, oxy-gen or nitrogen atom of the substrate. Upon transfer of the methyl group, SAM is converted to S-adenosyl-homocysteine (SAH) (Figure 5) (Schubert, Blumenthal and Cheng, 2003).

Figure 5. SAM dependent methyltransfer reaction. The SAM dependent methyl-transferase catalyze the transfer of a methyl group from SAM to the substrate result-ing in SAH and methylated substrate.

+

+

+

+

S-adenosyl-homocysteine

Substrate-CH3

S-adenosyl-methionine

MethyltransferaseSubstrate

23

Restriction modification system in bacteria and methyltransferases In bacteria and archaea, the SAM-dependent DNA MTases and restriction endonucleases (REase) are parts of the restriction-modification (RM) sys-tems. The RM systems are part of a defense mechanism in which they dif-ferentiate their own DNA from the foreign DNA. Once recognized as for-eign, REase will digest the DNA. There are four major types of RM systems in bacteria and archaea. I will only describe the type I system, which is relevant for this section. The type I RM system consists of three subunits: REase (R), MTase (M) and site spe-cific (S) subunits. The S subunit consists of two-target recognition domains. These subunits can rearrange into two different complexes consisting of either: 2M+1S or 2R+2M+1S, if one strand of DNA is already methylated, these complexes methylate the second strand. If both strands are unmethylat-ed, the complexes cleave the DNA upstream or downstream of the recogni-tion site. In this way they overcome the threat from viruses and protect the host (Loenen et al., 2014; Liu et al., 2017).

Defense against the restriction modification systems Bacteriophages (phages) are the viruses that infect bacteria. They are diverse and found everywhere on the planet. Phages have evolved to defend against the bacterial RM systems by anti-restriction–modification (antiRM) systems. AntiRM systems work in various ways. One way is by virus-encoded DNA MTases. Another way is by stimulation of host DNA MTases, and another by unusual bases in phage DNA, like thymine bases are replaced with 5-hydroxymethyluracil (Kruger and Bickle, 1983). T3 and T7 phages have yet another antiRM system that targets the RM sys-tem. In T7 phages the gene 0.3 encodes ocr (overcome classical restriction), a protein that functions as an inhibitor for REase (Kruger and Bickle, 1983). In T3 phages, the gene 0.3 encodes for T3 SAM hydrolase (T3 SH) that hy-drolyzes the intracellular SAM. Depletion of SAM blocks the bacterial RM system, which needs this cofactor. Rudolf Hausmann first discovered this enzyme T3SH in 1967. Later, another study showed methyl 5’-thioadenosine (MTA) and homoserine to be the products of SAM hydrolysis (Gold and Hurwitz, 1964; Hausmann, 1967). However, some T3 SH mutants are unable to hydrolyze SAM although they still inhibit the type I RM system. There-fore, it is unclear if SAM hydrolase activity is necessary for the anti-restriction activity or it inhibits the RM system by binding to it (Spoerel, Herrlich and Bickle, 1979; Kruger and Bickle, 1983).

24

Discovery of novel bacteriophage SAM hydrolase enzyme In a recent study in collaboration with the research group of Prof. Dan I. Andersson (Uppsala University), three new enzymes were discovered from phage-DNA libraries. The phage DNA was constructed from environmental samples collected around Uppsala, Sweden. Three enzymes (Svi3-3, Svi3-7 and ORF1) were discovered during screening for the auxotrophic ΔilvA E. coli mutant. The ilvA gene encodes an enzyme threonine ammonia-lyase that converts L-threonine into α-ketobutyrate, which is further metabolized into isoleucine. The proteomics and transcriptomics results show that these phage enzymes up-regulate the methionine biosynthetic pathway by degrading SAM. The promiscuous activity of the MetB enzyme in the methionine bio-synthetic pathway rescues the ΔilvA strain by providing the α-ketobutyrate for isoleucine synthesis (Jerlström Hultqvist et al., 2018). One of the phage-encoded enzymes, Svi3-3, was expressed recombinantly and purified. The in vitro activity assay shows that it degrades SAM to MTA, albeit with lower activity than T3 SAM hydrolase. In paper IV, the project was aim to under-stand the structure of Svi3-3 and its ligand binding, and to further explore the reaction mechanism of the enzyme

25

Antibiotic resistance

Most antibiotics are derived from natural compounds produced by microor-ganisms that either kill or inhibit the growth of other microbes. They are widely used to treat bacterial infections (Waksman, 1947; Matzov, Bashan and Yonath, 2017). Penicillin, the first antibiotic, was discovered by Alex-ander Fleming in 1928 (Fleming, 1929). After that, a substantial number of antibiotics were discovered and the period during 1950-1960 is known as the golden age of antibiotic discovery. It revolutionized how we treat infectious diseases (reviewed in Aminov, 2010; Gould, 2016). Antibiotics with similar chemical properties and mechanism of action are classified into a class. More than 20 classes of antibiotic are in clinical use today (Coates, Halls and Hu, 2011).

Ribosomes and antibiotics The ribosome is one of the major targets for antibiotics in the bacterial cell. The antibiotics target various stages of the translation cycle and thereby in-hibit the protein synthesis. Structural and biochemical studies have helped us understand the molecular basis of how antibiotics inhibit the protein synthe-sis. However, these ribosome inhibitors have also been used as a tool to study the mechanisms of the translation cycle (reviewed in Wilson, 2014; Lin et al., 2018).

The antibiotics binding to the bacterial 70S ribosome can be divided based on which subunit they bind: 50S or 30S. The antibiotics that bind to the 30S subunit target functional centers like the decoding center, and the P and E sites. Some examples include:

• Inhibition of initiation by kasugamycin and pactamycin • Inhibition of aa-tRNA binding to the A-site by tetracycline • Increased error frequency during decoding by streptomycin and

paromomycin • Inhibition of translocation by spectinomycin and viomycin.

In contrast, the antibiotics that bind the 50S subunit target the peptidyl trans-ferase center, the peptide exit tunnel, and the association/dissociation of translation factors. Some examples include:

• Inhibition of peptide bond formation by chloramphenicol and linezolid

26

• Blocking the peptide exit tunnel by erythromycin • Inhibition of dissociation of elongation factor-G from the ribo-

some by fusidic acid

For a review of antibiotics binding to the ribosome, see (Wilson, 2014; Lin et al., 2018). In my thesis, I worked on two antibiotics, streptomycin and spectinomycin, which belongs to the aminoglycoside and aminocyclitol class respectively.

Aminoglycosides Aminoglycosides are a broad-spectrum class of ribosome-targeting antibiot-ics mainly produced by Actinomycetes. Streptomycin was the first aminogly-coside antibiotic to be discovered (1940) and was isolated from Streptomy-ces griseus. By 1944 it was in clinical use. Later, other aminoglycosides like neomycin, kanamycin, and gentamycin, were discovered (reviewed in Serio et al., 2018). The aminoglycosides are active on both Gram-positive and Gram-negative organisms. These drugs are clinically used in treatment of urinary tract infections, pneumonia, tuberculosis and other bacterial infec-tions. They were extensively used in clinic until the 1980s (reviewed in Krause et al., 2016). Aminoglycoside use gradually declined for three rea-sons. First, resistance against these drugs emerged in their clinical use. Se-cond, aminoglycosides cause toxicity, particularly nephrotoxicity and oto-toxicity. Third, new classes of antibiotics were introduced (Krause et al., 2016). However, aminoglycosides are still valuable antimicrobial agents that are being reconsidered. For example aminoglycosides are used along with another class of antibiotic polymyxins to treat Multidrug resistant (MDR) Gram-negative bacterial infections (reviewed in Durante-Mangoni et al., 2009; Serio et al., 2018). Aminoglycosides are made of amino sugars with characteristic glycosidic linkages to a dibasic aminocyclitol core. Streptomycin is a three-ring amino-glycoside with aminocyclitol O-linked to a disaccharide (Figure 6) (Waksman and Schatz, 1945). Spectinomycin is a tricyclic aminocyclitol consisting of actinamine fused to a sugar component through a glycosidic and hemiketal linkage. (Wiley, Argoudelis and Hoeksema, 1963; Lee et al., 2014) (Figure 6).

27

Figure 6. Chemical structure of antibiotics used in my study, the site of modification is shown in bold oblique character. Streptomycin consists of aminocyclitol A-ring; B- and C- rings are the amino sugars. Spectinomycin consists of diaminocyclitol moiety actinamine A-ring, fused to C-ring sugar through a B-ring.

Most aminoglycosides bind to the 16S rRNA of the 30S ribosomal subunit where they induce errors during the decoding or they block translocation during protein synthesis (Davies, Gorini and Davis, 1965). But some amino-glycosides, gentamicin and neomycin have a secondary binding site on 50S subunit that blocks ribosome recycling (Borovinskaya, Pai, et al., 2007). The first structural study of aminoglycosides (paramomycin, streptomycin, and spectinomycin—an aminocyclitol) bound to the 30S subunit was pub-lished in 2000 (Carter et al., 2000). Here I will focus on streptomycin and spectinomycin binding to ribosome in following section.

Ribosomal binding sites of streptomycin and spectinomycin At the A-site, aa-tRNA has to form Watson-Crick interactions at the first two bases in the codon. In the third or ‘wobble’ position, it is also allowed to form certain non-Watson-Crick interactions (G-U base pairs). These aatRNA are considered as cognate tRNAs. In contrast, the tRNAs that do not obey these requirements are referred to as near- and non-cognate tRNA (Ogle, Carter and Ramakrishnan, 2003). Streptomycin binds to the decoding center in the 30S subunit, where it stabilizes the ribosomal ambiguity (ram) or er-ror-prone state of ribosome (Lodmell and Dahlberg, 1997). At this state, the ribosome has higher affinity to tRNA even for near-cognate tRNA. Strepto-mycin makes contacts with helices 1, 18, 27 and 44 of the 16S rRNA (close to A-site Figure 3) and the S12 protein. Binding of streptomycin destabilizes helix 44 and its interactions. This makes the ribosome unable to discriminate between near-cognate and cognate aa-tRNA, thereby increasing the error frequency in the decoding step of translation (Figure 2) (Carter et al., 2000; Wilson, 2009; Demirci et al., 2013).

28

Spectinomycin binds to helix 34 of 16S rRNA (Figure 3), at the junction between the head and the body of the 30S ribosomal subunit. Spectinomycin binds to a partially swiveled state and restricts the movement of the head region with respect to the body thereby blocking translocation (Figure 2). (Carter et al., 2000; Borovinskaya, Shoji, et al., 2007; Mohan, Donohue and Noller, 2014)

Resistance to aminoglycosides Aminoglycoside resistance is mediated through various mechanisms. Here I discuss the most prevalent types.

Efflux-mediated resistance Efflux systems are protein transporters that pump toxic substrates from the cell to the outside environment. These protein transporters are expressed as a response to a stressor (e.g., an antibiotic). These pumps transport one or more substrates in all bacteria, but certain types of efflux pumps are associ-ated with multiple-drug resistant bacteria (Webber and Piddock, 2003). Here are three examples of efflux pumps involved in aminoglycoside resistance in Mycobacterium tuberculosis (Mtb). The Rv2333c efflux pump makes the bacteria intrinsically resistant to tetracycline and spectinomycin (Ramón-García et al., 2007). The proteins DrrAB and Rv1258cs confer resistance to several aminoglycosides including streptomycin (Aínsa et al., 1998; Choudhuri et al., 2002).

Target mutations and modifications

Chromosomal mutations in the 16S rRNA or the ribosomal proteins: When some bacteria are exposed to low or bacteriostatic amounts of amino-glycosides, because of the selective pressure they can gain resistance by mutations in their genes for rRNA and ribosomal proteins. Two studies re-port the development of this type of resistance to spectinomycin in Esche-richia coli. The spontaneous mutations occurred in the 16S rRNA genes upon exposure to spectinomycin whereupon they conferred resistance (Anderson, 1969). However, a similar observation was seen in the ribosomal protein S5, where spectinomycin-resistant mutants contained a single amino acid mutation from serine to proline (Funatsu, Nierhaus and Wittmann-Liebold, 1972). Mutations in both 16S rRNA and ribosomal protein S12 are also responsible for streptomycin resistance. The 16S rRNA mutations pre-vent streptomycin from binding, which in turn disrupts the error-prone state and leads to hyper-accurate ribosomes. Several mutations in the r-protein

29

S12 also lead to streptomycin resistance and varying degrees of hyper-accuracy during ribosomal translation (Montandon, Wagner and Stutz, 1986; Carter et al., 2000).

16S rRNA methyltransferases: Aminoglycoside-producing Actinomycetes bacteria like Streptomyces spp are inherently resistant to their own antibiotics. They harbor intrinsic 16S rRNA methyltransferases (MTases), which specifically modify the bases 1405 and 1408 of 16S rRNA. These modifications prevent their own antibiotic from inhibiting their ribosome (Cundliffe, 1989). Similar MTases are absent in pathogenic bacteria. An exception to this are certain clinical isolates that have acquired such MTases through horizontal gene transfer (reviewed in Wachino and Arakawa, 2012).

Enzymatic modification of aminoglycosides or aminoglycoside-modifying enzymes (AMEs) The most prevalent mechanism of aminoglycoside resistance is enzymatic modification, which prevents the binding of the modified antibiotic to the target. The origin of these enzymes is hypothesized to be via horizontal gene transfer from Actinomycetes. The basis of this hypothesis is that S. griseus, the streptomycin producer, expresses streptomycin 6-phosphotransferase, an enzyme that inactivates streptomycin (Shinkawa et al., 1985). However, there is another hypothesis about the origin of AMEs in Actinomycetes, where the AMEs are derived from genes that encode for enzymes involved in cellular metabolism. The mutations in metabolic genes would provide the ability to modify the aminoglycosides, thereby conferring resistance to these antibiotics (Piepersberg et al., 1988; Shaw et al., 1993). AMEs catalyze the modification of –OH or –NH2 groups of the aminocycli-tol ring, or sugar moieties of an aminoglycoside. The acetyltransferases (AACs) transfer an acetyl group from acetyl-CoA; phosphotransferases (APHs) phosphorylate using either ATP or GTP; and nucleotidyltransferases (ANTs) adenylate using ATP. These three classes of enzymes modify the same or different positions in the aminoglycosides, see a review in Ramirez and Tolmasky, 2010). Here I will discuss the enzymes related to streptomy-cin and spectinomycin modification.

Aminoglycoside N-acetyltransferases (AACs) The AAC class comprises the largest family of AMEs. They acetylate vari-ous aminoglycosides except streptomycin and spectinomycin. However, the bifunctional enzyme ANT(3′′)-Ii/AAC(6′)-IId from Serratia marcescens acetylates the 6′-amine of kanamycin and adenylates the 3′′- and 9-hydroxyl

30

of streptomycin and spectinomycin, respectively. This fused enzyme consists of two domains, one with adenyltransferase and the other with acetyltrans-ferase activity (Kim et al., 2006).

Aminoglycoside O-phosphotransferases (APHs) The APHs are the second largest family of AMEs after the AACs. These enzymes are structurally and functionally similar to kinases and catalyze ATP- or GTP-dependent phosphorylation at different positions in aminogly-cosides (Ramirez and Tolmasky, 2010; Shi and Berghuis, 2012). They are a structurally diverse group of enzymes, they confer resistance in many spe-cies of bacteria by modifying different positions in the aminoglycosides (Ramirez and Tolmasky, 2010). The ATP-dependent phosphorylation occurs at the 6" position of streptomycin with APH(6), 3″ position of streptomycin with APH(3″) and 9 position of spectinomycin with APH(9). There are four different enzymes in the APH(6) class, encoded by genes aph(6)-Ia to Id, all of which confer resistance to streptomycin. They were first discovered in chromosomes of Streptomyces griseus and then later in transposons (Ramirez and Tolmasky, 2010). In the APH(9) group, APH(9)-Ia enzyme from Legionella pneumophila is well-characterized structurally and functionally. It binds and modify spec-tinomycin and no other aminoglycosides (Thompson et al., 1998; Fong et al., 2010). The three crystal structures of APH(9)-Ia—the apo form, the complex with AMP, and the complex with ADP and spectinomycin—all have overall structures similar to the rest of the APH family members except for the binding site of the substrate. The enzyme changes conformation—from open to closed—upon binding to the ATP. In contrast, binding of spec-tinomycin induces no conformational change. One magnesium is bound be-tween the β-phosphate and position 9 of spectinomycin in the ternary com-plex of APH(9)-Ia. The spectinomycin in the active site has five different interactions with residues of APH(9)Ia on the three rings of the molecule. This shows that APH(9)-Ia has a stringent ligand recognition not found in the other APH family members. Mutagenesis studies confirm that the Asp-212 that interacts with the position 9 of spectinomycin is the catalytic resi-due for phosphorylation (Fong et al., 2010).

Aminoglycoside O-nucleotidyltransferases (ANTs): The ANTs are the least-studied, and fewest in number, of the enzymes in the AMEs. These enzymes transfer the AMP group from ATP to the hydroxyl group of aminoglycosides. The five classes of ANTs are classified on the basis of the modification positions, namely ANT(6), ANT(9), ANT(4′), ANT(2′′), and ANT(3′′). Among these, ANT(6) confers resistance only to streptomycin, and ANT(9) to spectinomycin (Ramirez and Tolmasky, 2010).

31

ANT(3′′)—which is also called AAD(3′′)(9)—modifies both streptomycin and spectinomycin (Hollingshead and Vapnek, 1985). The first crystal structure of AadA belonging to the ANT(3′′)-Ia family was solved from Salmonella enterica in our research group (Chen et al., 2015). AadA adenylates the 3′′ hydroxyl of the glucosamine ring of streptomycin and the 9 hydroxyl of the actinamine ring of spectinomycin (Figure 6). AadA consists of an N-terminal nucleotidyltransferase domain and a C-terminal α-helical domain. Small angle X-ray scattering (SAXS) experiments confirm that AadA is monomer in solution, as it is in the crystal structure. The ITC binding experiments shows that AadA binds to ATP at micromolar affinity and has to bind before the antibiotic. Our group used sequence analysis of AadA homologues to identify the residues involved in ligand binding and catalysis. Site directed mutagenesis experiments, combined with ITC bind-ing experiments, identified the residues critical for ATP and antibiotic (strep-tomycin and spectinomycin) binding. Meanwhile, an in vivo minimum inhib-itory concentration (MIC) study with wild type or mutant AadA showed that Glu-87 is the catalytic residue. The same study showed the critical residues that are involved in streptomycin and spectinomycin resistance (Chen et al., 2015). In my thesis, paper-I builds on the study of Chen et al.

Pathogens and antibiotic resistance Salmonella enterica are Gram-negative bacteria belong to Enterobacteriacae family. The food borne pathogen causes infections in humans and has a large impact on public health. S. enterica are ubiquitously present in all kinds of food, ranging from fresh foods to food products from animals. Due to exten-sive use of antibiotics for veterinary purposes and in farm feeds, there is a rise in bacterial antibiotic resistance. This has contributed to S. enterica de-veloping resistance to most antibiotics, and to outbreaks in the past decade (Nair, Venkitanarayanan and Johny, 2018). S. enterica isolates from out-breaks all over the world show the presence of aminoglycoside resistance genes that contain most of AMEs like AAC, APH and ANTs (reviewed in Alcaine, Warnick and Wiedmann, 2007). The study of S. enterica and its resistance enzymes will help in understanding the response to new antibiot-ics and to tackle future outbreaks. Enterococcus faecalis belongs to the genus Enterococci. These are Gram-positive bacteria that can survive harsh conditions and antimicrobial envi-ronments. Most of the clinical isolates of Enterococcus spp. are resistant to most antibiotics, with some resistance being acquired through horizontal gene transfer (García-Solache and Rice, 2019). Enterococcus spp. were first discovered in 1899 and originally believed to belong to the genus Strepto-coccus. It was only after detailed biochemical and culturing analysis in the

32

1980s that they were classified as a separate genus (Schleifer and Kilpper-Balz, 1984). Most of the hospital multi-drug resistance cases are caused by Enterococcus; that includes urinary tract infection, burn or surgical site wound infections and infections of implanted medical devices, hence these bacteria represent an imminent health threat to humans. Studying these spe-cies will help us understand their adaptability in healthcare environment, antibiotic resistance mechanisms and virulence (García-Solache and Rice, 2019).

33

Mechanism of adenyltransferases

Most adenyltransferase enzymes consist of two domains: a conserved N-terminal catalytic nucleotidyltransferase domain and a C-terminal, mostly α-helical, domain. All known ANTs—including ANT(4′)-Ia, ANT(4′′)-Ib, ANT(6)-Ia, ANT(3′′)9 and ANT(9)—have nucleotidyltransferase domains The adenyltransferases LinB and Lnu(A), also have nucleotidyltransferase domains, although they are active against different class of antibiotics (Cox et al., 2015). Yet another enzyme, DNA polymeraseβ, also has an N-terminal nucleotidyltransferase domain and similar architecture of the cata-lytic site. The conserved N-terminal nucleotidyltransferase domain allows these enzymes to catalyze similar reactions. The more variable C-terminal domain allows them to recognize different substrates. These observations support a divergent evolutionary relationship between adenyltransferases and DNA polymerases (Figure 7) (Cox et al., 2015; Stern et al., 2018).

Figure 7. The conserved nucleotidyltransferase (NT) fold in ANT enzymes (AadA: 4CSB, KNTase: 1KNY, ANT(2′′)-Ia: 4WQL) and DNA polymeraseβ ( DNA-polβ: 2FMS). The nucleotidyltransferase domain is shown in blue and variable C-terminal domain in red.

All adenyltransferases have similar active site architecture: three acidic resi-dues (Asp and Glu), two magnesium ions, a nucleotide, and a hydroxyl group from the substrate antibiotic (Cox et al., 2015). Until now, most of the ANTs that have been characterized are those that modify the aminoglyco-sides containing a 2-deoxystreptamine core (e.g., kanamycin, gentamycin,

34

tobramycin). In contrast, streptomycin has a streptidine core and spectino-mycin have a streptamine core (Fong et al., 2010). Adenylyltransferase ANT(2′′)-Ia from Klebsiella pneumoniae, confers re-sistance by modification of the 2′′ hydroxyl of kanamycin, gentamycin and tobramycin. Crystal structures of ANT(2′′)-Ia in its apo form and in complex with the drug kanamycin reveal the structural similarity with other adenyl-transferases and the conserved active site residues. A mutagenesis study confirmed Asp-86 to be the catalytic base. The active site contains two mag-nesium ions (MgI and MgII). The MgI

coordinates the α-phosphate of ATP and the MgII coordinates the β and γ phosphates of ATP and the 2′′-OH of the substrate. The proposed catalytic mechanism is that Asp-86 abstracts the proton from the 2′′-OH of the substrate, leading to a subsequent nucleophilic attack on the α-phosphate of ATP (Cox et al., 2015). Another well-studied adenyltransferase enzyme is the ANT(4′) kanamycin nucleotidyltransferase (KNTase) from Staphylococcus aureus. The ternary crystal structure of KNTase, bound with the ATP analogue AMPCPP and kanamycin, shows the protein to be a homodimer. They bind both substrates at the interface; both monomers contribute to both binding sites (Pedersen, Benning and Holden, 1995). The monomer consists of two domains: an N-terminal nucleotidyltransferase domain and a C-terminal α-helical domain. The two monomers wrap around each other to create a cavity and residues from both subunits form interactions with substrates. There are few contacts with the adenine ring, so the enzyme might also bind nucleotides other than ATP. The KNTase structure has a single magnesium ion that coordinates the β and γ phosphates of ATP, Glu-52 and Asp-50. Glu-145 from the other subunit is proposed to be a catalytic base because it is within hydrogen-bonding distance to the 4′-OH group of the aminoglycoside. The proposed mechanism of adenylylation is deprotonation of 4′-OH by Glu-145. This leads to a nucleophilic attack on the α-phosphate, thereby forming a nucleo-tidyl aminoglycoside, and subsequently releasing the pyrophosphate. The nucleotide and aminoglycoside in the active site are oriented for a single, in-line displacement reaction (Pedersen, Benning and Holden, 1995).

35

Summary of papers

Paper I: Structural studies of a dual-specificity adenylyltransferase AadA In this study the aim was to understand how AadA recognizes and modifies the two chemically distinct antibiotics, streptomycin and spectinomycin.

Structure of AadA with ATP or ATP and streptomycin

AadA structure with ATP and magnesium Cocrystallization of AadA with ATP and magnesium gave crystals with spacegroup P32 and two molecules in the asymmetric unit. Upon binding of ATP and magnesium, the C-terminal domain of AadA rotated 11° towards the N-terminal domain. The electron density for ATP and two Mg was seen deep in the interdomain cleft. The residues that interact with ATP and mag-nesium are conserved in other adenylyltransferases. In our structure, Asp-47, Asp-49 and Glu-87, along with phosphates of ATP, coordinate the two mag-nesium ions. Ser-36, Ser-46 and Tyr-231 make hydrogen bonds to the γ phosphate of ATP. Lys-205 and Arg-192 also make hydrogen bonds with the phosphates. The hydroxyl group of both ribose and adenine makes hydrogen bond with Asp-130 and Thr-196, respectively. AadA complex structure with ATP and streptomycin or dihydrostreptomycin Two structures were solved with crystals of the catalytically inactive mutant E87Q of AadA and ATP, in complex with streptomycin or dihydrostrepto-mycin. The overall structures of these two complexes are similar to AadA-ATP with an RMSD of 0.3 Å over 252 Cα atoms (262 amino acids in AadA-ATP structure). A complex structure of AadA, ATP, magnesium and streptomycin was pre-pared by soaking of the antibiotic into AadA-ATP crystals. The resulting 2.05-Å-structure shows electron density for all the ligands. Streptomycin binds to AadA at the entry of the interdomain cleft and is coordinated with both N- and C-terminal residues. The C-ring of streptomycin is closer to ATP and magnesium, whereas the B- and A-rings (paper I, Figure 3) extend out from the active site and towards the C-terminus. The catalytic base Glu-

36

87 is around 3.5 Å away from 3′′-OH of the streptomycin C-ring. The 6′′-OH position of the same ring forms a hydrogen bond with Asp-182 and stacking interactions with Trp-112. The B-ring of streptomycin has a 3′-OH group that forms a hydrogen bond with His-185, but overall fewer interactions than the C-ring. The aminocyclitol A-ring has multiple interactions with the C-terminus of AadA, including hydrogen bond with Trp-173, and also to the backbone carbonyl of Ala-177 and Asp-178.

The catalytic base Previous sequence analysis and minimum inhibitory concentration (MIC) experiments show that Glu-87 is the catalytic base in AadA (Chen et al., 2015). An in vitro activity assay with malachite green shows that both WT AadA and the mutant E87Q was active on streptomycin. In the enzymatic assay, the pyrophosphatase converts the pyrophosphate by-product of the reaction to phosphate that is detected with malachite green assay. In order to check if the phosphate production by the mutant was caused by streptomycin adenylylation, a chromatographic assay was performed. It showed that E87Q could not adenylate the streptomycin. This suggests that in absence of a cata-lytic base, water can act as a nucleophile and make AadA an ATP hydrolase enzyme.

Structure-based sequence alignment According to the Pfam database and the literature, 197 sequences were pre-dicted to have the same two domains as AadA. Of these, eight enzymes have been experimentally characterized including AadA. Based on structure-function analysis, we divided them into three classes:

• The enzymes active on both streptomycin and spectinomycin (the ANT(3′′)(9) enzymes);

• Those active only on spectinomycin (ANT(9) enzymes) • Those active on neither antibiotic.

The structure-based sequence alignment was performed with AadA and sev-en other enzymes (paper-I, Figure 5). The alignment shows that residues in contact with A-ring of streptomycin are conserved in ANT(3′′)(9) but not in ANT(9) enzymes. The ANT(3′′)(9) enzymes have Trp-173 and a two-amino acid insert at position 177-178 that forms backbone carbonyl interactions; these residues are involved in recognition of the A-ring of streptomycin. Based on this, we propose that Glu-87, Trp-112, Trp-173 and two-amino acid insert are necessary for the streptomycin activity of ANT(3′′)(9). Glu-87, Trp-112, Asp-182 and His/Asn-185 are important for spectinomycin activity. To test this hypothesis, I performed in vitro thermal shift assays on

37

AadA mutants. Meanwhile, our collaborator Joakim Näsvall (Uppsala Uni-versity) performed in vivo MIC experiments for the respective chromosomal mutants of the aadA gene.

Basis for streptomycin recognition A thermal shift assay was used to show the stabilization effects of ligand binding to AadA and its mutants W173A and D178A. Upon titration of lig-ands to WT AadA, the melting temperature (Tm) increases up to 8°C for streptomycin and 2°C for spectinomycin. W173A and D178A mutants have lower Tm compared to the WT AadA, showing their destabilizing effect. Furthermore, stabilization by streptomycin only happens for both mutants at higher concentration of streptomycin, indicative of a lower binding affinity (Figure 8). In contrast, the mutants titrated with spectinomycin have the same Tm as the WT AadA (Figure 8). Similar results were observed from the MIC values for W173A and D178A mutants. The streptomycin MIC values were reduced 10- and 5- fold for W173A and D178A mutants respectively compared to WT AadA . But for spectinomycin, MIC values for mutants were same as the WT AadA (Paper-I,Table-1 ).

Figure 8. Thermal stabilization (ΔTm) of WT and mutant (D178A & W173A) AadA in the presence of streptomycin (A) and spectinomycin (B).

Docking of spectinomycin to AadA Despite tremendous efforts, I did not succeed in obtaining a complex struc-ture with spectinomycin in the active site of AadA. Thus, spectinomycin was manually docked into WT AadA and E87Q, based on the hypothesis that the modified hydroxyl groups of both substrates would be aligned similarly in the active site. This docked structure was subjected to 40 ns MD simulations, which was done by our collaborator Hugo Gutierrez-de-Teran (Uppsala Uni-versity). The simulations showed the proposed binding mode was stable for WT AadA, while in the E87Q mutant spectinomycin moves away from the active site in less than 10 ns. This result corroborated previous ITC data showing that spectinomycin does not bind to the E87Q mutant (Chen et al., 2015).

38

Paper II: Structural recognition of spectinomycin by resistance enzyme ANT(9) ANT(9) enzymes that confer spectinomycin-specific resistance, have been previously characterized from Enterococcus faecalis, Staphylococcus aureus and Campylobacter jejuni (Murphy, 1985; LeBlanc, Lee and Inamine, 1991; Nirdnoy, Mason and Guerry, 2005). They share 29-37 % sequence identity with the dual specificity enzyme AadA that is active on streptomycin and spectinomycin. Our main research question here was: How does ANT(9) recognize spectinomycin and how does it differ from the dual specificity AadA? To find out, I expressed and purified all three ANT(9) enzymes, of which the E. faecalis (strain LDR55) (LeBlanc, Lee and Inamine, 1991) was successfully crystallized. This paper is about the structural studies of this enzyme in its apo-form and in complex with spectinomycin and ATP.

Overall structure of apo-ANT(9) E. faecalis ANT(9) consists of 263 amino acids and eluted as a monomer on the size-exclusion column. The protein was very unstable and gradually pre-cipitated, so, I immediately put it into crystallization trials or flash froze it in liquid nitrogen for later use. Initial crystallization trials produced thin, nee-dle-shaped crystals at room temperature in many conditions of the Morpheus crystallization screen (Gorrec, 2009). These needle-shaped crystals had poor diffraction. Nevertheless, similar crystallization experiments at 8°C resulted in hexagonal, prism-shaped crystals that diffracted to 1.9 Å resolution. They belong to spacegroup I4 and have one molecule in the asymmetric unit. The structure was solved by molecular replacement using the both domains of AadA-ATP structure as a template. The initial model of ANT(9) consisted of residues 8-154 and 166-255. The missing residues were added with Phenix Autobuild (Terwilliger et al., 2007) and later, by manual building with COOT (Emsley et al., 2010). Residues 195-205 and 226-239 were challeng-ing to fit into the model because of the weak electron density, which is re-flected by higher B-factors in this region (100 vs 68 for rest of protein). ANT(9) consists of two domains: an N-terminal nucleotidyltransferase do-main and a C-terminal α-helical domain.

The first structures of ANT(9) in complex with spectinomycin or ATP and spectinomycin Pre-grown apo-ANT(9) crystals were soaked with ATP, Mg, and spectino-mycin and resulting in two complex structures—one at pH 8.5 with only spectinomycin and the other at pH 7.5 with ATP, Mg, and spectinomycin.

39

The ANT(9) with spectinomycin (ANT(9)-spc) structure was refined to 2.8 Å resolution and shows strong difference density (Fo-Fc) for spectinomycin (Paper II Figure 2B). The ANT(9) structure with ATP, Mg and spectinomy-cin (ANT(9)-ATP-spc) at 3 Å resolution had a different space group, C121, and had two molecules in the asymmetric unit. The two chains of the mole-cules are nearly identical to each other, with an RMSD of 0.43 Å over 248 Cα-atoms, but have small differences between the molecules.

How different is the active site of ANT(9) from that of the dual-specific AadA? ATP binds deep in the interdomain cleft of ANT(9). The ATP phosphate interacts with Ser-35, Ser-45, Asp-46, Asp-48, Arg-190, Lys-203 and Arg-190. Asp-129 hydrogen bonds with the ATP ribose and Arg-190 form a cati-on–pi interaction with the adenine (Paper II Figure 2A). These ATP interac-tion residues are conserved between ANT(9) and AadA. We observed two magnesium ions (MgA and MgB) in chain A and one (MgB) in chain B. The tricyclic spectinomycin binds in the entry of the interdomain cleft. This L-shaped molecule has two co-planar rings (A & B) and a third one (C) that inserts into a hydrophobic pocket of ANT(9) (Paper II Figure 1D). The 9-OH of the A-ring in spectinomycin is close to the proposed catalytic base, Glu-86, and stacks against Trp-111. Asp-180 and Asn-183 are in hydrogen-bond distance to the 4a-OH in the C-ring of spectinomycin. Structural superposition of ANT(9) and AadA shows they are almost identi-cal at the N-terminal nucleotidyltransferase domain. The differences we no-tice are in the α-helical domain, between the α5-α6 helices (residue 160-185 region). The straight α5 helix and the short loop between α5-α6 means that ANT(9) would clash with a larger substrate like streptomycin (Paper II Fig-ure 3C). Spectinomycin is a rigid aminocyclic compound that makes fewer interactions in ANT(9). The larger streptomycin makes more interactions at different parts of the active site in AadA.

Can spectinomycin analogues bind to ANT(9) enzymes? Almost as soon as spectinomycin was discovered and put into clinical use, resistance became a problem. Development of spectinomycin analogues has therefore been going on since 1980. Recently, spectinamides and aminome-thyl spectinomycins (Lee et al., 2014; Bruhn et al., 2015; Butler et al., 2018) were developed. These new molecules have modifications at the 4-O group in the C-ring of spectinomycin to increase its binding on the ribosome and to escape the intrinsic efflux pump of M. tuberculosis. As seen in our structures

40

of ANT(9) in complex with spectinomycin, the 4-O group in the C-ring of the spectinomycin extends outside the active site (Figure 9). Hence, the mod-ified spectinomycin is still at risk of being modified by both the ANT(9) and ANT(3")(9) class of enzymes.

Figure 9. Structure of ANT(9) with ATP and spectinomycin (6XXO). Spectinomy-cin shown in cyan and ATP in magenta stick, two magnesium ions (MgA and MgB) are shown as green sphere. A black asterisk(*) shows the 4-O group in the C-ring of spectinomycin that extends away from the active site. The Fo-Fc omit maps shown for spectinomycin, ATP and magnesium ions contoured to 2.5σ.

41

Paper III: Substrate recognition by RlmF, a 23S rRNA methyltransferase The aim of this study was to understand how RlmF recognizes its substrate, A1618 of 23S rRNA, and to investigate the binding pocket of the cofactor SAM. Furthermore, we wanted to determine the cryo-EM structure of the 50S LiCl core particles, the RlmF substrate, to study how RlmF recognize its substrate.

Homology model and design of RlmF mutants A BLAST search of the E. coli RlmF sequence against the PDB resulted in a hit, the catalytic domain of human methyltransferase METTL16 (the struc-tures of METTL16 with RNA hairpin substrates PDB code 6DU4 (Doxtader et al., 2018) or the structures of METTL16 without RNA, PDB code 6B92 (Brown et al., 2016; Ruszkowska et al., 2018)). The 291-residue N-terminal METTL16 contains a Rossmann fold methyltransfease domain. RlmF and METTL16 share 28 % sequence identity over 297 residues. The multiple sequence alignment with two other representatives from bacterial RlmF se-quences shows the conserved sequences and motifs (Paper III Figure 1). The two conserved motifs we noticed from the multiple sequence alignment are motif I GXG (G111 – G113 in RlmF) and the motif IV NPPF. We performed the homology modeling using I-TASSER (Yang and Zhang, 2015) with the PDB entry 6DU4 (METTL16 in complex with MAT2A 3' UTR hairpin RNA) as a template. The homology model contains all 308 amino acids of RlmF. Even though METTL16 recognizes a different kind of RNA than RlmF, the modification site is N6 of adenosine. Based on complex structures of METTL16, we designed the mutants in RlmF to investigate the binding sites for cofactor and substrate. We compared the RlmF homology model with the METTL16 structure in complex with SAH (PDB code 6B92), to investigate the SAM binding pocket in RlmF. Based on the analysis we designed four mutants: Arg83Ala, to probe the interaction with the carboxyl group of SAM; Glu134Ala to probe the ribose hydroxyl interaction; Asn185Ala and Asn185Asp, to probe the interactions with the substrate adenine. Similarly, we compared the METTL16-RNA (MAT2A 3' UTR hairpin RNA; PDB code 6DU4) to investigate the substrate binding to RlmF. From this analysis, we designed two deletion mutants of RlmF: Δ9 (deleted 193-201 amino ac-ids) and Δ26 (deleted 193-218 amino acid). We also designed a double mu-tant, Lys285Ala/Arg288Ala, close to the C-terminus of RlmF, to remove the positive charge in this beta hairpin.

42

In vitro methylation assay and cofactor SAM binding studies RlmF can modify LiCl core particles but not naked 23S rRNA or 50S subu-nits (Sergiev et al., 2008). Based on this, we prepared 50S LiCl core particles from an RlmF knockout strain. We performed an in vitro methylation assay using tritium-labeled SAM for wild type RlmF and the mutants with LiCl core particles as substrate. Wild type RlmF showed methylation activity on this substrate and saturation after 10 minutes (Paper III, Figure 3A). The mutants Arg83Ala, Lys285Ala/Arg288Ala, and the Δ26 deletion mutant showed no or severely reduced methylation activity, while Glu134Ala, Asn185Asp, Asn185Ala and the Δ9 deletion mutant showed 2.5-6-fold re-duced activity compared to wild-type RlmF (Paper III, Figure 3B). A ther-mal shift assay was performed to assess the binding of cofactor SAM to wild type RlmF and mutants. For wild type RlmF, the SAM and SAH binding increased the melting temperature (Tm) by 3 and 5 degrees respectively. Without ligands, all RlmF mutants had higher thermal stability (Tm) than wild type. Interestingly, the Δ26 deletion mutant binds to SAM and SAH, but does not show methylation activity. This shows that the 190-220 region is crucial for substrate binding. All other active site mutants of RlmF did not show any temperature shift in presence of SAM or SAH.

43

Paper IV: Structure of SAM lyase Structure determination and SAXS studies My colleagues performed the expression and purification of full-length and an N-terminal-truncated variant of Svi3-3. The crystals formed only for the truncated and TEV-cleaved Svi3-3, in the presence of SAM or SAH. These co-crystals diffracted to 1.5-Å resolutions, in space group F4132. I contribut-ed in data collection of these crystals, solving the structures, SEC-SAXS data collection and analysis and thermal shift assay. Several attempts to solve the structure either by MAD or SAD phasing methods failed. The reso-lution of data was high enough resolution to try ab initio structure-solution methods, and solved using Arcimboldo (Rodríguez et al., 2009). We could build 130 out of 146 amino acids of the Svi3-3 structure; the last 16 residues of the C-terminal end were disordered. The structure of Svi3-3 with product MTA or substrate analogue is described in Paper IV. The small angle X-ray scattering (SAXS) experiment was performed for Svi3-3 with and without the substrate SAM. The results show that in both cases Svi3-3 is a trimer in solution and it does not need to dissociate to bind the substrate and release the product.

Thermal shift binding assays To study the ligand binding to Svi3-3, a thermal shift assay was performed. The change in melting temperature Tm of protein at different ligand concen-tration allowed us to obtain the binding curve. For Svi3-3, we could see up to 30°C increase in melting temperature in presence of substrate SAM, prod-uct MTA or substrate analogue SAH. The binding curve for all these ligands look close to identical (Figure 10), however during the experiment the SAM is degraded to MTA and stabiliza-tion is by binding of MTA. We tested the binding of other ligands such as homoserine-lactone, homoserine, ATP, ADP and AMP to Svi3-3, but none of these led to a shift in Tm. Interestingly, the E69Q mutant of Svi3-3 binds to SAM similarly as wild-type Svi3-3 but not to the substrate analogue SAH. A possible explanation for this is that SAH binds to Svi3-3 slightly differently than SAM. The con-formation observed in the SAH crystal structure would not fully represent the binding of SAM to Svi3-3.

44

Figure 10. Thermal shift assay data for WT and mutant E69Q Svi3-3 with different ligands. WT Svi3-3 with SAM (orange square), WT Svi3-3 with SAH (purple down-pointing-triangles), E69Q with SAM (blue circles) and E69Q with SAH (cyan up-pointing-triangles).

45

Conclusions

Papers I and II of this thesis provide insights into the antibiotic resistance mechanism of aminoglycoside modifying enzymes. The crystal structure of the aminoglycoside adenyltransferase AadA with ATP and magnesium shows that ATP binds deep in the inter-domain cleft. Upon binding to ATP, the C-terminal domain of AadA moves 11° towards the N-terminal domain. The structure of AadA with ATP and streptomycin reveals that streptomycin binds at the entry of interdomain cleft and forms interactions with both N- and C-terminal residues. Site-directed mutagenesis experiments, combined with in-vivo resistance and in vitro binding studies, revealed the residues critical for streptomycin recognition in AadA. Also, we could propose a catalytic mechanism—adenylation by AadA that uses Glu-87 as the catalytic base. ANT(9) from E. faecalis shows a similar overall structure to AadA, with an N-terminal nucleotidyltransferase domain and a C-terminal α-helical do-main. We obtained the first crystal structures of ANT(9) in complex with spectinomycin, and in complex with ATP and spectinomycin. The binding of ligands to the ANT(9) structure induces a change in the conformation of the C-terminal domain towards N-terminal domain. The 9- hydroxyl of spec-tinomycin that gets modified forms a very close hydrogen bond with the proposed catalytic base Glu-86. Structural superposition of ANT(9) and AadA complex structures reveal that the straight α5 helix and the short loop between α5-α6, streptomycin would clash with the enzyme (paper II, Figure 3C). This explains how AadA can accommodate and modify two chemically dissimilar antibiotics, while ANT(9) is specific for only spectinomycin. The modified hydroxyl groups of both antibiotics align at similar positions in the active site despite the differences in size and chemical structure of the two antibiotics. The active-site arrangements observed in ANT(9) is in agree-ment with reaction mechanism proposed for other aminoglycoside nucleo-tidyl transferase (ANT) enzymes. In Paper III, we studied recognition of cofactor SAM and substrate binding by E. coli RlmF. Multiple sequence alignment of RlmF sequence with hu-man METTL16 and close bacterial RlmF sequences shows conserved motifs for SAM binding (motif I) and catalysis (motif IV). We built a homology

46

model of RlmF using the human METTL16 methyltransferase structure as template. Then we made predictions for binding of SAM and RNA substrate to RlmF based on complex structures of hMETTL16 with SAH and RNA substrate. Based on these predictions we designed mutants of RlmF and per-formed in vitro methylation and thermal shift assays. The results show that the 190-220 region and Lys285 and Arg288 are critical for substrate recogni-tion. RlmF is slightly faster compared to other rRNA methyltransferases at in vitro methylation conditions. We used single particle cryo-EM to deter-mine structures of the RlmF substrate, 50S LiCl core particles. The struc-tures reveal that some of the sub-particles have an accessible and unfolded modification site. Paper IV presents that the newly discovered SAM-degrading enzyme, Svi3-3, is a trimer in solution. This was confirmed by SAXS experiments. Here we report the first crystal structure of a phage-encoded SAM-degrading en-zyme. Structures of apo Svi3-3, and in complex with its product MTA or substrate analogue SAH, show that it consists of a ferredoxin-like fold and has the active site at the subunit interface. The two structures with MTA and SAH are nearly identical. Importantly, the thermal shift assay demonstrated significant stabilization in the presence of MTA, SAM or SAH. However, the catalytically inactivated mutant of Svi3-3 does not bind to SAH any-more, while binds to MTA or SAM is like in the wild type protein.

47

Future perspectives

In most ANT enzymes that modify different aminoglycosides, the N-terminal nucleotidyltransferase domain and catalytic residues are conserved (Figure 7). In Papers I and II, we show that the structural differences in heli-cal C-terminal domain allow recognition of two different antibiotics. This leads to the possibility that the ANT enzymes may evolve their C-terminal domain to modify new antibiotics. Hence, design of any new antimicrobials should involve mapping of the active site contacts of these resistance en-zymes to the drug. This strategy could potentially prevent modification of the drug by resistance enzymes, while maintaining interactions with the tar-get (such as the ribosome).

In future studies for paper III; experiments to measure the binding affinity of LiCl core particles to RlmF and its mutant will be performed. This can be done using ITC (isothermal titration calorimetry) or MST (microscale ther-mophoresis). This will help us to elucidate how RlmF binds the 50S LiCl core particles. Also test the SAM binding to RlmF and its mutants with an-other method such as ITC. Ribosome assembly is a coordinated process and rRNA modification enzymes have to act fast enough to allow mature subu-nits to form. How each one of these enzymes recognizes their substrate rRNA is mostly unknown. The RlmF study may contribute to comprehensive understanding of RNA modification enzymes. In Paper IV, structural and biochemical characterization of other two phage-encoded SAM degrading-enzymes like ORF1 and Svi3-7 that were discovered along with Svi3-3 will advance our understanding of the anti-restriction system.

48

Populärvetenskaplig sammanfattning

Användandet av antibiotika för att behandla bakterieinfektioner har revolut-ionerat läkarvetenskapen men det har även lett till en ökning av antibiotika-resistens. Det faktum att bakterier utvecklat mekanismer genom vilka effek-ten av antibiotika motverkas har resulterat i att WHO har deklarerat att anti-biotikaresistens utgör ett globalt hot som kräver organiserade åtgärder (WHO, 2014). Många slags antibiotika slår mot olika delar av det cellulära maskineriet som ansvarar för att översätta det genetiska materialet till prote-iner – den s.k. ribosomen. En grupp antibiotika som kallas aminoglykosider slår mot ribosomens lilla subenhet. En resistensmekanism som bakterier utvecklat mot aminoglykosider utgörs av enzym-katalyserad modifiering av antibiotikamolekylen. Genom att studera hur dessa enzymer identifierar och modifierar sitt antibiotika-substrat kan vi uppnå en djupare förståelse av resistensmekanismer men det kan även hjälpa till i utvecklingen av nya ge-nerationens antimikrobiella läkemedel.

I den första delen av min avhandling (artikel I och II) fokuserar jag på två enzymer av typen aminoglykosidnukleotidyltransferaser (ANT): AadA från Salmonella enterica och ANT(9) from Enterococcus faecalis. AadA är ett enzym med dubbel specificitet som modifierar både antibiotikan streptomy-cin och spektinomycin, till skillnad från ANT(9) som bara modifierar spek-tinomycin. Genom att studera enzymernas strukturella och biokemiska egen-skaper undersökte jag följande frågeställningar: Hur kan AadA känna igen och modifiera två kemiskt olika antibiotikamolekyler? Vad är de strukturella likheterna och skillnaderna mellan AadA och ANT(9)?

Med hjälp av röntgenkristallografi löste jag kristallstrukturen av ANT(9) i komplex med (ko-substratet) ATP och antibiotikan spektinomycin. Vi löste även strukturen av AadA med ATP och streptomycin. Dessa strukturer hjäl-per oss att förklara hur AadA kan känna igen och modifiera båda slags anti-biotika medan ANT(9) bara kan modifiera spektinomycin. Tack vare struk-turerna som visar båda antibiotikamolekylerna i sina respektive aktiva säten kunde vi föreslå en katalytisk mekanism för denna klass ANT enzymer. Ti-digare studier har föreslagit spektinomycinanaloger som behandling mot multiresistent tuberkulos och andra infektioner. Våra strukturer förutspår dock att dessa spektinomycin-liknande molekyler fortfarande kan modifieras av AadA och ANT(9).

49

Den andra delen av min avhandling utforskar en kofaktor som är viktig för olika biokemiska processer, S-adenosylmetionin (SAM). SAM, liksom ATP, används i så många processer i cellen att de båda kan ses som en slags ”cell-valuta”. Valutor inom ekonomin visar på egenskaper såsom begränsad tillgång, hållbarhet, delbarhet, enhetlighet, begränsat utbud och universell acceptabilitet. På samma sätt som valutor i en ekonomi är SAM en molekyl vars kvantitet är tydligt reglerad, använd av många olika enzymklasser och evolutionärt konserverad. Dess metaboliska produkter används som pre-kursorer för andra reaktioner i cellen. I artikel III och IV har jag studerat två olika enzymer: Ett rRNA-metyltransferas (RlmF) och ett SAM-nedbrytande enzym från en bakteriofag. Båda enzymer använder SAM. RlmF är ett en-zym involverat i ribosombiogenesen som katalyserar metyleringen av A1618-basen i 23S rRNA genom att använda SAM som metyldonator. Mo-difikationens roll och hur metyltransferaser känner igen sina substrat-rRNA är dock okänd. I detta projekt förutspådde vi vilka aminosyror som är invol-verade i att binda kofaktorn SAM samt rRNA-substratet genom att jämföra med RlmF-homologen METTL16 MTase. Vi använde oss av riktad mutage-nes för att introducera punktmutationer och deletioner i RlmF, in vitro-metylering samt bindningsassay för att visa bindningen av SAM och sub-strat. Vi använde oss av kryo-elektonmikroskopi för att bestämma strukturen av kärnpartiklar efter LiCl-tvätt av 50S-ribosomer som imiterar det 50S-ribosomintermediat som utgör RlmFs substrat. Våra resultat visade att RlmF är det snabbaste av alla 23S rRNA-metyltransferaser in vitro. RlmF binder SAM i det konserverade bindningssätet och använder sig av liknande struk-turella element som METTL16 för att känna igen sitt substrat.

I artikel IV karakteriserade vi ett nyupptäckt SAM-nedbrytande enzym från en oidentifierad bakteriofag som isolerats från en damm i Uppsala (Svandammen). Det har rapporterats att den här typen av enzym är involve-rad i bakteriers antirestriktionsmekanism. I det här projektet bidrog jag med ligandbindningsstudier samt till att lösa kristallstrukturer av enzymet.

50

Acknowledgements

There are many people that I want to acknowledge for these wonderful years of my PhD.

Firstly, I am very grateful to my supervisor, Maria Selmer. Thanks for accepting as a student and always being there to help and support me. Thanks for your patience in teaching and training me since the time I was a Master student. You also allowed me to work independently in projects and guided me when I was stuck. My sincere and heartfelt thanks for trusting my abilities and making me better at doing science. I consider myself lucky to be part of the group and I really enjoyed all the discussions, meetings and get-togethers. Thanks to Terese Bergfors for helping me with everything during my PhD. I am privileged and honored to share office with you during all these years. Of course you are the best office mate and editor. There are no words to describe how nice a person you are and I learned a lot of things during these years from crystallization to travelling. Consider this as thank you for read-ing all my written texts and improving it. I would like to thank Avinash Punekar, for teaching me almost everything in the lab from day one during my master studies. You made me confident to learn, think and accept the challenges in science. The synchrotron trips were very enjoyable and I learned a lot about data collection. You are good teach-er. Thanks Stefan Knight for all the courses that you taught and all the interest-ing discussions about science. My sincere gratitude to Alwyn Jones—you are an inspiration for learning crystallography and I enjoyed all the talks about cricket. Thanks Sherry Mowbray for all the discussions. I would like to thank my co-supervisors Anthony Forster and Per Jemth. It was a pleasure discussing about project progress and receiving your support. I would like to express my sincere gratitude to past and present members of Selmer group members. It was great journey all along the way since I joined the lab as a master student and then as a PhD student.

51

Current members: Daniel, it was a great collaboration with the RlmF project. I enjoyed learning Cryo-EM technique with you and those excellent scien-tific discussions. Silvia, thanks for the all-good times in the lab and for tak-ing good care of the lab and everyone. Robert, thanks for all the fun during fika and our Friday after-work. It was fun to collaborate on making the mov-ie and you are a good director. Martin, thanks for your positive and happy vibes in the lab. Annika, thanks for being such a good friend and all the help in lab. It was great to plan and participate in all the running events—they were fun and inspiring. Annette, it’s been very cheerful and happy environment around you; thanks for all the social events and fun. Julia, it was great to have you around and thanks for the all the fika cakes and discussions. Dirk and Cisco, thanks for all the lunchtime, fika and after work discussions and fun. It was joyful to be around you both and very supportive. I am thankful to Grim, Josefine, Mangirdas and Malin for fika and discussions. I would like to thank the master students whom I supervised. Dennis, you are a motivated person and it was pleasure to supervise you. Alex, you did a good job with the project and I was happy to supervise you.

Past members: Ulrich, thanks for your support and help in the lab. Noth-ing can match your energy and ambition to work in a project. Thanks for all the inputs, critical comments and thinking. I really learned a lot from you. Thanks also for the fun times during the Nation visit and beer-club. Xiahuo, you are kind, helpful and all thanks for your thoughtful discussions about science and everything. Thanks for the fun time during our ‘game’ in the summer months and a wonderful visit in Amsterdam. Andrei, thanks for all the great times in the lab and those after-work evenings. Ana Stern, it was so much to fun to work with you and all our running events. Alena, it was good to have you around and nice working with you in the lab. Yang, thanks for the fun time and discussions. Roopa, it was great working with you on the methyltransferase project; thanks for the motivation and positive vibe while working in lab. I am thankful to Andras, Harald, Shiying and Nina. Maurice and Sanders, thanks for the great time, when we were hanging-out for board games in the evening and at the Nation. Members of structural biology: Adrian, thanks for all the help in lab. It was always nice to discuss science with you. Shabbir, it is nice to have you around in the lab. I would like to thank all the previous members of structur-al biology: Wimal, Christoffer, Li Zhang, Maiara, Hari, Sofia, Rebecca, Kai, Matthieu, Jens, Aysenur, Viktor and Samuel Flores I am very lucky to have wonderful colleagues as good friends too. Sanjee, Thanks for everything; you are very kind and down to earth person who is helping everyone all the time. It was fun to travel with you on all the syn-

52

chrotron and the other trips. Lu, thanks for the entire delightful trip to China from west to east, for all the discussions about China and the ‘game’. Wangshu, it was so cheerful and fun having you around the lab. Thanks for all the help and fun-filled synchrotron trips. Soneya, thanks for all the help and critical discussions. I would like to thank people from other groups in ICM department who I have interacted and helped me with various aspects: Suparna Sanyal, Se-bastian Deindl, Josefine Liljeruhm, Kenta Okamoto, Martin Svenda, Narayan, Chandra Mandava, Arindam, Laura Lehmann, Laura Gunn, Dirk Hasse, Carolin Seefeldt, Javier, Christoffer L, Masoud, Dilip, Saeid Karkehabadi and also previous and present members of the Kamerlin group: Miha, Paul, Ania, Klaudia, Alex, Beat and Adrian. I am thankful to the administrative personnel: Frida Nilsson, Åsa Hammarborg, Åsa Klef-borg, Akiko Cerenius and Helene Norlin. Staying in Uppsala for decades would not be boring if you are around the right people. Ana Oliveira, thanks for everything and all the help, support and enjoyable times. There are no words to describe them. Sudarsan and Kristine for all the help, support, travel, dinners and discussions. I am sure we will have many more trips to make. Silvana, thanks for the delightful talks and sharing thoughts about PhD difficult times. I would like to thank my Uppsala family—my PhD days would have been inadequate without you all. Hemanth, thanks for sharing the apartment, the good times and also being helpful in everything. Yashraj, you are such a wonderful flatmate—thanks all the fun during the TV and cooking time. Thanks for making all the celebrations and many other occasions joyful. All our festival celebrations, lunch and dinners would have been impossible without Punya and Smruthi. Even though I am an occasional visitor to the lunch group, it was fulfilling with all of you around. Phani, thanks for those dinners and celebration at your place. Malavika, thanks for the cheerful lunch group and all the help. Vivek, Anil and Oksana, thanks for all the gatherings and fun. Sangeetha, thanks for all the cooking at 83B and movie evenings. Thanks for all the great times to Pruthvi, Megha, Satwik and Chandra.

I am grateful for my friends who were there with me from day one in Uppsala. Prashanth, thanks for the fun trips, hiking, gym training and play-ing in freezing cold or scorching sun. Rashmi and Pooja for all the great memories during the Master’s time. Thanks for all the help and always being available—Karthik, Sagar, Srinivas and Sethu. I thank all my friends and previous colleagues in India for their constant support. Thanks to Padmini Sudarshana, Shiva Prakash, Sudheendra and

53

everyone else I worked with for being in touch with me all these years and your help. Special thanks to Sandeep, Shashank and Kiran for everything and for being there during my good and bad times. My sincere thanks to all my teachers, since high school to Bachelor studies and to H G Nagendra for the inspiration and motivation to pursue structural biology. Lastly, I thank my family. Thanks to Pramod for visiting me in Uppsala and all the support. Amma and Appa, thanks for your encouragement and un-derstanding. This would not have been possible without your endless love and support. I would like to thank my brothers, Sandeepa and Santhosha, for everything and all my family members.

54

References

Agris, P. F., Koh, H. and Söll, D. (1973) ‘The effect of growth temperatures on the in vivo ribose methylation of Bacillus stearothermophilus transfer RNA’, Archives of Biochemistry and Biophysics, 154(1), pp. 277–282. doi: 10.1016/0003-9861(73)90058-1.

Aínsa, J. A. et al. (1998) ‘Molecular cloning and characterization of Tap, a putative multidrug efflux pump present in Mycobacterium fortuitum and Mycobacterium tuberculosis’, Journal of Bacteriology. doi: 10.1128/jb.180.22.5836-5843.1998.

Alcaine, S. D., Warnick, L. D. and Wiedmann, M. (2007) ‘Antimicrobial resistance in nontyphoidal Salmonella’, Journal of Food Protection, 70(3), pp. 780–790. doi: 10.4315/0362-028X-70.3.780.

Aminov, R. I. (2010) ‘A brief history of the antibiotic era: Lessons learned and challenges for the future’, Frontiers in Microbiology, 1(134), pp. 1–7. doi: 10.3389/fmicb.2010.00134.

Anderson, P. (1969) ‘Sensitivity and Resistance to Spectinomycin in Escherichia coli.’, Journal of bacteriology, 100(2), pp. 939–47. Available at: http://www.ncbi.nlm.nih.gov/pubmed/16559073%0Ahttp://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=PMC250178.

Baldridge, K. C. and Contreras, L. M. (2014) ‘Functional implications of ribosomal RNA methylation in response to environmental stress’, Critical Reviews in Biochemistry and Molecular Biology, 49(1), pp. 69–89. doi: 10.3109/10409238.2013.859229.

Ban, N. et al. (2000) ‘The complete atomic structure of the large ribosomal subunit at 2.4 Å resolution’, Science. doi: 10.1126/science.289.5481.905.

Boccaletto, P. et al. (2018) ‘MODOMICS: A database of RNA modification pathways. 2017 update’, Nucleic Acids Research. doi: 10.1093/nar/gkx1030.

Borovinskaya, M. A., Shoji, S., et al. (2007) ‘A steric block in translation caused by the antibiotic spectinomycin’, ACS Chemical Biology, 2(8), pp. 545–552. doi: 10.1021/cb700100n.

Borovinskaya, M. A., Pai, R. D., et al. (2007) ‘Structural basis for aminoglycoside inhibition of bacterial ribosome recycling’, Nature Structural and Molecular Biology, 14, pp. 727–32. doi: 10.1038/nsmb1271.

Brown, J. A. et al. (2016) ‘Methyltransferase-like protein 16 binds the 3’-terminal triple helix of MALAT1 long noncoding RNA’, Proceedings of the National Academy of Sciences of the United States of America, 113(49), pp. 14013–14018. doi: 10.1073/pnas.1614759113.

Bruhn, D. F. et al. (2015) ‘Aminomethyl spectinomycins as therapeutics for drug-resistant respiratory tract and sexually transmitted bacterial infections’, Science Translational Medicine, 7(288), pp. 1–12. doi: 10.1126/scitranslmed.3010572.

Butler, M. M. et al. (2018) ‘Aminomethyl spectinomycins as therapeutics for drug-resistant gonorrhea and chlamydia coinfections’, Antimicrobial Agents and Chemotherapy, 62(5), pp. 1–10. doi: 10.1128/AAC.00325-18.

55

Cantoni, G. L. (1952) ‘The nature of the active methyl donor formed enzymatically from l-methionine and adenosinetriphosphate’, Journal of the American Chemical Society. doi: 10.1021/ja01131a519.

Carter, A. P. et al. (2000) ‘Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics’, Nature, 407, pp. 340–348. doi: 10.1038/35030019.

Chen, S. S. et al. (2012) ‘Measuring the dynamics of E. coli ribosome biogenesis using pulse-labeling and quantitative mass spectrometry’, Molecular BioSystems. doi: 10.1039/c2mb25310k.

Chen, Y. et al. (2015) ‘Structure of AadA from Salmonella enterica: A monomeric aminoglycoside (3’’)(9) adenyltransferase’, Acta Crystallographica Section D: Biological Crystallography, 71(9), pp. 2267–2277. doi: 10.1107/S1399004715016429.

Choudhuri, B. S. et al. (2002) ‘Overexpression and functional characterization of an ABC (ATP-binding cassette) transporter encoded by the genes drrA and drrB of Mycobacterium tuberculosis’, Biochemical Journal, 367(1), pp. 279–285. doi: 10.1042/BJ20020615.

Coates, A. R., Halls, G. and Hu, Y. (2011) ‘Novel classes of antibiotics or more of the same?’, British Journal of Pharmacology, 163(1), pp. 184–194. doi: 10.1111/j.1476-5381.2011.01250.x.

Connolly, K., Rife, J. P. and Culver, G. (2008) ‘Mechanistic insight into the ribosome biogenesis functions of the ancient protein KsgA’, Molecular Microbiology, 70, pp. 1062–1075. doi: 10.1111/j.1365-2958.2008.06485.x.

Cox, G. et al. (2015) ‘Structural and molecular basis for resistance to aminoglycoside antibiotics by the adenylyltransferase ANT(2″)-Ia’, mBio, 6(1), pp. 1–9. doi: 10.1128/mBio.02180-14.

Cundliffe, E. (1989) ‘How antibiotic-producing organisms avoid suicide’, Annu. Rev. Microbiol, 43, pp. 208–233.

Davies, J., Gorini, L. and Davis, B. D. (1965) ‘Misreading of RNA codewords induced by aminoglycoside antibiotics.’, Molecular pharmacology, 1(1), pp. 93–106.

Davis, J. H. and Williamson, J. R. (2017) ‘Structure and dynamics of bacterial ribosome biogenesis’, Philosophical Transactions of the Royal Society B: Biological Sciences, 372(1716). doi: 10.1098/rstb.2016.0181.

Decatur, W. A. and Fournier, M. J. (2002a) ‘rRNA modifications and ribosome function’, Trends in Biochemical Sciences. doi: 10.1016/S0968-0004(02)02109-6.

Decatur, W. A. and Fournier, M. J. (2002b) ‘rRNA modifications and ribosome function’, Trends in Biochemical Sciences, 27(7), pp. 344–51. doi: 10.1016/S0968-0004(02)02109-6.

Decatur, W. A. and Fournier, M. J. (2003) ‘RNA-guided nucleotide modification of ribosomal and other RNAs’, Journal of Biological Chemistry. doi: 10.1074/jbc.R200023200.

Demirci, H. et al. (2013) ‘A structural basis for streptomycin-induced misreading of the genetic code’, Nature Communications. Nature Publishing Group, 4(January), pp. 1355–1358. doi: 10.1038/ncomms2346.

Doxtader, K. A. et al. (2018) ‘Structural Basis for Regulation of METTL16 , an S- Article Structural Basis for Regulation of METTL16 , an S-Adenosylmethionine Homeostasis Factor’, Molecular Cell, 71, pp. 1001–1011. doi: 10.1016/j.molcel.2018.07.025.

56

Durante-Mangoni, E. et al. (2009) ‘Do we still need the aminoglycosides?’, International Journal of Antimicrobial Agents, 33(3), pp. 201–205. doi: 10.1016/j.ijantimicag.2008.09.001.

Emsley, P. et al. (2010) ‘Features and development of Coot’, Acta Crystallographica Section D: Biological Crystallography, 66(4), pp. 486–501. doi: 10.1107/S0907444910007493.

Fleming, A. (1929) ‘On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae’, British Journal of Experimental Pathology, 10(3), pp. 226–236.

Fong, D. H. et al. (2010) ‘Structure of the antibiotic resistance factor spectinomycin phosphotransferase from Legionella pneumophila’, Journal of Biological Chemistry, 285(13), pp. 9545–9555. doi: 10.1074/jbc.M109.038364.

Funatsu, G., Nierhaus, K. and Wittmann-Liebold, B. (1972) ‘Ribosomal proteins. XXII. Studies on the altered protein S5 from a spectinomycin-resistant mutant of Escherichia coli’, Journal of Molecular Biology, 64(1), pp. 201–206. doi: 10.1016/0022-2836(72)90329-4.

Gana, R. et al. (2013) ‘Structural and functional studies of S-adenosyl-L-methionine binding proteins: A ligand-centric approach’, BMC Structural Biology. doi: 10.1186/1472-6807-13-6.

García-Solache, M. and Rice, L. B. (2019) ‘The enterococcus: A model of adaptability to its environment’, Clinical Microbiology Reviews, 32(2), pp. 1–28. doi: 10.1128/CMR.00058-18.

Gold, M. and Hurwitz, J. (1964) ‘THE ENZYMATIC METHYLATION OF RIBONUCLEIC ACID AND DEOXYRIBONUCLEIC’, The Journal of biological chemistry, 239, pp. 3866–3874.

Golovina, A. Y. et al. (2012) ‘The last rRNA methyltransferase of E. coli revealed: The yhiR gene encodes adenine-N6 methyltransferase specific for modification of A2030 of 23S ribosomal RNA’, RNA, 18(9), pp. 1725–1734. doi: 10.1261/rna.034207.112.

Gorrec, F. (2009) ‘The MORPHEUS protein crystallization screen’, Journal of Applied Crystallography, 42, pp. 1035–1042. doi: 10.1107/S0021889809042022.

Gould, K. (2016) ‘Antibiotics: From prehistory to the present day’, Journal of Antimicrobial Chemotherapy. doi: 10.1093/jac/dkv484.

Hausmann, R. (1967) ‘Synthesis of an S-Adenosylmethionine-cleaving Enzyme in T3-infected Escherichia coli and Its Disturbance by Co-infection with Enzymatically Incompetent Bacteriophage’, Journal of Virology, 1(1), pp. 57–63. doi: 10.1128/jvi.1.1.57-63.1967.

Helser, T. L., Davies, J. E. and Dahlberg, • James E. (1971) ‘Change in Methylation of 16S Ribosomal RNA Associated with Mutation to Kasugamycin Resistance in Escherichia coli’, Nature, 233, pp. 12–14. doi: 10.1038/233187a0.

Hollingshead, S. and Vapnek, D. (1985) ‘Nucleotide sequence analysis of a gene encoding a streptomycin/spectinomycin adenyltransferase’, Plasmid, 13(1), pp. 17–30. doi: 10.1016/0147-619X(85)90052-6.

Jerlström Hultqvist, J. et al. (2018) ‘A bacteriophage enzyme induces bacterial metabolic perturbation that confers a novel promiscuous function’, Nature Ecology and Evolution, 2, pp. 1321–1330. doi: 10.1038/s41559-018-0568-5.

Kim, C. et al. (2006) ‘Characterization of the bifunctional aminoglycoside-modifying enzyme ANT(3″)-Ii/AAC(6′)-IId from Serratia marcescens’, Biochemistry, 45(27), pp. 8368–8377. doi: 10.1021/bi060723g.

57

Kozbial, P. Z. and Mushegian, A. R. (2005) ‘Natural history of S-adenosylmethionine-binding proteins’, BMC Structural Biology, 5(19). doi: 10.1186/1472-6807-5-19.

Krause, K. M. et al. (2016) ‘Aminoglycosides: An overview’, Cold Spring Harbor Perspectives in Medicine, 6(6), pp. 1–18. doi: 10.1101/cshperspect.a027029.

Kruger, D. H. and Bickle, T. A. (1983) ‘Bacteriophage survival: Multiple mechanisms for avoiding the deoxyribonucleic acid restriction systems of their hosts’, Microbiological Reviews, 47(3), pp. 345–360. doi: 10.1128/mmbr.47.3.345-360.1983.

LeBlanc, D. J., Lee, L. N. and Inamine, J. M. (1991) ‘Cloning and nucleotide base sequence analysis of a spectinomycin adenyltransferase AAD(9) determinant from Enterococcus faecalis’, Antimicrobial Agents and Chemotherapy, 35(9), pp. 1804–1810. doi: 10.1128/AAC.35.9.1804.

Lee, R. E. et al. (2014) ‘Spectinamides: A new class of semisynthetic antituberculosis agents that overcome native drug efflux’, Nature Medicine, 20(2), pp. 152–158. doi: 10.1038/nm.3458.

Lin, J. et al. (2018) ‘Ribosome-Targeting Antibiotics: Modes of Action, Mechanisms of Resistance, and Implications for Drug Design’, Annual Review of Biochemistry, 87(1), pp. 451–478. doi: 10.1146/annurev-biochem-062917-011942.

Liu, Y. P. et al. (2017) ‘Structural basis underlying complex assembly and conformational transition of the type I R-M system’, Proceedings of the National Academy of Sciences of the United States of America, 114(42), pp. 11151–11156. doi: 10.1073/pnas.1711754114.

Lodmell, J. S. and Dahlberg, A. E. (1997) ‘A conformational switch in Escherichia coli 16S ribosomal RNA during decoding of messenger RNA’, Science, 277(5330), pp. 1262–1267. doi: 10.1126/science.277.5330.1262.

Loenen, W. A. M. et al. (2014) ‘Type i restriction enzymes and their relatives’, Nucleic Acids Research, 42(1), pp. 20–44. doi: 10.1093/nar/gkt847.

Lu, S. C. (2000) ‘S-Adenosylmethionine’, International Journal of Biochemistry and Cell Biology, pp. 391–395. doi: 10.1016/S1357-2725(99)00139-9.

Malone, T., Blumenthal, R. M. and Cheng, X. (1995) ‘Structure-guided analysis reveals nine sequence motifs conserved among DNA amino-methyltransferases, and suggests a catalytic mechanism for these enzymes’, Journal of Molecular Biology, 253(4), pp. 618–632. doi: 10.1006/jmbi.1995.0577.

Mato, J. M. et al. (1997) ‘S-adenosylmethionine synthesis: Molecular mechanisms and clinical implications’, Pharmacology and Therapeutics. doi: 10.1016/S0163-7258(96)00197-0.

Matzov, D., Bashan, A. and Yonath, A. (2017) ‘A Bright Future for Antibiotics?’, Annual Review of Biochemistry, 86(1), pp. 567–583. doi: 10.1146/annurev-biochem-061516-044617.

McCarthy, B. J., Britten, R. J. and Roberts, R. B. (1962) ‘The Synthesis of Ribosomes in E. coli: III. Synthesis of Ribosomal RNA’, Biophysical Journal, 2, pp. 57–82. doi: 10.1016/S0006-3495(62)86841-6.

Meyer, K. D. and Jaffrey, S. R. (2017) ‘ Rethinking m 6 A Readers, Writers, and Erasers ’, Annual Review of Cell and Developmental Biology, 33(1), pp. 319–342. doi: 10.1146/annurev-cellbio-100616-060758.

Mohan, S., Donohue, J. P. and Noller, H. F. (2014) ‘Molecular mechanics of 30S subunit head rotation’, Proceedings of the National Academy of Sciences of the United States of America, 111(37), pp. 13325–13330. doi: 10.1073/pnas.1413731111.

58

Montandon, P. E., Wagner, R. and Stutz, E. (1986) ‘E. coli ribosomes with a C912 to U base change in the 16S rRNA are streptomycin resistant.’, The EMBO Journal, 5, pp. 3705–3708. doi: 10.1002/j.1460-2075.1986.tb04703.x.

Murphy, E. (1985) ‘Nucleotide sequence of a spectinomycin adenyltransferase AAD(9) determinant from Staphylococcus aureus and its relationship to AAD(3″) (9)’, MGG Molecular & General Genetics, 200, pp. 33–39. doi: 10.1007/BF00383309.

Nair, D. V. T., Venkitanarayanan, K. and Johny, A. K. (2018) ‘Antibiotic-resistant Salmonella in the food supply and the potential role of antibiotic alternatives for control’, Foods, 7(10), p. 167. doi: 10.3390/foods7100167.

Nirdnoy, W., Mason, C. J. and Guerry, P. (2005) ‘Mosaic structure of a multiple-drug-resistant, conjugative plasmid from Campylobacter jejuni’, Antimicrobial Agents and Chemotherapy, 49(6), pp. 2454–2459. doi: 10.1128/AAC.49.6.2454-2459.2005.

Noeske, J. et al. (2015) ‘High-resolution structure of the Escherichia coli ribosome’, Nature Structural and Molecular Biology, 22, pp. 336–341. doi: 10.1038/nsmb.2994.

Ogle, J. M., Carter, A. P. and Ramakrishnan, V. (2003) ‘Insights into the decoding mechanism from recent ribosome structures’, Trends in Biochemical Sciences, 28(5), pp. 259–266. doi: 10.1016/S0968-0004(03)00066-5.

Pedersen, L. C., Benning, M. M. and Holden, H. M. (1995) ‘Structural Investigation of the Antibiotic and ATP-Binding Sites in Kanamycin Nucleotidyltransferase’, Biochemistry, 34(41), pp. 13305–13311. doi: 10.1021/bi00041a005.

Pennisi, E. (1999) The Race to the Ribosome Structure. Available at: https://science.sciencemag.org/content/285/5436/2048.full.

Piepersberg, W. et al. (1988) ‘Antibiotic resistance by modification: Many resistance genes could be derived from cellular control genes in actinomycetes - A hypothesis.’, Actinomycetologica, 2(2), pp. 83–98. doi: 10.3209/saj.2_83.

Pletnev, P. et al. (2020) ‘Comprehensive Functional Analysis of Escherichia coli Ribosomal RNA Methyltransferases’, Frontiers in Genetics, 11(February), pp. 1–16. doi: 10.3389/fgene.2020.00097.

Polikanov, Y. S. et al. (2015) ‘Structural insights into the role of rRNA modifications in protein synthesis and ribosome assembly’, Nature Structural and Molecular Biology, 22, pp. 342–344. doi: 10.1038/nsmb.2992.

Popova, A. M. and Williamson, J. R. (2014) ‘Quantitative analysis of rRNA modifications using stable isotope labeling and mass spectrometry’, Journal of the American Chemical Society, 136(5), pp. 2058–2069. doi: 10.1021/ja412084b.

Punekar, A. S. et al. (2013) ‘Structural and functional insights into the molecular mechanism of rRNA m6A methyltransferase RlmJ.’, Nucleic acids research, (9), pp. 1–12. doi: 10.1093/nar/gkt719.

Ramirez, M. S. and Tolmasky, M. E. (2010) ‘Aminoglycoside modifying enzymes’, Drug Resistance Updates, 13(6), pp. 151–171. doi: 10.1016/j.drup.2010.08.003.

Ramón-García, S. et al. (2007) ‘Contribution of the Rv2333c efflux pump (the Stp protein) from Mycobacterium tuberculosis to intrinsic antibiotic resistance in Mycobacterium bovis BCG’, Journal of Antimicrobial Chemotherapy, 59(3), pp. 544–547. doi: 10.1093/jac/dkl510.

Rodríguez, D. D. et al. (2009) ‘Crystallographic ab initio protein structure solution below atomic resolution’, Nature Methods, 6, pp. 651–653. doi: 10.1038/nmeth.1365.

59

Ruszkowska, A. et al. (2018) ‘Structural insights into the RNA methyltransferase domain of METTL16’, Scientific Reports, 8(1), pp. 1–13. doi: 10.1038/s41598-018-23608-8.

Schleifer, K. H. and Kilpper-Balz, R. (1984) ‘Transfer of Streptococcus faecalis and Streptococcus faecium to the genus Enterococcus nom. rev. as Enterococcus faecalis comb. nov. and Enterococcus faecium comb. nov.’, International Journal of Systematic Bacteriology, 34(1), pp. 31–34. doi: 10.1099/00207713-34-1-31.

Schlünzen, F. et al. (2001) ‘Structural basis for the interaction of antibiotics with the peptidyl transferase centre in eubacteria’, Nature, 413, pp. 814–821. doi: 10.1038/35101544.

Schmeing, T. M. and Ramakrishnan, V. (2009) ‘What recent ribosome structures have revealed about the mechanism of translation.’, Nature. Macmillan Publishers Limited. All rights reserved, 461(7268), pp. 1234–42. doi: 10.1038/nature08403.

Schubert, H. L., Blumenthal, R. M. and Cheng, X. (2003) ‘Many paths to methyltransfer: A chronicle of convergence’, Trends in Biochemical Sciences, 28, pp. 329–335. doi: 10.1016/S0968-0004(03)00090-2.

Selmer, M. et al. (2006) ‘Structure of the 70S ribosome complexed with mRNA and tRNA’, Science, 313(5795), pp. 1935–1942. doi: 10.1126/science.1131127.

Sergiev, P. V. et al. (2011a) ‘Modifications of ribosomal RNA: From enzymes to function’, in Ribosomes. Springer Vienna, pp. 97–110. doi: 10.1007/978-3-7091-0215-2_9.

Sergiev, P. V. et al. (2011b) ‘Modifications of ribosomal RNA: From enzymes to function’, Ribosomes, pp. 97–110. doi: 10.1007/978-3-7091-0215-2_9.

Sergiev, P. V. et al. (2016) ‘N6-Methylated Adenosine in RNA: From Bacteria to Humans’, Journal of Molecular Biology. Elsevier Ltd, 428(10), pp. 2134–2145. doi: 10.1016/j.jmb.2015.12.013.

Sergiev, P. V. et al. (2018) ‘Structural and evolutionary insights into ribosomal RNA methylation’, Nature Chemical Biology. Nature Publishing Group, 14(3), pp. 226–235. doi: 10.1038/nchembio.2569.

Sergiev, P. V et al. (2008) ‘The ybiN gene of Escherichia coli encodes adenine-N6 methyltransferase specific for modification of A1618 of 23 S ribosomal RNA, a methylated residue located close to the ribosomal exit tunnel.’, Journal of molecular biology, 375(1), pp. 291–300. doi: 10.1016/j.jmb.2007.10.051.

Serio, A. W. et al. (2018) ‘Aminoglycoside Revival: Review of a Historically Important Class of Antimicrobials Undergoing Rejuvenation’, EcoSal Plus, Domain 11, pp. 1–20. doi: 10.1128/ecosalplus.esp-0002-2018.

Shajani, Z., Sykes, M. T. and Williamson, J. R. (2011) ‘Assembly of Bacterial Ribosomes In vitro reconstitution’, Annual review of biochemistry. doi: 10.1146/annurev-biochem-062608-160432.

Shaw, K. J. et al. (1993) ‘Molecular genetics of aminoglycoside resistance genes and familial relationships of the aminoglycoside-modifying enzymes’, Microbiological Reviews, 57(1), pp. 138–163. doi: 10.1128/mmbr.57.1.138-163.1993.

Shi, K. and Berghuis, A. M. (2012) ‘Structural basis for dual nucleotide selectivity of aminoglycoside 2″-phosphotransferase IVa provides insight on determinants of nucleotide specificity of aminoglycoside kinases’, Journal of Biological Chemistry, 287(16), pp. 13094–13102. doi: 10.1074/jbc.M112.349670.

60

Shinkawa, H. et al. (1985) ‘Molecular cloning and expression in Streptomyces lividans of a streptomycin 6-phosphotransferase gene from a streptomycin-producing microorganism’, FEBS Letters, 181, pp. 385–389. doi: 10.1016/0014-5793(85)80298-2.

Siibak, T. and Remme, J. (2010) ‘Subribosomal particle analysis reveals the stages of bacterial ribosome assembly at which rRNA nucleotides are modified’, Rna, 16(10), pp. 2023–2032. doi: 10.1261/rna.2160010.

Spoerel, N., Herrlich, P. and Bickle, T. A. (1979) ‘A novel bacteriophage defence mechanism: The anti-restriction protein’, Nature, 278, pp. 30–34. doi: 10.1038/278030a0.

Stern, A. L. et al. (2018) ‘Structural mechanism of AadA, a dual-specificity aminoglycoside adenylyltransferase from Salmonella enterica’, Journal of Biological Chemistry, 293(29), pp. 11481–11490. doi: 10.1074/jbc.RA118.003989.

Stojković, V. et al. (2020) ‘Assessment of the nucleotide modifications in the high-resolution cryo-electron microscopy structure of the Escherichia coli 50S subunit’, Nucleic Acids Research, 48(5), pp. 2723–2732. doi: 10.1093/nar/gkaa037.

Terwilliger, T. C. et al. (2007) ‘Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard’, Acta Crystallographica Section D: Biological Crystallography, 64(1), pp. 61–69. doi: 10.1107/S090744490705024X.

Thammana, P. and Held, W. A. (1974) ‘Methylation of 16S RNA during ribosome assembly in vitro’, Nature, 251, pp. 682–686. doi: 10.1038/251682a0.

Thompson, P. R. et al. (1998) ‘Spectinomycin kinase from Legionella pneumophila: Characterization of substrate specificity and identification of catalytically important residues’, Journal of Biological Chemistry, 273(24), pp. 14788–14795. doi: 10.1074/jbc.273.24.14788.

Tu, D. et al. (2005) ‘Structures of MLSBK antibiotics bound to mutated large ribosomal subunits provide a structural explanation for resistance’, Cell, 121, pp. 257–70. doi: 10.1016/j.cell.2005.02.005.

Voorhees, R. M. et al. (2009) ‘Insights into substrate stabilization from snapshots of the peptidyl transferase center of the intact 70S ribosome’, Nature Structural and Molecular Biology, 16(5), pp. 528–533. doi: 10.1038/nsmb.1577.

Wachino, J. I. and Arakawa, Y. (2012) ‘Exogenously acquired 16S rRNA methyltransferases found in aminoglycoside-resistant pathogenic Gram-negative bacteria: An update’, Drug Resistance Updates, 15, pp. 133–148. doi: 10.1016/j.drup.2012.05.001.

Waksman, S. A. (1947) ‘What is an antibiotic or an antibiotic substance?’, Mycologia, 39(5), pp. 565–569. doi: 10.2307/3755196.

Waksman, S. A. and Schatz, A. (1945) ‘Streptomycin–Origin, Nature, and Properties*††Journal Series Paper of the Department of Microbiology of the New Jersey Agricultural Experiment Station, Rutgers University.’, Journal of the American Pharmaceutical Association (Scientific ed.). Elsevier Masson SAS, 34(11), pp. 273–291. doi: 10.1002/jps.3030341102.

Webber, M. A. and Piddock, L. J. V. (2003) ‘The importance of efflux pumps in bacterial antibiotic resistance’, Journal of Antimicrobial Chemotherapy, 51(1), pp. 9–11. doi: 10.1093/jac/dkg050.

WHO (2014) Antimicrobial resistance. Global report on surveillance, World Health Organization. doi: 10.1007/s13312-014-0374-3.

61

Wiley, P. F., Argoudelis, A. D. and Hoeksema, H. (1963) ‘The Chemistry of Actinospectacin. IV. The Determination of the Structure of Actinospectacin’, Journal of the American Chemical Society, 85, pp. 2652–2659. doi: 10.1021/ja00900a028.

Wilson, D. N. (2009) ‘The A-Z of bacterial translation inhibitors’, Critical Reviews in Biochemistry and Molecular Biology, 44(6), pp. 393–433. doi: 10.3109/10409230903307311.

Wilson, D. N. (2014) ‘Ribosome-targeting antibiotics and mechanisms of bacterial resistance’, Nature Reviews Microbiology. Nature Publishing Group, 12(1), pp. 35–48. doi: 10.1038/nrmicro3155.

Wilson, D. N. and Nierhaus, K. H. (2008) ‘The weird and wonderful world of bacterial ribosome regulation.’, Critical reviews in biochemistry and molecular biology. Informa UK Ltd UK, 42(3), pp. 187–219. doi: 10.1080/10409230701360843.

Wimberly, B. T. et al. (2000) ‘Structure of the 30S ribosomal subunit’, Nature. doi: 10.1038/35030006.

Yang, J. and Zhang, Y. (2015) ‘I-TASSER server: New development for protein structure and function predictions’, Nucleic Acids Research, 43(W1), pp. W174–W181. doi: 10.1093/nar/gkv342.

Yusupov, M. M. et al. (2001) ‘Crystal structure of the ribosome at 5.5 Å resolution’, Science. doi: 10.1126/science.1060089.

Zengel, J. M. and Lindahl, L. (1994) ‘Diverse Mechanisms for Regulating Ribosomal Protein Synthesis in Escherichia coli’, Progress in Nucleic Acid Research and Molecular Biology, 47, pp. 331–370. doi: 10.1016/S0079-6603(08)60256-1.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1933

Editor: The Dean of the Faculty of Science and Technology

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

Distribution: publications.uu.seurn:nbn:se:uu:diva-408674

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2020