Mechanisms and Dynamics of Mecillinam Resistance in Escherichia Coli

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1375

Mechanisms and Dynamics of Mecillinam Resistance in Escherichia coli

ELISABETH THULIN

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6206 ISBN 978-91-513-0090-0 UPPSALA urn:nbn:se:uu:diva-330856 2017 Dissertation presented at Uppsala University to be publicly examined in A1:111a, BMC, Husargatan 3, Uppsala, Friday, 24 November 2017 at 09:00 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conducted in English. Faculty examiner: Professor Laura Piddock ( Institute of Microbiology and Infection, University of Birmingham, UK).

Abstract Thulin, E. 2017. Mechanisms and Dynamics of Mecillinam Resistance in Escherichia coli. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1375. 69 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0090-0.

The introduction of antibiotics in healthcare is one of the most important medical achievements with regard to reducing human morbidity and mortality. However, bacterial pathogens have acquired antibiotic resistance at an increasing rate, and due to a high prevalence of resistance to some antibiotics they can no longer be used therapeutically. The antibiotic mecillinam, which inhibits the penicillin-binding protein PBP2, however, is an exception since mecillinam resistance (MecR) prevalence has remained low. This is particularly interesting since laboratory experiments have shown that bacteria can rapidly acquire MecR mutations by a multitude of different types of mutations. In this thesis, I examined mechanisms and dynamics of mecillinam resistance in clinical and laboratory isolates of Escherichia coli. Only one type of MecR mutations (cysB) was found in the clinical strains, even though laboratory experiments demonstrate that more than 100 genes can confer resistance Fitness assays showed that cysB mutants have higher fitness than most other MecR mutants, which is likely to contribute to their dominance in clinical settings. To determine if the mecillinam resistant strains could compensate for their fitness cost, six different MecR mutants (cysB, mrdA, spoT, ppa, aspS and ubiE) were evolved for 200-400 generations. All evolved mutants showed increased fitness, but the compensation was associated with loss of resistance in the majority of cases. This will also contribute to the rarity of clinical MecR isolates with chromosomal resistance mutations. How MecR is mediated by cysB mutations was previously unclear, but in this thesis I propose and test a model for the mechanism of resistance. Thus, inactivation of CysB results in cellular depletion of cysteine that triggers an oxidative stress response. The response alters the intracellular levels of 450 proteins, and MecR is achieved by the increase of two of these, the LpoB and PBP1B proteins, which rescue the cells with a mecillinam-inhibited PBP2. Mecillinam is used for UTI treatments and to investigate mecillinam resistance in a more host- like milieu, MecR strains were grown in urine and resistance was examined. Interestingly, this study showed that neither laboratory, nor clinical cysB mutants are resistant in urine, most likely because the cysteine present in the urine phenotypically reverts the bacteria to susceptibility. These findings suggest that mecillinam can be used to treat also those clinical strains that are identified as MecR in standard laboratory tests, and that testing of mecillinam susceptibility in the laboratory ought to be performed in media that mimics urine to obtain clinically relevant results. In summary, the work described in this thesis has increased ourgeneral knowledge of mecillinam resistance and its evolution. Hopefully this knowledge can be put to good use in clinical settings to reduce the negative impact of antibiotic resistance.

Keywords: Mecillinam, Antibiotic resistance, Escherichia coli, Urinary tract infections, Fitness, Penicillin binding proteins, cysteine biosynthesis

Elisabeth Thulin, Department of Medical Biochemistry and Microbiology, Box 582, Uppsala University, SE-75123 Uppsala, Sweden.

ISSN 1651-6206 ISBN 978-91-513-0090-0 urn:nbn:se:uu:diva-330856 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-330856)

If there is one thing the history of evolution has taught us it’s that life will not be contained. Life breaks free, it expands to new territories and crashes through barriers, painfully, maybe even dangerously, but, uh… well, there it is. - Dr. Ian Malcolm

For Måns, Alvar and Irma

Members of the committee

Faculty examiner Professor Laura Piddock Institute of Microbiology and Infection University of Birmingham, UK

Members of the evaluation committee Professor Roland Möllby Department of Microbiology, Tumor and Cell Biology Karolinska Institute, Sweden

Professor Staffan Svärd Department of Cell and Molecular Biology Uppsala University, Sweden

Docent Göte Swedberg, Department of Medical Biochemistry and Microbiology Uppsala University, Sweden

Chairman of the thesis defence Professor Staffan Johansson Department of Medical Biochemistry and Microbiology Uppsala University, Sweden

List of Papers

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

I Thulin, E., Sundqvist, M., Andersson, D. (2015) Amdinocillin (mecillinam) resistance mutations in clinical isolates and laboratory- selected mutants of Escherichia coli. Antimicrobial Agents and Chemotherapy, 59:1718–1727

II Knopp, M., Thulin, E., Ekstrand, E., Andersson, D. (2017) Com- pensatory evolution in mecillinam-resistant Escherichia coli. Manu- script.

III Thulin, E., Andersson, D. (2017) An oxidative stress-induced bypass mechanism confers mecillinam resistance in Escherichia coli. Submitted manuscript.

IV Thulin, E., Thulin, M., Andersson, D. (2017) Reversion of high- level mecillinam resistance to susceptibility in Escherichia coli during growth in urine. EBioMedicine, 23:111–118

Reprints were made with permission from the respective publishers.

Work not included in this thesis:

Lofton, H., Pränting, M., Thulin, E., Andersson, D. (2013) Resistance mechanisms to antimicrobial peptides LL-37, CNY100HL and Wheat Germ Histones. PLOS ONE, 8:e68875

Contents

Introduction ...... 13 Antibiotics ...... 14 Antibiotic resistance ...... 16 Mechanisms of resistance ...... 17 Spread of resistance ...... 20 The cost of being resistant ...... 21 β-lactam antibiotics ...... 23 The β-lactam target – the cell wall ...... 23 Penicillin-binding proteins ...... 25 β-lactam resistance ...... 29 β-lactams in therapeutic use ...... 29 Mecillinam ...... 33 Mecillinam resistance ...... 34 The importance of cysteine biosynthesis in mecillinam resistance ..... 36 Escherichia coli – the accidental pathogen ...... 38 Urinary tract infections ...... 39 Present investigations ...... 42 Paper I: Amdinocillin (Mecillinam) resistance mutations in clinical isolates and laboratory-selected mutants of Escherichia coli ...... 42 Paper II: Compensatory evolution in mecillinam-resistant Escherichia coli ...... 43 Paper III: An oxidative stress-induced bypass-mechanism confers mecillinam resistance in Escherichia coli ...... 45 Paper IV: Reversion of high-level mecillinam resistance to complete susceptibility in Escherichia coli during growth in urine ...... 48 Concluding remarks ...... 51 Svensk sammanfattning ...... 53 Acknowledgements ...... 55 References ...... 59

Abbreviations

AUM artificial urine medium bp base pair DNA deoxyribonucleic acid DTT dithiothreitol EAU European Association of Urology ESBL extended spectrum β-lactamase ESCMID European Society of Clinical Microbiology and Infectious Diseases GTase glycosyltransferase HGT horizontal gene transfer HMW high molecular weight ICE integrative conjugative element IDSA Infectious Diseases Society of America kb kilobase LMW low molecular weight Mec mecillinam MecR mecillinam-resistant MecS mecillinam-susceptible MHB Mueller Hinton broth MIC minimal inhibitory concentration MRSA methicillin-resistant Staphylococcus aureus NAG or GlcNAc N-acetylglucosamine NAM or MurNAc N-acetylmuramic acid NDM New Delhi metallo-β-lactamase PBP penicillin binding protein PG peptidoglycan ppGpp guanosine tetraphosphate RNA ribonucleic acid RRDR rifampin resistance-determining region spp. species pluralis TPase transpeptidase UTI urinary tract infection VRE vancomycin-resistant Enterococci WHO World Health Organization

Introduction

"The thoughtless person playing with penicillin treatment is morally responsible for the death of the man who succumbs to infection with the penicillin- resistant organism." - Alexander Fleming, 1945

The discovery and development of antibiotics is undoubtedly one of the largest breakthroughs in the history of medicine. They were true miracle drugs that made it possible to cure previously deadly diseases, which saved many millions of lives. Thus, when the first systematically used antibiotic – a sulfonamide – was introduced in healthcare in the 1930s it was truly the beginning of a new era. The era of antibiotics opened previously impossible fields of medicine such as transplantation and other major surgery, cancer treatments, and treatment of severe burns. In addition, it meant treatment for many infections that are not necessarily lethal (although might become so if the infection is not cured in time) for instance pneumonia and urinary tract infections (UTIs). Throughout the mid 20th century, more and more antibiotics were developed and put into use, not only in health care, but also in agriculture. The result was an unprecedented shift in the direction of selection on the bacterial population, and as stated by Darwin, selection is the process that drives evolution (Darwin, 1859). In fact, human use and misuse of antibiotics these last 70 years could be described as a grand unintentional evolution experiment. Sure as fate, the first reports of antibiotic resistance came shortly after the discovery of antibiotics. Due to the antibiotic resistance development in pathogenic bacteria, what is often referred to as the “post-antibiotic era” is now drawing closer. This will have dire consequences and diseases we have become used to easily curing with antibiotics may once more have deadly outcomes. The need for new types of antibiotics is urgent, and keeping the ones that we already have working for as long as possible is essential to buy time for developing new ones. Thus, detailed knowledge about the antibiotic resistance mechanisms and how and when to use which antibiotic is necessary. The work presented in this thesis is focused on resistance to the antibiotic mecillinam (Mec), which has been used for treatment of UTIs for some 40 years. The starting point of the projects that are presented in this thesis was two contradictory observations, i) that the added results of a large number of laboratory studies have shown that well over 100 genes can confer resistance

13 when mutated, meaning that the mutation frequency to mecillinam resistance (MecR) is very high and, ii) that MecR is uncommon in clinical isolates, despite several decades of therapeutic use.

Antibiotics The term “antibiotics” was first used by Selman Waksman, further developing concept of “antibiosis” that had been used already in 1889 (Waksman, 1947). His definition of antibiosis was: "inhibiting the growth or the meta- bolic activities of bacteria and other micro-organisms by a chemical substance of microbial origin”. In the beginning of the antibiotic era the term was exclusively used for antibiotics produced by other organisms, while synthetic drugs were not included. This has changed and today several terms, including antibiotics, antimicrobials and antibacterials, are used (sometimes interchangeably) for these types of drugs. In this thesis the term antibiotics will be used for any type of drug that targets bacteria, irrespective of its origin. Antibiotics are bacteriostatic, bactericidal or bacteriolytic. Bacterio- static antibiotics inhibit growth of the bacterial cells, bactericidal antibiotics kill the bacteria, and bacteriolytics not only kill the bacteria, but in the process of doing so also lyse the cells (Figure 1).

addition of removal of antibiotic antibiotic no antibiotic bacteriostatic log number of bacteria log number bacteriocidal

Time Figure 1. Mode of action of bactericidal and bacteriostatic antibiotics.

The first scientist to work systematically with antibiotics was Paul Ehrlich, who introduced the concept of selective toxicity, referring to the ability of a compound to kill or inhibit a pathogenic microorganism without negatively affecting the host. In the early 1900s Ehrlich tested large numbers of chemicals in the search for the “magic bullet” that would hit only the pathogens

14 and not the host. He successfully developed the drug Salvarsan that was used as an anti-syphilis drug. In the 1930s Gerhard Domagk conducted a large scale screening of chemicals to find anti-streptococcal agents, which resulted in the development of the sulfonamides, or sulfa drugs (Domagk, 1935). However, the true breakthrough of antibiotics came with penicillin. The antibacterial property of penicillin was famously discovered by Alexander Fleming in 1928 when the producer organism Penicillium chrysogenum (a mold) was found in his petri dish, causing inhibition of the bacteria that he was growing (Fleming, 1929). A large-scale production process of the compound was developed by a group of scientists lead by chemist Howard Flo- rey in the early years of World War II. As a result, penicillin was used to treat the military personnel of the Allied forces during the latter part of the war. After the war Penicillin G also became available to the general public. Penicillin was not only more effective than the sulfa drugs against some specific streptococcal infections, but also had a broader range of activity, targeting Staphylococci and Pneumococci as well (Walsh, 2003). As stated by Ehrlich, a successful antibiotic drug should be selective in toxicity, affecting only the pathogen and not the host. The selectivity can be achieved either by the drug targeting a cellular structure that is unique to the bacteria, like the cell wall, or by targeting a cellular mechanism that is present but significantly different in the host, for example protein biosynthesis (Figure 2). There are three major cellular mechanisms that are targets for antibiotics: i) cell wall synthesis, ii) protein biosynthesis and iii) Nucleic acid synthesis, elongation or repair. Protein synthesis: Folate synthesis: 50s: 30s: Trimethoprim RNA synthesis: Macrolides Aminoglycosides Chloramphenicol Tetracyclines Sulfonamides Rifampicin Clindamycin

PABA 50s

Dihydropteroic acid

DNA Ribosome gyrase Dihydrofolic acid RNAP purines

30s Tetrahydrofolic acid

Cell wall: DNA structure: Cell membrane: β-lactams Vancomycin Quinolones Colistin Fosfomycin Polymyxins Bacitracin Nitrofurantoin Nitrimidazole

Figure 2. The bacterial cellular targets for antibiotics.

15 Examples of antibiotic classes that target each mechanism are β-lactams and vancomycin (cell wall), tetracyclines, macrolides and aminoglycosides (protein synthesis), and rifamycins and fluoroquinolones (nucleic acid synthesis and structure). There are also a number of antibiotics with other targets: i) folate biosynthesis inhibitors (sulfonamides and trimethoprim), ii) lipid biosynthesis (platensimycin) and iii) cytoplasmic membrane structure and function (peptide antibiotics) (Madigan et al., 2008; Walsh, 2003).

Antibiotic resistance The first reports of antibiotic resistance in pathogens followed soon after the introduction of antibiotics in therapy. For example, sulfonamides were introduced in the 1930s and resistance was reported in the 1940s, penicillin was discovered in 1928 and resistance observed in 1946, and so on, see Figure 3 (Walsh, 2003; Diaz Högberg et al., 2010). It is easy to think of resistance as something that emerged as a result of the introduction of antibiotics in healthcare. However, most antibiotics we use originate from natural com- pounds produced by microorganisms and the natural producers of antibiotics are resistant to their own antibiotics (Aminov, 2009; Fischbach and Walsh, 2009). The natural producers of antibiotics are mainly soil-living bacteria and fungi, and bacteria that co-inhabit the same niches as the natural producers have also developed resistance due to millennia of selection.

antibiotic discovery antibiotic resistance first described

Oxazolidone

Lipopeptides

Trimethoprim Quinolones Glycopeptides Macrolides Chloramphenicol Tetracycline Aminoglycosides Penicillin

1930 1940 1950 1960 1970 1980 1990 2000 2010

Figure 3. Antibiotic resistance timeline. The time of discovery and first description of resistance for some important antibiotic groups.

16 One example of this is a study that screened the resistance patterns of the culturable microbiome in samples from a Mexican cave estimated to have been cut-off from the outside world for approximately 4 million years. In the screen, strains that were resistant to all major classes of antibiotics could be found (Bhullar et al., 2012). A similar result is found in a recently published paper in which the researchers screened historic Staphylococcus aureus samples that had been collected before the clinical introduction of methicillin (a semi-synthetic type of penicillin). They found that a very small subset of strains were resistant, even though they could never have been exposed to methicillin (Harkins et al., 2017). The conclusion that can be drawn from this is that whatever new antibiotics we will discover, the potential resistance mechanism for it is already out there waiting to be selected for. Antibiotic resistance in a bacterial strain can be intrinsic, adaptive or acquired. The resistance is considered intrinsic when caused by an inherent structure or function of that particular bacterial species (Cox and Wright, 2013). An example is the bacterial phylum Tenericutes, whose members do not have a cell wall, and are thus not susceptible to antibiotics targeting that particular structure (Krieg et al., 2011). Another example is the intrinsic resistance to the biocide triclosan that is a common feature of the genus Pseudomonas. Triclosan inhibits the enoyl-ACP reductase FabI, that is part of the fatty acid synthesis, but the members of the Pseudomonas genus have an additional enoyl-ACP reductase (FabV), which is not sensitive to triclosan (Zhu et al. 2010). Adaptive resistance refers to when some spontaneous mutation of a bacterial gene confers resistance. Typically this can be a mutation in a gene that codes for the target of the antibiotic. Moreover, horizontal gene transfer (HGT) provides the means for acquired resistance, where a resistance mechanism from one strain is transferred to another. In many cases the intrinsic resistance genes have spread from for example antibiotic producers or their co-inhabitant bacteria to end up in pathogens (Stokes and Gillings, 2011).

Mechanisms of resistance Resistance to an antibiotic can be achieved in several ways, and the main mechanisms of antibiotic resistance are: i) inactivation of the antibiotic for example by enzymatic modifications, ii) regulation of efflux/influx of antibiotics, which can be achieved by decreasing the permeability of the bacterial membrane or pumping out the antibiotic from the cell, and iii) modification, replacement or overproduction of the antibiotic target (Walsh, 2003; Blair et al., 2014), see Figure 4. These main mechanisms are described next.

17 Regulation of efflux/influx

pump Modification or replacement of target Inactivation or modification of antibiotic Overproduction of target

X+Y XY

enzyme

Figure 4. The main routes to antibiotic resistance.

Target modification, replacement or overproduction As previously discussed, an antibiotic acts by inhibiting an essential cellular function in the bacteria. The inhibition of this cellular function, or target is often achieved by specific binding. If the bacteria acquire a mutation in the target-coding gene, the result can be lowered affinity for the antibiotic, which makes the bacteria resistant. A good example of this is resistance to the RNA synthesis inhibitor rifampicin that targets the RNA polymerase. Rifampicin resistance is easily acquired by chromosomal mutations in a region of rpoB, the gene that encodes the β-subunit of the RNA polymerase (Ramaswamy and Musser, 1998). This part of rpoB in Mycobacterium tuberculosis (codons 507–533) is known as the rifampin resistance- determining region (RRDR). Another way to attain resistance is by replacing the target. If the cellular function inhibited by the antibiotic can be upheld, either by an alternative pathway or structure, the cell is no longer susceptible to the antibiotic. The pathogen MRSA (methicillin-resistant Staphylococcus aureus) owes its resistance to such a mechanism. Methicillin, like other β-lactams, inhibits the function of penicillin binding proteins (PBPs). However, MRSA have acquired a PBP that is insensitive to methicillin (Beck et al., 1986; Pinho et al., 2001). A third type of target-based resistance mechanism is overexpression of the target of the antibiotic. The mechanism is simple, but effective, and acts by expressing high levels of the target so the antibiotic molecules are out- competed on a molecule for molecule bases. Clinically, this is a resistance mechanism known from Enterococcus faecium that achieve resistance to benzylpenicillin by overproduction of PBP5 (Fontana et al., 1994).

18 Antibiotic inactivation or modification Disabling the antibiotic by destroying or changing it is an efficient way of attaining resistance. The most notable example of destruction of antibiotic is the enzymatic inactivation of β-lactams by β-lactamases (Livermore, 1995; Llarrull et al., 2010). The members of this large and diverse group of enzymes cleave the core structure of β-lactam antibiotics, rendering them inactive (Bush and Bradford, 2016). In particular the so-called Extended Spec- trum β-lactamases (ESBLs) are emerging as a major antibiotic resistance threat (Bush, 2013; Paterson and Bonomo, 2005). Moreover, the most important and prevalent aminoglycoside resistance mechanism is enzymatic modification of different –OH or –NH2 groups in the aminoglycoside core structure, which result in non-functional antibiotics (Ramirez and Tolmasky, 2010). The aminoglycoside modification is cata- lyzed by three kinds of modifying enzymes; nucleotidyltranferases, phos- photransferases, and acetyltransferases.

Regulation of antibiotic influx/efflux A general resistance mechanism is to keep the intracellular levels of the antibiotic low, thereby keeping the target unaffected. This can be achieved in two distinct ways: i) by reducing the influx of antibiotic to the cell, or ii) by increasing the efflux of the antibiotic that has entered the cell (Delcour, 2009; Webber and Piddock 2002). Decreased influx and increased efflux are common causal factors behind intrinsic as well as acquired antibiotic resistance. The true master of influx/efflux-mediated antibiotic resistance is the opportunistic pathogen Pseudomonas aeruginosa, which is an important cause of nosocomial infections (Breidenstein et al., 2011; Hancock, 1998; Yordanov and Strateva, 2009). The species is intrinsically resistant to a wide variety of antibiotics, such as aminoglycosides, fluoroquinolones and β- lactams due to low permeability of its membrane (Nicas and Hancock, 1983). In addition, P. aeruginosa also has several intrinsic and acquired β- lactamases and efflux pumps that boost its antibiotic resistance abilities (Breidenstein et al., 2011). For example Masuda et al. showed that the interplay between the intrinsic AmpC β-lactamase and the MexAB-OprM pump system confer resistance to several β-lactams (Masuda et al., 1999). Moreover, resistance to the clinically important tetracycline family of antibiotics is often mediated by the expression of efflux pumps. There are seven classes of Tet pumps, of which the majority belong to the major facilitator superfamily (MFS) of secondary transporters (Guillaume et al., 2004; Thaker et al., 2009). Overexpression of the Tet pumps is toxic to the cell, entailing that expression needs strict regulation. Consequently, many Tet pumps are regulated by a transcriptional repressor called TetR (Guay and Rothstein, 1993; Hinrichs et al., 1994).

19 Spread of resistance All use of antibiotics will result in an increased selection pressure on the acquisition of antibiotic resistance (Aminov, 2009). Before human use of antibiotics, this selection pressure was only present in environments inhabit- ed by the natural antibiotic producers, mainly in soil (Benveniste and Davies, 1973). In contrast, since the introduction of more and more antibiotics in healthcare during the last 70 years, the selection pressure is now on the pathogenic bacteria. In addition, the use of antibiotics in agriculture is a contributing factor to increased selection pressure. The high selection pressure will increase the spread of resistance, especially since bacteria have an alternative route of dissemination of genetic material transfer: horizontal (or lateral) gene transfer (HGT) (Perry and Wright, 2013). HGT makes it possible for bacteria to not only pass genetic material (such as resistance genes) on to their progeny, but also to their neighbors. HGT is mediated in three different ways (Figure 5); i) transformation, which is the uptake of genetic material into the cell directly through the cell membrane, ii) conjugation, that is dependent on cell-to-cell contact, during which DNA is transferred from one strain to another, and iii) transduction, which is the spread of genetic material by phages (bacterial viruses) (Madigan et al., 2008).

Transformation Transduction

Conjugation

Figure 5. The different routes of horizontal gene transfer; transformation, transduction and conjugation.

However, a successful HGT event requires both transfer of DNA and integration of the DNA into the replication process of the receiving cell (Stokes and Gillings, 2011). Out of these three transfer mechanisms, only conjugation can result in stable inheritance to progeny through the plasmid´s auton- omous replication. In contrast, DNA received by transformation or transduction needs to be incorporated into the genome of the recipient bacteria. The integration can be mediated by transposition, site-specific recombination or homologous recombination (Thomas and Nielsen, 2005). The frequency of integration of the horizontally transferred DNA is increased by the existence of mobile genetic elements such as transposons, common regions of insertion sequences, integrative and conjugative elements, gene cassettes, mobi-

20 lized integrons, and the previously mentioned plasmids (Stokes and Gillings, 2011). Mobile genetic elements encode their own transfer and maintenance functions, which provide them with the means to spread rapidly within a bacterial population. When a resistance gene is coupled to a mobile genetic element, it greatly increases the risk of resistance dissemination. The resistance gene will provide a selective advantage and the mobile genetic element the possibility to spread this advantage. The proof of this being a very competitive combination is seen in the large number of resistance genes that are linked to mobile genetic elements in antibiotic resistant bacterial isolates (Perry and Wright, 2013). The most frequent mechanism of HGT for antibiotic resistance genes is conjugation of plasmids and integrative conjugative elements (ICEs or conjugative transposons) (Thomas and Nielsen, 2005). This is easily under- standable, considering that plasmids and ICEs can carry out both transfer and integration into the replicative process of the receptor. The emergence of resistance plasmids, which often carry an array of several different resistance genes, cause severe problems in clinical settings (Barlow, 2009; Carattoli, 2013). These plasmids have evolved on numerous occasions as a response to the selective pressure imposed on the bacterial population by antibiotics. Bacteria that carry the resistance plasmids will have the ability to survive and reproduce in a wider range of environments (Lawrence, 1999; 2000) and they are the cause of the multi-resistance seen in an increasing number of pathogens (Carattoli, 2013).

The cost of being resistant Antibiotics target important cellular functions and any resistance mutations are likely to occur in genes involved in the same function, which should interfere with the growth and physiology of the cell. Consequently, mutations that confer resistance to antibiotics usually come with a fitness cost, which has been shown in numerous experimental resistance studies (Anders- son and Hughes, 2010; Andersson and Levin, 1999; Hughes and Andersson, 2015). However, there are also examples of cost-free and even beneficial resistance mutations. In experimental settings, fitness of antibiotic resistant as well as susceptible bacteria is measured as differences in growth rates, survival and competitive performance (Andersson and Levin, 1999). The relative fitness of a resistant bacterial strain in absence and presence of the antibiotic is an important factor in the evolutionary success of the strain. One would think that a high fitness cost of a resistance mutation would automatically be a major hurdle for bacteria to maintain antibiotic resistance; unfortunately this is not the case because several processes will act to maintain the resistance.

21 The fitness cost of a resistance mutation is one of several factors that affect the probability of that mutation persisting in a bacterial population (An- dersson and Hughes, 2011; Criswell et al., 2006; Ramadhan, 2005). Other factors are; (i) the antibiotic selective pressure on the mutation (Hughes and Andersson, 2016; Sommer et al., 2017), (ii) co-selection of resistance (Enne et al., 2001; Yates et al., 2006), and (iii) the possibility to acquire second-site compensatory mutations (Lenski, 1998; Levin et al., 2000; Melnyk et al., 2014). The result of such compensatory mutations is a strain with higher fitness and maintained or lost resistance (Figure 6).

AbS AbR Fitness AbR

Sensitive Resistant

Figure 6. The fitness cost of a mutation conferring antibiotic resistance (AbR) can be compensated by second site-mutations. Such mutations can either revert to a susceptible phenotype (AbS) or maintain the resistant phenotype.

An alarming example of the importance of selective pressure on competi- tiveness of antibiotic resistant strains was published in 2011 (Gullberg et al., 2011). Gullberg et al. showed that extremely low concentrations of antibiotics, as low as 1/200 of the minimal inhibitory concentration (MIC) of a susceptible strain is enough to select for a resistant strain in competition experiments. Such low concentrations of antibiotics have been detected in environmental samples (Kümmerer, 2009). Thus, a resistance mechanism can be selected for and maintained even if antibiotic levels are very low. Several studies have also shown that once bacteria have acquired resistance, the resistance is not necessarily lost when the bacteria is transferred to an antibiotic free environment. Instead resistant bacteria can readily gain fitness by acquiring compensatory mutations, which reduce the cost and allow the bacteria to maintain the resistance, Figure 6 (Andersson and Hughes, 2010; Andersson and Levin, 1999). Another important contributor to resistance maintenance is co-selection of the resistance gene(s) with an unrelated gene (Andersson, 2003; 2006; Enne, 2004). One interesting exam-

22 ple of co-selection is that mutations in the same molecular mechanisms can result in both antibiotic and heavy metal resistance (Baker-Austin et al., 2006). As previously discussed, an important contribution to co-selection of antibiotic resistance is mobilized genetic elements, especially plasmids that can contain large arrays of different resistance genes (Barlow, 2009; Carat- toli, 2013). These stabilizing processes that cause retainment of antibiotic resistance also seem to be acting in the clinical setting. One example of such stabiliza- tion is provided by an intervention study where the use of trimethoprim for treatment of UTI’s was reduced by 85% for two years. However no significant decrease in community trimethoprim-resistant isolates was seen, most likely due to the low cost of trimethoprim resistance combined with co- selection for other antibiotics (Sundqvist et al., 2010). Another example is a retrospective study of isoniazid-resistant M. tuberculosis isolates carrying inhA mutations, causing the resistance. The isolates were found to have compensated for the fitness cost conferred by the inhA mutations through the acquisition of additional mutations in the promoter region of the ahpC gene (Sherman et al., 1996).

β-lactam antibiotics Mecillinam belongs to the large and very important antibiotic family of β- lactams. One β-lactam, Penicillin G, was the first really successful antibiotic, and β-lactams remain the most prescribed group of antibiotics to this day (Bush and Bradford, 2016). Two β-lactam groups, the penicillins and the cephalosporins, make up over 50% of the antibiotic consumption worldwide (Center for Disease Dynamics and Policy, 2015).

The β-lactam target – the cell wall β-lactams inhibit the synthesis of the bacterial cell wall, also known as peptidoglycan (PG) or murein. Peptidoglycan is a polymer made up of amino acids and the sugars N-acetylglucosamine (GlcNAc or NAG) and N- acetylmuramic acid (MurNAc or NAM) and its function is to provide structural integrity to the cell (Madigan et al., 2008). The murein cell wall is a structure that is unique to bacteria. The lack of peptidoglycan in eukaryotes is of great advantage for the use of β-lactams for treatment of bacterial infections, since that ensure a low toxicity to the host. The great majority of bacteria have some form of cell wall and the structure is so fundamental that we use it to define the main bacterial groups: Gram-negative bacteria and Gram-positive bacteria. The groups owe their name to a staining method developed by the Danish bacteriologist Hans

23 Gram-negative cell envelope

Lipopolysaccharide

Outer membrane

Porin

Peptidoglycan

Penicillin binding protein Inner membrane

Membrane proteins

Figure 7. The cell envelope of Gram-negative bacteria.

Christian Gram (Claus, 1992), after which bacteria will display different color depending on which kind of cell wall they have. The difference between Gram-negative and Gram-positive bacteria is their outermost part, the cell envelope. Like all other organisms, a bacterial cell has a membrane that encloses the cytosol, and the cell wall is located outside of that membrane. Gram-positive bacteria have a thick cell wall, while Gram-negative bacteria have a thin cell wall and an additional cell membrane outside of it (Figures 7 and 8). Since the cell envelope is built up differently in the two groups, the surface structures also differ greatly, with lipopolysaccharide being the main Gram- negative surface structure and lipoteichoic acid the main Gram-positive surface structure (Figures 7 and 8). The amino acid content of the cell wall is also different for Gram-negative and Gram-positive bacteria, with the Gram- negative peptidoglycan containing L-alanine, D-glutamic acid, meso- diaminopimelic acid and D-alanine, and the Gram-positive peptidoglycan containing L-alanine, D-glutamine, L-lysine, and D-alanine (Tang, 2015). Additionally, there is a group of bacteria that have a very different cell envelope, the Mycobacteria (including the pathogen Mycobacterium tuberculosis). The mycobacterial cell wall is comprised of several linked layers; closest to the membrane is peptidoglycan, which is linked to galactofuran, which in turn is linked to arabinofuran, which is finally linked to the outermost layer of long fatty acids called mycolic acids. (Brennan, 2003; Alsteens et al., 2007). Some bacteria also lack a cell wall, e.g. all species of the phylum Tenericutes (Krieg et al., 2011).

24 Gram-positive cell envelope

Surface proteins

Lipoteichoic acid

Peptidoglycan

Teichoic acid

Penicillin binding protein

Cell membrane

Membrane proteins

Figure 8. The cell envelope of Gram-positive bacteria.

The differences in cell envelope build-up are of great importance for the susceptibility of a bacterium to certain antibiotics. For example some antibiotics, like vancomycin, are not able to penetrate the outer membrane of Gram-negative bacteria due to the sheer size of the molecule (Walsh et al. 1996). Gram-negative bacteria are also intrinsically resistant to the lipopep- tide daptomycin, which is an efficient antistaphylococcal drug. The Gram- negative cytoplasmic membrane has a lower proportion of anionic phospho- lipids than its Gram-positive counterpart. This results in insufficient Ca2+- mediated insertion of daptomycin into the Gram-negative membrane, which is then not inhibited (Randall et al. 2012).

Penicillin-binding proteins β-lactams are defined by a common structure, their chemical warhead – the β-lactam ring. This structure gives them their antibacterial property, i.e. inhibition of the peptidoglycan by the specific binding of PG biosynthesis enzymes known as PBPs (Typas et al., 2011). PBPs are acyl serine transfer- ases, and β-lactams bind covalently to their active site serine. The serine residue participates in the transfer of an acyl compound to an amino group or water (Figure 9). When a β-lactam binds to a PBP, it forms a covalent bond, acting as an irreversible inhibitor of PBP transpeptidase activity (Tipper and Strominger, 1965). After PG synthesis inhibition, the cell wall is damaged by a drug-induced imbalance between the cell wall synthases and hydrolases, which ultimately results in lysis of the bacterial cell.

25 R S NH2 N COOH O

PBP PBP

D-Ala D-Ala

R S NH2 O C-O-Ser-PBP N II O H Donor peptide O PBP-Ser +

HOOC Acceptor peptide 2HN X

PBP

NH2 GlcNAc MurNAc L-Ala D-Glu meso-Dap D-Ala

Figure 9. The mechanism of β-lactam antibiotics. β-lactams bind to the active site serine of PBPs since they are analogs of D-Ala-D-Ala, which is the terminal amino acid on the NAM/NAG-subunits of the PG. The binding of a PBP and a β-lactam nucleus is irreversible, which causes inhibition of the PG synthesis.

26 Every bacterial species has its own distinctive set of PBPs, and generally a species has between three and ten different PBPs (Georgopapadakou and Liu, 1980). E. coli have in total 13 PBPs, including two β-lactamases (Table 1). There are two groups of PBPs: the high molecular weight (HMW) PBPs and the low molecular weight (LMW) PBPs that include some β-lactamases (Scheffers and Pinho, 2005). The HMW PBPs (divided into PBP classes A and B) are required for cell wall synthesis (Typas et al., 2011). PG is synthesized in two important steps by enzymes within multiprotein complex ma- chineries (Sauvage and Terrak, 2016): i) polymerization of the glycan chains by glycosyltransferases (GTases) and, ii) crosslinking of the peptides by DD- transpeptidases (DD-TPases). The LMW (class C) PBPs are involved in peptidoglycan hydrolysis and structure regulation The LMW PBPs control the cellular shape and have either carboxypeptidase or endopeptidases enzyme activity (Typas et al., 2011).

Table 1. E. coli peptidoglycan synthesis and structure enzymes. Function Enzymatic activity Name(s) Cellular role Peptidoglycan GTases and PBP1A The major peptidoglycan syn- synthesis dd-TPases PBP1B thases. (class A PBPs) PBP1C Predicted GTase only. dd-TPases PBP2 Cell elongation (class B PBPs) PBP3 Cell division GTase MtgA Cell division Regulation of dd-CPases PBP5, PBP4B, Proposed regulatory role in pepti- peptidoglycan (class C PBPs) PBP6, PBP6B, doglycan synthesis by removal of structure excess pentapeptide donors Peptidoglycan dd-EPases PBP4, PBP7 Septum cleavage and biofilm for- hydrolysis (class C PBPs) mation (PBP7) . PBP4 also has dd- CPase activity. β-lactam β-lactamase AmpC, AmpH Affects cell shape hydrolysis (class C PBPs)

Class A PBPs There are three E. coli class A PBPs (aPBPs); PBP1A, PBP1B and PBP1C, all of which exhibit both GTase and TPase activities (Egan et al., 2015). The aPBPs have sequence homologies, with conserved sequence motifs charac- teristic of transglycosylases and transpeptidases (Höltje, 1998). The cellular function of PBP1A and PBP1B is fairly well understood, and they are considered the main murein-synthetizing enzymes. The role of PBP1C remains to be elucidated, but it has been suggested that it might be a monofunctional GTase in vivo (Schiffer and Holtje, 1999). In addition, PBP1C has a very different β-lactam specificity, despite sequence homology with the two other aPBPs. PBP1A and PBP1B are semi-redundant, as E. coli survives deletion of either, but the deletion of both is lethal (Yousif and Broome-Smith, 1985). PBP1A and PBP1B do however have differences in their activities, PBP1A being part of the elongasome and PBP1B of the divisome.

27 Two recent studies showed that activation by OM lipoproteins are essential for the PG synthase activity of both PBP1A and PBP1B (Paradis-Bleau et al., 2010; Typas et al., 2010). The TPase function of PBP1A and PBP1B require direct interaction with LpoA and LpoB respectively (Typas et al., 2010). Deletion of LpoA or LpoB has the same effect as deletion of their cognate PBP; consequently a deletion of both is lethal.

Class B PBPs: The two monofunctional TPases that make up the E. coli class B PBPs (bPBPs) are PBP2 and PBP3, which are both essential. Since the bPBPs can only carry out TPase activity, each of them is in a molecular partnership with an aPBP: PBP2 interacts with PBP1A and PBP1B with PBP3 (Banzhaf et al., 2012; Bertsche et al., 2006). PBP2 is part of the elongasome (or Rod- system) which is a dynamic complex of proteins that is responsible for the rod shape of the E. coli cell (Blaauwen et al., 2008). PBP3 in turn, is part of a similar protein complex, the divisome, that carries out the cell division (Sauvage et al., 2008). PBP2 and PBP3 are major contributors to the function of respective protein complex and the depletion of either PBP results in distorted growth phenotypes; PBP2 depletion results in inhibition of elongation, causing the cells to become spherical, and PBP3 depleted cells are unable to divide, which results in the cells growing in filaments (Spratt, 1975).

LpoA LpoB PBP3 PBP1A PBP2 PBP1B

To l - P a l FtsQLB

MreC MreC RodA FtsNFtsW Z ZZ BBBB Z Z FtsAZZ B Z FtsA FtsA MreBBB FtsZZ Elongasome Divisome Figure 10. The proteins that make up the complexes responsible for elongation of the rod-shaped cell and cell division in E. coli (Typas et al., 2011; Cho et al., 2014).

28 β-lactam resistance Resistance to β-lactams can be achieved in many ways, and there are known examples of all the main mechanisms of resistance – target modification/replacement, inactivation of antibiotic and regulation of antibiotic efflux/influx. First and foremost, to have an effect, a β-lactam needs to reach its PBP target, and the low permeability of the outer membrane (OM) of Gram- negative bacteria provides some protection. In E. coli and other members of the Enterobacteriaceae family, β-lactams can only reach the periplasm through the non-specific OM porins, proteins that form water-filled channels for the transport of nutrients (Poole, 2002). β-lactam resistance is conferred by mutations leading either to a decreased amount of porins in the membrane or by replacing the porins with narrower or substrate-specific channels (Ad- ler et al., 2012; Doumith et al., 2009; Garcia-Sureda et al., 2011). In general, the level of β-lactam resistance conferred by loss or change of porins is quite low, but in combination with PBP mutations, increased drug efflux or the acquisition of β-lactamases, very high level resistance can be reached (Adler et al., 2012; Nikaido, 2001; Tangden et al., 2013). The clinically most important β-lactam resistance mechanism in Gram- negative pathogens is enzymatic inactivation of the antibiotics by β- lactamases (Bush, 2013). β-lactamases cleave the β-lactam ring, which ren- ders the drug inactive. Even though most β-lactamases have a limited range, almost each and every β-lactam can be inactivated by at least one of these enzymes. The most notable example of pathogens owing their resistance to β-lactamases are those that produce so called Extended Spectrum β- lactamases (ESBL), carbapenemases and metallo-β-lactamases (Paterson and Bonomo, 2005). One of these ESBL enzymes, NDM-1, has the enzymatic capacity to inactivate all β-lactams except aztreonam (Nordmann et al., 2011). It is hardly surprising that in their new report Prioritization of Patho- gens to Guide Research and Development of New Antibiotics, the WHO identifies ESBL-producing Enterobacteriaceae as one of the critical priority pathogen groups, for which new therapeutic antibiotic alternatives are ur- gently needed (WHO 2017). Gram-positive pathogens also produce β-lactamases, but the clinically most important β-lactam resistance is PBP modification, e.g. the previously described accessory PBP (PBP2a), found in S. aureus, with very low affinity for β-lactams (Pinho et al., 2001; Frère, 1995).

β-lactams in therapeutic use The β-lactam family of antibiotics has several subgroups; penicillins, cephalosporins, carbapenems, monobactams and β-lactamase inhibitors. Clinically relevant β-lactams are described in Figure 11.

Penicillins

Natural Penicillinase resistant Amino- Extended spectrum penicillins penicillins penicillins penicillins Penicillin G, Methicillin, Oxacillin Ampicillin, Mecillinam, Carbenicillin, Penicillin V Amoxicillin Piperacillin Cephalosporins

1st generation 2nd generation 3rd generatgenerationion 4th ggeneratione 5th generation Cephalosporin C, Cefuroxime, Cefotaxime, Cefepime, Ceftobiprole, Cefazolin, Cefoxitin. Ceftazidime, Cefpirome Ceftaroline, Cephalotin Cefaclor Ceftriaxone Ceftolozane Carbapenemsp Monobactams β-lactamasetamase ininhibitorshibitors

Imipenem, Aztreonam, Clavulanic acid, Meropenem, Nocardicin A, Sulbactam, Ertapenem, Tabtoxinine Tazobactam Doripenem

Figure 11. The different subgroups in the β-lactam antibiotic family.

All the medically important β-lactams were originally derived from natural secondary metabolites from different organisms, even if a great degree of manipulation of the chemical structures of different β-lactams has been done in laboratories since their discoveries (Demain and Elander, 1999). Fungi are the natural producers of the two largest β-lactam groups; penicillins by the family Penicillium and cephalosporins by the family Acremonium (Fleming 1929; Bryskier et al. 1994). Cephalosporins are also produced by bacterial Streptomyces species. Natural carbapenems are produced by bacteria as Streptomyces cattleya and the plant-pathogenic Erwinia species, while monobactams originate from Chromobacterium violaceum and clavams from Streptomyces clavuligerus (Walsh, 2003; Reading and Cole, 1976).

30 The plethora of β-lactams was developed for different reasons. Initially the aim was to increase the spectrum of bacterial pathogens that were susceptible to β-lactams but as the problem of antibiotic resistance emerged, β- lactams that were resistant to β-lactamases were also needed.

Penicillins Penicillins are characterized by two structures that are attached to their β- lactam ring, a thiazoldine ring and a side chain. There are four penicillin groups; natural penicillins, penicillinase-resistant penicillins, aminopenicillins and extended spectrum penicillins (Miller, 2002). The natural penicillins are only active against a narrow spectrum of Gram-positive bacteria and are susceptible to penicillinases. Resistance to penicillins in pathogens was discovered shortly after the introduction of the antibiotic, and a new synthetic class of penicillins that was penicillinase-resistant was developed (Diaz Högberg et al., 2010; Medeiros, 1997). These penicillinase-resistant penicillins have a slightly less narrow spectrum of target bacteria than the natural penicillins and are efficient in treating staphylococcal infections. They owe their resistance to penicillinases to the addition of a bulky side chain that prevents the enzyme from reaching the β-lactam ring (Batchelor et al., 1959). The aminopenicillins (the most important of them being ampicillin) were developed as broad-spectrum penicillins and are active against Gram- negative bacteria as well as Gram-positive bacteria (Peterson, 2006). They have a signature amino group added onto the penicillin molecule. Ami- nopenicillins are also used therapeutically in combinations with β-lactamase inhibitors against β-lactamase producing bacteria (Jones, 1988). The combination greatly increases the effect of the aminopenicillin, since the β- lactamases are rendered inactive by the β-lactamase inhibitors (Bush, 1988). The extended spectrum penicillins are a diverse group of penicillins, resulting from further manipulations of the ampicillin molecule (Fu and Neu, 1978; Lund and Tybring, 1972).

Cephalosporins The cephem ring, to which the β-lactam ring is fused with a six-membered dihydrothiazine ring, is the defining structure of the cephalosporins. The first one, Cephalosporin C, was originally isolated from the fungus Acremonium chrysogenum (Newton and Abraham, 1956). A large number of semi- synthetic cephalosporins have been developed since, and are now divided into five “generations” to keep the different groups separated (Bush and Bradford, 2016; Sonawane, 2008). First-generation cephalosporins have a moderate spectrum of activity, being effective against Gram-positive cocci as well as many Enterobacteriaceae genera, but not other Gram-negative bacteria (Fernandes et al., 2013). Importantly, the first-generation cephalosporins are sensitive to enzymatic inactivation by β-lactamases (Bush and

31 Bradford, 2016). The second-generation cephalosporins on the other hand, are less active against the Gram-positive cocci. Instead they have increased activity against Haemophilus influenzae and anaerobic bacteria and are more resistant to β-lactamases (Prober, 1998; Bennett and Brown, 2009). Third- generation cephalosporins have an extended spectrum of activity compared to the earlier generations of the drug; they are very efficient against Entero- bacteriaceae, some even against P. aeruginosa and Bacteroides fragilis (Paladino et al., 2008). But, they are usually less active against staphylococci than the first-generation cephalosporins (Bush and Bradford, 2016). The third-generation cephalosporins are also somewhat more stable against β- lactamases. The fourth-generation cephalosporins are zwitterionic broad- spectrum antibiotics that are efficient for treating infections caused by both Gram-positive and Gram-negative pathogens (Hancock and Bellido, 1992). They are usually resistant against the β-lactamases that degrades third- generation cephalosporins. The fifth-generation cephalosporins were developed to be predominantly active against problematic resistant pathogens such as methicillin-resistant Staphylococcus aureus (MRSA), vancomycin- resistant Enterococci (VRE) and Pseudomonas spp. (Fernandes et al., 2013).

Carbapenems The carbapenem β-lactams have potent inhibiting activity on a very broad spectrum of bacterial phyla. They are efficient in treating problematic pathogens such as the Gram-positive Enterococci and Gram-negative P. aeruginosa (Bush and Bradford, 2016). Carbapenems also have high activity against gut anaerobes (Walsh, 2003). As they are stable against most serine-based β- lactamases (but not to metallo-β-lactamases), they are last resort drugs, mainly used to treat multi-resistant bacteria in hospital settings. In the last few years, carbapenem resistance prevalence has increased in clinical settings, which is alarming since they have been used as the last resort drug for bacterial strains that already were resistant to other antibiotics (Potron et al., 2015).

Monobactams The monobactam structure consists of the single β-lactam ring (Sykes et al., 1981). The only monobactam in clinical use is aztreonam, which is active against Gram-negative bacteria and is used for treatments of pseudomonal infections (Hellinger and Brewer, 2011). Aztreonam is also quite resistant to β-lactamases.

β-lactamase inhibitors These β-lactams have poor affinities for the bacterial PBP's and cannot be used as antibiotics. Instead, a β-lactamase inhibitor is used in combination with another β-lactam for treatment of β-lactamase-producing bacteria (Bush and Bradford, 2016). The β-lactamase inhibitors act as suicide substrates,

32 binding to the !-lactamase and giving other !-lactams the chance to act on the bacterial PBPs. Examples of combinations that are frequently used are: amoxicillin-clavulanic acid (called Augumentin) and ampicillin-sulbactam (called Sultamicillin) (Bush, 1988).

Mecillinam In clinical practice, mecillinam is exclusively used for treatment of uncomplicated urinary tract infections (Nicolle, 2000). Due to low oral bioavaila- bility, mecillinam is administered in the form of the pro-drug pivmecillinam (Andrews et al. 1976). Mecillinam is in many ways an ideal antibiotic for treatment of UTIs: it has a narrow spectrum of susceptible bacteria and is primarily affecting the main UTI pathogen E. coli. Mecillinam reaches high levels in urine and has excellent pharmacokinetic properties (Gambertoglio et al., 1980; Roholt et al., 1975). In addition, the mecillinam molecule is stable to most !-lactamases (Brenwald and Andrews, 2006; Sougakoff and Jarlier, 2000). Mecillinam (Figure 12) belongs to the extended spectrum penicillins and is a 6!-amidinopenicillanic acid, meaning that the !-lactam nucleus is linked to the side chain by an amidino group instead of an amido group. The drug was developed by Leo Pharmaceutical Products in the 1970s (Lund and Tybring, 1972; Tybring and Melchior, 1975). It is an unu- sual penicillin in that it is predominantly active against Gram-negative bacteria, especially those belonging to the Enterobacteriaceae family, with the exception of P. aeruginosa and some Klebsiella and Proteus species (Lund and Tybring, 1972).

Figure 12. The structure of mecillinam.

In addition, mecillinam also shows high efficiency of eradication of the Gram-positive uropathogen Staphylococcus saprophyticus when used therapeutically (Monsen et al., 2013). This is unexpected, since standard laboratory antimicrobial susceptibility tests (ASTs) indicate that S. saprophyticus is intrinsically MecR. A likely explanation for the observed therapeutic effect when treating S. saprophyticus is the extremely high levels that mecillinam reaches in urine. Levels of Mec in urine can reach over 1000 mg/L during

33 treatment (Anderson and Adams, 1979; Kerrn et al., 2003; Roholt et al., 1975). Mecillinam exclusively binds to PBP2 and inhibits its activity, which is the TPase step of peptidoglycan synthesis during elongation of the rod- shaped cell. As a result of PBP2 inhibition, the affected cells become spherical, are unable to divide and ultimately lyse (Spratt, 1975; 1977).

Mecillinam resistance The prevalence of clinical resistance to mecillinam has remained low, despite that the drug has been used therapeutically for 40 years. In a number of European countries, the proportion of mecillinam-resistant UTI isolates has been stable at approximately 5% over the years 2000 to 2014 (Kahlmeter et al., 2015). However, the mutation frequency of mecillinam resistance in laboratory experiments has repeatedly been observed to be high, which is due to the mutational target being very large (Paper I). To this date, well over 100 genes have been described to be able to confer resistance to mecillinam when mutated (Lai et al., 2017; Paper I). In short, mecillinam resistance has been found to be caused by mutations in a wide range of cellular functions such as: i) cell elongation and division, (Iwaya et al., 1978; Matsu- zawa et al., 1989; Tamaki et al., 1980; Vinella et al., 2000; 1993; Wachi et al., 1987), ii) important regulatory systems such as the Rcs system (Costa and Antón, 2001; Laubacher and Ades, 2008) and cya/crp in combination with lipopolysaccharide mutations (Antón, 1995; Aono et al., 1979; D’Ari et al., 1988), iii) aminoacyl-tRNA synthetases and during amino acid starvation (Bouloc et al., 1992; Vinella et al., 1992), and iv) cysteine biosynthesis (Cos- ta and Antón, 2006; Oppezzo and Antón, 1995). Two of the first types of MecR mutations that was discovered were mutations in the gene coding for PBP2, mrdA and the neighboring gene, mrdB (or rodA) (Tamaki et al., 1980). Mutations in mrdA or mrdB give E. coli a coc- coid phenotype in addition to the mecillinam resistance; the cells grow and divide as spheres instead of the normal rod-shape (Matsuzawa et al. 1973; Iwaya et al., 1978; Tamaki et al., 1980; Matsuzawa et al., 1989). The resistance mechanism of an mrdA (PBP2) mutant can easily be explained as target modification. The function of RodA has long been unknown, but was recently elucidated. RodA is likely to be the GTase that is in a functional partnership with the TPase PBP2 in peptidoglycan synthesis during cell elongation (Cho et al. 2016; Meeske et al., 2016). However, the understanding of the underlying resistance mechanisms in the numerous MecR mutants with mutations in other genes eluded scientist for many years, but has recently been unraveled. The foundation to the understanding of MecR was provided by the extensive work done by D’Ari and colleagues in the 1990s. They showed that increased intracellular levels of

34 the bacterial signaling molecule ppGpp conferred MecR (Vinella et al., 1992). Elevated levels of ppGpp is the cellular signal for the stringent response (Figure 13) – a shift in the bacterial cell from growth to synthesis when nutrients become scarce and thus limiting for growth (Magnusson et al., 2005). Intracellular levels of a mix of ppGpp and pppGpp are increased and thus signal to the cell to stop the synthesis of rRNA and tRNA and the building of new ribosomes. Instead the focus of the intracellular machinery is turned to biosynthesis of amino acids (Chatterji and Ojha, 2001; Traxler et al., 2008).

GDP + pp i GTP + ATP Ribosome SpoT RelA ppGpp pppGpp

GppA ppGpp RNAP

ppGpp + ppi

Figure 13. The stringent response. The signal molecule pppGpp is primarily synthesized by RelA, but also to a lesser degree by SpoT – the protein that is mainly responsible for ppGpp hydrolysis. RelA binds to the ribosomal protein L11 and in response to starvation and stress starts to synthesize pppGpp. GppA then converts pppGpp to ppGpp that can bind to RNA polymerase (RNAP). This result in a shift of the transcription machinery from ribosomal genes to biosynthesis genes and different stress genes.

When the concentration of ppGpp in the cell is high, PBP2 is no longer essential and normally lethal mutations in mrdA, the gene coding for PBP2, become non-lethal (Navarro et al., 1998). The increase in ppGpp levels in the cell allows the spherical cells to not only survive, but also divide as spheres (Navarro et al., 1998). This is confirmed by the observations that bacteria over-expressing the main ppGpp synthetase RelA and bacteria growing in poor media (conditions that both increase intracellular ppGpp levels) are more resistant to mecillinam (Joseleau-Petit et al., 1994; Vinella et al., 1992). Many of the known MecR mutations have been confirmed to confer resistance through elevated ppGpp levels, for example mutations in spoT, rpoB, galE, aroK and several tRNA synthetase genes (Antón, 1995; Vinella and D'Ari, 1994; Vinella et al., 1996; 1992). Further work of the D’Ari group revealed the upregulation of expression of the cell division gene

35 ftsZ in response to the increased ppGpp levels as essential for the resistance phenotype (Vinella et al., 2000; 1993). Increased cellular levels of FtsZ make PBP2 non-essential in E. coli. However, simply over-expressing ftsZ do not confer resistance. A recent study by Cho et al. gave more insight to the underlying mechanism of mecillinam resistance. They provided evidence that mecillinam has a dual adverse effect on the bacterial cell, not only inhibiting the transpeptidase (TP) activity of PBP2, but also causing the activity of the Rod system to become toxic due to a futile cycle of PG synthesis and degradation performed by the Rod- complex and Slt (lytic transglycosylase) (Cho et al., 2014). The futile cycle of the formation of un-crosslinked glycans rapidly degraded by Slt is not specific to mecillinam, but seems to be a common feature of β-lactams. This would infer that only by inactivating the Rod-system and in addition rendering PBP2 non-essential for growth, a strain could become MecR. In Paper III, we extend the understanding of MecR mechanisms by presenting a model for how mutations in cysteine biosynthesis genes confer resistance. The surprisingly low prevalence of clinical isolates in relation to the very high MecR mutation frequency was addressed in Paper I, in which we found that all of the clinical isolates studied had the MecR-conferring mutation in the same gene (cysB). Other types of chromosomal MecR mutants do not seem to be able to establish in clinical settings (Paper I). As described earlier, mecillinam is resistant to most β-lactamases and unlike resistance to other β-lactams there has been no β-lactamase known to confer resistance in MecR clinical isolates. However, the emergence of a novel resistance plasmid might change this. In a study of French MecR strains, Birgy et al. describe a resistance plasmid with a mutation increasing the expression of the β- lactamase TEM-1 resulting in a high-level expression penicillinase (HEP) phenotype (Birgy et al., 2017). They show that the HEP plasmid confers resistance in almost 80% of clinical MecR isolates examined in the study. The more mecillinam is used, the more the selection pressure for a plasmid conferring MecR will inevitability increase. Perhaps this is the start of a development that will lead to mecillinam becoming unusable in the future.

The importance of cysteine biosynthesis in mecillinam resistance Oppezzo and Antón were the first to describe mecillinam resistance conferred by mutations in cysteine biosynthesis genes. They showed that several mutations destroying the function of the cysB and cysE genes resulted in MecR phenotype in Salmonella enterica serovar. Typhimurium (Oppezzo and Antón, 1995). As described in Paper I, our screening of isolates from a MecR E. coli UTI collection showed that all isolates had mutations in the cysB gene, and that the cysB mutations were responsible for the MecR phenotype.

36 CysB is a transcription factor that regulates expression of several genes that are involved in sulfur and sulfonate metabolism. It positively regulates most cysteine biosynthesis genes and cross-membrane transporters of thiosulfate, sulfate and cystine (Kredich, 2008; van der Ploeg et al., 2001). CysB is negatively regulated by itself and by high cysteine levels and positively regulated by N-acetyl-L-serine, which acts as an inducer for the CysB- dependent transcription (Jagura-Burdzy and Hulanicka, 1981). High levels of N-acetyl-L-serine cause CysB-facilitated upregulation of several genes, including most cysteine biosynthesis genes and genes coding for importers of sulfate and of thiosulfate (Kredich, 2008; Sirko et al., 1990). The amino acid cysteine is a key molecule in the cell. It is obviously needed as a building block in proteins, where the cysteine residues form disulfide bonds. But cysteine residues are also sites that can undergo several redox states, which can change the enzymatic activity or binding activity of a protein (Reddie and Carroll, 2008). In addition, the L-cysteine/L-cystine shuttle system (which is located in the inner membrane of E. coli) provides reducing equivalents to the periplasm (Figure 14), thereby removing periplasmic H2O2 (Ohtsu et al., 2010).

Thiosulfate pathway Sulfate pathway

2- 2- S2O3 SO4 Cystine Cysteine

OM To l C Peroxidation

H2O H2O2 S O 2- SO 2- Periplasm 2 3 4 Cystine Cysteine

CysP Spb TcyJ

IM TcyP TcyL EamA EamB CysT CysT CysW CysW

CysA CysA TcyN Cysteine/Cystine Cytoplasm 2- 2- shuttle system S2O3 SO4 CysDN Cystine NrdH,Trxs, Grxs or GSH? Cysteine NADPH APS Serine CysC CysE PAPS O-acetylserine CysH

2- SO3 CysIJ (CysG)

S2- CysM CysK

NADPH NrdH,Trxs, Grxs or GSH? Degradation S-sulfocysteine Cysteine Pyruvate, H2S, ammonia

SO 2- 3 Figure 14. Cysteine biosynthesis in E. coli. Cysteine is synthesized by two pathways, one using thiosulfate and the other using sulfate as substrate. Both pathways need O-acetylserine that is converted from serine by the CysE enzyme. Located in the inner membrane is the Cysteine/Cystine shuttle system which transfers reducing equivalents from the cytoplasm to the periplasm (Ohtsu et al., 2010; 2015).

37 However, cysB and cysE mutations only confer mecillinam resistance when levels of cysteine are sufficiently low. If enough cysteine is added to the growth medium, the resistance phenotype is lost and if there is no cysteine present in the medium, these mutants cannot grow since strains with non- functional CysB or CysE are cysteine auxotrophs (Oppezzo and Antón, 1995; Paper I). Moreover, the addition of Na2S2O3, Na2SO3 or Na2S have the same effect as cysteine on the cysB mutant, but the MecR phenotype of a cysE mutant is instead reverted by N-acetyl-L-serine and O-acetyl-L-serine (Oppezzo and Antón, 1995). Moreover, reversion of the MecR phenotype of cysB mutants is also achieved by the addition of cystine or glutathione, which are readily converted to cysteine in the bacterial cell. Interestingly, a recent study by Hufnagel et al. showed that addition of the reductant DTT reverts the MecR phenotype of cysE mutants (Hufnagel, 2016). The authors hypothesized that the cysE mutation causes a hyper-oxidation of the cell, similar to our model in Paper III.

Escherichia coli – the accidental pathogen E. coli was first described by (and later given its name from) Theodor Esche- rich in 1885 (Friedmann, 2014). The species is a member of the large, heter- ogeneous and clinically important bacterial family Enterobacteriaceae. The- se Gram-negative rods are ubiquitously distributed in all kinds of ecological niches, such as soil, water and the commensal microbiota of most animals. Some Enterobacteriaceae species are always associated with disease in humans, notable examples being Yersina pestis that cause bubonic plague, Salmonella typhi that cause typhoid fever and Shigella that cause gastrointestinal infections. Other species are part of our normal microbial flora, but can become pathogens under certain conditions, e.g. E. coli, Proteus mirabilis and Klebsiella pneumoniae, all three of which can cause urinary tract infections (UTIs). An interesting feature of the E. coli species is that it inhabits two very different niches, the primary niche being the intestine of warm-blooded animal and reptile hosts and the second being the environmental niche of soil and water (Tenaillon et al., 2010). The host-living E. coli strains also have very versatile lifestyles, including mutualism, commensalism, opportunistic pathogenesis and even specialized pathogenesis. The large differences in lifestyle between different strains of E. coli have greatly influenced the need for ad- aptation and thereby the acquisition of genes by HGT, making the E. coli genome highly dynamic (Dobrindt et al., 2003). Indeed, one of the characteristics of the species E. coli is its very large pan-genome, which is the sum of all the genes that are present in at least one strain of the E. coli species (Ras- ko et al., 2008). Genes that are present in all or the great majority of all the

38 strains of a species and are therefore “species-defining” are referred to as the core genome. The E. coli core genome consists of between 1500 to 3000 genes, depending on the strictness in definition of core genes (Kaas et al., 2012). In contrast the pan-genome of E. coli is considered open, meaning that genes are still being added to the pan-genome as new strains are whole- genome sequenced, but it has been estimated to contain over 45000 genes (Snipen et al., 2009). The genes acquired to adapt to different habitats and lifestyles include genes that increase the ability to colonize or infect, genes that code for tox- ins, genes promoting the acquisition of different nutrients, and importantly, genes that confer resistance to antibiotics. The ability to readily acquire all of these genes has surely made the E. coli species a jack-of-all-trades, but individual strains can be true masters in their niche. As described before, E. coli is normally a harmless inhabitant of the ver- tebrate gut but can cause disease in humans either by a commensal strain ending up in the wrong part of the body (for examples in the urinary tract, causing a UTI) or by colonization and subsequent infection by a more viru- lent pathogenic E. coli strain (Murray et al., 2008). The pathogenic E. coli strains cause a broad range of diseases in humans, such as: i) Gastroenteritis, which is caused by several different E. coli pathotypes (Kaper et al., 2004). The pathotypes are often named by the specific symptoms they cause, e.g. enteropathogenic (EPEC), enterotoxigenic (ETEC), enterohemorrhagic (EHEC), enteroinvasive (EIEC) and enteroaggregative (EAEC). Another type of E. coli that causes gastroenteritis is Shigella, which was originally thought to be its own species. However, due to the clinical significance of Shigella, the nomenclature distinction is maintained (Pupo et al., 2000). ii) UTIs are not only caused by commensal strains, but also by more specialized uropathogenic E. coli (UPEC) that typically have acquired adhesins and hemolysin (Brzuszkiewicz et al., 2006; Kaper et al., 2004). iii) Septicemia, which is caused by E. coli that usually originates from UTIs or gastrointestinal infections (Zhen et al., 2010). iv) Gram-negative neonatal meningitis, for which meningitis/sepsis- associated E. coli (MNEC) is the most common cause (Kaper et al., 2004). MNEC meningitis have fatality rates of 15–40%, and many survivors suffer from severe neurological defects (Unhanand et al., 1993).

Urinary tract infections Urinary tract infection (UTI) is one of the most common bacterial diseases in humans. UTIs are defined by colonization and proliferation of bacteria in the bladder (cystitis), urethra (urethritis) and sometimes also the kidneys (pyelo- nephritis) (Kaper et al., 2004). E. coli is the most predominant UTI pathogen and cause up to 80-90% of all UTIs (Salvatore et al., 2011). However, a

39 wide range of other pathogens also cause UTIs, for example Enterococcus spp., Proteus mirabilis, Klebsiella pneumoniae, Pseudomonas spp., Strepto- coccus spp., Staphylococcus spp. and several species of the yeast genus Candida (Nicolle, 2013). The urinary tract of healthy people is normally sterile. A great contributing factor to the sterility is the recurring total empty- ing of the bladder that reduces colonization. However, bacteria can gain access to the urinary tract in a number of ways, for example: i) acquisition of virulence factors such as adhesins that facilitate colonization, ii) colonization when normal infection barriers are penetrated (e.g. by urethral catheters or mucosal trauma), and iii) underlying patient factors, such as immunocom- promisation or structural abnormalities of the urinary tract (Kaper et al., 2004; Nicolle, 2013). UTIs are divided into uncomplicated UTIs, which mainly affect the lower urinary tract, and, complicated UTIs, which are associated with symptoms in the upper urinary tract or compromised general conditions of the patient (Salvatore et al., 2011). Uncomplicated UTIs almost only affect women, who have a lifetime incidence of about 50% (Foxman, 2002), while complicated UTIs affect both sexes alike. Complicated UTIs are among the most common sources of sepsis, and even though urosepsis can be life- threatening, the prognosis for the patient when treated correctly is usually better than sepsis caused by an infection with another source (Nicolle, 2013; Zhen et al., 2010). Since UTIs often cause morbidity of the affected and can become life threatening if not treated, antibiotic therapy is commonly needed. UTI treatments are usually very effective, as successful antimicrobial therapy will sterilize the urine within minutes (Nicolle, 2013). In addition, UTIs are among the most common bacterial infections in humans. Consequently, about 15% of all prescribed antibiotics in the community are for the treatment of UTIs (Salvatore et al., 2011). This makes treatment of UTIs an important factor in the overall antibiotic resistance development, and the choic- es of which antibiotics to use therapeutically and the compliance of the patients will matter greatly. Which antibiotic to use is decided on the basis of which pathogen is the cause of the UTI, the severity of the infection and also the local availability and resistance prevalence of different antibiotics (Gupta et al., 2011; Nicolle, 2013). Clinical resistance development to traditional antibiotics used for UTI treatment has resulted in a shift in which antibiotics are considered suitable for UTI therapy. The Infectious Diseases Society of America (IDSA) and the European Society for Microbiology and Infectious Diseases (ESCMID) guidelines state that the regional antibiotic resistance percentages of the causative bacteria must be < 20% to consider a drug suitable for treatment of a lower UTI and must be < 10% for treatment of an upper UTI. As a result, trimethoprim, trimethoprim-sulphamethoxazole, amoxicillin, co-amoxiclav, ciprofloxacin and other fluoroquinolones are no longer considered suitable

40 as UTI drugs (Gupta et al., 2011; Kahlmeter et al., 2015). The European Association of Urology (EAU), IDSA and ESCMID recommends fosfomycin trometamol, nitrofurantoin macrocrystal or pivmecillinam as drugs of first choice for uncomplicated UTIs (Gupta et al., 2011; 2007; Nicolle, 2000). For complicated UTIs, the decision on therapeutics is more complex and must be done on a patient to patient basis. Treatment options include fluoroquinolones, aminoglycosides alone or in combinations with β-lactams (such as amoxicillin or a second or third generation cephalosporin), or an extended-spectrum penicillin (Geerlings et al., 2013).

41 Present investigations

“One sometimes finds what one is not looking for.” – Alexander Fleming

Paper I: Amdinocillin (Mecillinam) resistance mutations in clinical isolates and laboratory-selected mutants of Escherichia coli A great number of genes can confer mecillinam resistance when mutated. These genes code for proteins involved in peptidoglycan synthesis and re- modeling, cell elongation and division, regulatory systems such as the stringent response, the Rcs-system and cya/crp, and cysteine biosynthesis (Antón, 1995; Aono et al., 1979; Costa and Antón, 2001; 2006; Laubacher and Ades, 2008; Oppezzo and Antón, 1995; Tamaki et al., 1980; Vinella et al., 2000; 1992; 1993; Wachi et al., 1987). Due to the very large mutational target of the many different genes that can confer MecR, mecillinam resistance is selected with a high frequency in laboratory experiments. In contrast, mecillinam resistance is rare in clinical settings, with a stable prevalence of about 5% of UTI isolates, despite therapeutic use over four decades (Kahlmeter et al., 2015). In Paper I, we compared MecR mutants that were selected in laboratory experiments with those isolated from UTI patients, with the aim of understanding the paradox of MecR being frequently selected for in the laboratory, but uncommon in clinical settings. We characterized the different MecR strains in terms of level of resistance and fitness (growth rate) and identified which mutation was responsible for resistance in each strain. We found that clinical and in vitro MecR isolates differed in two major ways; (i) the fitness levels of the clinical isolates were significantly higher than those of the in vitro isolates, and (ii) while MecR was conferred by many different mutated genes in the laboratory isolated strains, mutations in only one gene was found to be the cause of resistance in the clinical isolates, namely cysB (Figure 15). When assessing the fitness (growth rate in standard medium and in urine) of strains with MecR mutations in different genes we found that cysB mutants have a higher relative fitness than most other types of MecR mutants when examined in MHB. Importantly, the relative fitness difference was even more pronounced in urine. The environment of the uri-

42 nary bladder is relatively harsh and a high enough growth rate is important for bacterial survival.

cysB

del aa * ** * FS(4) aa * FS FS FS IS aa aa - amino acid change, * - stop codon, FS - frame shift, del - deletion, IS - IS insertion Figure 15. Positions in the gene cysB that were found to be mutated in clinical MecR E. coli strains isolated from UTI patients. The cysB mutations were responsible for the resistance phenotype in all MecR strains that were examined.

Our results clearly show that the mutational spectrum of clinical mecillinam resistance is very narrow. We suggest that even though a great number of MecR mutations will occur with a high frequency, the majority will be lost due to selection of a high enough growth rate. As a result, one of the contributing factors for only cysB mutants being prevalent in clinical samples is likely to be that most other MecR mutants are not fit enough to survive in clinical settings.

Paper II: Compensatory evolution in mecillinam- resistant Escherichia coli The clinical prevalence of mecillinam-resistant isolates is low, despite the numerous MecR conferring genes that have been discovered in laboratory experiments. In Paper I, we showed that all clinical isolates had mutations only in one of all known MecR genes (cysB), explaining the difference in MecR incidence between laboratory and clinical settings. Fitness assays in MHB and in urine showed that the great majority of MecR mutations confer a substantial fitness cost. However, the cysB mutants (both laboratory and clinical) generally had higher fitness than strains with other types of MecR mutations. Based on these results, we hypothesized that most non-cysB MecR mutants were not fit enough to survive in the harsh environment of the urinary bladder. However, it is well established that the fitness cost conferred by a resistance mutation can be compensated by second-site mutations. The result of such fitness compensation is a strain that has either maintained or lost it´s resistance phenotype. The focus of Paper II, was to determine if the low fitness of MecR strains could be genetically compensated and if so, what would happen to the resistance phenotype. To this end, we conducted an evolution experiment: eight lineages of six different MecR strains with mutations in mrdA, ppa, ubiE, aspS, spoT, and cysB respectively were cycled in MHB without mecillinam for 200-400 generations. Subsequently, the fitness

43 and Mec susceptibility was assessed by growth rate measurements and MIC assays respectively. All lineages of all mutants had increased fitness compared to the parental strain. Importantly, the increase in fitness was associated with loss of resistance in all compensated mutants except for the mrdA mutant lineages, which retained the resistance phenotype. In addition, the mrdA mutant lineages did not reach as high relative fitness levels as the compensated strains originating from the other MecR mutants (Figure 16). 1.2

1.0 spoT aspS ∆cysB ppa ubiE 0.8

0.6 ∆cysB mrdA aspS spoT 0.4 ppa ubiE mrdA 0.2 Mean relative growth rate growth Mean relative MecR MecS 0

Figure 16. The mean relative fitness levels and MecR/MecS phenotype of the compensated lineages of MecR strains with mutations in the cysB, mrdA, ppa, ubiE, aspS, and spoT genes.

Local sequencing of the resistance-conferring genes showed that the ppa, ubiE, and aspS mutants all had compensated by acquiring intragenic mutations. To find the mutations responsible for compensation in the spoT, cysB, and mrdA compensated mutants, the compensated lineages originating from these mutants were whole genome sequenced. SpoT is the ppGpp hydrolase and all spoT lineages had acquired compensatory mutations in the gene coding for the ppGpp synthetase RelA. The compensated cysB had acquired mutations in other genes involved in cysteine biosynthesis; tcyP, cysP and cysK. Compensation of fitness cost of the mrdA mutation was more complex. The compensated strains all had combinations of point mutations in nlpI and pgsA, deletions of sppA, and a large duplication of about 300 kb that was found in all lineages. Thus, compensation of the fitness cost of mecillinam resistance is generally associated (5/6 cases) with a complete loss of resistance, suggesting that mecillinam resistance and high fitness are often incompatible with each other. This finding provides an additional potential explanation for why the frequency of mecillinam resistance has remained low in clinical isolates.

44 Paper III: An oxidative stress-induced bypass- mechanism confers mecillinam resistance in Escherichia coli Since the work of Oppezzo and Antón some twenty years ago, it has been known that mutations that inactivate CysB (the major regulator of sulphur metabolism including cystein biosynthesis) and the cysteine synthase CysE confer MecR (Oppezzo and Antón, 1995). They also showed that resistance to mecillinam is conditional for cys mutants. When grown in media containing cysteine, or something that can readily be reverted to cysteine, the MecR phenotype of cys mutants is lost. In Paper I we identified cysB mutations as the major cause of resistance in clinical MecR E. coli isolated from UTI patients. Still, the underlying mechanism of resistance for cys mutants has not been fully understood. It was, however, clear that mutations in the cysB gene do not cause resistance by increased intracellular levels of ppGpp (a signalling system known as the stringent response), like a great number of other MecR conferring mutations (Lai et al., 2017; Oppezzo and Antón, 1995; Vinella et al., 1992). In Paper III, we used a combination of global protein profiling, genetic screening and mecillinam susceptibility assays in different conditions to elucidate how MecR is conferred by a cysB mutation. The relative global protein levels were analysed in the cysB mutant when grown in MHB, MHB + cysteine and MHB + ascorbic acid. The pro- teome analysis of a cysB mutant showed that more than 450 individual proteins are at a level of more than 2x or less than 0.5x of the wild type. Interestingly, many of these proteins are products of genes that are known to be over-expressed under oxidative stress conditions. Why starvation for cysteine results in oxidative stress can be explained by the fact that cysteine provides reducing equivalents in the cell, for example by the cysteine/cystine shuttle system located in the E. coli inner membrane (Ohtsu et al., 2010). The hypothesis that cys-mutants become hyperoxidated has previously been suggested by Hufnagel, from his work with cysE mutants (Hufnagel, 2016). As expected, the addition of cysteine during the growth of the cysB mutant when prepared for medium global protein profiling resulted in the great majority of the proteins that were differentially expressed reverting back to wild type levels. The addition of the reductant ascorbic acid to the growth medium of the cysB mutant did not have an as pronounced effect as addition of cysteine, as only a subset of the differentially expressed proteins were reverted to wild type levels. Since cysteine is a key molecule in the cell (Reddie and Carroll, 2008), the changed levels of many proteins upon cysteine starvation is not surprising and the effect of oxidative stress is probably only one of several responses. Therefore, the difference in reversion of protein levels between when a cysB mutant is provided with cysteine and pro-

45 vided with ascorbic acid can likely be explained by “total” reversion of cysteine starvation and reversion of only the oxidative stress response that is the result of the low intracellular levels of cysteine. The higher levels of proteins associated with oxidative stress in the cysB mutant prompted us to consider that the cysB mediated mecillinam resistance might be an effect of hyperoxidation. We therefore investigated the mecillinam susceptibility of the cysB strain in different conditions. Expect- edly, the MecR phenotype was reverted by cysteine as well as by cystine and glutathione, both of which can readily be reverted to cysteine in the cell (Anderson, 1998; Ohtsu et al., 2010). Furthermore, supplementing the growth medium with either of the reductants ascorbic acid or DTT and growth under anaerobic conditions also reverted the resistance, which indicate that the MecR conferred by cysB mutations is an effect of oxidative stress. To further elucidate the resistance mechanism, we guessed that we were looking for a protein that fulfilled two criteria: i) that it would be found at a different level in the cysB mutant compared to the wild type in the global protein profiles and ii) that the levels of the protein should revert back to wild type in the profiles generated by the addition of either cysteine or ascorbic acid during growth. One such protein was LpoB (the activator of PBP1B), which was expressed at 2x the levels of wild type in the cysB mutant. PBP1B also had a similar expression pattern, but the relative protein levels were lower (1.3x). To analyse the effect of overexpression of these proteins on the susceptibility to Mec, we constructed two sets of strains: i) cysB strains with deletions of lpoB and mrcB (the gene coding for PBP1B) and, ii) strains with chromosomal duplications of lpoB or mrcB in the wild type E. coli MG1655 background. Deletion of either of lpoB and mrcB in the cysB strain resulted in loss of the MecR phenotype. Moreover, MIC assay results showed that a strain with a duplication of lpoB has a Mec MIC of 24 mg/L, which is the same as a cysB strain. The strain with the mrcB duplication also had an increased MIC of 1 mg/L, compare to the wild type 0.125 mg/L. These results fit remarkably well with a recently published study on the mecillinam resistome from Lai et al., in which they predict that a class A PBP (the class to which PBP1B belongs) should be at the heart of the mecillinam resistance mechanism (Lai et al., 2017). To conclude, we show in Paper III that oxidative stress, PBP1B and its activator LpoB are involved in MecR conferred by inactivation of CysB: see model in Figure 17. These results contribute to elucidating the resistance mechanism of the major type of clinical MecR strains.

46 Wild type, MecS ∆cysB mutant, MecR

H O H2O H O + LpoB 2 2 + 2 2 H2O2 LpoB PBP1b PBP1b Cystine Cysteine Cystine Cysteine + H O Peroxidation Peroxidation 2 2

H2O2 +

H O Cystine Cysteine Cystine Cysteine 2 2 H2O2 Cysteine/Cystine Cysteine/Cystine shuttle system shuttle system

H2O2 H2O2 H2O2 H2O2 PG PG H O H2O2 2 2

+ Mec + LpoB + LpoB PBP1b H2O2 + PBP1b

+ H2O2

H2O2

Cysteine Cysteine 2 O 2 H Cysteine Cysteine H2O2 O 2 H H2O2

Peroxidation Peroxidation

Cysteine/Cystineshuttle system Cystine Cystine

Cystine Cystine

H2O2 H2O2

H2O2

H2O2 PG

Lysis Growth

Figure 17. The proposed model for how cysB mutations confer mecillinam resistance. In a ∆cysB mutant the cysteine level is reduced and as a result the cysteine/cystine shuttle system (Ohtsu et al., 2010; 2015) cannot provide the periplasm with enough reducing agents, resulting in an oxidative stress response which chang- es the expression levels of over 450 proteins, including an increase in LpoB and PBP1B levels. The increase in LpoB and PBP1B allows for cell wall biosynthesis via an alternative pathway and bypasses the need for a functional PBP2.

47 Paper IV: Reversion of high-level mecillinam resistance to complete susceptibility in Escherichia coli during growth in urine Mecillinam is almost exclusively used to treat UTIs, thus to study MecR in a more host-like environment we used urine as a growth medium for mecillinam-resistant and -susceptible bacteria. We grew a number of different MecR strains in urine supplemented with mecillinam. Interestingly, cysB mutants were not resistant, as they are in the standard laboratory medium for resistance measurements, Mueller Hinton broth (MHB). In contrast, all other types of MecR mutants tested kept their resistant phenotype in urine. This was especially intriguing, since our previous study (Paper I), had shown that cysB mutants were the cause of resistance in all MecR UTI isolates in the collection examined.

-Mec

+Mec

Urine samples collected from 48 individuals. 36 urine samples metabolome analyzed. Growth curves +/- Mec analyzed in Bioscreen. Figure 18. The experimental set up of the study of mecillinam susceptibility of cysB mutants grown in urine.

In Paper IV, we wanted to explore the generality of loss of MecR of cysB mutants grown in urine. To this end, we examined mecillinam susceptibility of clinical and laboratory-isolated cysB mutants and the laboratory wild type strain when grown in urine obtained from 48 different healthy volunteers. Both of the cysB mutant strains showed complete phenotypic susceptibility in the majority of the urine samples, having the same growth profile as the wild type. However, the concentration of mecillinam necessary for inhibition of the strains was not the same in different urine samples. Based on which concentration of mecillinam killed the strains grown in them, the urine samples were divided into three groups: Urinelow, in which strains were killed at 0.25 mg/L, Urinehigh, in which strains grew at 0.25 mg/L but died at 1 mg/L, and Urineint (urine samples yielding intermediate growth profile). Examples of growth in the urine samples from the Urinelow and Urinehigh groups are seen in Figure 19.

48 DA5438 (wt strain) DA28439 (cysB KO) DA24686 (clinical cysB mut) − 1.4 low -Mec − 1.8 -Mec -Mec − 2.2 ln(OD) in urine +0.25 − 2.6 + 0.25 +0.25

− 1.4 -Mec high -Mec -Mec − 1.8 +0.25 +0.25 +0.25 − 2.2

ln(OD) in urine +1 +1 +1 − 2.6

0 100 200 300 400 0100200300400 0100200300400

Time (minutes) Figure 19. Examples of growth in urines from the Urinelow and Urinehigh groups. When no mecillinam is added, all three strains (wild type and the constructed and clinical cysB mutants) grow. When 0.25 mg/L Mec is added to the urine, all three strains are killed in Urinelow, but maintain growth in Urinehigh. By the addition of 1 mg/L Mec, all strains are also killed in Urinehigh.

Global metabolome analysis was done on the urine samples and the association between the mecillinam susceptibility patterns of the bacteria and me- tabolite levels was studied. As already discussed in Papers I-III, cysB mutants are susceptible to mecillinam when provided with enough cysteine. In accordance, the metabolome analysis showed that all urine samples contained cysteine as well as cystine and glutathione, which can readily be converted to cysteine in the bacterial cell. Furthermore, we had previously seen that the degree of reversion to susceptibility varied between the two identified groups Urinelow and Urinehigh. The metabolome analysis revealed that the difference between these two groups was osmolality. A Mann-Whitney test showed that there was a two- fold and highly significant difference in osmolality between the Urinelow group (low osmolality) and the Urinehigh group (high osmolality) (p = 0.0027). To further investigate the impact of osmolality on MecR in urine, we conducted a set of experiments assessing the MIC when increasing and decreasing the osmolality of the Urinehigh and Urinelow urine groups: i) dilution of urine Urinehigh, ii) concentration of Urinelow and iii) addition of sucrose to Urinelow. The results clearly showed that an increase of osmolality was corre- lated to an increased MIC for mecillinam. Since all urine samples contained cysteine, we could not see the effect on growth of cysB mutants in urine without it. Instead, we used artificial urine

49 medium (AUM) agar plates with different concentrations of cysteine for MIC assays (Brooks and Keevil, 1997). As expected, the cysB mutants (that are auxotrophic for cysteine) could not grow on AUM without cysteine. But as the cystein concentration was increased, the strains grew and higher cysteine concentration was associated with lower MIC. This is the first example describing conditional resistance where a genetically stable antibiotic resistance can be phenotypically reverted to susceptibility by metabolites present in urine. These findings have important clinical implications for using mecillinam to treat UTIs. They also suggest that mecillinam might also be used to treat those strains that show a MecR phenotype in laboratory tests.

50 Concluding remarks

“These results underscore a critical reality: antibiotic resistance already exists, widely disseminated in nature, to drugs we have not yet invented.” - Brad Spellberg

Antibiotics are true medical miracles, which have enabled unprecedented advances in healthcare. But as Fleming, Ehrlich and Domagk were making their big discoveries; they also started the selection for antibiotic resistance in pathogens. Just as the antibiotics had been around for millions of years, so had the resistance mechanisms. Early on in the antibiotic era, there were warnings about the likelihood of development of antibiotic resistance, even by Fleming himself. Unfortunately these warnings were disregarded and the miracle drugs were squandered. Obviously, all antibiotic use will lead to increased selection for antibiotic resistance, but the unnecessary use has hastened the development. We are now at a tipping point. A recent report from WHO is meant serve as a guide for which pathogens to prioritize in the development of new antibiotics for drug-resistant bacterial infections, and there are some new antibiotics in the development pipe-line (WHO, 2017). In addition, there are a great number of scientists working on alternative solutions to combating bacterial infections, solutions that do not include antibiotics. Hopefully we will never reach the post-antibiotic era, but to make sure that we do not reach it, we will need to be several steps ahead of the bacteria. If the selection pressure is high enough, the pathogens will inevitably evolve to resist anything we expose them to. But research, including the studies described in this thesis, can give us the upper hand and by using the knowledge gained from it, we can make better decisions on what to use to treat which bacterial disease. A good example of an antibiotic that is used correctly is mecillinam. Firstly, it is a narrow spectrum antibiotic, which means that the selective pressure will be put on the targeted pathogen and not a larger proportion of the total bacterial population. Secondly, it reaches very high concentrations in the site of the targeted infection (urinary bladder). Finally, in important aspects, laboratory experiments do not predict the clinical end result very well in the case of mecillinam: the antibiotic is much more efficient than predicted from bacteriological tests (Monsen et al., 2013), and despite high

51 mutation frequency to MecR the clinical resistance development has been very slow. Are there perhaps other antibiotic drug leads out there with the same characteristics? There could be antibiotics that appear “useless” in laboratory settings but might work just fine as therapeutics. We will not know until we try them in the relevant milieu. To be ahead of the bacterial pathogens, we have to continuously find and develop new functioning antibiotics and can leave no stone unturned in our search for them. In the different papers included in this thesis I have: provided likely ex- planations as to why the clinical MecR incidence is low despite the high mutation frequency (Papers I, II and IV), elucidated the MecR mechanism of cysB mutants (Paper III) and examined the susceptibility to mecillinam in bacteria growing in the more host-like environment of urine (Paper IV). In my opinion, the most important insight from my work is that it pinpoints the great importance of combining information on the specific antibiotic, resistance mechanisms, bacterial pathogen and site of infection when assessing if a compound can be a functional antibiotic in clinical settings. At first glance, the very high mutation frequency to MecR seen in laboratory experiments would probably result in the assumption that clinical resistance would emerge very fast upon the introduction of mecillinam in UTI therapy. In hindsight we know that this did not happen and the work presented in this thesis has contributed to the understanding of why. Hopefully this knowledge can also be put to good use when considering new antibiotics.

52 Svensk sammanfattning

Upptäckten och utvecklingen av antibiotika är en av den medicinska veten- skapen stora landvinningar. Under början och mitten av 1900-talet utfördes ett gediget arbete i mikrobiologiska laboratorier runt om i världen och många nya antibiotika togs fram. Idag är det svårt att föreställa sig en värld utan antibiotika, men det är tyvärr stor sannolikhet att vi måste. När man använ- der ett antibiotikum så främjar man samtidigt livsbetingelserna för bakterier som är resistenta mot det, jämfört med de bakterier som inte är det. Inom varje population (i detta fall bakterier) finns det en genetisk variation. Vissa bakterier har genvarianter (mutationer) som gör dem mer motståndskraftiga mot antibiotika. Antibiotikaanvändning leder alltså till att dessa bakterier får en fördel och ökar i antal jämfört med andra, vilket ofelbart leder till ökad antibiotikaresistens. Tyvärr har vi människor slösat bort det medicinska mi- rakel som antibiotika faktiskt är, delvis genom överanvändande inom sjuk- vården, men kanske främst genom användande inom livsmedelsproduktion. Resistensutvecklingen skulle ha skett i vilket fall, men det onödiga använ- dandet har påskyndat den avsevärt. Problemet görs också värre av att det har utvecklats väldigt få helt nya antibiotika sedan 1980-talet. Resistenta bakterier som orsakar sjukdom sprids lätt inom sjukvården och det finns bakteriestammar som är resistenta mot alla antibiotika som finns att uppbringa. Detta har såklart stor påverkan på infektionsmönster och årligen beräknas 25000 personer i EU avlida som en direkt följd av antibiotikaresistens. Mänskligheten är i stort behov av nya antibiotika, men vi måste också använda de antibiotika som fortfarande är verksamma på bästa möjliga sätt för att köpa tid att hitta nya sätt att bekämpa infektioner på.

Denna avhandling innehåller resultatet av mitt arbete med att förstå hur re- sistens mot antibiotikan mecillinam uppstår och hur det påverkar dess kliniska användande. Mecillinam är en vidareutveckling av den klassiska antibiotikan penicillin och används endast för att behandla urinvägsinfektioner. Grunden för forskningen jag har utfört under min doktorandtid är en paradox: det är väldigt lätt att få fram mecillinamresistenta bakterier i laboratori- eförsök, men trots att mecillinam har använts inom sjukvården i över 40 år har den kliniska resistensutvecklingen varit väldigt långsam. Förekomsten av mecillinamresistenta bakterier hos patienter med urinvägsinfektion ligger stabilt runt 5 % i de allra flesta länder, oavsett hur mycket mecillinam som använts.

53 Genom mitt arbete har jag kunnat bidra till ökad förståelse för hur mecillinamresistens uppstår och varför det är ovanligt hos bakterier som orsakar sjukdom. Anledningen till att det är lätt att hitta mecillinamresistenta bakterier i laboratorieexperiment är att det finns så många olika mutationer som ger mecillinamresistens. I den första artikeln i avhandlingen så visar vi att kliniska stammar endast har en sorts mutationer, i en gen som heter cysB. Utan cysB-genen kan bakterier inte tillverka aminosyran cystein. Vi kunde även se att de andra typerna av mutationer som orsakar mecillinamresistens medför att bakterierna växer långsamt. Antagligen är det den dåliga tillväx- ten som gör att de inte kan infektera och därmed inte hittas i patientprover. I den andra artikeln går vi vidare och ser om vi kan öka tillväxthastighet- en hos sex olika mecillinamresistenta mutanter, genom att låta dem växa och utvecklas utan mecillinam. Resultatet blev att alla de olika mecillinamresistenta stammarna ökade sin tillväxthastighet, men de allra flesta tappade på grund av detta resistensen. Det verkar som att egenskaperna mecillinamresistens och hög tillväxthastighet är oförenliga för bakterierna, vilket ytterli- gare förklarar varför mecillinamresistens är så ovanligt i den kliniska miljön. Den tredje artikeln behandlar våra försök att förstå hur de kliniskt viktiga cysB-mutationerna faktiskt gör bakterierna mecillinamresistenta. I artikeln visar vi att cysB-mutationer orsakar en stress i bakteriecellen, och att om- ställningen som cellen gör för att hantera stressen ökar mängden av vissa proteiner i cellen. Vi ser att en ökning av dessa proteiner (LpoB och PBP1B) är det som ger mecillinamresistens. I den fjärde artikeln lägger vi fram belägg för att cysB-stammar faktiskt inte är mecillinamresistenta i urin. Detta beror på att urin innehåller höga nivåer av cystein. Det kan alltså gå att behandla bakterier som ser ut att vara mecillinamresistenta i tester som utförs på kliniska laboratorier med mecillinam. Denna artikel är mycket viktig, för den visar att man kan felbedöma om en bakterie är resistent eller inte, beroende av hur den får växa när man gör resistenstestet. Avslutningsvis så har min forskning ökat den generella förståelsen av mecillinamresistens, gett förklaringar till varför det är ovanligt med mecillinamresistenta bakterier i den kliniska miljön, samt visat hur viktigt det är att resistenstestning utförs på rätt sätt för att kunna bedöma om ett visst antibiotika är lämpligt för behandling av en patient eller inte.

54 Acknowledgements

First and foremost I want to thank my supervisor Dan for giving me the opportunity to pursue my PhD in your research group. I truly could not have wished for a better place to be these last seven years. I have greatly appreci- ated your generosity and humor, but most of all that you have trusted me to work independently and have had high expectations on the outcome. I am impressed by the way you have built your group, creating not only a great research environment but also encouraging us to have a lot of fun. You also understand the importance of ceremony and the publication champagne with all that it entails is pure brilliance!

Many thanks to my co-supervisor Diarmaid, who always gave me good advise when I needed it. I want you to know what a great teacher I think you are. It was your courses back when I was still on the Bachelors program that made me realize what kind of biologist I wanted to be when I grew up.

I must acknowledge that I have actually had what probably is best described as a “skugghandledare” – Linus. Thank you for always taking the time to talk both science and life with me. I also really have to acknowledge the help that I got with bioinformatics when revising Paper I, when I had given birth to Irma some week earlier. Due to being very tired for obvious reasons, I forgot to acknowledge you in the paper and I provided you with some Talisker as a compensation, but now it´s official! I really did enjoy teaching on your course as well.

My co-authors: Martin, who started the MecR story, Michi, the great com- pensator, you just have so much grit and I foresee an amazing scientific ca- reer for you, and Emelie, it has been so great knowing you are aptly han- dling the lab work while I am cooped up writing the thesis. Måns – I´ll get back to you!

Life as a PhD student is hard at times, but all the lovely people in the DA lab made it so much easier, obviously because of all of the scientific input provided over the years, but also because of all the fun we have! I especially want to thank the two fairy godmothers of the lab – Ulrika, who is the rock of the lab, and the reason I lived to tell the tale about going home from Vemdalen, and Karin, my wonderful lab bench neighbor with whom I have

55 shared so much in talks whilst doing lab work and who has helped me with so many “little things” over the years that they have added up to a huge debt which I am afraid I have to pay at some point... Many thanks to the DA gi- ants on whose shoulders I´ve been standing; Maria, who was my supervisor when I first came the lab and is now my good friend, Sanna, who is undoubtedly one of the best scientists I know, and Chris, Song, Peter and Anna Z who all took so good care of me when I was a “baby-scientist”. The Post Docs in my life: Jocke, who is always keen to discuss results and hy- potheses, Hervé – my E. coli go-to-guy – thank you for letting me rant about my experiments, Amira – Spain was great!, Jon, for being such a nice, car- ing colleague and person in general, Jessica KS who was my early morning coffee buddy and without whom ASM Microbe 2016 would not have been half as fun!, Omar for being such a great office mate, always listening to whatever I feel the need to ventilate and taking out plates for me during weekends, Lionel, (not a Post Doc, but you don´t fit my categories!), it was so great that you decided to join the D7:3 corridor. I will always have a special place in my heart for my fellow “smådoktorander”; Marius, my sassy and warm-hearted PhD twin who surpassed me long ago, Erik G – the Pan- cake General, co-founder of the Milk Alliance, fellow Marvel fan and the only person who like eating indoors as much as I do, Honeybadger Marlen who is sorely missed during lab parties, and Hava, with whom I shared all those dual PhD/parent experiences. I of course want to thank those that came after us; Anna K, who is one of the kindest people I know, Lisa A – I cher- ish all of our office chats on important and trivial things, Erik L for the lunch-time company and for the pEL plasmids, Fredrika – I have enjoyed our cross lab bench talks, keep playing music in the lab!, Hind, for brighten- ing every day in the lab with your laugh, Jenny for early morning company and for proof-reading the thesis, Isa, who is truly the queen of compliments, Greta the calm, humoristic cat lady, Dennis – I´ll see you by the coffee ma- chine, and Po/Bosse – I´m so glad you decided to stay in Sweden.

The past & present members of the DH Lab who amazingly made the D7:3 corridor even better when you joined us there; Disa, Marie, Franzi, Gerrit, Lisa PA, Doug, Eva, Kavita and Linnéa, but most importantly Jessica B, with whom I had the great joy to run in parallel with throughout gymnasiet, Bachelor, Master and PhD and that I really miss having around everyday, and last but not least, Rachel, who is finally making her Uppsala comeback.

Other BMC people that have been important are; at IMBIM: Göte & Nizar, it was great sharing the corridor with you, Pernilla, who I tought was my examiner, Göran, who really was my examiner, Jens, who helped me get a great looking thesis cover, and the fantastic TA-staff who made the research possible: Nalle, Olav, Lasse, Barbro, Erika, Alexis, Rehné, Malin S, Ma- lin R, Veronica, Susanne and Eva; the ICM people: Elin & Britta (a.k.a

56 the pancake girls), the lovely people in the SK group – it was great sharing seminars and conference experiences with you!, and Andrea & Yani who were with us in the D7:3 corridor briefly.

Many thanks to Värmlands nation, for sponsoring my USA conferences with Uddeholm grants and all of the great memories from my student years!

My friends outside of science: The New Years/Last of April/Halloween gang: Kriffe & Camilla, Bobo & Linda, Martina & Jonas and Emma & Krister, it is time to add another holiday to our list! The Remembrance crew, especially The SL Mateusz, who is one of the best people in the world.

The Eriksson families in Leksand/Sifferbo/Gävle. It is such a privilege to have an extra family as you. There is nothing better for my soul than to go to Dalarna and enjoy the views of lake Siljan, good food, your great company and seeing all the kids playing and growing.

My mother Margaretha, who was my first Biology teacher long before I started any school. As I recall, you were also the person that introduced me to the concept of antibiotic resistance, after I had been prescribed penicillin(?) for an eye infection. Thank you for always being there for me. My father Åke, who always encouraged me to get a doctoral degree, even before I knew what kind of scientist I wanted to be. I am not sorry I did not become an engineer like you. I am, however, glad I inherited some of your romantic personality. My sisters: Linnéa, you do the same things I do (nations, PhDs etc.), but you always do it your way, and often way better than me, Veroni- ka, who I am happy is following a little bit in my footsteps becoming a BMA, and the sister I chose – Monica, who is my person in thick and thin.

Måns – there is so much to say. This thesis and everything else in my life would have not have been possible without you. I mean, where would I have been without my pet statistician, my barista, my bearded “mattisrövare”, my programmer, my Indian cook, my gardener, my equal Marvel fan, my chauf- fer, my baker, my board game adversary, my brewer, my mushroom forager, my comedian, my fellow Mannschaft fan, my “cute, nerdy pop-boy”, my best friend, and my love? Thank you for carrying me through all the rough parts of life and helping me to celebrate the best. Team Thulin FTW! But more than anything, thank you for being the best father I could ever imagine.

Alvar och Irma, nu har jag äntligen skrivit klart min doktorsbok. Men det finns ingenting jag kan åstadkomma som gör mig stoltare eller gladare än vad ni gör mig varje dag. Älskar er.

References

Adler, M., Anjum, M., Andersson, D.I., and Sandegren, L. (2012). Influence of acquired β-lactamases on the evolution of spontaneous carbapenem resistance in Escherichia coli. J Antimicrob Chemother 68, 51–59. Alsteens, D., Verbelen, C., Dague, E., Raze, D., Baulard, A.R., and Dufrêne, Y.F. (2007). Organization of the mycobacterial cell wall: a nanoscale view. Pflugers Arch - Eur J Physiol 456, 117–125. Aminov, R.I. (2009). The role of antibiotics and antibiotic resistance in nature. Envi- ron Microbiol 11, 2970–2988. Anderson, J.D., and Adams, M.A. (1979). Urinary excretion of mecillinam by volunteers receiving film-coated tablets of pivmecillinam hydrochloride. Chemo- therapy 25, 1–4. Anderson, M.E. (1998). Glutathione: an overview of biosynthesis and modulation. Chem Biol Interact 111-112, 1–14. Andersson, D.I. (2003). Persistence of antibiotic resistant bacteria. Curr Opin Mi- crobiol 6, 452–456. Andersson, D.I. (2006). The biological cost of mutational antibiotic resistance: any practical conclusions? Curr Opin Microbiol 9, 461–465. Andersson, D.I., and Hughes, D. (2010). Antibiotic resistance and its cost: is it possible to reverse resistance? Nature Rev Microbiol 8, 260–271. Andersson, D.I., and Hughes, D. (2011). Persistence of antibiotic resistance in bacterial populations. FEMS Microbiol Revi 35, 901–911. Andersson, D.I., and Levin, B.R. (1999). The biological cost of antibiotic resistance. Curr Opin Microbiol 2, 489–493. Andrews, J., Kendall, M. J., and Mitchard, M. (1976). Factors influencing the ab- sorption and disposition of mecillinam and pivmecillinam in man. Br J Clin Pharmacol 3, 627–632. Antón, D.N. (1995). Resistance to mecillinam produced by the cooperative action of mutations affecting iipopolysaccharide, spoT, and cya or crp genes of Salmonel- la typhimurium. Mol Microbiol 16, 587–595. Aono, R., Yamasaki, M., and Tamura, G. (1979). High and selective resistance to mecillinam in adenylate cyclase-deficient or cyclic adenosine 3, 5- monophosphate receptor protein-deficient mutants of Escherichia coli. J Bacte- riol 137, 839–845. Baker-Austin, C., Wright, M.S., Stepanauskas, R., and McArthur, J.V. (2006). Co- selection of antibiotic and metal resistance. Trends Microbiol 14, 176–182. Banzhaf, M., van den Berg van Saparoea, B., Terrak, M., Fraipont, C., Egan, A., Philippe, J., Zapun, A., Breukink, E., Nguyen-Distèche, M., Blaauwen, den, T., et al. (2012). Cooperativity of peptidoglycan synthases active in bacterial cell elongation. Mol Microbiol 85, 179–194. Barlow, M. (2009). What antimicrobial resistance has taught us about horizontal gene transfer. Methods Mol Biol 532, 397–411.

59 Batchelor, F.R., Doyle, F.P., Nayler, J.H.C., and Rolinson, G.N. (1959). Synthesis of penicillin: 6-aminopenicillanic acid in penicillin fermentations. Nature 183, 257–258. Beck, W.D., Berger-Bächi, B., and Kayser, F.H. (1986). Additional DNA in methicillin-resistant Staphylococcus aureus and molecular cloning of mec-specific DNA. J Bacteriol 165, 373–378. Bennett P.N., and Brown M.J. (2009). Clinical pharmacology 9th ed. Churchill Liv- ingston: Toronto. Benveniste, R., and Davies, J. (1973). Aminoglycoside antibiotic-inactivating enzymes in actinomycetes similar to those present in clinical isolates of antibiotic- resistant bacteria. Proc Natl Acad Sci USA 70, 2276–2280. Bertsche, U., Kast, T., Wolf, B., Fraipont, C., Aarsman, M.E.G., Kannenberg, K., Rechenberg, von, M., Nguyen-Distèche, M., Blaauwen, den, T., Höltje, J.-V., et al. (2006). Interaction between two murein (peptidoglycan) synthases, PBP3 and PBP1B, in Escherichia coli. Mol Microbiol 61, 675–690. Bhullar, K., Waglechner, N., Pawlowski, A., Koteva, K., Banks, E.D., Johnston, M.D., Barton, H.A., and Wright, G.D. (2012). Antibiotic resistance is prevalent in an isolated cave microbiome. PLoS ONE 7, e34953. Birgy, A., Delecourt, M., Geslain, G., Desselas, E., Caseris, M., Magnan, M., Mari- ani-Kurkdjian, P., Bidet, P., and Bonacorsi, S. (2017). A combination of mecillinam and amoxicillin/clavulanate can restore susceptibility of high-level TEM- 1-producing Escherichia coli to mecillinam. J Antimicrob Chemother 72, 1911– 1914. Blaauwen, den, T., de Pedro, M.A., Nguyen-Distèche, M., and Ayala, J.A. (2008). Morphogenesis of rod-shaped sacculi. FEMS Microbiol Rev 32, 321–344. Blair, J.M.A., Webber, M.A., Baylay, A.J., Ogbolu, D.O., and Piddock, L.J.V. (2014). Molecular mechanisms of antibiotic resistance. Nature Rev Microbiol 13, 42–51. Bouloc, P., Vinella, D., and D'Ari, R. (1992). Leucine and serine induce mecillinam resistance in Escherichia coli. Mol Gen Genet 235, 242–246. Breidenstein, E.B.M., la Fuente-Núñez, de, C., and Hancock, R.E.W. (2011). Pseu- domonas aeruginosa: all roads lead to resistance. Trends Microbiol 19, 419– 426. Brennan, P.J. (2003). Structure, function, and biogenesis of the cell wall of Myco- bacterium tuberculosis. Tuberculosis 83, 91–97. Brenwald, N.P., and Andrews, J. (2006). Activity of mecillinam against AmpC β- lactamase-producing Escherichia coli. J Antimicrob Chemother 58, 223. Brooks, T. and Keevil, C.W. (1997). A simple artificial urine for the growth of urinary pathogens. Lett Appl Microbiol 24, 203–206. Bryskier, A., Aszodi, J., and Chantot, J.-F. (1994). Parenteral cephalosporin classifi- cation. Expert Opin Invest Drugs 3, 145–171. Brzuszkiewicz, E., Brüggemann, H., Liesegang, H., Emmerth, M., Olschläger, T., Nagy, G., Albermann, K., Wagner, C., Buchrieser, C., Emody, L., et al. (2006). How to become a uropathogen: comparative genomic analysis of extraintestinal pathogenic Escherichia coli strains. Proc Natl Acad Sci USA 103, 12879– 12884. Bush, K. (1988). Beta-lactamase inhibitors from laboratory to clinic. Clin Microbiol Rev 1, 109–123. Bush, K. (2013). Proliferation and significance of clinically relevant β-lactamases. Ann N Y Acad Sci 1277, 84–90. Bush, K., and Bradford, P.A. (2016). β-Lactams and β-Lactamase Inhibitors: An Overview. Cold Spring Harb Perspect Med 6, a025247.

60 Carattoli, A. (2013). International Journal of Medical Microbiology. Int J Med Mi- crobiol 303, 298–304. Center for Disease Dynamics, Economics & Policy. (2015). State of the World's Antibiotics, 2015. CDDEP: Washington, D.C. Chatterji, D., and Ojha, A.K. (2001). Revisiting the stringent response, ppGpp and starvation signaling. Curr Opin Microbiol 4, 160–165. Cho, H., Uehara, T., and Bernhardt, T.G. (2014). Beta-lactam antibiotics induce a lethal malfunctioning of the bacterial cell wall synthesis machinery. Cell 159, 1300–1311. Cho, H., Wivagg, C.N., Kapoor, M., Barry, Z., Rohs, P.D.A., Suh, H., Marto, J.A., Garner, E.C., and Bernhardt, T.G. (2016). Bacterial cell wall biogenesis is mediated by SEDS and PBP polymerase families functioning semi-autonomously. Nat Microbiol 1, 1–8. Claus, D. (1992). A standardized Gram staining procedure. W J Microbiol Biotech- nol 8, 451–452. Costa, C.S., and Antón, D.N. (2001). Role of the ftsA1p promoter in the resistance of mucoid mutants of Salmonella enterica to mecillinam: characterization of a new type of mucoid mutant. FEMS Microbiol Lett 200, 201–205. Costa, C.S., and Antón, D.N. (2006). High-level resistance to mecillinam produced by inactivation of soluble lytic transglycosylase in Salmonella enterica serovar Typhimurium. FEMS Microbiol Lett 256, 311–317. Cox, C., and Wright, G. (2013). Intrinsic antibiotic resistance: Mechanisms, origins, challenges and solutions. Int J Med Microbiol 303, 287–292. Criswell, D., Tobiason, V.L., Lodmell, J.S., and Samuels, D.S. (2006). Mutations Conferring Aminoglycoside and Spectinomycin Resistance in Borrelia burgdor- feri. Antimicrob Agents Chemother 50, 445–452. Darwin, C. (1859). On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray: London. Delcour, A.H. (2009). Outer membrane permeability and antibiotic resistance. Bio- chim Biophys Acta 1794, 808–816. Demain, A.L., and Elander, R.P. (1999). The beta-lactam antibiotics: past, present, and future. Antonie Van Leeuwenhoek 75, 5–19. Diaz Högberg, E., Heddini, A., Cars, O. (2010). The global need for effective antibiotics: challenges and recent advances. Trends Pharmacol Sci 31, 509-515. Dobrindt, U., Agerer, F., Michaelis, K., Janka, A., Buchrieser, C., Samuelson, M., Svanborg, C., Gottschalk, G., Karch, H., and Hacker, J. (2003). Analysis of genome plasticity in pathogenic and commensal Escherichia coli isolates by use of DNA arrays. J Bacteriol 185, 1831–1840. Domagk, G. (1935). Ein Beitrag zur Chemotherapie der bakteriellen Infektionen. Dtsch med Wschr 7, 250–253. Doumith, M., Ellington, M.J., Livermore, D.M., and Woodford, N. (2009). Molecu- lar mechanisms disrupting porin expression in ertapenem-resistant Klebsiella and Enterobacter spp. clinical isolates from the UK. J Antimicrob Chemother 63, 659–667. D’Ari, R., Jaffe, A., Bouloc, P., and Robin, A. (1988). Cyclic AMP and cell division in Escherichia coli. J Bacteriol 170, 65–70. Egan, A.J.F., Biboy, J., van't Veer, I., Breukink, E., and Vollmer, W. (2015). Activi- ties and regulation of peptidoglycan synthases. Phil Trans R Soc B 370, 20150031. Enne, V.I. (2004). Enhancement of host fitness by the sul2-coding plasmid p9123 in the absence of selective pressure. J Antimicrob Chemother 53, 958–963.

61 Enne, V.I., Livermore, D.M., Stephens, P., and Hall, L. (2001). Persistence of sul- phonamide resistance in Escherichia coli in the UK despite national prescribing restriction. Lancet 357, 1325–1328. Fernandes, R., Amador, P., and Prudêncio, C. (2013). β-Lactams: chemical structure, mode of action and mechanisms of resistance. Rev Med Microbiol 24, 7– 17. Fischbach, M.A., and Walsh, C.T. (2009). Antibiotics for emerging pathogens. Sci- ence 325, 1089–1093. Fleming, A. (1929). On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae. Brit J Experim Pathol 10, 226–236. Fontana, R., Aldegheri, M., and Ligozzi, M. (1994). Overproduction of a low- affinity penicillin-binding protein and high-level ampicillin resistance in Enter- ococcus faecium. Antimicrob Agents Chemother 38, 1980–1983. Foxman, B. (2002). Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Amer J Med 113, 5–13. Frère, J.-M. (1995). Beta-lactamases and bacterial resistance to antibiotics. Mol Microbiol 16, 385–395. Friedmann, H.C. (2006). Escherich and Escherichia. Adv Appl Microbiol 60 134– 184. Fu, K.P., and Neu, H.C. (1978). Piperacillin, a new penicillin active against many bacteria resistant to other penicillins. Antimicrob Agents Chemother 13, 358– 367. Gambertoglio, J.G., Barriere, S.L., Lin, E.T., and Conte, J.E. (1980). Pharmacoki- netics of mecillinam in health subjects. Antimicrob Agents Chemother 18, 952– 956. Garcia-Sureda, L., Domenech-Sanchez, A., Barbier, M., Juan, C., Gasco, J., and Alberti, S. (2011). OmpK26, a novel porin associated with carbapenem resistance in Klebsiella pneumoniae. Antimicrob Agents Chemother 55, 4742– 4747. Geerlings, S.E., van Nieuwkoop, C., and van Haarst, E. (2013). SWAB Guidelines for Antimicrobial Therapy of Complicated Urinary Tract Infections in Adults. Secretariat SWAB. Georgopapadakou, N.H., and Liu, F.Y. (1980). Penicillin-binding proteins in bacteria. Antimicrob Agents Chemother 18, 148–157. Guay, G.G., and Rothstein, D.M. (1993). Expression of the tetK gene from Staphy- lococcus aureus in Escherichia coli: comparison of substrate specificities of TetA (B), TetA (C), and TetK efflux proteins. Antimicrob Agents Chemother 37, 191–198. Guillaume, G., Ledent, V., Moens, W., and Collard, J.-M. (2004). Phylogeny of efflux-mediated tetracycline resistance genes and related proteins revisited. Mi- crob Drug Resist 10, 11–26. Gullberg, E., Cao, S., Berg, O.G., Ilbäck, C., Sandegren, L., Hughes, D., and An- dersson, D.I. (2011). Selection of resistant bacteria at very low antibiotic concentrations. PLoS Pathog 7, e1002158. Gupta, K., Hooton, T.M., and Roberts, P.L. (2007). Short-course nitrofurantoin for the treatment of acute uncomplicated cystitis in women. Arch Intern Med 167, 2207–2212.

62 Gupta, K., Hooton, T.M., Naber, K.G., Wullt, B., Colgan, R., Miller, L.G., Moran, G.J., Nicolle, L.E., Raz, R., Schaeffer, A.J., et al. (2011). International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelo- nephritis in women: A 2010 update by the Infectious Diseases Society of Amer- ica and the European Society for Microbiology and Infectious Diseases. Clin In- fect Dis 52, e103–e120. Hancock, R.E., and Bellido, F. (1992). Factors involved in the enhanced efficacy against gram-negative bacteria of fourth generation cephalosporins. J Antimi- crob Chemother 29, 1–6. Hancock, R. (1998). Resistance mechanisms in Pseudomonas aeruginosa and other nonfermentative gram-negative bacteria. Clin Infect Dis 27, S93–S99. Harkins, C.P., Pichon, B., Doumith, M., Parkhill, J., Westh, H., Tomasz, A., de Lencastre, H., Bentley, S.D., Kearns, A.M., and Holden, M.T.G. (2017). Methi- cillin-resistant Staphylococcus aureus emerged long before the introduction of methicillin into clinical practice. Genome Biol 18, 1–11. Hellinger, W.C., and Brewer, N.S. (2011). Carbapenems and monobactams: imipenem, meropenem, and aztreonam. Mayo Clin Proc 74, 420–434. Hinrichs, W., Kisker, C., Düvel, M., and Müller, A. (1994). Structure of the Tet repressor-tetracycline complex and regulation of antibiotic resistance. Science 264, 418–420. Höltje, J.-V. (1998). Growth of the stress-bearing and shape-maintaining murein sacculus of Escherichia coli. Microbiol Mol Biol Rev 62, 181–203. Hufnagel, D. (2016). Environmental and molecular control of the uropathogenic Escherichia coli extracellular matrix. Dissertation from University of Michigan, USA. Hughes, D., and Andersson, D.I. (2015). Evolutionary consequences of drug resistance: shared principles across diverse targets and organisms. Nature Rev Gen 16, 459–471. Hughes, D., and Andersson, D.I. (2016). Evolutionary trajectories to antibiotic resistance. Annu Rev Microbiol 71, 579–96. Iwaya, M., Jones, C.W., Khorana, J., and Strominger, J.L. (1978). Mapping of the mecillinam-resistant, round morphological mutants of Escherichia coli. J Bacte- riol 133, 196–202. Jagura-Burdzy, G., and Hulanicka, D. (1981). Use of gene fusions to study expression of cysB, the regulatory gene of the cysteine regulon. J Bacteriol 147, 744– 751. Jones, R.N. (1988). In vitro evaluations of aminopenicillin/beta-lactamase inhibitor combinations. Drugs 35 Suppl 7, 17–26. Joseleau-Petit, D., Thévenet, D., and D'Arl, R. (1994). ppGpp concentration, growth without PBP2 activity, and growth‐rate control in Escherichia coli. Mol Micro- biol 13, 911–917. Kaas, R.S., Friis, C., Ussery, D.W., and Aarestrup, F.M. (2012). Estimating variation within the genes and inferring the phylogeny of 186 sequenced diverse Escherichia coli genomes. BMC Genomics 13, 577. Kahlmeter, G., Åhman, J., and Matuschek, E. (2015). Antimicrobial resistance of Escherichia coli causing uncomplicated urinary tract infections: A European update for 2014 and comparison with 2000 and 2008. Infect Dis Ther 4, 417– 423. Kaper, J.B., Nataro, J.P., and Mobley, H.L.T. (2004). Pathogenic Escherichia coli. Nature Rev Microbiol 2, 123–140.

63 Kerrn, M.B., Frimodt-Møller, N., and Espersen, F. (2003). Urinary concentrations and urine ex-vivo effect of mecillinam and sulphamethizole. Clin Microbiol In- fect 10, 54–61. Kredich, N.M. (2008). Biosynthesis of cysteine. EcoSal Plus 2013; doi:10.1128 514–527. Krieg, N.R., Ludwig, W., Whitman, W., Hedlund, B.P., Paster, B.J., Staley, J.T., Ward, N., and Brown, D. (2011). Phylum XVI Tenericutes. Bergey's Manual of Systematic Bacteriology 4, 567–734. Kümmerer, K. (2009). Antibiotics in the aquatic environment – A review – Part I. Chemosphere 75, 417–434. Lai, G.C., Cho, H., and Bernhardt, T.G. (2017). The mecillinam resistome reveals a role for peptidoglycan endopeptidases in stimulating cell wall synthesis in Esch- erichia coli. PLoS Genet 13, e1006934. Laubacher, M.E., and Ades, S.E. (2008). The Rcs phosphorelay is a cell envelope stress response activated by peptidoglycan stress and contributes to intrinsic antibiotic resistance. J Bacteriol 190, 2065–2074. Lawrence, J. (1999). Selfish operons: the evolutionary impact of gene clustering in prokaryotes and eukaryotes. Curr Opin Genet Devel 9, 642–648. Lawrence, J.G. (2000). Clustering of antibiotic resistance genes: beyond the selfish operon. ASM News 66, 281–286. Lenski, R.E. (1998). Bacterial evolution and the cost of antibiotic resistance. Int. Microbiol. 1, 265–270. Levin, B.R., Perrot, V., and Walker, N. (2000). Compensatory mutations, antibiotic resistance and the population genetics of adaptive evolution in bacteria. Genet- ics 154, 985–997. Livermore, D.M. (1995). Beta-lactamases in laboratory and clinical resistance. Clin Microbiol Rev 8, 557–584. Llarrull, L.I., Testero, S.A., Fisher, J.F., and Mobashery, S. (2010). The future of the β-lactams. Curr Opin Microbiol 13, 551–557. Lund, F., and Tybring, L. (1972). 6β-amidinopenicillanic acids – a new group of antibiotics. Nature 236, 135–137. Madigan, M.T., Martinko, J.M., Dunlap, P.V., and Clark, D.P. (2008). Brock Biolo- gy of microorganisms 12th edn. Pearson Benjamin Cummings: San Francisco. Magnusson, L.U., Farewell, A., and Nyström, T. (2005). ppGpp: a global regulator in Escherichia coli. Trends Microbiol 13, 236–242. Masuda, N., Gotoh, N., Ishii, C., Sakagawa, E., Ohya, S., and Nishino, T. (1999). Interplay between chromosomal beta-lactamase and the MexAB-OprM efflux system in intrinsic resistance to beta-lactams in Pseudomonas aeruginosa. An- timicrob Agents Chemother 43, 400–402. Matsuzawa, H., Hayakawa, K., Sato, T., and Imahori K. (1973). Characterization and genetic analysis of a mutant of Escherichia coli K-12 with rounded morphology. J Bacteriol 115, 436–442. Matsuzawa, H., Asoh, S., Kunai, K., Muraiso, K., Takasuga, A., and Ohta, T. (1989). Nucleotide sequence of the rodA gene, responsible for the rod shape of Escherichia coli: rodA and the pbpA gene, encoding penicillin-binding protein 2, constitute the rodA operon. J Bacteriol 171, 558–560. Medeiros, A.A. (1997). Evolution and dissemination of beta-lactamases accelerated by generations of beta-lactam antibiotics. Clin Infect Dis 24, S19–S45. Meeske, A.J., Riley, E.P., Robins, W.P., Uehara, T., Mekalanos, J.J., Kahne, D., Walker, S., Kruse, A.C., Bernhardt, T.G., and Rudner D.Z. (2016). SEDS proteins are a widespread family of bacterial cell wall polymerases. Nature 537, 634–638.

64 Melnyk, A.H., Wong, A., and Kassen, R. (2014). The fitness costs of antibiotic resistance mutations. Evol Appl 8, 273–283. Miller, E.L. (2002). The penicillins: a review and update. J Midwifery Womens Health 47, 426–434. Monsen, T.J., Holm, S.E., Ferry, B.M., and Ferry, S.A. (2013). Mecillinam resistance and outcome of pivmecillinam treatment in uncomplicated lower urinary tract infection in women. APMIS 122, 317–323. Murray, P.R., Rosenthal, K.S., and Pfaller, M.A. (2008). Medical Microbiology 6th edition. Elsevier: Philadelphia. Navarro, F., Robin, A., D'Ari, R., and Joseleau-Petit, D. (1998). Analysis of the effect of ppGpp on the ftsQAZ operon in Escherichia coli. Mol Microbiol 29, 815–823. Newton, G.G., and Abraham, E.P. (1956). Isolation of cephalosporin C, a penicillin- like antibiotic containing D-alpha-aminoadipic acid. Biochem J 62, 651–658. Nicas, T.I., and Hancock, R.E. (1983). Pseudomonas aeruginosa outer membrane permeability: isolation of a porin protein F-deficient mutant. J Bacteriol 153, 281–285. Nicolle, L.E. (2000). Pivmecillinam in the treatment of urinary tract infections. J Antimicrob Chemother 46, 35–39. Nicolle, L.E. (2013). Urinary tract infection. Critical Care Clinics 29, 699–715. Nikaido, H. (2001). Preventing drug access to targets: cell surface permeability barriers and active efflux in bacteria. Sem Cell Develop Biol 12, 215–223. Nordmann, P., Poirel, L., Toleman, M.A., and Walsh, T.R. (2011). Does broad- spectrum β-lactam resistance due to NDM-1 herald the end of the antibiotic era for treatment of infections caused by Gram-negative bacteria? J Antimicrob Chemother 66, 689–692. Ohtsu, I., Wiriyathanawudhiwong, N., Morigasaki, S., Nakatani, T., Kadokura, H., and Takagi, H. (2010). The L-cysteine/L-cystine shuttle system provides reducing equivalents to the periplasm in Escherichia coli. J Biol Chem 285, 17479– 17487. Ohtsu, I., Kawano, Y., Suzuki, M., Morigasaki, S., Saiki, K., Yamazaki, S., Nonaka, G., and Takagi, H. (2015). Uptake of L-cystine via an ABC transporter contributes defense of oxidative stress in the L-cystine export-dependent manner in Escherichia coli. PLoS ONE 10, e0120619. Oppezzo, O.J., and Antón, D.N. (1995). Involvement of cysB and cysE genes in the sensitivity of Salmonella typhimurium to mecillinam. J Bacteriol 177, 4524– 4527. Paladino, J.A., Sunderlin, J.L., Singer, M.E., Adelman, M.H., and Schentag, J.J. (2008). Influence of extended-spectrum β-lactams on Gram-negative bacterial resistance. Am J Health Syst Pharm 65, 1154– 1159. Paradis-Bleau, C., Markovski, M., Uehara, T., Lupoli, T.J., Walker, S., Kahne, D.E., and Bernhardt, T.G. (2010). Lipoprotein cofactors located in the outer membrane activate bacterial cell wall polymerases. Cell 143, 1110–1120. Paterson, D.L., and Bonomo, R.A. (2005). Extended-spectrum β-lactamases: a clinical update. Clin Microbiol Rev 18, 657–686. Perry, J.A., and Wright, G.D. (2013). The antibiotic resistance “mobilome”: search- ing for the link between environment and clinic. Front Microbiol 4, 138. Peterson, L.R. (2006). Penicillins for treatment of pneumococcal pneumonia: does in vitro resistance really matter? Clin Infect Dis 42, 224–233.

65 Pinho, M.G., Filipe, S.R., de Lencastre, H., and Tomasz, A. (2001). Complementa- tion of the essential peptidoglycan transpeptidase function of penicillin-binding protein 2 (PBP2) by the drug resistance protein PBP2A in Staphylococcus aureus. J Bacteriol 183, 6525–6531. Poole, K. (2002). Outer membranes and efflux: the path to multidrug resistance in Gram-negative bacteria. Curr Pharm Biotechnol 3, 77–98. Potron, A., Poirel, L., and Nordmann, P. (2015). Emerging broad-spectrum resistance in Pseudomonas aeruginosa and Acinetobacter baumannii: Mecha- nisms and epidemiology. Int J Antimicrob Agents 45, 568–585. Prober, C.G. (1998). Cephalosporins: an update. Pediatr Rev 19, 118–127. Pupo, G.M., Lan, R., and Reeves, P.R. (2000). Multiple independent origins of Shi- gella clones of Escherichia coli and convergent evolution of many of their characteristics. Proc Natl Acad Sci USA 97, 10567–10572. Ramadhan, A.A. (2005). Survivability of vancomycin resistant enterococci and fitness cost of vancomycin resistance acquisition. J Clin Pathol 58, 744–746. Ramaswamy, S., and Musser, J.M. (1998). Molecular genetic basis of antimicrobial agent resistance in Mycobacterium tuberculosis: 1998 update. Tubercle and Lung Dise 79, 3–29. Ramirez, M.S., and Tolmasky, M.E. (2010). Aminoglycoside modifying enzymes. Drug Resist Upd 13, 151–171. Randall, C.P., Mariner, K.R., Chopra, I., and O'Neill, A.J. (2012). The target of daptomycin is absent from Escherichia coli and other Gram-negative pathogens. Antimicrob Agents Chemother 57, 637–639. Rasko, D.A., Rosovitz, M.J., Myers, G.S.A., Mongodin, E.F., Fricke, W.F., Gajer, P., Crabtree, J., Sebaihia, M., Thomson, N.R., Chaudhuri, R., et al. (2008). The pangenome structure of Escherichia coli: comparative genomic analysis of E. coli commensal and pathogenic isolates. J Bacteriol 190, 6881–6893. Reading, C. and Cole, M. (1976). Clavulanic Acid: a Beta-Lactamase-Inhibiting Beta-Lactam from Streptomyces clavuligerus. Antimicrob Agents Chemother 11, 852–857. Reddie, K.G., and Carroll, K.S. (2008). Expanding the functional diversity of proteins through cysteine oxidation. Curr Opin Chem Biol 12, 746–754. Roholt, K., Nielsen, B., and Kristensen, E. (1975). Pharmacokinetic studies with mecillinam and pivmecillinam. Chemotherapy 21, 146–166. Salvatore, S., Salvatore, S., Cattoni, E., Siesto, G., Serati, M., Sorice, P., and Torel- la, M. (2011). Urinary tract infections in women. Eur J Obstet Gynecol Reprod Biol 156, 131–136. Sauvage, E., and Terrak, M. (2016). Glycosyltransferases and transpeptidases/penicillin-binding proteins: valuable targets for new antibacterials. Antibiot- ics 5, 12. Sauvage, E., Kerff, F.D.R., Terrak, M., Ayala, J.A., and Charlier, P. (2008). The penicillin-binding proteins: structure and role in peptidoglycan biosynthesis. FEMS Microbiol Rev 32, 234–258. Scheffers, D.-J., and Pinho, M.G. (2005). Bacterial cell wall synthesis: new insights from localization studies. Microbiol Mol Biol Rev 69, 585–607. Schiffer, G., and Holtje, J.V. (1999). Cloning and characterization of PBP1C, a third member of the multimodular class A penicillin-binding proteins of Escherichia coli. J Biol Chem 274, 32031–32039. Sherman, D.R., Mdluli, K., Hickey, M.J., Arain, T.M., Morris, S.L., Barry, C.E., and Stover, C.K. (1996). Compensatory ahpC gene expression in isoniazid-resistant Mycobacterium tuberculosis. Science 272, 1641–1643.

66 Sirko, A., Hryniewicz, M., Hulanicka, D., and Böck, A. (1990). Sulfate and thiosulfate transport in Escherichia coli K-12: nucleotide sequence and expression of the cysTWAM gene cluster. J Bacteriol 172, 3351–3357. Snipen, L., Almøy, T., and Ussery, D.W. (2009). Microbial comparative pan- genomics using binomial mixture models. BMC Genomics 10, 385. Sommer, M.O.A., Munck, C., Toft-Kehler, R.V., and Andersson, D.I. (2017). Pre- diction of antibiotic resistance: time for a new preclinical paradigm? Nat Rev Microbiol doi: 10.1038/nrmicro.2017.75. Sonawane, V.C. (2008). Enzymatic modifications of cephalosporins by cephalosporin acylase and other enzymes. Critic Rev Biotechnol 26, 95–120. Sougakoff, W., and Jarlier, V. (2000). Comparative potency of mecillinam and other beta-lactam antibiotics against Escherichia coli strains producing different beta- lactamases. J Antimicrob Chemother 46 , 9–14. Spratt, B.G. (1975). Distinct penicillin binding proteins involved in the division, elongation, and shape of Escherichia coli K12. Proc Natl Acad Sci USA 72, 2999–3003. Spratt, B.G. (1977). The mechanism of action of mecillinam. J Antimicrob Chemother 3, 13–19. Stokes, H.W., and Gillings, M.R. (2011). Gene flow, mobile genetic elements and the recruitment of antibiotic resistance genes into Gram-negative pathogens. FEMS Microbiol Rev 35, 790–819. Sundqvist, M., Geli, P., Andersson, D.I., Sjolund-Karlsson, M., Runehagen, A., Cars, H., Abelson-Storby, K., Cars, O., and Kahlmeter, G. (2010). Little evidence for reversibility of trimethoprim resistance after a drastic reduction in trimethoprim use. J Antimicrob Chemother 65, 350–360. Sykes, R.B., Bonner, D.P., Bush, K., Georgopapadakou, N.H., and Wells, J.S. (1981). Monobactams – monocyclic beta-lactam antibiotics produced by bacteria. J Antimicrob Chemother 8, 1–16. Tamaki, S., Matsuzawa, H., and Matsuhashi, M. (1980). Cluster of mrdA and mrdB genes responsible for the rod shape and mecillinam sensitivity of Escherichia coli. J Bacteriol 141, 52–57. Tang, Y-W., Sussman, M., Liu, D., Poxton, I. and Schwartzman, J. (2015). Molecu- lar Medical Microbiology 2nd edition. Elsevier: London. Tenaillon, O., Skurnik, D., Picard, B., and Denamur, E. (2010). The population genetics of commensal Escherichia coli. Nat Rev Microbiol 8, 207–217. Thaker, M., Spanogiannopoulos, P., and Wright, G.D. (2009). The tetracycline resistome. Cell Mol Life Sci 67, 419–431. Thomas, C.M., and Nielsen, K.M. (2005). Mechanisms of, and barriers to, horizontal gene transfer between bacteria. Nat Rev Microbiol 3, 711–721. Tipper, D.J., Strominger, J.L. (1965). Mechanism of action of penicillins: A pro- posal based on their structural similarity to acyl-D-alanyl-D-alanine. Proc Natl Acad Sci 54, 1133–1141. Traxler, M.F., Summers, S.M., Nguyen, H.-T., Zacharia, V.M., Hightower, G.A., Smith, J.T., and Conway, T. (2008). The global, ppGpp-mediated stringent response to amino acid starvation in Escherichia coli. Mol Microbiol 68, 1128– 1148. Tybring, L., and Melchior, N.H. (1975). Mecillinam (FL 1060), a 6β- amidinopenicillanic acid derivative: bactericidal action and synergy in vitro. CORD Conf Proc 8, 271–276. Typas, A., Banzhaf, M., Gross, C.A., and Vollmer, W. (2011). From the regulation of peptidoglycan synthesis to bacterial growth and morphology. Nat Rev Micro- biol 10, 123–136.

67 Typas, A., Banzhaf, M., van den Berg van Saparoea, B., Verheul, J., Biboy, J., Nichols, R.J., Zietek, M., Beilharz, K., Kannenberg, K., Rechenberg, von, M., et al. (2010). Regulation of peptidoglycan synthesis by outer-membrane proteins. Cell 143, 1097–1109. Tängden, T., Adler, M., Cars, O., Sandegren, L., and Lowdin, E. (2013). Frequent emergence of porin-deficient subpopulations with reduced carbapenem susceptibility in ESBL-producing Escherichia coli during exposure to ertapenem in an in vitro pharmacokinetic model. J Antimicrob Chemother 68, 1319–1326. Unhanand, M., Mustafa, M.M., McCracken, G.H., and Nelson, J.D. (1993). Gram- negative enteric bacillary meningitis: a twenty-one-year experience. J Pediatr 122, 15–21. van der Ploeg, J., Eichhorn, E., and Leisinger, T. (2001). Sulfonate-sulfur metabolism and its regulation in Escherichia coli. Arch Microbiol 176, 1–8. Vinella, D., and D'Ari, R. (1994). Thermoinducible filamentation in Escherichia coli due to an altered RNA polymerase beta subunit is suppressed by high levels of ppGpp. J Bacteriol 176, 966–972. Vinella, D., Cashel, M., and D'Ari, R. (2000). Selected amplification of the cell division genes ftsQ-ftsA-ftsZ in Escherichia coli. Genetics 156, 1483–1492. Vinella, D., Gagny, B., Joseleau-Petit, D., D'Ari, R., and Cashel, M. (1996). Mecil- linam resistance in Escherichia coli is conferred by loss of a second activity of the AroK protein. J Bacteriol 178, 3818–3828. Vinella, D., D’Ari, R., Jaffe, A., and Bouloc, P. (1992). Penicillin binding protein 2 is dispensable in Escherichia coli when ppGpp synthesis is induced. The EMBO J 11, 1493. Vinella, D., Joseleau-Petit, D., Thévenet, D., Bouloc, P., and D’Ari, R. (1993). Peni- cillin-binding protein 2 inactivation in Escherichia coli results in cell division inhibition, which is relieved by FtsZ overexpression. J Bacteriol 175, 6704– 6710. Wachi, M., Doi, M., Tamaki, S., Park, W., Nakajima-Iijima, S., and Matsuhashi, M. (1987). Mutant isolation and molecular cloning of mre genes, which determine cell shape, sensitivity to mecillinam, and amount of penicillin-binding proteins in Escherichia coli. J Bacteriol 169, 4935–4940. Waksman, S.A. (1947). What is an antibiotic or an antibiotic substance? Mycologia 39, 565–569. Walsh, C. (2003). Antibiotics: actions, origins, resistance. ASM Press: Washington, DC. Walsh, C.T., Fisher, S.L., Park, I.-S., Prahalad, M., and Wu, Z. (1996). Bacterial resistance to vancomycin: five genes and one missing hydrogen bond tell the story. Chemistry & Biology 3, 21–28. Webber, M.A., and Piddock, L.J.V. (2002). The importance of efflux pumps in bacterial antibiotic resistance. J Antimicrob Chemother 51, 9–11. WHO (2017). Prioritization of pathogens to guide discovery, research and development of new antibiotics for drug-resistant bacterial infections, including tuberculosis. Geneva: World Health Organization; 2017 (WHO/EMP/IAU/2017.12). Yates, C.M., Shaw, D.J., Roe, A.J., Woolhouse, M.E.J., and Amyes, S.G.B. (2006). Enhancement of bacterial competitive fitness by apramycin resistance plasmids from non-pathogenic Escherichia coli. Biol Lett 2, 463–465. Yordanov, D., and Strateva, T. (2009). Pseudomonas aeruginosa – a phenomenon of bacterial resistance. J Med Microbiol 58, 1133–1148. Yousif, S.Y., and Broome-Smith, J.K. (1985). Lysis of Escherichia coli by β-lactam antibiotics: deletion analysis of the role of penicillin-binding proteins 1A and 1B. J Gen Microbiol 131, 2839–2845.

68 Zhen, W., Schorr, C., Hunter, K., and Dellinger, P. (2010). Contrasting treatment and outcome of septic shock: presentation on hospital floors versus emergency department. Chin Med J (Engl) 123, 3550–3553. Zhu, L., Lin, J., Ma, J., Cronan, J.E., and Wang, H. (2010). Triclosan resistance of Pseudomonas aeruginosa PAO1 is due to FabV, a triclosan-resistant enoyl-acyl carrier protein reductase. Antimicrob Agents Chemother 54, 689–698.

69 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1375 Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-330856 2017