Logo: to the web site of Uppsala University

Antibiotic susceptible bacteria can develop resistance to antibiotics through different genetic pathways. Among the most important are mutations occurring within the genome, and the acquisition through horizontal genetic transfer (HGT) of additional genes that can reduce susceptibility to antibiotics, located in mobile genetic elements such as plasmids. The discovery of pathogenic strains of Escherichia coli and Klebsiella pneumoniae that have a chromosome with at least 20% (> 1 Mb) of DNA originating from a foreign strain suggest that an alternative pathway remains unexplored: the conjugative transfer of large areas of chromosomal DNA, generating bacterial strains with hybrid chromosomes. The projects presented in this thesis focused on studying different aspects about the generation of chromosomal hybrids.  

We first designed and implemented a conjugation method that allows for the selection of bacteria with hybrid chromosomes from large populations and found that a large proportion of clinical strains of E. coli are capable of mobilizing their chromosome and can generate hybrids. Genome sequencing showed that several Mb of foreign DNA could be integrated into the recipient’s chromosome. Multiple resistance genes and alleles could be acquired simultaneously and without direct selection are maintained stably by the hybrid. Furthermore, we showed that any one conjugative mobile element, be it plasmid or ICE, was itself sufficient to mobilize the chromosome. 

In the second project, we studied interspecies hybrids of K. pneumoniae clinical isolates and E. coli laboratory strains. We found chromosomal hybrids to be stable in the absence of selective pressure and fit for growth. 

Third, we initiated a study to identify biological barriers that inhibit the formation of chromosomal hybrids in nature. We used both phenotypic selection and transposon generated mutants of the clinical E. coli strain ATCC25922 to identify genes which, when inactivated, lower the conjugation barrier to the formation of hybrids.  

Finally, we explored whether E. coli could maintain prototrophy in the absence of HGT by acquiring compensatory mutations within its genome. We found that compensatory bypass mechanisms in mutants with gene deletions causing auxotrophy were accessible only to a handful of pathways for amino acid synthesis. 

These results highlight the importance of HGT in bacterial survival and evolution. 

Pathogenic multidrug-resistant bacteria with hybrid chromosomes have emerged as a significant global healthcare threat. These include the pandemic Escherichia coli ST1193, the product of homologous recombination events involving two phylogenetically distant strains of E. coli, in which mutant alleles of the widely separated genes, gyrA and parC, generating high-level fluoroquinolone resistance were acquired. The mechanisms and frequency of hybrid formation are poorly understood. We developed a robust hybrid selection procedure and applied it to 118 clinical UTI isolates of E. coli mixed with suitable recipient strains. Hybrids were selected from 39% of isolates. All hybrids were recombinants of donor and recipient chromosomal DNA (median length of donor DNA 367 kb), with 90% also acquiring conjugative mobile genetic elements (MGE) from the donor. We showed that individual conjugative plasmids, and integrative conjugative elements (ICE), from donors were sufficient to drive hybrid formation. These observations strongly support conjugative chromosomal DNA transfer as the major mechanism underlying hybrid formation. ICE are genome-integrated and passively propagated but when transferring to recipients they normally do so by excising and producing their own conjugation machinery. We found that ICE were responsible for the highest frequencies of hybrid chromosome formation. They could mobilize DNA around the full length of the chromosome, including the simultaneous acquisition of mutant variants of gyrA and parC, separated by ∼826 kb, generating highly fluoroquinolone-resistant bacteria in a single event. Bacterial hybrid chromosome formation driven by conjugative MGE may be an important and widespread mechanism in the emergence and evolution of high-risk bacterial pathogens.

Here, we describe the identification of an antibiotic class acting via LpxH, a clinically unexploited target in lipopolysaccharide synthesis. The lipopolysaccharide synthesis pathway is essential in most Gram-negative bacteria and there is no analogous pathway in humans. Based on a series of phenotypic screens, we identified a hit targeting this pathway that had activity on efflux-defective strains of Escherichia coli. We recognized common structural elements between this hit and a previously published inhibitor, also with activity against efflux-deficient bacteria. With the help of X-ray structures, this information was used to design inhibitors with activity on efflux-proficient, wild-type strains. Optimization of properties such as solubility, metabolic stability and serum protein binding resulted in compounds having potent in vivo efficacy against bloodstream infections caused by the critical Gram-negative pathogens E. coli and Klebsiella pneumoniae. Other favorable properties of the series include a lack of pre-existing resistance in clinical isolates, and no loss of activity against strains expressing extended-spectrum-beta-lactamase, metallo-beta-lactamase, or carbapenemase-resistance genes. Further development of this class of antibiotics could make an important contribution to the ongoing struggle against antibiotic resistance.

A new series of dual low nanomolar benzothiazole inhibitors of bacterial DNA gyrase and topoisomerase IV were developed. The resulting compounds show excellent broad-spectrum antibacterial activities against Gram-positive Enterococcus faecalis, Enterococcus faecium and multidrug resistant (MDR) Staphylococcus aureus strains [best compound minimal inhibitory concentrations (MICs): range, <0.03125–0.25 μg/mL] and against the Gram-negatives Acinetobacter baumannii and Klebsiella pneumoniae (best compound MICs: range, 1–4 μg/mL). Lead compound 7a was identified with favorable solubility and plasma protein binding, good metabolic stability, selectivity for bacterial topoisomerases, and no toxicity issues. The crystal structure of 7a in complex with Pseudomonas aeruginosa GyrB24 revealed its binding mode at the ATP-binding site. Expanded profiling of 7a and 7h showed potent antibacterial activity against over 100 MDR and non-MDR strains of A. baumannii and several other Gram-positive and Gram-negative strains. Ultimately, in vivo efficacy of 7a in a mouse model of vancomycin-intermediate S. aureus thigh infection was also demonstrated.

Background: The clinical-stage drug candidate EBL-1003 (apramycin) represents a distinct new subclass of aminoglycoside antibiotics for the treatment of drug-resistant infections. It has demonstrated best-in-class coverage of resistant isolates, and preclinical efficacy in lung infection models. However, preclinical evidence for its utility in other disease indications has yet to be provided. Here we studied the therapeutic potential of EBL-1003 in the treatment of complicated urinary tract infection and acute pyelonephritis (cUTI/AP).

Methods: A combination of data-base mining, antimicrobial susceptibility testing, time-kill experiments, and four murine infection models was used in a comprehensive assessment of the microbiological coverage and efficacy of EBL-1003 against Gram-negative uropathogens. The pharmacokinetics and renal toxicology of EBL-1003 in rats was studied to assess the therapeutic window of EBL-1003 in the treatment of cUTI/AP.

Findings: EBL-1003 demonstrated broad-spectrum activity and rapid multi-log CFU reduction against a phenotypic variety of bacterial uropathogens including aminoglycoside-resistant clinical isolates. The basicity of amines in the apramycin molecule suggested a higher increase in positive charge at urinary pH when compared to gentamicin or amikacin, resulting in sustained drug uptake and bactericidal activity, and consequently in potent efficacy in mouse infection models. Renal pharmacokinetics, biomarkers for toxicity, and kidney histopathology in adult rats all indicated a significantly lower nephrotoxicity of EBL-1003 than of gentamicin.

Interpretation: This study provides preclinical proof-of-concept for the efficacy of EBL-1003 in cUTI/AP. Similar efficacy but lower nephrotoxicity of EBL-1003 in comparison to gentamicin may thus translate into a higher safety margin and a wider therapeutic window in the treatment of cUTI/API.

Middle East respiratory syndrome coronavirus (MERS-CoV) causes a potentially fatal respiratory disease. Although it is most common in the Arabian Peninsula, it has been exported to 17 countries outside the Middle East, mostly through air travel. The Risk Assessment Guidelines for Infectious Diseases transmitted on Aircraft (RAGIDA) advise authorities on measures to take when an infected individual travelled by air. The aim of this systematic review was to gather all available information on documented MERS-CoV cases that had travelled by air, to update RAGIDA. The databases used were PubMed, Embase, Scopus and Global Index Medicus; Google was searched for grey literature and hand searching was performed on the EU Early Warning and Response System and the WHO Disease Outbreak News. Forty-seven records were identified, describing 21 cases of MERS that had travelled on 31 flights. Contact tracing was performed for 17 cases. Most countries traced passengers sitting in the same row and the two rows in front and behind the case. Only one country decided to trace all passengers and crew. No cases of in-flight transmission were observed; thus, considering the resources it requires, a conservative approach may be appropriate when contact tracing passengers and crew where a case of MERS has travelled by air.

Integrated conjugative elements can drive the transfer of chromosomal DNA from one Escherichia coli strain into another, generating strains with hybrid chromosomes. Several pandemic virulent multidrug-resistant strains of E. coli are chromosomal hybrids. Here we show that the frequency of hybrid formation can differ by orders of magnitude depending on the parental combination. However, the donor-recipient characteristics affecting hybridization efficiency are poorly understood. The laboratory workhorse strain E. coli MG1655 is significantly more proficient as a recipient in hybrid generation than E. coli ATCC25922, the strain used for quality control in antibiotic susceptibility testing. Focusing on these two strains we tested several hypotheses to identify barriers to hybrid formation. Among these, we found that relative conjugation efficiency correlated with hybrid formation frequency. To explore this, we used transposon mutagenesis to identify genes in ATCC25922 whose inactivation relieves a conjugation barrier. Among the first 1500 mutants tested we isolated and identified one that was 500-fold more conjugation-permissive than the parental strain. Whole genome sequencing identified this candidate ‘barrier’ gene as a protein-coding sequence on plasmid 3. Understanding the barrier mechanism will require further study. We are currently testing whether inactivation of this gene also increases chromosomal hybrid formation. The importance of this work is that it could significantly deepen our understanding of preferred pathways of HGT, something that will have broad application in both clinical and evolutionary biology.

Horizontal gene transfer (HGT) between bacterial species is an important contributor to the development of multidrug-resistant (MDR) clinical pathogens. One example is the transfer of MDR conjugative plasmids from Klebsiella pneumoniae into Escherichia coli occurring within co-infected human hosts. Here, we asked whether, in mixed-species bacterial populations, we could also detect the transfer of chromosomal DNA. To address this, we screened 50 K. pneumoniae clinical isolates for their ability to generate chromosomal hybrids with E. coli as recipient. We found that 7/50 (14%) generated chromosomal hybrids at frequencies above our limit of detection (~10-11) up to a frequency of 5x10-10. Genome sequence analysis of 28 hybrids revealed multiple underlying mechanisms of DNA transfer and recombination. These included two different mechanisms by which a region of the E. coli chromosome was replaced with K. pneumoniae chromosomal DNA (Hfr-like mechanisms, with up to 586 kb of K. pneumoniae DNA acquired), and three different mechanisms by which F’-like plasmids carrying K. pneumoniae chromosomal DNA were created and transferred into E. coli. Accordingly, clinical isolates of K. pneumoniae have the ability to transfer, not just plasmids, but also intrinsically non-mobile chromosomal DNA into E. coli to create novel hybrid strains. Several of the isolated hybrids acquired genes related to virulence or antibiotic resistance. Chromosomal hybrids were stable after evolution in rich media and had a 70-95% fitness compared to MG1655. The results provide insights into the mechanisms involved in creating bacteria with novel hybrid chromosomes, which may have important implications for bacterial evolution, and the spread of antibiotic resistance and virulence genes.

Escherichia coli is a prototroph and can synthesize all twenty proteinogenic amino acids when required to grow in minimal medium. There are approximately sixty protein-coding genes individually essential for amino acid synthesis. This is a large mutational target for the accumulation of detrimental mutations. E. coli can rewire biosynthetic pathways in response to mutational damage but the limits of this capacity are poorly understood. Here, to address evolutionary robustness, we asked whether and how the phenotypes of irreversible mutations causing auxotrophy could be suppressed or bypassed in the absence of horizontal gene transfer (HGT). Spontaneous suppressors could be selected for only ten of fifty-nine mutants tested (detection limit ~7x10^-11). Mechanisms of suppression included: regional amplifications; mutations increasing gene or operon expression; mutations relaxing enzyme specificity; and mutations causing biochemical pathway diversions. Overall, the data show that spontaneous suppression of auxotrophy caused by an irreversible mutation is an evolutionary survival mechanism relevant only to a minority of the genes essential for amino acid synthesis. As a consequence, the essential genetic foundations for amino acid prototrophy are expected to be degraded over time by mutations (Muller’s ratchet) and metabolic rewiring alone will be insufficient to counteract this effect. This implies that maintaining phenotypes, including prototrophy in E. coli, and potentially other bacterial species, is likely to be reliant on HGT of housekeeping genes to counteract the effects of inevitable mutational inactivation. Accordingly, chromosomal HGT in bacteria may be critical for survival across diverse environmental niches.