Date of Award

Summer 8-15-2015

Author's School

Graduate School of Arts and Sciences

Author's Department

Biology & Biomedical Sciences (Molecular Microbiology & Microbial Pathogenesis)

Degree Name

Doctor of Philosophy (PhD)

Degree Type

Dissertation

Abstract

Antibiotic resistance represents a grave threat to modern medicine’s control over infectious disease. Pathogens of the Enterobacteriaceae have proven particularly problematic as they can cause a wide variety of infections, and they can be, in some cases, resistant to all antibiotics recommended for use against them. A major part of the threat posed by the Enterobacteriaceae is their ability to exchange resistance genes by horizontal gene transfer (HGT). HGT has allowed some Enterobacteriaceae to quickly accumulate resistance against diverse antibiotics, and then to spread their resistance gene collection to other pathogenic strains. I explore three aspects of how HGT has affected the spread of resistance in the Enterobacteriaceae, with particular focus on the β-lactamase resistance genes: 1) how two, recently discovered β-lactamases have spread between different strains and species in the Enterobacteriaceae; 2) the extent to which other phyla in the microbiota can contribute new resistance genes to the Enterobacteriaceae; and 3) how whole genome sequencing can be used as a clinical diagnostic to detect antibiotic resistance in the Enterobacteriaceae.

There are two major methods by which a resistance gene may increase in prevalence: clonal expansion in a successful strain, or HGT between strains. To determine the contributions of each of these strategies to the success of the KPC and NDM-1 β-lactamases, we use whole genome sequencing and plasmid sequencing to identify the genetic context of the genes encoding each protein. We found that although these genes can sometimes be found in multiple members of the same clade, they are also encoded by much more distant strains and species. Similarly, we identified a few pairs of plasmids with high sequence identity where one carried NDM-1 and the other a KPC, but overall the similarity between any two plasmids was low, even between plasmids carrying the same β-lactamase. This suggests that HGT is playing a large role in the spread of these genes, and that it is mediated not by a single plasmid but by a diverse array of plasmids.

To determine the potential for the Enterobacteriaceae to gain antibiotic resistance genes from other members of the microbiota, we applied a parametric measure of HGT to all of the resistance genes from a set of 457 known microbiota strains. Although resistance genes were significantly more likely to have undergone HGT than phylogenetic marker genes, the distance between the two groups was small. We tested the ability of Escherichia coli, a member of the Enterobacteriaceae, to utilize resistance genes from other phyla using functional metagenomics, pulling genes from a collection of 76 known microbiota strains. We found only a few examples of genes from other phyla that could be used by E. coli, again suggesting a minor role for inter-phyla HGT in the antibiotic resistance of the Enterobacteriaceae.

One way to counter the spread of antibiotic resistance is through antibiotic stewardship coupled with active monitoring of resistance genes in pathogens. One barrier to this is that the current standard antibiotic resistance diagnostic in hospitals takes two days to determine the resistance of an isolate, and it does not identify the gene causing resistance for most patients. Whole genome sequencing has been proposed as an alternative method that could return results more quickly, and with gene information. We used a set of 78 Enterobacteriaceae clinical isolates to compare two approaches for determining antibiotic susceptibility of a pathogenic isolate from its genome sequence, a rules-based algorithm and a machine-learning algorithm. We found that both algorithms performed similarly, with approximately 90% of predictions matching what was found by in vitro phenotyping, but that the machine-learning algorithm showed potential for extension to more specific predictions.

In sum, I explore the scope of how HGT contributes to the increasing prevalence of antibiotic resistance in Enterobacteriaceae, and one method that can be used to boost containment efforts.

Language

English (en)

Chair and Committee

Gautam Dantas

Committee Members

Carey-Ann Burnham, Barak Cohen, Jeffrey Gordon, Scott Hultgren, David Wang,

Comments

Permanent URL: https://doi.org/10.7936/K7RJ4GND

Download

Included in

Biology Commons

Share

COinS