Antimicrobial resistance (AMR) is one of the largest threats facing modern-day healthcare and society in the coming decades. AMR genes are widely disseminated on genetic vehicles called plasmids, leading to resistant bacteria in many environments. Development of new antibiotics is inefficient, and stewardship of existing antibiotics is often ineffective. One promising novel approach to reduce AMR in bacteria is the delivery of genes coding for CRISPR-Cas9, which can specifically cleave a target sequence of choice – and in this way can be utilised to kill bacteria or remove their resistance plasmids. The general concept of such CRISPR delivery tools has been proven to be effective under laboratory conditions, however antibiotic resensitisation is more complex when targeting natural plasmids in mixed microbial communities. In this thesis, I aimed to develop a CRISPR delivery tool that can reach various species of bacteria embedded in microbial communities and resensitise these to antibiotics, allowing successful treatment using existing antibiotic drugs. In the first chapter, I reviewed the role which plasmids play in the AMR crisis by horizontal transfer of resistance genes. I summarised various approaches of counteracting this, with a focus on CRISPR-mediated AMR plasmid removal. In the second chapter, I engineered a broad host-range plasmid pKJK5 to encode CRISPR-Cas9 (pKJK5::Cas). I showed that this plasmid can be used to block target AMR plasmid uptake in Escherichia and Pseudomonas isolates. In the third chapter, I utilised pKJK5::Cas’ conjugative ability to remove a target AMR plasmid from recipient bacteria, which depended on pKJK5::Cas conjugation efficiency and CRISPR targeting efficiency. In the fourth chapter, I investigated removal of the broad host-range conjugative plasmid RP4 by pKJK5::Cas. I found that presence of toxin-antitoxin systems and target plasmid incompatibility can interfere with the use of pKJK5::Cas. In the fifth chapter, I assayed pKJK5::Cas transfer and maintenance in a synthetic bacterial community. Surprisingly, pKJK5::Cas maintenance and fitness of its host was dependent on community context where the plasmid became lost from a Variovorax host strain in presence of Stenotrophomonas growth partners. Finally, I offer concluding remarks on my data where I speculated under which conditions target plasmid removal may be successful in such a community context.Medical Research Council (MRC)