Membranes confer cells with individual identity and capacity to regulate their response to their environment. A critical aspect of having a membranous partition is the ability to transport substances into and out of cells as part of life-sustaining functions. In pathogenic bacteria, transporters aid infection and survival in the host. Two such transporters in Gram-negative bacterial species are the MacA-MacB-TolC (MacAB-TolC) antibiotic efflux pump and the Type I Secretion System (T1SS), responsible respectively for antibiotic resistance and export of protein virulence factors. To pass the Gram-negative envelope in a one-step translocation process, both machines use a tripartite system, consisting of outer membrane protein TolC, a periplasmic adapter protein (MacA or haemolysin D (HlyD) in the T1SS), and an inner membrane protein (MacB or haemolysin B (HlyB) in the T1SS). Both use the power of ATP-hydrolysis to export their substrates. Here, I utilise computational and experimental approaches to elucidate the mechanism of function for both machines. I conduct molecular dynamics (MD) simulations of membrane embedded HlyB component of the T1SS with and without its haemolysin A (HlyA) substrate as in silico experiments. I also conduct MD simulations with and without substrate for a related peptidase. I show that substrate recognition is via conserved charge-charge interactions. I also show that HlyB has an asymmetric preferential interaction with cardiolipin when its substrate is present, which is not seen in the peptidase simulations. I propose that this preference is part of the mechanism of transport, with cardiolipin providing energy via the proton-motive force. I test this hypothesis through flow cytometry detection of labelled substrate trapped T1SS in a mixed population of cells, by comparing parental MG1655 Escherichia coli with a cardiolipin deficient MG1655 strain. I found that the cardiolipin deficient strain has reduced T1SS levels compared to its parent. To aid structural studies, I optimise the expression of the T1SS using a flow cytometry based sequential design strategy where conditions are iteratively tested via detection of substrate trapped T1SS and updated until no more improvement can be made. I also test purification strategies for single-particle cryo-electron microscopy studies. Finally, I apply further bioinformatic approaches and synthesise my computational and experimental results to propose a mechanism of transport and suggest future experimental tests. I conduct MD simulations of MacB in membrane with and without a trapped lipid. I show that this trapped lipid locks MacB into an open state, allowing for substrate entry into the pump. I contextualise the results by comparing MD simulations to MacB-like structures and propose a revised mechanism of transport as a function of its free-energy landscape. Lastly, I explore the use of cryo-electron tomography (cryo-ET) as a method to obtain in vivo structural insights. I show that the use of “ghost” partially lysed E. coli can produce high-contrast specimens for tomography. I collect a tomographic dataset of “ghost” MacAB-TolC containing cells and apply subtomogram averaging. Preliminary results suggest that MacAB-TolC forms an array in cells, and that MacB is structurally flexible, likely in its nucleotide-binding domain. Together, these studies of the MacAB-TolC efflux pump and the T1SS shed light on their function and suggest new avenues of research to explore in order to fulfil the goal of finding novel inhibitors.