Plasticity of the β-barrel assembly machinery investigated by NMR

Abstract

The ß-barrel assembly machinery (BAM) is a protein complex in the outer membrane of Gram-negative bacteria that mediates folding and insertion of outer membrane proteins (OMPs). These OMPs fulfill critical roles for survival of the bacteria, serving e.g. as porins that allow diffusion of nutrients into the periplasm or as virulence factors. The structures of the BAM components are known, yet it is unclear how BamA functions and what role the lipoproteins BamB-E play. Solid-state NMR (ssNMR) spectroscopy is very suitable to study membrane proteins in their native environment, i.e. in lipid bilayers. In this thesis, we described ssNMR studies on the dynamics of BamA and its interactions with the lipid bilayer and the lipoproteins BamCDE. BamA consists of a membrane-embedded ß-barrel domain and five periplasmic POTRA domains, which have been shown to play a role in substrate interactions. We discovered that POTRA 1-5 were relatively rigid in context of full-length BamA reconstituted in proteoliposomes and did not experience global motion on the µs-ms timescale or faster. In addition, BamA could accommodate to lipid bilayers of different hydrophobic thickness, whereas electron microscopy of the NMR samples revealed that BamA distorts the bilayer. To obtain more residue-specific information on the transmembrane (TM) domain of BamA, we described different reverse and forward labeling strategies leading to the first ssNMR chemical shift assignments for the membrane-embedded BamA TM domain. Interestingly, we succeeded to tentatively assign a number of residues from the long extracellular loop 6 which contains a conserved motif that is essential for ß-barrel assembly, demonstrating that it is relatively rigid. We went on to investigate local dynamics within the POTRA 5 (P5) domain, which is positioned at the interface of the BamA ß-barrel and the lipid bilayer and is required for interaction with BamD. We found that a highly conserved patch of residues on P5 experienced local conformational exchange on the µs-ms timescale. BamD is essential for viability in Escherichia coli and binds BamA through the P5 domain, but the molecular basis for the interaction has not been revealed. When we co-reconstituted BamA with the entire BamCDE lipid-anchored sub-complex, a strong interaction between the components was observed that was mediated by P5 residues that mapped to the dynamic POTRA 5 region identified before. Docking yielded several possible models for the BamA-BamD interaction in which electrostatic interactions played a dominant role. Finally, proton-detected methods bear the promise of higher sensitivity and resolution in ssNMR. We introduced a new approach termed “proton-clouds”, in which fully protonated amino acids are incorporated in a deuterated background. We showed that the proton line widths of proton-cloud ubiquitin were significantly better than for a fully protonated sample. In addition, we could employ the remaining proton-proton dipolar couplings for magnetization transfer by means of spin diffusion, which yielded a number of contacts among the proton-clouds. We also prepared a proton-cloud sample of membrane-embedded BamA and demonstrated efficient magnetization transfer. Perspectives for future research include substrate interaction studies and investigation of the BAM complex in its native cellular environment.