Role of Rab5 in synaptic vesicle recycling

During synaptic transmission, NT-filled synaptic vesicles are released by Ca2+-triggered exocytosis at the active zone. Following exocytosis, SV membrane is immediately re-internalized and synaptic vesicles (SVs) are regenerated by a local recycling mechanism within the presynaptic terminal. It is debated whether an endosomal compartment is involved in this recycling process. In contrast, it is well known from cultured mammalian cells, that endocytic vesicles fuse to the early sorting endosome. The early endosome is a major sorting station of the cell where cargo is send into the degradative pathway to late endosome and lysosome or towards recycling. Each trafficking step is mediated by a certain protein of the Rab family. Rab proteins are small GTPases belonging to the Ras superfamily. They accumulate at their target compartments and have thereby been used as markers for the different endocytic organelles in cultured mammalian cells. Rab5 controls trafficking from the PM to the early endosome and has thereby been used as marker for this compartment. A second marker is based on the specific binding of the FYVE zinc finger protein domain to the lipid PI(3)P that is specifically generated at the early endosomal membrane. This study used the Drosophila NMJ as a model system to investigate the SV recycling process. In particular, three questions were addressed: First, is an endosomal compartment present at the synapse? Second, do SVs recycle through an endosome? Third, is Rab5 involved in SV recycling? We used GFP fusions of Rab5 and 2xFYVE to visualize endosomal compartments at the presynaptic terminal of Drosophila third instar larval NMJs. Furthermore, the endosomes are located within the pool of recycling SVs, labeled with the styryl-dye FM5-95. Using the temperature-sensitive mutation in Dynamin, shibirets, we showed that SV recycling involves trafficking through an intermediate endosomal compartment. In cultured mammalian cells, interfering with Rab5 function by expressing the dominant negative version, Rab5SN causes the fragmentation of the endosome and the accumulation of endocytic vesicles. In contrast, when Rab5 is overexpressed enlarged endosomal compartments were observed. In Drosophila, the endosomal compartment was disrupted when loss of function and dominant negative mutants of Rab5 were expressed. In addition, at the ultrastructural we observed an accumulation of endocytic vesicles in Rab5S43N expressing terminals and enlarged endosomes when Rab5 was overexpressed. Furthermore, interfering with Rab5 function using the dominant negative Rab5S43N caused a decrease in the SV recycling kinetics as shown by FM1-43 experiments. In contrast, overexpression of Rab5 or GFP-Rab5 caused an increase in the FM1-43 internalization rate. Finally, standard electrophysiological techniques were used to measure synaptic function. We found that the Rab5-mediated endosomal SV recycling pathway generates vesicles with a higher fusion efficacy during Ca2+-triggered release, compared to SVs recycled when Rab5 function was impaired. We therefore suggest a model in which the endosome serves as organelle to control the SV fusion efficacy and thereby the synaptic strength. Since changes in the synaptic strength are occuring during learning and memory processes, controlling endosomal SV recycling might be a new molecular mechanism involved in learning and memory.

The periplasmic domain of the barrel assembly machinery protein A (BamA) from Escherichia coli assists folding of outer membrane protein A

The assembly of outer membranes of the cell wall of Gram-negative bacteria and of various organelles of eukaryotic cells requires the evolutionarily conserved β-barrel-assembly machinery (BAM) complex. This thesis describes the biochemical and biophysical properties of the periplasmic domain of the β-barrel assembly machinery protein A (PD-BamA) of the E. coli BAM complex, its effect on insertion and folding of the Outer membrane protein A (OmpA) into lipid bilayers and the identification of regions of PD-BamA that may be involved in protein-protein interactions. The secondary structure of PD-BamA in mixed lipid bilayers, analyzed by Circular dichroism (CD) spectroscopy, contained less β-sheet at an increased content of phosphatidylglycerol (PG) in the lipid membrane. This result showed membrane binding, albeit only in the presence of negatively charged lipids. Fluorescence spectroscopy demonstrated that PD-BamA only binds to lipid bilayers containing the negatively charged DOPG, confirming the results of CD spectroscopy. PD-BamA did not bind to zwitterionic but overall neutral lipid bilayers. PD-BamA bound to OmpA at a stoichiometry of 1:1. PD-BamA strongly facilitated insertion and folding of OmpA into lipid membranes. Kinetics of PD-BamA mediated folding of OmpA was well described by two parallel folding processes, a fast folding process and a slow folding process, differing by 2-3 orders of magnitude in their rate constants. The folding yields of OmpA depended on the concentration of lipid membranes and also on the lipid head groups. The presence of PD-BamA resulted in increased folding yields of OmpA in negatively charged DOPG, but PD-BamA did not affect the folding kinetics of OmpA into bilayers of zwitterionic but overall neutral lipids. The efficiency of folding and insertion of OmpA into lipid bilayers strongly depended on the ratio PD-BamA/OmpA and was optimal at equimolar concentrations of PD-BamA and OmpA. To examine complexes of unfolded OmpA with PD-BamA in more detail, site-directed spectroscopy was used to explore contact regions in both, PD-BamA and OmpA. Similarly, contact regions were also investigated for another protein complex formed by PD-BamA and the lipoprotein BamD. The obtained data suggest, that the site of interaction on PD-BamA for OmpA might be oriented towards the exterior environment away from the preceding POTRA domains, but that PD-BamA is oriented with its short α-helix α1 of POTRA domain 5 towards the C-terminal end of BamD.

On the function of the Dictyostelium Argonaute A protein (AgnA) in epigenetic gene regulation

With molecular biology methods and bioinformatics, the Argonaute proteins in Dictyostelium discoideum were characterized, and the function of the AgnA protein in RNAi and DNA methylation was investigated, as well as cellular features. Also interaction partners of the PAZ-Piwi domain of AgnA (PAZ-PiwiAgnA) were discovered. The Dictyostelium genome encodes five Argonaute proteins, termed AgnA/B/C/D/E. The expression level of Argonaute proteins was AgnB/D/E > AgnA > AgnC. All these proteins contain the characteristic conserved of PAZ and Piwi domains. Fluorescence microscopy revealed that the overexpressed C-terminal GFP-fusion of PAZ-PiwiAgnA (PPWa-GFP) localized to the cytoplasm. Overexpression of PPWa-GFP leaded to an increased gene silencing efficiency mediated by RNAi but not by antisense RNA. This indicated that PAZ-PiwiAgnA is involved in the RNAi pathway, but not in the antisense pathway. An analysis of protein-protein interactions by a yeast-two-hybrid screen on a cDNA library from vegetatively grown Dictyostelium revealed that several proteins, such as EF2, EF1-I, IfdA, SahA, SamS, RANBP1, UAE1, CapA, and GpdA could interact with PAZ-PiwiAgnA. There was no interaction between PAZ-PiwiAgnA and HP1, HelF and DnmA detected by direct yeast-two-hybrid analysis. The fluorescence microscopy images showed that the overexpressed GFP-SahA or IfdA fusion proteins localized to both cytoplasm and nuclei, while the overexpressed GFP-SamS localized to the cytoplasm. The expression of SamS in AgnA knock down mutants was strongly down regulated on cDNA and mRNA level in, while the expression of SahA was only slightly down regulated. AgnA knock down mutants displayed defects in growth and phagocytosis, which suggested that AgnA affects also cell biological features. The inhibition of DNA methylation on DIRS-1 and Skipper retroelements, as well as the endogenous mvpB and telA gene, observed for the same strains, revealed that AgnA is involved in the DNA methylation pathway. Northern blot analysis showed that Skipper and DIRS-1 were rarely expressed in Ax2, but the expression of Skipper was upregulated in AgnA knock down mutants, while the expression of DIRS-1 was not changed. A knock out of the agnA gene failed even though the homologous recombination of the disruption construct occurred at the correct site, which indicated that there was a duplication of the agnA gene in the genome. The same phenomenon was also observed in ifdA knock out experiments.

Structural insights into the allosteric activation mechanism of cGMP-dependent protein kinase I

Cyclic GMP-dependent protein kinase (PKG) is a key transducer in the NO-cGMP signaling pathway. In this line, PKG has been considered an important drug target for treating hypertensive cardiovascular and pulmonary diseases. However, the investigation of PKG’s allosteric activation mechanism has been hampered by a lack of structural information. One of the fundamental questions on the cGMP-dependent activation of PKG is how the enzyme can distinguish cGMP over cAMP and selectively respond to cGMP. To ensure proper signaling, PKG must have developed unique features to ensure its activation upon the right activation signal. In this thesis, the cGMP-selective activation mechanism of PKG was studied through determining crystal structures of three truncated constructs of the regulatory domain [CNB-A (92-227), CNB-B (271-369), and CNB-A/B (92-351)] of PKG Iβ in the absence or presence of cyclic nucleotides. Herein, two individual CNB domain structures with biochemical data revealed that the C-terminal CNB domain (CNB-B) is responsible for cGMP selectivity, while the N-terminal CNB-domain (CNB-A) has a higher binding affinity for both cGMP and cAMP without showing any selectivity. Based on these crystal structures, mutagenesis studies were performed in which the critical residues for cyclic nucleotide selectivity and activation were identified. Furthermore, we discovered that the conformational changes of the C-terminal helix of the CNB-B that bridges between the regulatory and catalytic domains including the hydrophobic capping interaction are crucial for PKG activation. In addition, to observe the global conformation of the activated R-domain, I solved a co-crystal structure of the CNB-A/B with cGMP. Although a monomeric construct was crystallized, the structure displays a dimer. Strikingly, the CNB-A domain and its bound cGMP provide a key interface for this dimeric interaction. Using small angle X-ray scattering (SAXS), the existence of the cGMP-mediated dimeric interface within the CNB domains was confirmed. Furthermore, measuring cGMP-binding affinities (EC50) of the dimeric interface mutants as well as determining activation constants (Ka) revealed that the interface formation is important for PKG activation. To conclude, this thesis study provides a new mechanistic insight in PKG activation along with a newly found interface that can be targeted for designing PKG-specific activity modulators.