Refine
Document Type
- Doctoral Thesis (2)
Language
- English (2)
Has Fulltext
- yes (2)
Is part of the Bibliography
- no (2)
Institute
- Pharmazie (2) (remove)
LmrA is a member of the ATP Binding Cassette (ABC) transporter family of membrane proteins and a structural and functional homologue of P-glycoprotein1, 2. ABC-transporters share a common architecture of two transmembrane domains and two nucleotide binding domains. The NBDs are highly conserved in this transporter family whereas the TMDs are highly diverse3. The TMDs recognize the substrate and the NBDs bind and hydrolyze ATP and thus contribute the energy for substrate translocation. ABC transporters as a protein family transport a high number of substrates including peptides, nutrients, ions, bile acids, lipids and other lipophilic compounds. LmrA is a multidrug transporter that recognizes a number of hydrophobic substrates including fluorescent dyes and antibiotics1, 4-6. LmrA is a native protein of the gram-positive bacterium Lactococcus lactis. In this thesis, L. lactis was used as a homologous expression host for the preparation of LmrA for a variety of experiments. Wildtype LmrA as well as a number of cysteine mutants were successfully expressed in L. lactis, purified and subsequently characterized by a variety of biochemical assays (Chapter 4). LmrA can be expressed to very high amounts in L. lactis. The purification and reconstitution were optimized for the requirements of solid-state NMR experiments in this thesis. For the first time, an ABC transporter has been reconstituted in synthetic lipids to a ratio of up to 1:150 (mol/mol). LmrA was shown to be active under magic angle spinning conditions with these reconstitution ratios. By taking advantage of the slower ATP hydrolysis by LmrA ΔK388 (lysine deletion in the Walker A motif), a real-time 31P solid-state NMR ATPase assay was established (Chapter 5). This assay allowed, for the first time, the investigation of all phosphor nuclei during the ATP hydrolysis cycle of a membrane protein simultaneously and in real time7. This assay has been successfully adapted to investigate both ATP hydrolysis and substrate phosphorylation of diacylglycerol kinase (together with S. Wollschlag) and ATP hydrolysis at high temperatures of the thermophilic ABC transporter ABC1 from Thermos thermophilus (together with A. Zutz). In the course of this thesis, the gene for LmrA has been cloned into expression vectors suitable for Escherichia coli and the heterologous expression of LmrA was established (Chapter 4). The functionality of the heterologously expressed protein has been investigated and compared to L. lactis LmrA. In these experiments, LmrA was shown to yield a distinct multidrug resistance phenotype in its E. coli host and to show secondary active multidrug transport in the absence of ATP and presence of a proton gradient [Hellmich et al, in prep] (Chapter 4). Previously, it had been shown that LmrA acts as a seconadary active transporter when the NBDs are truncated8. The overexpression in minimal and defined medium and the purification of LmrA from E. coli have been optimized. Isotope labeling for ssNMR has been established and the first multinuclear ssNMR experiments have been carried out on a functional ABC transporter (Chapter 8). ABC transporters couple two cycles: upon ATP binding, the NBDs dimerize, hydrolyze the ATP, subsequently release Pi and ADP and finally dissociate. During this cycle, conformational changes are relayed to the TMDs which utilize the energy from ATP binding and/or hydrolysis to translocate the respective substrate. The prehydrolysis state can be trapped by beryllium fluoride, whereas the post-hydrolysis state of this cycle can be trapped by vanadate9-12. Trapping protocols for these reagents were successfully established for LmrA in this thesis (Chapter 4). This allowed for the investigation of different catalytic states by both ssNMR and EPR. A general 19F labeling protocol for membrane proteins has been established in the course of this thesis and successfully applied to proteorhodopsin (together with N. Pfleger)13 and LmrA (chapter 6). Single cysteine mutants of LmrA that line out the dimer interface have been labeled with a fluorine label for ssNMR. In the apo state, the 19F labeling indicates highly flexible transmembrane domains, a finding that is supported by 13C ssNMR and EPR measurements. The addition of drugs has a different effect on different positions within the LmrA dimer, therefore indicating that different drugs are recognized at a different position within the protein. For P-glycoprotein and LmrA it has been previously shown by biochemical methods that different drug binding sites co-exist. For a 19F label attached at position 314 (LmrA E314C), the spectra showed two distinct peaks with similar populations. This could hint towards a structural asymmetry within the LmrA dimer that might also be reflected in the alternating ATP hydrolysis at the NBDs. E314 has been specifically implicated with drug transport. Thus, structural asymmetry at this position might be functionally relevant for guiding a substrate through the transporter. Structural asymmetry within a homodimeric ABC transporter has also been shown for BtuCD, the E. coli vitamin B12 importer14. In addition, the conserved glutamates in EmrE, a small multidrug resistance protein, were shown to be asymmetric in the drug bound state15. Both, uniformly 13C/15N labeled as well as selectively amino acid type labeled LmrA has been investigated in different conformational states. Interestingly, significant dynamic changes in the b-sheet regions of LmrA (confined to the NBDs) were observed in the pre-hydrolysis (beryllium fluoride) and transition state (vanadate trapped) state. These were interpreted as the transition from a domain in fast conformational exchange in the apo state to one of intermediate exchange in the nucleotide bound state. A significant change in NBD mobility upon nucleotide binding was previously also shown with 2H ssNMR on LmrA16. By EPR it was shown that LmrA in both the vanadate and BeFx trapped states displays a significantly higher rigidity and therefore defined distances, whereas the apo state resembled a “floppy” protein with no preferred distance distribution. This concurs with data obtained from 19F ssNMR with fluorine labeled single-cysteine mutants. Here, in agreement with the EPR data, a higher label (and possibly) protein mobility was observed in the apo state displaying rather broad line widths. Upon trapping with vanadate, the line widths of the majority of fluorine-labeled mutants decreased due to an enhanced protein rigidity and a more homogenous environment of the fluorine labels. A similar observation was made when increasing the temperature that can be explained due to higher protein flexibility at increased temperatures. Solution NMR was employed to investigate the isolated soluble NBD of LmrA (Chapter 9). First 2D and 3D spectra were successfully obtained and could be utilized for a preliminary assignment of a significant fraction of residues. Additionally, binding of ATP and ADP in absence and presence of magnesium was investigated. Finally, the effects of peptides emulating the coupling helices of the full-length transporter on the soluble NBD were investigated. Strikingly, binding of one of these peptides only occurred in the presence of nucleotides (whereas the other showed no binding at all) hinting towards a tightly coupled regulation of the NBD and TMD during the substrate translocation/ATP hydrolysis cycle based on nucleotide binding.
The role of the Ca2+-dependent protease calpain in the diabetes-associated platelet hyperreactivity
(2012)
Platelets from diabetic patients are characterised by hyperreactivity resulting in exaggerated adhesion, aggregation and thrombus formation which contribute to the development of cardiovascular complications known to be one of the main causes of diabetes-related mortality. One of the mechanisms suggested to be involved in the diabetes-related platelet hyperactivation is the increased [Ca2+]i which leads to the overactivation of Ca2+-dependent proteases, the calpains. Among the calpain isoforms expressed in platelets the two ubquitiously expressed μ- and m-calpain are thought to play an important role in physiological and pathophysiological processes. Particularly μ-calpain is known to be involved in many steps of physiological platelet activation such as aggregation, adhesion, secretion, and signalling. However, we could show that diabetes was associated with an enhanced activation of both μ- and m-calpain in platelets
In the first part of the study we focussed on the characterization of the molecular mechanism regulating calpain activity. Indeed, although Ca2+ is considered to be the main regulator of the proteolytic activity of the conventional calpains, other mechanisms such as the presence of phospholipids and phosphorylation have been reported to affect their activity. Since most studies reported the phosphorylation of m-calpain we were interested to see whether μ-calpain activity might be also affected by phosphorylation. We could show that the activity of μ-calpain was enhanced by the PKC activator PMA suggesting its possible regulation by phosphorylation. However, whether PKC directly targeted μ-calpain remains unclear. Given that substrate recognition is important for a protease to process its substrate and since no common consensus could be attributed to calpain substrates, our next interest was to understand the mechanism regulating the recognition of its substrates by calpain. Since phosphorylation has been reported to protect different proteins from calpain degradation we investigated whether the calpain substrate CD31 could be phosphorylated in platelets and whether this could affect its recognition by calpain. Although we could show that the tyrosine phosphorylation of CD31 was increased after activation of platelets by thrombin and that this effect was attenuated in platelets from diabetic patients, tyrosine phosphorylation of CD31 seemed to have no effect on its sensitivity to calpain-mediated proteolysis.
After the analysis of the mechanism regulating calpain activity as well as its interaction with its substrates, our next interest was the identification of new calpain substrates in platelets. Since a previous study from our group showed that PPARγ agonists could indirectly reverse the diabetes-associated calpain activation we performed DIGE analysis of platelet samples from diabetic patients before and after PPARγ agonist treatment. Using this approach we could identify four novel calpain substrates in platelets: Integrin-linked kinase (ILK), α parvin, CLP36 and septin-5. Next, we assessed the effect of calpain-mediated cleavage on the function of these newly identified proteins. We could show that μ-calpain was essential for the dissociation of ILK from the IPP complex and its activation while m-calpain-mediated cleavage led to its cleavage and inactivation. Functionally, we also showed that μ-calpain was involved in platelet adhesion while m-calpain was important for spreading.
The next protein we analysed was septin-5, a small GTPase known to regulate platelet degranulation by association with other septins and syntaxin-4. We found that the interaction between septin-5 and syntaxin-4 was inhibitory for platelet degranulation. We could demonstrate that the μ-calpain-mediated cleavage dissociated septin-5 from syntaxin 4 and led to increased secretion of platelet α-granules. Next, we investigated the in vivo role of calpain in the diabetes-associated platelet hyperreactivity. We induced diabetes in mice and could reproduce calpain activation in platelets such as that found in human. Indeed, calpain activation in murine platelets also led to the cleavage of several calpain substrates including ILK and septin-5. Moreover, platelets from diabetic mice demonstrated an increased aggregation and thrombus formation in vivo. Treatment of the animals with the calpain inhibitor A-705253 (30 mg/kg/day for 10 days) significantly restored platelet function and substrate cleavage. In conclusion, in this part of the study, we could show that the increased calpain-dependent α-granule secretion and platelet adhesion may account for the enhanced vascular proliferation and thrombus formation in diabetes and calpain inhibition represents a promising way to prevent atherothrombosis development.
In the last part of the study we analysed another enzyme known to play a crucial role in diabetes, the AMPK which is an energy-sensing kinase known to be impaired in diabetes. We could show that the two catalytic subunits AMPK α1 and α2 are expressed in platelets. The AMPKα2 seemed to be the subunit involved in platelet activation since AMPKα2-deficient mice demonstrated a defect in clot retraction and the stabilization of the thrombus while the animals showed a normal bleeding time. Mechanistically, we showed in platelets that the upstream kinase of AMPKα2 is LKB1 which was activated by thrombin stimulation via a PI-3K-dependent pathway. AMPKα2 then phosphorylated the Src-family kinase Fyn, which is responsible for the phosphorylation of its substrate β3 integrin on Tyr747. These data indicate that AMPKα2, by affecting Fyn phosphorylation and activity, plays a key role in platelet αIIbβ3 integrin signalling, leading to clot retraction and thrombus stability. Although the effect of diabetes in the AMPK-dependent pathway could not be investigated we assume that the dysregulation of this pathway may account for the thrombus destabilization and enhanced embolization encountered in diabetes.