Doctoral Thesis
Refine
Year of publication
Document Type
- Doctoral Thesis (282) (remove)
Language
- English (282) (remove)
Has Fulltext
- yes (282)
Is part of the Bibliography
- no (282)
Keywords
- Membranproteine (6)
- NMR-Spektroskopie (5)
- RNA (4)
- ABC-Transporter (3)
- Biochemie (3)
- Crystallography (3)
- Kristallographie (3)
- Photosynthese (3)
- Proteinfaltung (3)
- Biochemistry (2)
Institute
- Biochemie und Chemie (282) (remove)
Nitric oxide (NO) represents a short-lived mediator that pivotally drives keratinocyte movements during cutaneous wound healing. In this study, we have identified p68 DEAD box RNA helicase (p68) from a NO-induced differential keratinocyte cDNA library. Subsequently, we have analyzed regulation of p68 by wound-associated mediators in the human keratinocyte cell line HaCaT. NO, serum, growth factors and pro-inflammatory cytokines were potent inducers of p68 expression in the cells. p68 was constitutively expressed in murine skin, but rapidly down-regulated upon injury. The down-regulation appeared to be transient, as p68 protein expression increased again after the inflammatory phase of repair. However, p68 protein expression did not completely disappear during wound inflammation, as immunohistochemistry and cell fractiona tion analysis revealed a restricted localization of p68 in keratinocyte nuclei of the developing epithelium. In line, cultured human (HaCaT) and murine (PAM 212) keratinocyte cell lines showed a nuclear localization of the helicase. Moreover, confocal microscopy revealed a strong localization of p68 protein within the nucleoli of the keratinocytes. Functional analyses demonstrated that p68 strongly participates in keratinocyte proliferation and gene expression. Keratinocytes that constitutively overexpressed p68 protein were characterized by a marked increase in serum-induced proliferation and vascular endothelial growth factor (VEGF) expression, whereas down-regulation of endogenous p68 using small interfering RNA (siRNA) markedly attenuated serum-induced proliferation and VEGF expression. Altogether, our results suggest a tightly controlled expression and nucleolar localization of p68 in keratinocytes in vitro and during skin repair in vivo that functionally contributes to keratinocyte proliferation and gene expression.
The detailed mechanism of the 20 S proteasome from Thermoplasma acidophilum is unknown. Substrates are degraded processively to small fragments without the release of intermediates, but the basis for this unique degradation mode remains obscure. The proteasome is a molecular machine, but how the different nanocompartments interplay and whether more than one substrate can be treated simultaneously has not been elucidated yet. To address these questions we had to disable the functionality of one aperture in order to dissect whether the other pore can compensate for the loss. As it is challenging to introduce mutations solely around one pore aperture of the highly symmetrical construct, we chose a novel approach by unique orientation of the proteasome at interfaces. For this purpose we purified recombinant 20 S proteasomes, where hexahistidine tags were fused either around the entrances or at the sides. According to electron microscopic studies we immobilized these constructs uniformly either end-on or side-on at metal-chelating interfaces (lipid vesicles, lipid monolayers and self-assembled thiol monolayers). Degradation of small fluorogenic peptides and large proteins like casein was analyzed. Small substrates were degraded with comparable activity by free and immobilized proteasomes, irrespective of their orientation. Thus it can be assumed that peptides can pass the sealed entrance of the 'dead-end' proteasome. However, larger substrates like fluorescently labeled casein were processed near the temperature optimum by side-on immobilized and soluble proteasomes with threefold activity compared to end-on immobilized proteasomes. Hence it can be concluded that one pore is sufficient for substrate entry and product release. In other words, the pore and antechamber can fulfil a triple function in the import and unwinding of substrates and the egress of products. With means of surface plasmon resonance the exact substrate/proteasome stoichiometry could be determined to ~1 for 'dead-end' proteasomes and ~2 for side-on immobilized (active and inactive) proteasomes. Most importantly, a fit with the Hill equation revealed positive cooperativity for side-on immobilized (Hill coefficient ~2) in contrast to end-on immobilized proteasomes (Hill coefficient ~1). Thus in case of soluble proteasomes two substrates bind presumably in opposite antechambers with positive cooperativity. The off-rate of casein as substrate is twofold for the active side-on immobilized proteasome in comparison to the end-on immobilized proteasome. The exact 2:1 stoichiometry of the off-rates equals the ratio of exit pathways amenable in case of side-on orientated versus 'dead-end' immobilized proteasomes. Thus crevices along the cylindrical body of the 20 S proteasome seem not to participate in the egress of small products. An inactive proteasome mutant displays a concentration-dependent off-kinetic against casein. Accordingly, the off-rate of the bisubstrate:proteasome complex can be attributed around half the value of the monosubstrate:proteasome complex. Consequently, substrates exit the inactive proteasome via the route of access due to obstruction of the trans side with an entering substrate. Hence the active proteasomes have to chop substrates down to small fragments prior to release through both pores. Thus the processive degradation mode might result from positive binding cooperativity. The on-rate constants for casein suggested that substrate association represents a two-step process comprising a rate-limiting translocation step and a fast binding step. As fluorescence cross-correlation revealed that two substrates can be co-localized in the proteasome and bind successively with increasing affinity (KD,1 = 8 µM versus KD,2 = 700 nM), an allosteric transition in the proteasome can be assumed. Combining our results with the data from other research groups led to a mechanistic model for the 20 S proteasome. Accordingly, the first substrate undergoes a slow translocation step, binds in the antechamber and diffuses subsequently to the catalytic centers, where it is degraded. By switching on the catalytic activity, the pores at both termini are dilated via conformational changes. Hence entry of the second substrate into the proteasome is facilitated due to omission of the rate-determining translocation step. The second substrate is either accommodated in the antechamber before it is processed (alternating degradation) or, most probably, is directly threaded into the central cavity (simultaneous degradation). As effusing peptides compete with entering proteins for binding in the antechamber, the pores are kept in an open state. After finishing digestion the pores are closed and a new degradation cycle can be reinitiated. In summary, substrate association with the proteasome underlies an ordered alternating binding mechanism in contrast to the random mode of degradation. Thus the two-stroke engine offers the advantage of speeding up degradation without enhancing complexity.
The transporter associated with antigen processing (TAP) plays a pivotal role in the adaptive immune response against virus-infected or malignantly transformed cells. As member of the ABC transporter family, TAP hydrolyzes ATP to energize the transport of antigenic peptides from the cytosol into the lumen of the endoplasmic reticulum. TAP forms a heterodimeric complex composed of TAP1 and TAP2 (ABCB2/3). Both subunits contain a hydrophobic transmembrane domain and a hydrophilic nucleotide-binding domain. The aim of this work was to study the ATP hydrolysis event of the TAP complex and gain further insights into the mechanism of peptide transport process. To analyze ATP hydrolysis of each subunit I developed a method of trapping 8- azido-nucleotides to TAP in the presence of phosphate transition state analogs followed by photocross-linking, immunoprecipitation, and high-resolution SDS-PAGE. Strikingly, trapping of both TAP subunits by beryllium fluoride is peptide-specific. The peptide concentration required for half-maximal trapping is identical for TAP1 and TAP2 and directly correlates with the peptide-binding affinity. Only background levels of trapping were observed for low affinity peptides or in the presence of the herpes simplex viral protein ICP47, which specifically blocks peptide binding to TAP. Importantly, the peptideinduced trapped state is reached after ATP hydrolysis and not in a backward reaction of ADP binding and trapping. In the trapped state, TAP can neither bind nor exchange nucleotides, whereas peptide binding is not affected. In summary, these data support the model that peptide binding induces a conformation that triggers ATP hydrolysis in both subunits of the TAP complex within the catalytic cycle. The role of the ABC signature motif (C-loop) on the functional non-equivalence of the NBDs was investigated. The C-loops of TAP transporter contain a canonical C-loop (LSGGQ) for TAP1 and a degenerated ABC signature motif (LAAGQ) for TAP2. Mutation of the leucine or glycine (LSGGQ) in TAP1 fully abolished peptide transport. TAP complexes with equivalent mutations in TAP2 showed however still residual peptide transport activity. To elucidate the origin of the asymmetry of the NBDs of TAP, we further examined TAP complexes with exchanged C-loops. Strikingly, the chimera with two canonical C-loops showed the highest transport rate whereas the chimera with two degenerated C-loops had the lowest transport rate, demonstrating that the ABC signature motifs control the peptide transport efficiency. All single-site mutants and chimeras showed similar activities in peptide or ATP binding, implying that these mutations affect the ATPase activity of TAP. In addition, these results prove that the serine of the C-loop is not essential for TAP function, but rather coordinates, together with other residues of the C-loop, the ATP hydrolysis in both nucleotide-binding sites. To study the coupling between the ATP binding/hydrolysis and the peptide binding, the putative catalytic bases of the TAP complex were mutated to generate the so-called EQ mutants. The mutations did not influence the peptide-binding ability. Dimerization of the NBDs of EQ mutants upon ATP binding does not alter the peptide binding property. At 27°C, both ATP and ADP could induce the loss of peptide-binding ability (Bmax) only in the variants bearing a mutated TAP2. Further studies are required to deduce at which stage in the catalytic cycle the peptide-binding site is affected. In addition, mutation of the putative catalytic base of both subunits showed a magnesium-dependent peptide transport activity, demonstrating these mutants did not abolish the ATP hydrolysis. Thus, the function of this acidic residue as the catalytic base is not likely to be universe for all ABC transporters.
Ligands of Iron-Sulphur Cluster N2: In this work the ubiquinone reducing catalytic core of NADH:ubiquinone oxidoreductase (complex I) from Y. lipolytica was studied by a series of point mutations replacing conserved histidines or arginines in the 49-kDa subunit. Although the missing 4th ligand of cluster N2 could not be found in the 49-kDa subunit of complex I, it was clearly demonstrated that iron-sulphur cluster N2 resides directly on the interface between the PSST and 49-kDa subunits. The results presented in this work show that residues in the 49-kDa subunit have strong influence on this redox centre and also on catalytic activity. The strong influence of Arg-141 and His-226 residues in 49-kDa subunit on this cluster can be deducted from complete loss of N2 signals in EPR spectra such as in case of mutants H226A and R141A. In the case of mutant H226M the EPR signal from cluster N2 was shifted and cluster N2 even lost the pH dependence of its redox midpoint potential and became more similar to the other so called 'isopotential' clusters. Specifically in the case of mutants R141M and R141K the characteristic signature of cluster N2 became undetectable in EPR spectra. However, specific dNADH:DBQ oxidoreductase activity that could be inhibited with the specific complex I inhibitors DQA and rotenone was not absolutely abolished but rather reduced. These reductions in complex I activity did not correspond to similar reductions in the specific EPR signal of cluster N2 as it was observed in the His-226 mutant series. No indications could be found that these mutations had modified the magnetic properties of cluster N2, resulting in different EPR spectra. From these observations it could be concluded that both mutants R141K and R141M virtually or entirely lack iron-sulphur cluster N2. The rates in complex I activity could be reconciled with electron transfer theory: After removal of a single redox centre in a chain, electron transfer rates are predicted to be still much faster than steady-state turnover of complex I. These results from mutants R141K, R141M and also the result from mutant H226M that protons are being pumped even if the redox midpoint potential of cluster N2 is not pH dependent questions the prominent role in the catalytic mechanism of complex I that has been ascribed to cluster N2. Histidine 91 and 95 were found to be absolutely essential for activity of complex I since in both mutants complex I was fully assembled and artificial NADH:HAR activity was parental whereas complex I specific dNADH:DBQ activity was abolished. The signal from cluster N2 in EPR spectra was parental for all His-91 and -95 mutants. Mutations at the C-terminal arginine 466 affected ubiquinone affinity and inhibitor sensitivity but also destabilised complex I. All these results provide further support for a high degree of structural conservation between the 49-kDa subunit of complex I and the large subunit of water soluble [NiFe] hydrogenases. Remodelling of Human Pathogenic 49-kDa Mutations in Y. lipolytica: Y. lipolytica has been proven a good system for studying complex I properties and thus also for studying defects that occur in humans. In this work pathogenic mutations in the 49-kDa subunit of complex I were recreated and studied. The P232Q mutant showed non-assembly of complex I and this is probably the cause why this mutation was lethal in patients. The mutants R231Q and S416P were parental for the content, artificial and also specific complex I activity, Km for DBQ and IC50 for DQA. From these results we can conclude that these two residues Arg-228 and Ser-413 in mammalian cells have specific structural importance for the 49-kDa subunit even if they are not directly involved in catalytic process.
The endothelin B receptor belongs to the rhodopsin-like G-protein coupled receptors family. It plays an important role in vasodilatation and is found in the membranes of the endothelial cells enveloping blood vessels. During the course of this work, the production of recombinant human ETB receptor in yeast, insect and mammalian cells was evaluated. A number of different receptor constructs for production in the yeast P. pastoris was prepared. Various affinity tags were appended to the receptor N-and C-termini to enable receptor detection and purification. The clone pPIC9KFlagHisETBBio, with an expression level of 60 pmol/mg, yielded the highest amount of active receptor (1.2 mg of receptor per liter of shaking culture). The expression level of the same clone in fermentor culture was 17 pmol/mg, and from a 10L fermentor it was possible to obtain 3 kg of cells that contained 20-39 mg of the receptor. For receptor production in insect cells, Sf9 (S. frugiperda) suspension cells were infected with the recombinant baculovirus pVlMelFlagHisETBBio. The peak of receptor production was reached at 66 h post infection, and radioligand binding assays on insect cell membranes showed 30 pmoL of active receptor /mg of membrane protein. Subsequently, the efficiency of different detergents in solubilizing the active receptor was evaluated. N-dodecyl-beta-D-maltoside (LM), lauryl-sucrose and digitonine/cholate performed best, and LM was chosen for further work. The ETB receptor was produced in mammalian cells using the Semliki Forest Virus expression system. Radioligand binding assays on membranes from CHO cells infected with the recombinant virus pSFV3CAPETBHis showed 7 pmol of active receptor /mg of membrane protein. Since the receptor yield from mammalian cells was much lower than in yeast and insect cells, this system was not used for further large-scale receptor production. After production in yeast and insect cells, the ETB receptor was saturated with its ligand, endothelin-1, in order to stabilize its native form. The receptor was subsequently solubilized with n-dodecyl-beta-D-maltoside and subjected to purification on various affinity matrices. Two-step affinity purification via Ni2+-NTA and monomeric avidin proved the most efficient way to purify milligram amounts of the receptor. The purity of the receptor preparation after this procedure was over 95%, as judged from silver stained gels. However, the tendency of the ETB receptor produced in yeast to form aggregates was a constant problem. Attempts were made to stabilize the active, monomeric form of the receptor by testing a variety of different buffer conditions, but further efforts in this direction will be necessary in order to solve the aggregation problem. In contrast to preparations from yeast, the purification of the ETB receptor produced in insect cells yielded homogeneous receptor preparations, as shown by gel filtration analysis. This work has demonstrated that the amounts of receptor expressed in yeast and insect cells and the final yield of receptor, isolated by purification, represent a good basis for beginning 3D and continuing 2D crystallization trials.
In the recent years, high-resolution conditions have been established in solid-state NMR by the combination of magic angle spinning, state-of-the-art r.f. pulse schemes and the introduction of ultra-high magnetic fields. Similar to what is now routine in solution-state NMR, this has opened the way for structure determination by HR-SSNMR methods. Complete structural or dynamical characterization of the biomolecule of interest is most easily achieved if multiple or even uniformly [13C, 15N]-labeled versions are studied. In a first step, experiments that allow the complete assignment of the 13C and 15N resonances have been recently designed. To date, nearly complete chemical shift assignments were reported for two well-ordered proteins, the ±-spectrin SH3 domain and the Crh protein. The SSNMR analysis of the later protein has been presented in Section 4.1. For SSNMR applications, not the molecular size or solubility, but the spectral resolution can be of crucial importance. Experimental parameters and sample inherent conditions such molecular disorder may reduce the overall spectral dispersion. In these circumstances, techniques that allow for spectral simplification without the need of elaborated biochemical procedures (of isotopelabeling) are of special importance. In Section 2, several spectral editing methods have been proposed. These methods not only select resonances due to changesin the physical and chemical environment of the nucleus but they can also directly probe molecular properties such as dynamics and conformational heterogeneity. Once the chemical shifts are available for the biomolecule of interest, methods that permit to obtain structural restraints can be applied. In the case of multiply isotope labeled proteins, such techniques can in principle result in multiple structural parameters. In Section 3.1, we have shown that, similar to solution-state NMR, secondary chemical shifts can be readily employed to study the local backbone conformation. Inaddition, distance constraints between protons may be encoded in high-resolution on rare spins like 13C and 15N and measured. Finally, carbon-carbon constraints may be probed by employing frequency selective r.f. pulse schemes. These dihedral and distance constraints may subsequently lead to the determination of protein secondary to tertiary structure from a single protein sample. In Section 4.2,we have shown that high-affinity ligand binding to membrane proteins can be investigated with solid-state NMR. Here, the neuropeptide neurotensin which binds to the Gprotein coupled receptor NTS1 in sub-nanomolar affinity was investigated.Except for the case of rhodopsin, there is currently no information on the high-resolution structure of any other GPCR or a corresponding high-affinity ligand.Our SSNMR results identify, for the first time, a distinct binding mode of neurotensin that could be of considerable relevance for further pharmacological studies. As exemplified in section 4.3, HR-SSNMR based structural studies can also assist in refining existing (X-ray or solution-state NMR) membrane-protein structures. The presented results provide, for the first time, direct experimental evidence for a double occupancy of the Q0 binding site in the ubiquinone-bc1 complex and may provide the basis for the complete 3D structural determination of the ubiquinone binding pocket. Advancements regarding sample preparation (for example, including modular labeling, in vitro expression and intein technology) and improvements in NMR hardware instrumentation could open up new areas of solid-state NMR research such as the investigation of large protein-protein complexes or the complete 3D characterization of larger membrane proteins. Solid-state NMR studies of multiply-labeled biomolecules will furthermore profit from improved procedures for calculating 3D structures, in particular in the presence of ambiguousor a limited number of structural constraints. Unlike X-ray crystallography, protein motion does not hinder solid-state NMR methods. In fact, complementary to solution-state NMR, it may provide a very efficient means to study protein folding, flexibility and function under biologically relevant conditions. Hand in hand with solution-state techniques and crystallographic methods, solid-state NMR could provide insight into protein function and the chemistry of life with unprecedented accuracy and flexibility.
Mitogen activated protein kinases (MAPKs) are found in all eukaryotic cells and represent crucial elements in the signal transduction from the plasma membrane to the nucleus. Although a broad variety of extracellular stimuli activate MAPKs, they evoke very distinct cellular responses. The amplitude and duration of MAPK activation determine signal identity and ultimately cell fate. A tight and finely tuned regulation is therefore critical for a specific cellular response. The role and the regulation of extracellular signal-regulated kinase 5 (ERK5), a MAPK with a large and unique C-terminal tail, were studied in different cellular systems. The study highlights two aspects of ERK5 regulation: control of the phosphorylation state and regulated protein stability. In analogy to other MAPKs ERK5 is activated by dual phosphorylation of threonine and tyrosine residues in its activation motif. A first part of the study concentrates on whether and how the protein tyrosine phosphatase PTP-SL is involved in the downregulation of the ERK5 signal. The direct interaction of both proteins is shown to result in mutual modulation of their enzymatic activities. PTP-SL is a substrate of ERK5 and, independent of its phosphorylation, binding to the kinase enhances its catalytic phosphatase activity. On the other hand, interaction with PTP-SL does not only downregulate enzymatic ERK5 activity but also effectively impedes its translocation to the nucleus. The second part of this study focuses on the interaction of ERK5 with c-Abl and its oncogenic variants Bcr/Abl and v-Abl. In this study these tyrosine kinases are demonstrated to regulate ERK5 by two mechanisms: first, by induction of kinase activity and secondly, by stabilisation of the ERK5 protein. Stabilisation involves the direct interaction of unique ERK5 domains with Abl kinases and is independent of MAPK cascade activation. The level of ERK5 and its intrinsic basal activity – rather than its activation – are essential for v-Abl-induced transformation as well as for survival of Bcr/Abl-positive leukaemia cells. Stabilisation of ERK5 thus contributes to cell survival and should therefore be considered as an additional aspect in therapy of chronic myeloid leukaemia. Taken together, the results obtained in this study demonstrate that diverse pathways regulate ERK5 signalling by affecting kinase activity, localisation and protein stability. While the phosphatase PTP-SL is involved in negative regulation of ERK5, Abl kinases potently activate ERK5 and increase its half-life. Protein stabilisation thus is presented as a novel mechanism in the regulation of MAPKs.
Periplasmic Sud protein encoded by the Wolinella succinogenes catalyses the transfer of bound polysulfide-sulfur to the active site of the membrane bound polysulfide reductase. The homodimeric protein consists of 131 residues per monomer, each with one cysteine residue in the active site. Polysulfide-sulfur is covalently bound to the catalytic Cys residues of the Sud protein. In order to understand the structure-function relationship of this protein, the features of its solution structure determined by heteronuclear multidimensional NMR techniques are reported here. The first step of structure determination leads to resonance assignments using 15N/13C/2H- and 15N/13C-labeled protein. The sequential backbone and side chain resonance assignments have been successfully completed. Structure calculations were carried out using the ARIA program package. The structure is based on 2688 NOE-derived distance restraints, 68 backbone hydrogen bond restraints derived from 34 slow-exchanging backbone amide protons and 334 torsion angle restraints obtained from the TALOS program as well as 158 residual dipolar coupling restraints for the refinement of relative vector orientations. The three-dimensional structure of the Sud protein was determined with an averaged rootmean- square deviation of 0.72 Å and 1.28 Å for the backbone and heavy atoms, respectively, excluding the terminal residues. Without the poorly defined segment between residues 90-94 the average r.m.s.d. value drops down to 0.6 Å and 1.14 Å. The ensemble refined with residual dipolar coupling (rdc) restraints shows good convergence. The r.m.s.d. value for the backbone heavy atoms, excluding residues 90- 94, drops down from 0.97 to 0.66 for the rdc-refined ensemble. The relative orientation of the two monomers in the protein structures refined with residual dipolar coupling restraints are also different from those without residual dipolar coupling restraints. The structure determination of the dimeric protein has been hampered by the high molecular mass (30 kDa), severe peak degeneracy, and by the small number of experimental intermonomer NOEs (relative orientation problem of two monomers). For the resonance assignments of aliphatic side chain, many resonances were ambiguously assigned because of severe overlap of signals. The Sud dimer protein contains 17 Lys, 14 Leu and one His tag for each monomer. It complicated the resonance assignments. The conventional 3D 15N-separated TOCSY HSQC experiment failed because of the large molecular weight which results in line broadening and hence made the resonance assignments of side chains more difficult. The determined structure contains a five-stranded parallel ß-sheet enclosing a hydrophobic core, a two-stranded anti-parallel ß-sheet and seven a-helices. The dimer structure is stabilized predominantly by hydrophobic residues. Sud catalyses the transfer of the polysulfide-sulfur to cyanide, similar to rhodanese encoded by Azotobacter vinelandii (Bordo et al., 2000). The two proteins are similar in the active site environment primarily owing to the main-chain conformation of the active-site loop with the cysteine residue and with respect to the surrounding positively charged residues. The active-site loop (residues 89-95) in the Sud protein appears to be flexible, reflected by few assigned proton resonances of residues 90-94 in the active site. Despite their similarity in function and their similar structure in active site, the amino acid sequences and the folds of the two proteins are remarkably different. The negatively charged polysulfide interacts with positively charged R46, R67, and R94 and hence may be stabilized in structure. The mutation of one of the three arginines that are also conserved in rhodanese from A. vinelandii leads to a loss of sulfur-transfer activity. The polysulfide chain extends from inside of Sud protein to outside, where Sud may form contacts with polysulfide reductase. These contacts provide the possible polysulfide-sulfur transfer from Sud protein to the active site of polysulfide reductase.
The light-harvesting chlorophyll a/b protein complex (LHC-II) is the major collector of solar energy in all plants and it binds about half of the chlorophyll in green plants. LHCII is a trimer in the photosynthetic membrane; each monomer consists of 232 amino acids, binds and orients a minimum of 12 chlorophyll molecules and three caroteinoids (two luteins and one neoxanthin) for light-harvesting and energy transfer. Although, the structure of LHC-II has been determined at 3.4 Å resolution by electron microscopy of two-dimensional crystals (Kühlbrandt et al., 1994), this is not sufficient to allow a complete understanding of the mechanism of energy transfer from LHC-II to the reaction centre, since the effective resolution in the z dimension is 4.9 Å. In fact, the chemical difference between Chl a and Chl b, which has a formyl group instead of the methyl group at the 7-position in the chlorin ring, is too small to be detected at this level of resolution. In addition, the orientation of the chlorophyll tetrapyrroles have not been determined unambiguously. This information is essential for a detailed understanding of the energy transfer within the complex and to the reaction centres of photosystem II and I (PSII and PSI). X-ray crystallography of three dimensional (3D) crystals may yield a more complete structure at high resolution. 3D crystals have been grown from LHC-II isolated from pea leaves using a standard purification procedure (Burke et al., 1978). The thylakoid membranes are solubilised in Triton X-100 and further purified by sucrose gradient ultra centrifugation. The LHC-II fraction is salt precipitated and pellets resuspended at the chlorophyll a/b ratio 2.8 mg/ml in 0.9 % Nonyl-glucoside. Crystals are currently obtained by vapour diffusion in hanging drops. These crystals are thin hexagonal plates, have a fairly large unit cell and diffract quite weakly. The high level of the background is due both to the detergent, necessary for protein solubilisation, and lipids, required for the trimer and crystals formation. However, three data sets, each from one single crystal have been collected up to 3.2 Å resolution over a rotation range of 135°. The crystals were exposed to a very highly collimated and brilliant beam (ID-14 EH1 at ESRF, Grenoble, France) and were kept under a stream of cold nitrogen to prevent radiation damage. Data were successfully integrated using the program XDS by Kabsch (1993). The crystals were found to belong to the space group P6 22 3 and have unit cell dimensions of a=128.45, b=128.45, c=135.32, a= ß=90º, ?=120. The solution of the phase problem was tackled by molecular replacement using, as a search model, the LHC-II structure solved by electron cryo-microscopy studies of twodimensional crystals (Kühlbrandt et al. 1994). Three different programs were tested: the most used AMoRe (Navaza et al., 1994) and the brute force based program Brute (Fujinaga
Diese Zusammenfassung ist in zwei Abschnitte gegliedert. Im Abschnitt 6.1. wird die physiologische Bedeutung der Glutamatrezeptoren (GluR) und ihr biologischer Hintergrund kurz erklärt. Am Ende dieses Abschnitts wird der Stand der Strukturanalyse des GluR-B Ionenkanals zu Beginn des Projektes zusammengefasst. Im nachfolgenden Abschnitt 6.2. sind die wesentlichen Ergebnisse der hier vorgelegten Arbeit zusammengefasst. 6.1. Die Bedeutung von Glutamatrezeptoren - Stand der Strukturanalyse zum Beginn dieser Arbeit Die Kommunikation zwischen Nervenzellen erfolgt vorwiegend an hochspezialisierten Kontaktstellen den chemischen Synapsen. Der enge Raum zwischen sendender und empfangender Nervenzelle wird auch als synaptischer Spalt bezeichnet. Der Prozess der synaptischen Übertragung beruht auf der präsynaptischen Freisetzung von chemischen Botenstoffen, sogenannten Neurotransmittern in den synaptischen Spalt. Die Aminosäure L- Glutamat (Glu) ist der wichtigste erregende Neurotransmitter im menschlichen Gehirn und Rückenmark. Dementsprechend bedeutend ist die Rolle der ionotropen Glutamatrezeptoren (iGluRs), die sie bei der elektrochemischen Erregungsübertragung am synaptischen Spalt spielen (Seeburg, 1993), (Hollmann and Heinemann, 1994), (Dingledine et al., 1999). Die Freisetzung von Neurotransmittern wird durch ein elektrisches Signal (Aktionspotential) ausgelöst, das sich entlang der Nervenfaser, dem Axon, bis zur Nervenendigung, der Synapse, fortpflanzt. Nach der Freisetzung diffundieren die Neurotransmitter durch den synaptischen Spalt und binden an sogenannte Rezeptoren. Ionotrope Glutamatrezeptoren sind Ionenkanäle, die in die Membran der nachgeschalteten (postsynaptischen) Nervenzelle eingebaut sind. Sie zählen deshalb zu den Membranproteinen. Als ligandgesteuerte kationenselektive Ionenkanäle machen Glutamatrezeptoren (GluRs) die postsynaptische Membran nach Aktivierung durch Ligandbindung für bestimmte Kationen durchlässig. Der Einstrom von Ionen bewirkt eine Änderung des Membranpotentials. Die Stärke der synaptischen Übertragung ist lebenslang modulierbar; die sogennante synaptische Plastizität wird als eine entscheidende Grundlage für die Erklärung von Lernen und Gedächtnis angesehen. Drei synthetische Agonisten aktivieren die GluRs selektiv und wurden deshalb für die Klassifizierung der ionotropen Glutamatrezeptoren herangezogen. Bei den Agonisten handelt es sich um -Amino-3-hydroxy-5-methyl-4-isoxazol-4-propionat (AMPA), Kainat and N- Methyl-D-Aspartat (NMDA). Die ersten beiden Subtypen werden auch als non-NMDA- Rezeptoren zusammengefasst. Die Aktivierung und Desensitivierung der non-NMDA Rezeptoren ist schneller als die der NMDA-Rezeptoren. Aus molekularbiologischer Sicht (siehe Kapitel 1.3.2.) zeigen die drei Klassen der ionotropen Glutamatrezeptoren eine beträchliche Diversität. So gibt es vier verschiedene Unterheiten vom AMPA-Subtyp, nämlich GluR-A, GluR-B, GluR-C und GluR-B. In dieser Arbeit steht die Strukturanalyse eines aus GluR-B Untereinheiten bestehenden AMPA-Rezeptors im Vordergrund. (Die weitere Unterteilung der NMDA- und Kainatrezeptoren kann dem Kapitel 1.3.2. auf Seite 6 entnommen werden.) Bestimmte Abschnitte der Aminosäurensequenz von Glutamatrezeptoren sind durch hydrophobe Bereiche gekennzeichnet ((M1-M4) in Abbildung 6.1.A (A.)). Das durch verschiedene Untersuchungen etablierte Modell der Glutamatrezeptor-Topologie zeigt 3 Transmembrandomänen (M1, M3 und M4) und eine Membranschleife (M2) (Hollmann et al., 1994), (Kuner et al., 1996). Der Aminoterminus ist extrazellulär, der Carboxyterminus hingegen intrazellulär. Daraus ergibt sich die in Abbildung 6.1.A (B.) abgebildete Topologie (Paas, 1998). S1 und S2 kennzeichnen die Ligandbindungsdomäne. Glutamatrezeptoren (GluR) sind Oligomere, die sich mit grosser Wahrscheinlichkeit aus vier Untereinheiten (Rosenmund et al., 1998), (Ayalon and Stern-Bach, 2001) zusammensetzen (siehe Kapitel 1.3.3.). Die Zusammenlagerung verschiedener Untereinheiten zu einem funktionellen Kanal setzt voraus, dass die Untereinheiten zum gleichen Subtyp gehören, d.h. AMPA Untereinheiten können nur mit anderen AMPA Untereinheiten einen Ionenkanal bilden. Das gleiche gilt für die Zusammensetzung von NMDA und Kainat-Rezeptoren. Das Modell eines tetrameren Glutamatrezeptors ist im Bild C. der Abbildung 6.1.A zu sehen. Die Bestimmung der Quartärstruktur eines vollständigen Glutamatrezeptors ist bislang nicht veröffentlicht. Die strukturelle Analyse von Proteinen erfordert die Isolierung von reinem und funktionellem Protein. Im Vergleich zu den meisten löslichen Proteinen erfordert die Isolierung von Membranproteinen oft besonderer Optimierung. Falls das Vorkommen des Proteins in natürlichem Gewebe gering ist, so kann die strukturelle Analyse durch rekombinante Expression in einem geeigneten Wirtsorganismus zugänglich gemacht werden. Die Isolierung von Milligramm-Mengen eines rekombinanten homomeren GluR-B Rezeptors aus dem entsprechenden Baculovirusexpressionssystem (Keinänen et al., 1994) wurde in unserem Labor etabliert (Safferling et al., 2001) und wurde im ersten Jahr dieses Projektes fortgeführt. Durch zonale Ultrazentrifugation konnte gezeigt werden, dass die molekulare Masse des GluR-B Proteinkomplexes ca. 495 kD beträgt. Dieser Wert liegt in der Nähe des theoretischen Molekulargewichts eines tetrameren Ionenkanals, dessen Molmasse sich aus vier GluR-B Untereinheiten (104 kD) und einer Detergenzmizelle von ca. 63-97 kD zusammensetzt (Safferling et al., 2001). Die elektronenmikroskopische Analyse des Proteinkomplexes von W. Tichelaar aus unserer Gruppe erfolgte 1999 durch Negativfärbung. Für die Strukturanalyse mit Hilfe der Software IMAGIC wurden 10 000 Proteinteilchen selektiert. Das Ergebnis der Bildrekonstruktion ist in der folgenden Abbildung 6.1.B gezeigt. Die projezierten Dimensionen des Models entsprechen einem Molekül mit den Dimensionen 17 nm × 11 nm × 14 nm. Das Model zeigt keine ausgezeichnete Symmetrie, die auf die Stöchiometrie des GluR hinweisen könnte. Das Molekül zeigt mit Färbemittel gefüllte Vertiefungen und innere Strukturen, die vielleicht an der Ionenleitung beteiligt sind. 6.2. Funktionelle und strukturelle Charakterisierung des GluR-B Ionenkanals In der Fortsetzung des oben beschriebenen Projektes wurden für die rekombinante Expression desselben Rezeptors (GluR-B homomer) stabil transformierte Insektenzellen eingesetzt. Dazu wurde die für die GluR-B Untereinheit kodierende und in Plasmiden enthaltene DNA in Insektenzellen transformiert (siehe APPENDIX A.2.2.). Im Vergleich zu dieser auf Dauerhaftigkeit angelegten Integration der Rezeptor DNA wird die Proteinexpression beim Baculovirusexpressionssystem durch Infektion mit rekombinanten Baculoviren initiiert. Der Vergleich zeigte, dass die mit Baculoviren erzielten Ausbeuten bei GluR-B etwa doppelt so hoch waren als bei stabil transformierten Zellen. Allerdings fallen bei stabil transformierten Zellen die eventuellen Nachteile der viralen Belastung auf die zellulären Sekretionsprozesse weg. Im Verlauf der elektronenmikroskopischen Analyse von baculoviral erzeugtem GluR-B Protein hat sich gezeigt, dass Proteine viralen Ursprungs unter Umständen selbst doppelt aufgereinigte GluR-B Proben verunreinigen können (siehe APPENDIX A.2.1.). Dieser Punkt ist bei einer Einzelbildverarbeitung von grosser Relevanz, falls die virusspezifischen Proteinverunreinigungen eine ähnliche Grösse haben wie das eigentliche Zielprotein. Das Hauptziel dieser Arbeit war es, das Potenzial stabil transformierter Insektenzellen für die Expression von homomeren GluR-B Ionenkanälen zu bewerten und dabei die Stöchiometrie der Untereinheiten in diesem Ionenkanal aufzuklären. Zu diesem Zweck wurden biochemische und elektronenmikrosopische Techniken eingesetzt. Zur Isolierung des GluR-B Ionenkanals aus stabil transformierten Insektenzellen wurde das bestehende Aufreinigungsprotokoll für die Affinitätchromatographie an immobilisierten Metallionen (IMAC) (Safferling et al., 2001) optimiert, indem das Chargenverfahren durch das Durchflussverfahren ersetzt wurde (zur genaueren Erklärung der Optimierung siehe RESULTS 4.1.2.). Abbildung 6.C zeigt ein silbergefärbtes Gel mit den Eluaten der IMAC und Eluaten der abschliessenden Affinitätschromatographie mit immobilisiertem M1-Antikörper. Die auf den Bahnen 5-8 aufgetragen GluR-B Proben wurden auch für die Einzelteilchenanalyse mittels Elektronenmikroskopie verwendet. Die Ligandbindungsaktivität von GluR-B wurde durch Filterbindungsexperimente mit dem Radioliganden [3H]-AMPA vor und nach der Isolierung aus den Membranfragmenten bestimmt. Die KD-Werte sind für beide Proben ähnlich gross. Der Bmax-Werte ist für die aufgereinigte Probe wie erwartet sehr viel (mehr als 200×) höher. Die Ergebnisse der Ligandbindungsexperimente sind im Kapitel 4.2.1 tabellarisch zusammengefasst. Die oligomere Struktur des isolierten Ionenkanals wurde durch Quervernetzungsexperimente (Cross-linking) und Einzelteilchenanalyse von negativ gefärbten Proteinmolekülen bewertet. Die Quervernetzungsexerimente selbst erbrachten kein eindeutiges Ergebnis im Hinblick auf oligomere Struktur des komplett zusammengesetzten Rezeptors. Kontrollexperimente mit dem Lysat vom Rattenhippocampus zeigten, dass mit DTSSP ein geeigneter Cross-Linker verwendet wurde (siehe RESULTS 4.3.2.). Neben einem aus 4 Banden bestehenden Muster (siehe RESULTS 4.3.1.) lieferten die Quervernetzungsexperimente mit isoliertem GluR-B aber einen deutlichen Hinweis auf die Stabilität von dimeren GluR-B Strukturen, die im Einklang mit einer jüngst veröffentlichten Arbeit stehen (Ayalon and Stern-Bach, 2001). Diese Veröffentlichung liefert zusätzliche (Armstrong et al., 1998) Hinweise auf die Bedeutung von Dimeren in der Glutamatrezeptorstruktur und postuliert, dass sich ein kompletter Glutamaterezeptor aus einem Dimer-Paar zusmmensetzt, wobei die Dimere zuerst gebildet werden. Die nachfolgende Abbildung 6.2.B zeigt negativ gefärbte GluR-B Ionenkanäle bei einer 46000× Vergrösserung. Die Aufnahme stammt von einem Philips EM 400 Elektronenmikroskop. Für die 3D Rekonstruktion wurden 500 der in Abbildung 6.2.B gezeigten Rezeptormoleküle ausgewählt. Dieser relativ kleine Datensatz besteht aus GluR-B Ionenkanälen deren Präservierung in Uranylacetat als besonderes vielversprechend eingeschätzt wurde. Dieser positive Effekt wurde auf die Verwendung frisch von einer Wasseroberfläche aufgefischter Kohlefilme zurückgeführt (siehe RESULTS 4.4.3.3.). Während der Klassifizierung dieses Datensatzes fiel auf, dass die beim Band-Pass-Filtern für die niedrigen Frequenzen gesetzten Cut-offs einen deutlichen Einfluss auf die erste Klassifizierung der unterschiedlichen zweidimensionalen Ansichten des Proteinkomplexes haben (siehe RESULTS 4.4.3.4.). Aus diesem Grund wurde der gleiche Datensatz mit 5 verschiedenen low-frequency cut-offs (LFCO) gefiltert (siehe Table 4.4.3.4.) und getrennt klassifiziert. Von den 5 resultierenden Klassifikationen wurden 3 (LFCO 0,005, 0,03 und 0,05) für die weiterführende 3D Rekonstruktion ausgewählt. Die Evaluierung der resultiernden 3D Modelle ergab, dass der mit einem LFCO von 0,03 gefilterte Datensatz eine Klassifikationen erlaubte, die zu einem 3D Modell (Modell GluR-BII/a siehe RESULTS Figure 4.4.3.4.H) führte, das im Vergleich zu den beiden anderen Rekonstruktionen konsistenter war. Am stärksten spricht für dieses Modell die Übereinstimmung der Input-Projektionen mit den Reprojektionen der 3D Rekonstruktion (siehe siehe RESULTS Figure 4.4.3.4.H). Zur Verfeinerung des Modells GluR-BII/a wurden die beiden Projektionen mit der höchsten Standardabweichung vom Klassendurchschnitt (class average) eliminiert. Die verbleibenden 11 Projektionen bildeten die Input-Projektionen für die Berechung eines verfeinerten Modells, GluR-BII/b, das auf einer neuen Zuordnung der Euler-Winkel beruht. Das Ergebnis dieser Berechung ist in der nachfolgenden Abbildung gezeigt. Das Modell in Abbildung 6.2.C zeigt einen zentralen Kanal und hat die Dimensionen 18 nm × 14 nm × 11 nm. Die Stöchiometrie der Untereinheiten ist aus dem Modell, das mit grosser Wahrscheinlichkeit einen komplett zusammengesetzten GluR darstellt, nicht ablesbar. Ebensowenig zeigt das Modell eine eindeutig vierzählige oder fünfzählige Symmetrie. Allerdings ist die erkennbare zweizählige Symmetrie im Einklang mit dem vorgeschlagenen Pair-of-Dimer Modell (Ayalon and Stern-Bach, 2001), das auf eine teramere Struktur des oligomeren Ionenkanals schliessen lässt. Die Ergebnisse dieser Arbeit zeigen, dass stabil transifzierte Insektenzellen eine durchaus geeignete Quelle für GluR-B Ionenkanäle sind. Nachteilig sind die geringen Ausbeuten. Allerdings kann durch weitere Selektion der Zellen die GluR Expression noch gesteigert werden (siehe APPENDIX A.2.2.). Bei höheren GluR-B Ausbeuten könnte zukünftig auch die Detektion des Rezeptors in vitrifizierten Proben in Verbindung mit Kryo-Elektronen- mikroskopie und auch die 2D-Kristallisation gelingen. Die während dieses Projekts gemachten Kristallisationsexperimente (siehe APPENDIX A.3.) und Kryo-Experimente mit GluR-B Protein aus dem Baculovirusexpressionssystem (siehe RESULTS 4.4.1. und 4.4.2.) ergaben negative Ergebnisse. Das Potential der Kryo-Methode konnte allerdings in Kontrollexperimenten mit Tabak-Mosaik-Virus (TMV) gezeigt werden. Kryo-Daten von GluR-B würden die Berechnung eines genaueren Strukurmodells erlauben. Die Reprojektionen des hier besprochenen Strukturmodells GluR-BII/b aus der Abbildung 6.2.C könnten als Referenzen für das Alignment der vitrifizierten GluR Ionenkanäle dienen. Für das langfristige Ziel der Rekonstituition des Rezeptors in Liposomen sollte die Delipidierung des Membranproteins während der Aufreinigung möglichst reduziert werden. Hier erscheinen zwei Ansätze sinnvoll. Die Aufreinigung des Proteins in einem Schritt durch die Erweiterung des tags am Carboxyterminus von nur 6 auf 10 Histidin-Reste. Ausserdem gibt es Hinweise, dass die Anwesenheit von Lipiden während der Aufreinigung für seine Rekonstituierbarkeit förderlich ist (Huganir and Racker, 1982).