Surfing the free energy landscape of flavodoxin folding
Bollen, Y.J.M. - \ 2004
Wageningen University. Promotor(en): Sacco de Vries, co-promotor(en): Carlo van Mierlo. - [S.l.] : S.n. - ISBN 9789085040644 - 165
eiwitten - moleculaire structuur - azotobacter vinelandii - proteins - molecular conformation - azotobacter vinelandii
The research described in this thesis has been carried out to obtain a better understanding of the fundamental rules describing protein folding. Protein folding is the process in which a linear chain of amino acids contracts to a compact state in which it is active. Flavodoxin from Azotobacter vinelandii is chosen as the representative of the group of OC-? parallel proteins. Flavodoxins are small monomeric proteins that contain a non-covalently bound FMN cofactor. The ?-? parallel topology is characterised by a five-stranded parallel ?-sheet surrounded by ?-helices at either side of the sheet. The doubly-wound topology is a rather popular fold: it belongs to the five most common observed folds, together with the ???-barrel, Rossman, thiamin-binding and P-loop hydrolase folds. In contrast to most protein folds, this topology is shared by many (i.e. nine) protein superfamilies. These nine superfamilies exhibit little or no sequence similarity and comprise a broad range of unrelated proteins with different functions like catalases, chemotactic proteins, lipases, esterases, and flavodoxins. By studying the folding behaviour of ?. vinelandii flavodoxin insight can be gained into how this large group of proteins folds.
First, the equilibrium (un)folding of apoflavodoxin from ?. vinelandii (i.e. flavodoxin in the absence of the FMN cofactor) is investigated. Apoflavodoxin is structuraUy identical to holoflavodoxin except for some dynamic disorder in the flavin-binding region. A molten globule-like intermediate is shown to populate during denaturant-induced equilibrium unfolding of apoflavodoxin (Chapter 2).
Subsequently, the folding and unfolding kinetics of the 179-residue A. vinelandii apoflavodoxin have been followed by stopped-flow experiments monitored by fluorescence intensity and anisotropy (Chapter 2). The denaturant concentration dependence of the folding kinetics is complex. Under strongly unfolding conditions, the kinetics can be described by a single rate constant. When this unfolding rate constant is plotted against the denaturant concentration, a change in the slope is observed. This, together with the absence of an additional unfolding process reveals the presence of two consecutive transition states on a linear pathway that surround a high-energy on-pathway intermediate.
Under refolding conditions, two folding processes are observed. The slowest of these two processes is the one that is populated most, and it becomes faster with increasing denaturant concentration. This means that an unfolding step is rate-limiting for folding of the majority of apoflavodoxin molecules. This, together with the absence of a 1ag in the formation of native molecules, means that the intermediate that populates during refolding is off-pathway.
The experimental data obtained on apoflavodoxin folding are consistent with the linear four-state folding mechanism I1 "=>unfolded apoflavodoxin t=>I2<=>native apoflavodoxin. The off-pathway intermediate I1 is the one that populates during refolding and that also populates during denaturant-induced equilibrium unfolding of apoflavodoxin. I2 is the unstable intermediate that is observed during kinetic unfolding.
The presence of such on-pathway and off-pathway intermediates in the folding kinetics of proteins with an ?-? parallel topology is predicted from simulations of Go-like protein models. In addition, two kinetic folding intermediates, one on-pathway and the other off-pathway, seern to be present under specific experimental conditions during the folding of all proteins with an ?-? parallel topology that have been investigated. The appearance of folding intermediates in this class of proteins is apparently governed by protein topology (Chapter 3).
Next, the local dynamics of apoflavodoxin have been studied by hydrogen deuterium exchange detected by heteronuclear NMR spectroscopy (Chapter 4). The use of native state hydrogen deuterium exchange detected by NMR spectroscopy leads to the identification of four partially unfolded forms (PUFs) of apoflavodoxin in which some non-native interactions apparently play a role. The rates of interconversion of these PUFs with native apoflavodoxin are determined. These rates are inconsistent with the PUFs being on a direct folding route between native and globally unfolded apoflavodoxin. PUFl and PUF2 are on an unfolding route starting from native apoflavodoxin that does not 1ead to the globally unfolded state of the protein. PUF3 and PUF4 are on a non-productive folding route starting from globally unfolded apoflavodoxin. A common free energy barrier separates both PUF3 and PUF4 from unfolded apoflavodoxin. This barrier has the same height as the one determined from stopped-flow kinetic folding studies that separates the known off-pathway apoflavodoxin folding intermediate I1 from the productive folding route. Therefore a single energy barrier is proposed to separate both PUF3 and PUF4 as well as I1 from the productive folding route. All three species thus need to unfold before productive folding of apoflavodoxin can occur (Chapter 4).
The influence of the presence of the non-covalently bound flavin mononucleotide (FMN) cofactor on the global stability and on the kinetic folding of A. vinelandii holoflavodoxin (i.e. flavodoxin in presence of the FMN co-factor) are reported in Chapter 5. The denaturant-induced equilibrium (un)folding data of flavodoxin in the presence and absence of FMN are excellently described by a model in which only native apoflavodoxin binds to FMN. As the intermediate I1 populates during apoflavodoxin equilibrium (un)folding, the holoflavodoxin equilibrium (un)folding model consists of four species: unfolded apoflavodoxin, the apoflavodoxin folding intermediate I1, native apoflavodoxin and holoflavodoxin molecules. Cofactor binding to apoflavodoxin is shown to affect the protein stability in a theoretically predictable manner.
Despite that many proteins require the binding of a ligand to be functional, the kinetic role of ligand-binding during folding is poorly understood. FMN binding to native apoflavodoxin occurs with two kinetically observable rate constants at all denaturant and protein concentrations studied, as is shown in Chapter 5. These two rate constants arise from two conformationally differing apoflavodoxin species, which most likely exist due to the binding of inorganic phosphate to the FMN phosphate binding site of a fraction of the A. vinelandii apoflavodoxin molecules.
In Chapter 5 it is also shown that excess FMN does not accelerate flavodoxin folding, and FMN does not act as a nucleation site for flavodoxin folding. During kinetic folding of holoflavodoxin formation of native apoflavodoxin precedes ligand binding. Even under strongly denaturing conditions, global unfolding of holoflavodoxin occurs only after release of its FMN. The model that describes A. vinelandii apoflavodoxin kinetic folding, which includes the stable off-pathway intermediate I1 and a high-energy on-pathway intermediate I2, can now be extended to describe kinetic holoflavodoxin folding: I, + FMN<=>unfolded apoflavodoxin + FMN<=>I2 + FMN<=>native apoflavodoxin + FMN<=>holoflavodoxin (Chapter 5).
Finally, in Chapter 6 native state WD exchange combined with NMR spectroscopy is used to probe the influence of FMN binding on the stability of A, vinelandii flavodoxin against local, subglobal and global unfolding. Almost the entire flavodoxin backbone is substantially more rigid in holoflavodoxin than in apoflavodoxin. No areas are detected in flavodoxin where FMN binding results in an increase of the local dynamics. Occasional release of FMN from holoflavodoxin results in the population of apoflavodoxin. Until FMN is rebound, these apoflavodoxin molecules behave as described in Chapter 4. Consequently, they will adopt the previously described partially unfolded forms (PUFs). At least three out of the four partially unfolded forms that apoflavodoxin occasionally adopts under native conditions are inaccessible to holoflavodoxin. Holoflavodoxin can form these partially unfolded conformations only when FMN is released.
AIl observations described in this thesis are used to create a schematic free energy landscape of folding ofA. vinelandii flavodoxin. This schematic energy landscape provides insight into how a protein molecule that adopts the ?-? parallel topology surfs from its unfolded state to its characteristic folded state in which it is active.
As described in Chapter 3 of this thesis, the appearance of both on- and off-pathway intermediates during the folding ofA. vinelandii apoflavodoxin appears to be governed by its ?-? parallel topology. Folding kinetics of other ?-? parallel proteins than the ones mentioned in this thesis need to be determined to verify this hypothesis.
An interesting question is why intermediate I1 that A. vinelandii apoflavodoxin populates during its denaturant-mduced equilibrium (un)folding is off the direct folding route. This question may be resolved by studying the structure of intermediate I1 using among others multidimensional NMR experiments. In addition, the investigation of possible residual structure within unfolded apoflavodoxin can inform about the origin of the kinetic partitioning of folding apoflavodoxin molecules into two routes, one leading to native apoflavodoxin, the other one leading to the molten globule-like intermediate I1.
The kinetic model for A. vinelandii apoflavodoxin folding presented in this thesis implies that apoflavodoxin molecules once they have formed the intermediate I1 need to unfold before folding to the native state can proceed. Studying the folding behaviour of single A. vinelandii apoflavodoxin molecules using sensitive fluorescence techniques can reveal to what extent an apoflavodoxin molecule has to unfold in the latter process. To date, only the folding kinetics of small proteins that fold in one step have been studied by single molecule detection techniques. Studying the folding ofindividual apoflavodoxin molecules can reveal the general dynamics involved in the partitioning of individual protein molecules into two separate folding trajectories.
Finally, it will be highly interesting to study the folding behaviour of proteins in their natural environment. As pointed out in the first chapter of this thesis, the high concentration of biomacromolecules in cells is bound to influence the latter behaviour. Therefore, studying the influence macromolecular crowding agents have on folding flavodoxin molecules will be of great interest. In this thesis, a solid, and strongly necessary, basis is laid for the future perspective of the in vivo investigation of flavodoxin folding in the living cell.
The Nifl PAS domain: Insight into its structure and function
Hefti, M.H. - \ 2003
Wageningen University. Promotor(en): Sacco de Vries, co-promotor(en): Jacques Vervoort. - [S.I.] : S.n. - ISBN 9789058088093 - 116
stikstoffixatie - azotobacter vinelandii - chemische structuur - nitrogen fixation - azotobacter vinelandii - chemical structure
Azotobacter vinelandii is an aerobic soil-dwelling organism with a wide variety of metabolic capabilities which include the ability to fix atmospheric nitrogen by converting it to ammonia. The biosynthesis of ammonia is controlled by 15 to 20 different nif gene products. The activation of nif gene expression by the regulatory enhancer binding protein NifA is controlled by the sensor flavoprotein NifL in response to changes in oxygen or nitrogen levels. NifL contains a PAS domain, which is an ubiquitous motif present in all kingdoms of life. PAS domains are involved in many regulatory signal transduction processes in a large variety of organisms and they function as on-off switches.
In this research the structure and function of this domain has been studied extensively using NMR and X-ray spectroscopy. Comparison of 1000 PAS protein sequences with predicted three dimensional structures resulted in a redefinition of this intriguing sensory domain.
Studies on 2-oxoacid dehydrogenase multienzyme complexes of Azotobacter vinelandii
Bosma, H.J. - \ 1984
Landbouwhogeschool Wageningen. Promotor(en): C. Veeger, co-promotor(en): A. de Kok. - Wageningen : Bosma - 127
azotobacter vinelandii - oxidoreductasen - synthese - moleculaire structuur - azotobacter vinelandii - oxidoreductases - synthesis - molecular conformation
In this thesis, some studies on the pyruvate dehydrogenase and 2-oxoglutarate dehydrogenase multienzyme complexes of Azotobacter vinelandii are described; the emphasis strongly lies on the pyruvate dehydrogenase complex.A survey of the literature on 2-oxoacid dehydrogenase complexes is given in chapter 1. It appears that the A.vinelandii pyruvate dehydrogenase complex resembles the complexes from other gram-negative bacteria with respect to its composition and working mechanism. The A.vinelandii complex is however much smaller than the pyruvate dehydrogenase complexes isolated from other sources.Chapter 2 describes the procedure that has been optimized for the isolation of the pyruvate dehydrogenase and 2-oxoglutarate dehydrogenase multienzyme complexes (PDC and OGDC respectively) from A.vinelandii. In comparison to the previous isolation procedure, several advantages exist. The A.vinelandii PDC is obtained as an essentially pure three-component complex, in a high yield (40-50%). 80% of the losses can be accounted for by discarded side-fractions, which indicates that the complex is hardly inactivated during its purification. The specific activity of the final preparation is about two times higher (15-19 U/mg) than previously could be obtained. From these observations we conclude that the formerly observed "fourth component" of A.vinelandii PDC was a mere contaminant. With the revised procedure, the 2-oxoglutarate dehydrogenase complex (OGDC) is obtained in a high yield (40-50%), free from contaminants. In the "old" procedure this complex was irreversibly inactivated by the action of protamine sulfate.In chapter 3 some observations on the A.vinelandii OGDC are reported. The molecular mass of this complex is of the order of 2.4 to 3.2 MDa, as determined by laser light-scattering measurements. The three component enzymes have the same molecular masses as have been reported for the OGDC's of Escherichia coli and pig-heart. The activity of the complex is regulated by its substrates in an analogues way as has been reported for the E.coli complex, and we therefore conclude that the A.vinelandii complex probably strongly resembles the OGDC of E.coli. In this chapter, an isolation procedure for the lipoamide dehydrogenase component is described, and it is shown that the lipoamide dehydrogenase components of the A.vinelandii PDC and OGDC probably are identical.The association behaviour of the A.vinelandii pyruvate dehydrogenase complex is described in chapter 4. From sedimentation and light-scattering studies we conclude that a monomer-dimer equilibrium exists for this complex; the molecular mass of the monomer has been estimated that 800 kDa. In this thesis, this monomer-dimer mixture is referred to as the 18 S form of the complex. Upon addition of polyethylene glycol 6000 and MgCl 2 , the 18 S form of the complex aggregates into a large structure, resembling the pyruvate dehydrogenase complex of E.coli with respect to its sedimentation, coefficient (56 S) and its appearance on electron micrographs. The isolated dihydrolipoyl transacetylase component of A.vinelandii PDC has a molecular mass of 2 MDa, and on electron micrographs it resembles the dihydrolipoyl acetyltransferase component of E.coli. It is concluded that this large structure probably is composed of 32 subunits. Upon the binding of the pyruvate dehydrogenase and lipoamide dehydrogenase components, this large particle dissociates into the smaller structures that are characteristic for the intact A.vinelandii complex. The small (18 S) and the large (56 S) forms of the (sub)complexes are in slow equilibrium, and this equilibrium can be perturbed by high hydrostatic pressure. From light-scattering measurements at varying pressures it is concluded that the 56 S form of the complex probably is an octamer of the 800 kDa monomers.The measurements concerning the chain-stoichiometry of A.vinelandii PDC are described in chapter 5. A novel method for the determination of chain-ratios was developed, based on the covalent modification of lysine residues in the three component enzymes with trinitrobenzene sulfonic acid. With this technique, an average chain ratio of 1.3:1:0.5 (pyruvate hydrogenase: dihydrolipoyl acetyl transferase:lipoamide dehydrogenase) was found for the isolated A.vinelandii PDC. In combination with the results of chapter 4, it is concluded that A.vinelandii PDC is based on a tetrameric dihydrolipoyl acetyltransferase core, to which the periferal components are bound in a non-covalent way. The complex can be reconstituted from its individual components, and from these reconstitution experiments it follows that the complex has maximal activity when three pyruvate dehydrogenase dimers and one lipoamide dehydrogenase dimer are bound to the dihydrolipoyl transacetylase tetramer.In chapter 6, the results of acetylation experiments are given. It is shown that the reductive acetylation of the lipoyl groups probably is the rate-limiting step in the reaction sequence of the A.vinelandii pyruvate dehydrogenase complex. In so-called servicing experiments, an extensive exchange of acetyl groups between individual (monomeric) pyruvate dehydrogenase complex particles is found. This phenomenon (inter-core transacetylation) has until now only been observed for the A.vinetandii complex. It is shown that the inter-core transacetylation occurs when two monomeric particles are associated. Although the transacetylation reactions show large effects in the servicing experiments, these reactions are however too slow to be of physiological importance. The servicing experiments also show that the large " E.coli -like" isolated dihycirolipoyl acetyltransferase component is composed of rather independently operating tetramers, i.e. the large form of the A.vinelandii PDC does not function as a large entity.In chapter 7, the results of the three preceding chapters are summarized and translated into a three-dimensional model of the molecular organisation of the A.vinelandii PDC. The merits of this model are discussed in relation to the generally accepted model for the pyruvate dehydrogenase complex of E.coli. It is suggested that the pyruvate of Azotobacter vinelandii could represent the morphological subunit of the larger structure that is found in Escherichia coli and perhaps in other gramnegative bacteria. It is concluded that further experiments have to be performed, in which the complexes of the two organisms are directly compared. to establish whether such a unifying model does exist.