Why do alpha-beta parallel proteins, like flavodoxins, form misfolded off-pathway intermediates?

The question: “Why do α-β parallel proteins, like flavodoxins, form misfolded off-pathway intermediates?" is the main subject of this thesis. A. vinelandii apoflavodoxin is chosen as protein of interest as it is a representative of α-β parallel proteins, which are widely prevalent in nature. The folding behavior of A. vinelandii apo- and holoflavodoxin has been studied extensively during the past years. Both denaturant-induced equilibrium and kinetic (un)folding of apoflavodoxin have been characterized in detail using GuHCl as denaturant 1-8. An off-pathway intermediate plays a major role during apoflavodoxin folding and is also observed during the kinetic folding of other proteins with an α-β parallel topology of which the folding mechanism has been studied 9. Approximately 90% of folding molecules fold via off-pathway intermediate Ioff, which is a relatively stable species that needs to unfold to produce native protein and thus acts as a trap 3. Residual structure in the unfolded state of apoflavodoxin probably facilitates formation of this species. In chapter 2 detailed information about unfolded apoflavodoxin is revealed by heteronuclear NMR spectroscopy. In 6.0 M GuHCl apoflavodoxin behaves as a random coil as is shown by far-UV CD and by 1H-15N R2 relaxation rates. Upon lowering denaturant concentration the amount of residual structure in apoflavodoxin increases. Chemical shift deviations between unfolded apoflavodoxin in 3.4 and 6.0 M GuHCl reveal in unfolded apoflavodoxin in 3.4 M GuHCl the presence of three transiently formed α-helices and of one structured region that is neither an α-helix nor a β-sheet. One of these transiently formed α-helices is non-native, and a part of this helix becomes a β-strand in native apoflavodoxin. Four regions with restricted flexibility on the (sub)nanosecond time scale are revealed by 1H-15N R2 relaxation rates of unfolded apoflavodoxin in 3.4 M GuHCl. These four regions coincide with the ordered regions found by chemical shift analysis and match with regions of large AABUF (average area buried upon folding), which is correlated with hydrophobicity 10. Chemical shift deviations upon substitution of a glutamine residue with a more hydrophobic cysteine residue on position 48, in the middle of the non-native α-helix in unfolded apoflavodoxin, show that this non-native helix has hydrophobic interactions with all other ordered regions in unfolded apoflavodoxin. Formation of native and non-native helices in unfolded apoflavodoxin and subsequent docking of these helices leads to formation of a compact off-pathway intermediate. The formation of this off-pathway intermediate is discussed in chapter 3. Backbone amide resonances of unfolded apoflavodoxin are followed in a series of 1H-15N HSQC spectra acquired at concentrations of GuHCl between 4.05 M and 1.58 M. Analysis of cross peak disappearance of unfolded backbone amides made it possible to determine midpoints of unfolding of 68 backbone amides. Residues were grouped in five different groups according to their midpoint of unfolding. The group with the highest Cm value forms the folding core of the molten globule of apoflavodoxin in presence of GuHCl. This folding core roughly coincides with the regions with restricted flexibility in unfolded apoflavodoxin. The core is gradually extended upon decreasing denaturant concentration, but part of apoflavodoxin’s molten globule remains random coil in the denaturant range investigated. The formation of the off-pathway intermediate of apoflavodoxin is non-cooperative and involves a series of distinct transitions in contrast to the cooperative formation of native apoflavodoxin 7. In addition, chemical shifts of the amides of unfolded apoflavodoxin could be tracked over the denaturant range investigated. Analysis of the chemical shift changes shows that structure formation within virtually all parts of the unfolded protein precedes folding to the molten globule. The results presented in this chapter, together with those reported on the molten globule of α-lactalbumin 11, show that helical molten globules apparently fold in a non-cooperative manner. To investigate long-range interactions in unfolded apoflavodoxin that lead to formation of this off-pathway intermediate, in chapter 4 use is made of site-directed spin labeling. For this purpose, glutamine at position 48, which resides in a non-native α-helix of unfolded apoflavodoxin, is replaced by a cysteine. This replacement enables covalent attachment of two different nitroxide spin labels, MTSL and CMTSL. Due to this amino acid replacement stability of native apoflavodoxin against unfolding decreases and attachment of the nitroxide spin label MTSL leads to a further decrease in stability. Replacement of Gln48 by Cys48 decreased flexibility of the ordered regions in unfolded apoflavodoxin in 3.4 M GuHCl, due to increased hydrophobic interactions. Interactions are detected between the MTSL spin label attached to Cys69 and region Ser40 - Leu62 of unfolded apoflavodoxin in 6.0 M GuHCl. These non-specific hydrophobic interactions between nitroxide spin labels and hydrophobic patches of unfolded apoflavodoxin perturb the unfolded protein. Our observations show that in 6.0 M GuHCl spin-labeled apoflavodoxin is less random coil than C69A apoflavodoxin is. Thus, care needs to be taken in the use of spin labels for the study of the conformational and dynamic properties of unfolded proteins. In 3.4 M GuHCl the attached CMTSL spin label induces the presence of two distinct states in unfolded apoflavodoxin. In one of these states, the spin label attached to residue 48 has persistent contact with residue Leu78. The spin label data show that non-native contacts exist between transiently ordered structured elements in unfolded apoflavodoxin. Full population of the molten globule-like folding state of apoflavodoxin is possible through covalent introduction of just a single extra oxygen atom in the protein, achieved by replacing Phe44 with Tyr44 through site-directed mutagenesis (chapter 5). This replacement leads to significant destabilization of native apoflavodoxin, as is demonstrated by GuHCl-induced equilibrium (un)folding and thermal unfolding experiments. Decreasing salt concentration destabilizes native apoflavodoxin even further. As a result, the native state of F44Y apoflavodoxin is hardly populated. Instead, in absence of denaturant, virtually all protein molecules exist as molten globule-like folding intermediate. Direct characterization of this intermediate by far-UV CD is possible, it is shown that the molten globule has a totally different topology: it is helical and lacks the parallel β-sheet of native apoflavodoxin. Full population of the molten globule state of F44Y apoflavodoxin enables use of H/D exchange for the characterization at the residue level by NMR spectroscopy of apoflavodoxin’s molten globule folding intermediate. In chapter 6, interrupted H/D exchange is used to detect the stable core of apoflavodoxin’s molten globule in absence of denaturant. Exchange rates could be determined for 68 backbone amides. Amide protons of residues Lys16 – Phe25 are poorly protected against exchange, and structure formed in this region is very unstable. In chapter 4 chemical shift data and Cm-values showed that these residues belong to the most unstable part of apoflavodoxin’s molten globule, as they remain random coil down to a GuHCl concentration of 1.58 M. Leu110 to Val125 have the highest protection factors against H/D exchange and form the single stable core of apoflavodoxin’s molten globule in absence of denaturant. The residues of this molten globule, which have the highest midpoints against unfolding by GuHCl, roughly coincide with those residues that are transiently ordered in unfolded apoflavodoxin. Only one of the four regions mentioned is significantly protected against exchange in this intermediate. This suggests that this helix is better buried in apoflavodoxin’s molten globule compared to the other helices. Hydrophobic interactions of this helix with the other ordered parts of the molten globule, although loose in nature, cause context-dependent stabilization of this helix against unfolding. The helical molten globule contains thus a single stable core. Non-native docking of helices in apoflavodoxin’s molten globule prevents formation of the parallel β-sheet of native apoflavodoxin. Hence, to produce native α-β parallel protein molecules, the off-pathway species needs to unfold. Discussion Formation of non-native secondary and tertiary structure in unfolded protein is the answer to the question: “Why do α-β parallel proteins, like flavodoxins, form misfolded off-pathway intermediates?” The presence of non-native secondary structure elements in unfolded proteins is probably a widespread phenomenon. However, subsequent formation of folding intermediates that contain these non-native structure elements is likely but rarely reported. In this thesis, it is proven for the first time that formation of native and non-native helices within an unfolded α-β parallel protein and subsequent non-native docking of these structured regions leads to formation of a compact helical off-pathway intermediate. One of the helices (residues Leu110 to Val125) forms a stable core in the molten globule in absence of denaturant. Hydrophobic interactions of this helix with the other ordered parts of the molten globule cause its context-dependent stabilization. Non-native docking of the helices prevents formation of the parallel β-sheet of native protein. To produce native α-β parallel protein molecules, the off-pathway species needs to unfold and as a result non-native interactions and nonnative secondary structure are disrupted. This thesis shows that acquisition of native-like topology is not necessarily the general result of the initial collapse in protein folding. Rather than directing productive folding, conformational pre-organization in the unfolded state of an α-β parallel type protein promotes off-pathway species formation. The data presented in this thesis indicate that especially proteins that contain domains with an α-β parallel topology seem susceptible to off-pathway intermediate formation. A single polypeptide sequence can code for monomeric protein folds that are largely different under native-like conditions. The amino acid sequence of apoflavodoxin codes for the α-β parallel topology of the native state, as well as for a helical protein species. Upon a mild change of conditions, topological switching between both folds occurs and a monomeric protein species with a distinct fold becomes energetically most favorable. Topological switching between unrelated protein structures is likely a general phenomenon in the protein structure universe.