ASSESSING THE ROLE OF STABILITY IN FIBRIL FORMATION IN APOLIPOPROTEIN A-I VARIANT L178H USING UREA AND GUANIDINE DENATURATION A Thesis Presented to the faculty of the Department of Chemistry California State University, Sacramento Submitted in partial satisfaction of the requirements for the degree of MASTER OF SCIENCE in Chemistry (Biochemistry) by Trang Thanh Thi Duong SUMMER 2012 © 2012 Trang Thanh Thi Duong ALL RIGHTS RESERVED ii ASSESSING THE ROLE OF STABILITY IN FIBRIL FORMATION IN APOLIPOPROTEIN A-I VARIANT L178H USING UREA AND GUANIDINE DENATURATION A Thesis by Trang Thanh Thi Duong Approved by: __________________________________, Committee Chair Linda Roberts, Ph.D. __________________________________, Second Reader Mary McCarthy Hintz, Ph.D. __________________________________, Third Reader Jeffrey Mack, Ph.D. ____________________________ Date iii Student: Trang Thanh Thi Duong I certify that this student has met the requirements for format contained in the University format manual, and that this thesis is suitable for shelving in the Library and credit is to be awarded for the thesis. __________________________, Department Chair Linda Roberts, Ph.D. Department of Chemistry iv ___________________ Date Abstract of ASSESSING THE ROLE OF STABILITY IN FIBRIL FORMATION IN APOLIPOPROTEIN A-I VARIANT L178H USING UREA AND GUANIDINE DENATURATION by Trang Thanh Thi Duong Apolipoprotein (apo) A-I is a major component of high density lipoprotein (HDL) which removes and transports cholesterol from peripheral cells to the liver to be excreted in a process known as reverse cholesterol transport (RCT). Plasma apo A-I levels are correlated with reduced cardiovascular disease (CVD) risk (1,2), presumably due to the role of apo A-I in RCT, as well as apo A-I’s demonstrated antiinflammatory and anti-oxidant properties (3). However, mutations in apo A-I result in fibril formation, which can lead to severe diseases including atherosclerosis, neuropathy, nephropathy, and peptic ulcer disease (4,5). These mutations cluster in two regions, within the first 100 amino acid residues of the protein and in a small cluster spanning residues 173-178. The two clusters appear to be associated with tissue specific deposition of fibrils (6). Two full-length fibril-forming proteins with a mutation in either the N-terminal region (G26R) or the C-terminal region (L178H) have been studied. The G26R mutation leads to the formation of amyloid fibrils with increased β-strand structure (7) that causes hereditary systemic amyloidosis, resulting in neuropathy, nephropathy and visceral amyloidosis. The L178H mutation results in helical fibril formation leading to fibril deposits in the larynx and in cardiac tissue (6). A first step in understanding differences between the two mutation clusters and how v they are manifested in vivo is to explore the properties of variants from each cluster. To that end, we are exploring the structure and stability of L178H. A common feature of fibril-forming proteins is that they are generally less stable, and therefore more prone to misfolding, than their wild-type (WT) counterparts (8-10). The goal of this thesis project is to analyze the stability of L178H in comparison to that of wild-type apo A-I using equilibrium solvent denaturation. In this work, the protein stability and cooperativity of unfolding were measured through the intrinsic fluorescence of tryptophan residues. From intrinsic fluorescence measurements, the Gibbs free energy of protein unfolding (ΔG), effectiveness of unfolding (m), and the midpoint of denaturation (D1/2) were obtained. The results indicate L178H is less stable and unfolds less cooperatively than WT protein. In addition, the stability of L178H under all conditions tested is dominated more by hydrophobic interactions than electrostatic interactions compared to WT apo A-I. These factors are likely responsible for the propensity of this mutant to form fibrillar structures in vivo. ___________________________________, Committee Chair Linda Roberts, Ph.D. ___________________________________ Date vi ACKNOWLEDGMENTS I would like to thank the people who have helped me complete this thesis. First, I thank my parents who have raised me and made me the person I am today. Next, thanks to my committee members Dr. Roberts, Dr. McCarthy, and Dr. Mack, who helped me through the whole thesis process. Thanks to the faculty and staff of the chemistry department who helped me with various questions and problems: Dr. Dixon, Dr. Baker, Dr. Gherman, Dr. Crawford, Ted, John, Michelle, Kathleen. Thanks to my research group members and friends for their help and support: Aurelia, Himali, Roman, Soraya, Megan, Tanya, Jessica, Jason, Kris, Yad, Holly, Sonya, Lauren. vii TABLE OF CONTENTS Page Acknowledgements ...................................................................................................... vii List of Tables.................................................................................................................. x List of Figures ............................................................................................................... xi Chapter 1. INTRODUCTION ................................................................................................... 1 1.1 Atherosclerosis ............................................................................................. 1 1.2 Role of apo A-I and HDL Function ............................................................. 2 1.3 Apo A-I Structure......................................................................................... 3 1.4 Amyloid ..................................................................................................... 18 1.5 Apolipoproteins and Amyloid .................................................................... 23 1.6 Apo A-I Amyloid ....................................................................................... 25 1.7 The Current Study ...................................................................................... 27 2. EXPERIMENTAL ................................................................................................. 28 2.1 Materials ..................................................................................................... 28 2.2 Methods ...................................................................................................... 29 3. RESULTS AND DISCUSSION ............................................................................ 38 3.1 Protein Purification .................................................................................... 38 3.2 Equilibrium Solvent Denaturation ............................................................. 40 3.3 Intrinsic Fluorescence Measurements ........................................................ 42 viii 3.4 Comparison of Experimental Data to Literature Values ............................ 44 3.5 Stability of Initial Structure ........................................................................ 45 3.6 Stability Over 14 Day Period ..................................................................... 52 4. CONCLUSIONS AND FUTURE WORK ............................................................ 58 Appendix A: List of Abbreviations .............................................................................. 60 Appendix B: Experimental Data .................................................................................. 64 References .................................................................................................................... 66 ix LIST OF TABLES Tables Page Table 1.1. The eight classes of steric zippers ............................................................... 21 Table 1.2. Apo A-I mutations associated with amyloidosis......................................... 26 Table 3.1 Comparison of experimental values to literature values for WT protein. .... 45 Table 3.2. Comparison of L178H to WT at 4⁰C ......................................................... 47 Table 3.3. Comparison of L178H to WT after incubation at 37⁰C .............................. 48 x LIST OF FIGURES Figures .........................Page Figure 1.1. Atherosclerosis ............................................................................................ 1 Figure 1.2. Putative amphipathic helices in apo A-I protein.......................................... 5 Figure 1.3. Secondary structure of apo A-I based on different methods ....................... 8 Figure 1.4. Lipid-free apo A-I tertiary structure .......................................................... 10 Figure 1.5. Stabilizing interactions of the lipid-free ∆(185-243) apo A-I dimer ......... 12 Figure 1.6. Twisting of double belt structure ............................................................... 14 Figure 1.7. Monomer, dimer, and tetramer form of ∆(1-43) apo A-I .......................... 16 Figure 1.8. Proposed mechanism for lipid binding by apo A-I................................... 17 Figure 1.9. Overview of amyloid assembly mechanism .............................................. 19 Figure 1.10. Underlying structure of amyloid .............................................................. 20 Figure 1.11. β-sheet stacking interfaces ....................................................................... 20 Figure 1.12. Clustering of single amino acid mutations in apo A-I. ............................ 27 Figure 3.1. SDS-PAGE analysis of the integrity of WT and L178H protein............... 39 Figure 3.2. Structures of urea and guanidine hydrochloride. ....................................... 40 Figure 3.3. Emission spectra of WT apo A-I in varying denaturant concentrations .... 43 Figure 3.4. A) Denaturation curve for WT apo A-I monitored by intrinsic fluorescence xxxxxxxxxB) Determination of ∆GH2O of WT apo A-I by linear extrapolation ........ 44 Figure 3.5. Urea induced unfolding of WT and L178H apo A-I at 4⁰C ...................... 46 Figure 3.6. Denaturation of WT and L178H apo A-I after incubation at 37⁰C ........... 48 xi Figure 3.7. Location of the L178H and R173C mutations ........................................... 51 Figure 3.8. Percent of reduction in thermodynamic value of L178H and R173C dddddddddcompared to WT at 4⁰C ............................................................................. 52 Figure 3.9. Change in free energy after incubation at 4⁰C ......................................... 54 Figure 3.10. Change in free energy after incubation at 37⁰C ..................................... 55 xii 1 Chapter 1 INTRODUCTION 1.1 Atherosclerosis Atherosclerosis is a condition in which fat, cholesterol, calcium, and other materials form deposits inside artery walls that can harden to form plaques (Figure 1.1) (11). Atherosclerosis can affect many different organs such as the lungs, heart, kidneys, brain, intestines and limbs. The plaques building up in these arteries can lead to myocardial infarction (heart attack), stroke, coronary heart disease (CHD), abdominal aortic aneurysm, kidney disease, mesenteric artery ischemia, peripheral artery disease, renal artery stenosis, hypertension and thoracic aortic aneurysm. CHD is the number one killer of men and women in the United States each year (12). In 2007, CHD killed 406,351 people. In 2008, it was estimated that 16.3 million Americans 20 years or older had CHD (13). In addition, in 2007, 132,968 Americans died of heart attack and 135,952 died of stroke. These numbers indicate that atherosclerosis is a significant health concern for Americans. Figure 1.1. Atherosclerosis. Artery narrowed by plaque buildup, leading to heart attack or stroke (14). 2 The risk factors associated with atherosclerosis, in particular CHD, are high triglyceride levels, increased low density lipoprotein (LDL) cholesterol levels, and decreased high density lipoprotein (HDL) cholesterol levels (15). HDL, commonly referred to as the “good cholesterol,” has been shown to inversely correlate with CHD risk, independent of triglyceride and LDL levels (16). That is, the more HDL present in the plasma, the lower the risk for CHD. 1.2 Role of apo A-I and HDL Function Apolipoprotein (apo) A-I is present in all HDL particles, and constitutes approximately 70-100% of HDL proteins (17,18). The intestines and liver synthesize and secrete apo A-I into the plasma as a lipid-poor protein. Immediately after secretion, apo A-I acquires phospholipids and unesterified cholesterol from macrophages with the help of ATP-binding cassette transporter A1 (ABCA1) to form discoidal HDL particles. This nascent HDL acquires more phospholipids and unesterified cholesterol from other peripheral cells such as skeletal myocytes, adipocytes, and skin fibroblasts. In addition, HDL receives phospholipids from other lipoproteins by activity of the phospholipid transfer protein (PLTP). Lecithin-cholesterol acyltransferase (LCAT) converts the cholesterol into a cholesteryl ester by transferring a fatty acid from a phospholipid to free cholesterol. The cholesteryl ester forms the hydrophobic core of HDL, which turns the nascent HDL into spherical (mature) HDL. Cholesteryl ester transport protein (CETP) transports the mature HDL to the liver where scavenger receptor class BI (SR- 3 BI) promotes uptake of cholesterol from HDL to eventually be excreted in bile and feces. This process is known as reverse cholesterol transport (RCT). As a primary factor in RCT, apo A-I has important anti-atherosclerotic properties. Furthermore, apo A-I possesses both anti-microbial and anti-inflammatory properties (19). Studies comparing apo A-I, HDL and LDL cholesterol level as markers for cardiovascular mortality show apo A-I can be a powerful predictor of CHD compared to HDL level (1, 2). In fact, there is a stronger correlation between CHD and apo A-I level than between CHD and HDL level. 1.3 Apo A-I Structure Since apo A-I is the main component of HDL, even compared to lipids, and the primary factor in RCT, it is important to understand its structure at various points along the RCT pathway. In human plasma, apo A-I exists in multiple lipid-free and lipidbound states. Approximately 5-10% of apo A-I in circulating plasma exists in a lipidfree state, while the rest is distributed among various lipid-bound states (20). Depending on solution conditions, lipid-free apo A-I may exist in oligomeric forms of dimers, trimers, tetramers and higher order oligomers. The physiological distribution of the oligomer forms is unknown and may include amyloid or other fibrillar forms that contribute to disease. Due to its importance in RCT and human health, the multiple forms of apo A-I have received intensive study over the past 30 years. Although much headway has been made in the past 10 years, significant gaps in understanding lipid-free and lipid-bound structures remain. With the increased attention on apo A-I amyloid 4 formation, there is an increasing awareness of the importance of apo A-I selfassociation in solution and in binding to lipid. To fully understand how apo A-I forms fibrils, a complete understanding of apo A-I structure in all forms is needed. 1.3.1 Apo A-I The human apo A-I gene is located on chromosome 11 (6). The gene encodes a pre-propeptide containing 267 amino acids (residues), which is cleaved intracellularly to a 249 amino acid propeptide (21). After secretion into the plasma, the propeptide is cleaved to yield a 243 amino acid, 28.1 kilodalton (kDa) polypeptide protein that lacks glycosylation and disulfide linkages (20); the latter is due to the absence of cysteine residues. The primary structure of apo A-I has been extensively analyzed using computer programs developed to identify and characterize the signature motif of all apolipoproteins, the amphiphatic helix (22). The computer models led to the identification of amphipathic helix classes found predominantly in apolipoproteins: G*, A, and Y (Figure 1.2). The G* helix, which is characterized by random radial arrangement of positively and negatively charged residues on the polar face of the helix (22), lies in residues 1-43 of the N-terminus of apo A-I. Residues 44-243 contain ten amphipathic α-helices containing either 11 or 22 residues belonging to class A or class Y (Figure 1.2). The class A amphipathic helix is associated with strong lipid-binding properties and is defined by positively charged residues at the polar-non polar interface and negatively charged residues at the center of the polar face. Class Y helices contain two negative residue clusters on the polar face and three positive residue clusters 5 forming a Y motif. In general, the lipid affinity for the three amphipathic helix classes is A>Y>G*. Figure 1.2. Putative amphipathic helices in apo A-I protein (22). Based on the proposed placement of amphipathic helices in apo A-I by these computer programs, Mishra et al. synthesized the peptides for each putative helix to probe their lipid-associating properties (23). The order of lipid affinity was determined to be helix 9-10 >1-2>4-5 (Figure 1.2); the other helices do not possess much affinity. Researchers subjected the full-length protein to point mutations and or deletion mutations to analyze apo A-I based on its lipid binding affinity and ability to efflux cholesterol. The resulting data suggests helices 1, 9, and 10 function to initiate the binding of lipid, helix 4 and 5 promote HDL formation, and helix 6 and 7 activate LCAT (Figure 1.2) (24, 25). 1.3.2 Apo A-I Secondary Structure Extensive analysis of the primary sequence suggests the type and number of amphipathic helices in apo A-I. A number of experimental approaches have been used to determine the actual helix content. Collectively, these approaches indicate lipid-free apo A-I contains 50-60% helix (26-30), whereas lipid-bound contains approximately 6 70-80% helix (31-34). The placement of helices using different methods has given conflicting results, especially for the lipid-free protein as discussed below. The full-length apo A-I in solution has a flexible and adaptable structure, which makes it difficult to determine the lipid-free as well as lipid-bound secondary structure using direct methods such as nuclear magnetic resonance (NMR) and X-ray crystallography (X-RC). Thus, secondary structures were determined using both indirect methods on full-length and direct methods on the less flexible truncated apo A-I mutants. The location of helices in lipid-free apo A-I was first suggested by a mixture of indirect methods (MIM) which include: circular dichroism (CD), sedimentation velocity, ANS binding, fluorescence quenching and limited proteolysis of full-length and truncated protein (35). The data from MIM were used to create a structural model of lipid-free apo A-I that contains six helices near the N-terminus (residues 1-189) and no helices near the C-terminus (residues 190-243) (Figure 1.3). The structure contains 60% helix, generally consistent with CD analysis of lipid-free apo A-I. More than a decade after the MIM model, a second model was determined by hydrogen exchange and mass spectrometry (HX-MS; 26). This second model has five helices instead of six, and the placement of the helices is somewhat different from that of the MIM model. However, the HX-MS model has 51% helix content, similar to that of the MIM model and also consistent with CD analysis. A third model, generated from the structure of a truncated apo A-I (∆(185-243) apo A-I) that was determined by X-RC (36), has five helices near the N-terminus, but the placement of helices is different from both the MIM 7 and HX-MS model. The three secondary structural models are similar in that they all have helices that span residues 8-33 and residues 44-54 and no helices for residues 135141. A fourth model was proposed using data from electron paramagnetic resonance (EPR; 37). This structure has 53% helix content, consistent with CD analysis, but it is not in good agreement with the previous three models, having helices placed in the Cterminal region and containing four β-strands (12% β-strand content). The β-strand content from the EPR model is consistent with CD analysis and Fourier transform infrared spectroscopy (FTIR) data which indicate apo A-I contains 8-16% β-strand content (27-30). A small amount of β-strand structure could have important implications for fibril formation, and the EPR data has been used to propose a model of lipid-free apo A-I that contains β-strand stabilizing regions. Until a high resolution structure of the full-length protein is determined, the existence and role of β-strands, as well as absolute placement of helices, remains unclear. There are two secondary structural models proposed for lipid-bound apo A-I. The first model was based on X-RC data (38) of a truncated lipid-free protein (∆(1-43) apo A-I), but it is believed to represent the lipid-bound structure for three reasons: It has a different secondary structure (increased helical content) compared to full-length apo A-I, its secondary structure does not change when it binds to lipid, and it has a proteolytic cleavage pattern similar to that of full-length lipid-bound apo A-I. The second model was based on NMR data for the full-length protein in the presence of sodium dodecyl sulfate (SDS), a lipid-mimicking solution (39). Both structural models 8 differ in the exact placement and amount of α-helices (Figure 1.3), but they both place helices throughout the length of apo A-I. They also have helices in the C-terminus, absent in most of the proposed lipid-free structures. The full length NMR structure has 75% helical content, consistent with CD analysis of lipid-bound apo A-I. The truncated ∆(1-43) apo A-I has 93% helical content. Figure 1.3. Secondary structure of apo A-I based on different methods. The boxes denote α-helix, and the arrows denote β-sheet. The structural models are derived from A) combination of methods including CD, sedimentation velocity, ANS binding, fluorescence quenching, and limited proteolysis of truncated mutant (35), B) hydrogen exchange and mass spectrometry (26), C) X-ray crystallography of ∆(185-243) apo A-I (36), D) EPR of full-length apo A-I (37), E) X-ray crystallography of ∆(1-43) apo A-I (38), and F) NMR of apo A-I in SDS (39) In summary, due to apo A-I's inherent flexibility (and therefore intractability for analysis by X-RC and NMR), and despite decades of research there is still no single 9 unified model that researchers can agree upon for the secondary structure of apo A-I. The same can be said of the tertiary structure, although, as discussed below, there is some agreement on its general features. 1.3.3 Apo A-I Tertiary Structures In humans, the total serum apo A-I concentration is approximately 1 mg/ml, with the lipid-free concentration being less than 0.1 mg/ml (40). At concentrations <0.1 mg/ml, apo A-I self-association is disrupted and apo A-I is assumed to be monomeric. At ≥ 0.1 mg/ml, apo A-I is found to self-associate due to hydrophobic interaction between amphipathic helices, forming oligomers (41). Apo A-I self-associates in both lipid-free and lipid-bound states, but the mechanism of self-association is unknown. Tertiary structure models of the protein in the presence and absence of lipid must take into account both monomeric and oligomeric states especially in light of the importance of these states in amyloid formation. While the self-associated structure of lipid-bound apo A-I is fairly well understood, that of lipid-free apo A-I is not. Several models of lipid-free apo A-I tertiary structures have been proposed. Early studies using MIM (refer to section 1.3.2) led to one of the first models of lipidfree monomeric apo A-I (Figure 1.4A). The MIM data suggests monomeric apo A-I has two structural domains. The N-terminal domain (residues 1-189) forms a helix bundle and the C-terminal domain (residues 190-243) is mostly unstructured (35). Two separate subsequent studies, deletion mutagenesis (42) and Förster Resonance Energy Transfer (FRET; 43), also support the two domain structure. The deletion mutagenesis data 10 indicate that, when a part or the entire N-terminal domain is deleted, the number of residues that form helices decreases significantly, indicating the N-terminal domain forms some kind of helical structure. When the C-terminal domain is deleted, there is no change in the number of residues that form helices, indicating the C-terminus is unstructured. The FRET study measured the distance between W50-R83 and W50-R173 in apo A-I, which proved to be consistent with a compact N-terminal structure such as a helical bundle. Figure 1.4. Lipid-free apo A-I tertiary structure. A) Structure of monomer of full length apo A-I determined by MIM (35). B) Structure of dimer of ∆(185-243) apo A-I determined by X-RC (36). C) Structure of proposed monomer and dimer of full length Apo A-I determined by EPR (37) The X-RC study of the truncated ∆(185-243) apo A-I protein led to a model of the lipid-free dimeric protein (Figure 1.4B) (36). The X-RC model indicates the monomers in the dimer lie in an antiparallel position to each other and form a semicircular structure with a 17 Å height and 110 Å diameter (Figure 1.5). The dimer contains two N-terminal helix bundles. Each helix bundle consists of three helices 11 (residues 7-41, 55-65, 69-121) of one apo A-I molecule, bundling together with one helix (residues 143-185) of the second apo A-I molecule. The N-terminal helix bundle is held together by four stabilizing interactions: two by the N-terminal and C-terminal aromatic clusters and two by the N-terminal and C-terminal π-cation interactions. The aromatic clusters are composed of aromatic residues at each terminus, forming a hydrophobic interaction. The π-cation interactions at each terminus contain an aromatic residue and a cation residue, forming a π-cation bond. The N-terminal aromatic cluster involves nonpolar interactions between W8, F71 and W72 of one monomer and nearby leucines (14, 60, 64, 75 from the same monomer and 170, 174, 178, 181 from the second monomer), holding residues 44-66 and 67-77 close together (44). The Caromatic cluster consists of F33, F104, and W108 of one monomer and nearby aliphathic residues (V30, L38, L42, L44, L46 from the same monomer and V156, L159 from the second monomer), holding the N-terminal helix close to residues 99-121. The N-terminal π-cation interaction consists of W8 and R61 in each monomer positioning residues 55-66 close to residues 1-43 and covering the N-aromatic cluster (W8, F71, and W72) of the same monomer. The C-terminal π-cation interaction consists of K23 and W50 in each monomer, holding residues 55-66 toward residues 1-43. In addition to interactions stabilizing the helix bundles, the central hinge segment (residues 121-143) connects the two monomers via a leucine zipper composed of L122, 126, 137, and 141 from each monomer. Additional stabilization is provided by intramolecular salt bridges (E5-R9, R11-E15-R18, E16-R20, E4-R7, K88-D89, E120-R123) and intermolecular salt bridges (K96-E69, D89-R173-E92, E85-R177-D89). 12 Figure 1.5. Stabilizing interactions of the lipid-free ∆(185-243) apo A-I dimer. One apo A-I molecule is gray and the other is black (36). The EPR study of full-length apo A-I plus spin coupling data led to a different model of the lipid-free dimeric protein (Figure 1.4C) (37). The spin coupling data indicate that, for each monomer in the dimer, residues 24, 44, 64, 167, 217, and 226 align to form a helical bundle. The four β-strands serve as stabilizing structures. The two center β-strands (residues 102-115 and 130-148) form a hairpin with a 15 residue loop to stabilize the center of the protein. The N-terminal β-strand (residues 20-25) stabilizes the N-terminus to the rest of the protein. The C-terminal β-stand (residues 214-220) stabilizes the C-terminus to prevent it from unraveling into random coils in 13 the absence of lipids, through hydrophobic association with the rest of the protein. An important feature of this model is a loop spanning residues 130-139 that initiates lipid binding. The dimer is formed from the anti-parallel association of two monomers aligned at residues 24, 44, 64, 167, 217, and 226. The authors suggest the two center βstrands, which flank the loop in each monomer, convert to α-helices as the protein binds lipids, so that the protein can re-arrange itself into the lipid-bound conformation. The X-RC and EPR tertiary structural model of lipid-free dimeric apo A-I are substantially different. The X-RC model is of a truncated apo A-I instead of full-length, but X-RC has the advantage of being a high resolution method. The EPR method has lower resolution, but it has the advantage of being conducted on the full-length protein. The EPR model also accounts for the small amount of β-structure measured in lipid-free apo A-I by CD (29,30) and FTIR (28). Until the full-length structure is resolved by high resolution techniques such as X-RC and NMR, it is difficult to determine which of the two models is more accurate. Lipid-bound apo A-I exists in multiple forms of discoidal and spherical HDL. These HDL have various lipid compositions and proteins in addition to apo A-I, such as apo A-II and apo C, resulting in different HDL sizes and shapes (45). Lipid-bound apo A-I has been studied most often in reconstituted HDL (rHDL), which are created by combining lipid-free apo dimyristoylphosphatidylcholine A-I (DMPC) with or synthetic lipids, usually palmitoyloleoylphosphatidylcholine (POPC). The rHDLs range from 7-12 nm in diameter and contain 2-5 molecules of apo A-I (45). Currently, there is a lack of substantial data on spherical rHDL and rHDL 14 containing more than two apo A-I molecules; thus, the focus here will mainly be on discoidal rHDL containing two molecules of apo A-I. Indirect methods including crosslinking studies (45), FRET (46) fluorescence quenching (47), and computational experiments combined with NMR spectroscopy and electron microscopy (48) all support a double belt model for lipid-bound protein (Figure 1.6). In this model, two anti-parallel apo A-I molecules wrap around a lipid bilayer to form discoidal HDL. The structure is stabilized by extensive intermolecular salt bridges between the two apo A-I molecules, in which residue 121-143 of one molecule directly overlaps residues 121143 of the second molecule (49). Computational analysis suggest the size variation in discoidal HDL that contains only two apo A-I molecules is due to the incremental twisting or untwisting of the double belt structure (Figure 1.5;48). When the double belt structure is twisted, it resembles the shape of a saddle (“saddle-shaped” structure), but the salt-bridges linkage remains the same and this results in smaller HDL particles. Figure 1.6. Twisting of double belt structure. A) Conversion of saddle-shaped structure to double belt structure, the black arrows denote the N-terminus. B) Same as in A, but shown with lipid molecules in black (48) 15 The X-RC structure of ∆(1-43) apo A-I (Figure 1.7) was the first crystal structure of apo A-I to be reported, and it supports the double belt model (38). The ∆(143) apo A-I tertiary structure was obtained in lipid-free conditions. However, for the reasons discussed above, it is believed to represent lipid-bound structure (as discussed above). The X-RC data indicates that ∆(1-43) apo A-I protein assumes a tetrameric structure, formed from two dimers. It has an elliptical shape with dimensions 125 x 80 x 40 Å, suitable for wrapping around a lipid disc to form HDL. The monomers within the dimer lie antiparallel to each other, and the dimers in the tetramer are also in an antiparallel position. When the ∆(1-43) apo A-I monomers form a dimer (Figure 1.7), 21% of the surface of each monomer is no longer solvent accessible. Half of this 21% is due to hydrophobic interactions and 30% is due to intermolecular salt bridges formed between Arg/Lys and Asp/Glu residues. The dimer possesses both a hydrophobic and hydrophilic face. The structure of ∆(1-43) apo A-I strongly supports the double belt model. The hydrophobic interaction and intermolecular salt bridge found in the ∆(1-43) apo A-I dimer is also comparable to those discussed in the ∆(185-243) apo A-I dimer. When the dimers combine to form the tetramer, the buried surface area of each monomer increases to 42% and the lipid-binding regions of the protein are buried within the tetramer (Figure 1.7). It is suggested that because the X-RC structure of ∆(143) apo A-I was acquired in a lipid-free condition, the dimers combined to form tetramers to shield the hydrophobic lipid-binding region from water. Due to the buried lipid-binding regions, the authors believe the dimeric rather than the tetrameric structure of ∆(1-43) apo A-I represents the true lipid-bound structure. 16 Figure 1.7. Monomer, dimer, and tetramer form of ∆(1-43) apo A-I determined by XRC (38). 1.3.4 Lipid-free to Lipid-bound Mechanism Structural models of lipid-free and lipid-bound apo A-I can be used to propose a mechanism for how lipid-free apo A-I binds to lipid and forms HDL. Based mainly on the X-RC data of ∆(185-243) apo A-I, Mei and Atkinson proposed a mechanism in which lipid-free apo A-I converts to lipid-bound apo A-I (Figure 1.8). In plasma, apo AI is represented as monomers. As the monomers interact with lipids, they come in close contact to each other and form a dimer with a semicircular structure. The π-cation interactions that hold residues 44-66 in place serve as a gate for lipids to enter into the hydrophobic core of the protein. As lipids enter the hydrophobic core, the helices in the N-terminal helix bundle move away from each other to open up the bundle. This is due to the disruption of the four stabilizing interactions (Figure 1.6). The hydrophobic core is exposed, and the dimer is inserted into the lipid membrane surface. This in turn unhinges the N-terminal helix bundle and causes residues 77-99 to extend out, forming a double belt structure, resulting in discoidal HDL formation. The hydrophobic surface of the dimer faces the inside of the belt structure and the hydrophilic surface faces the 17 outside. This double belt structure is strongly supported by the X-RC structure of ∆(143) apo A-I. Figure 1.8. Proposed mechanism for lipid binding by apo A-I (36). Apo A-I is quite flexible because it needs to unfold and wrap itself around lipids to form HDL. However, this may leave the protein prone to misfolding, which can have several consequences, including lack of stable HDL formation or increased abnormal self-association, leading to the formation of insoluble aggregates or fibrils. A number of fibril-forming or amyloid variants of apo A-I have been identified, and because WT apo A-I is found in arterial plaque, it is important to understand the factors that can cause or accelerate improper folding in apo A-I. The propensity of protein in general to misfold and the research conducted to identify misfolding mechanisms is described in the next section. 18 1.4 Amyloid In 1854, Schleiden and Virchow coined the term “amyloid” to describe pale and waxy looking macroscopic tissue that produced a positive iodine staining similar to starch, found during organ autopsies (50). Thus, amyloids were initially thought to be carbohydrates. In 1859, amyloids were discovered to be proteins when their high nitrogen content was determined (50). Research conducted in 1927 found amyloids specifically bind congo red and produce an apple-green birefringence under polarized light suggesting their fibrillar structure, and the fibrillar structure was confirmed by electron microscopy in 1959. X-ray diffraction analysis in 1968 found the fibrils contain β-sheets running perpendicular to the fibril axis (50). Years of subsequent research indicated amyloids form from specific proteins, and these proteins become amyloidogenic upon denaturation of tertiary and quaternary structures in vitro. X-ray crystallography analysis of short amyloidogenic peptides led to a new definition for amyloids, which are protein fibers of varying length, consisting of repeating β-sheets running perpendicular to the fibril axis (cross-β-sheet) (51). Two or more stacked βsheets form a filament, two or more filaments form a protofibril, and two or more protofibrils form a fibril (Figure 1.9). The cross-β-sheet can either be parallel or antiparallel and has a characteristic X-ray fiber diffraction pattern with 4.7 Å meridional reflection and 6-11 Å equatorial reflection (Figure 1.10 C). The meridional reflection corresponds to the inter-β-strand spacing and the equatorial reflection corresponds to the distance between stacked β-sheets (Figure 1.10B). All fibrils have helical symmetry as well as translational symmetry parallel to the fibril axis. The β-sheet stacking 19 contains two types of interfaces: wet and dry. The wet interface has hydrogen bonds between the side chains and water molecules. The dry interface has complementary side chains interlocking to form a steric zipper and is empty of water molecules (Figure 1.11). The interactions that stabilize the fibril are both polar (alternating charges and hydrogen-bond ladders) and nonpolar (Van der Waals dispersion and aromatic stacking). Despite all these same characteristics in many amyloids, there are also important differences among the amyloids. Figure 1.9. Overview of amyloid assembly mechanism (52). 20 Figure 1.10. Underlying structure of amyloid. A) Electron micrograph of amyloid fibrils. B) Schematic diagram of the cross-β-sheets in a fibril. C) Diffraction pattern of amyloid fiber showing meridianal reflection (black dashed box) and equatorial reflection (white dashed box;51) Figure 1.11. β-sheet stacking interfaces. Pictured are six rows of β-sheets that run horizontally. The + signs represent water molecules. The big dashed box indicates one row of β-sheet. The small dashed box highlights the steric zipper formed by interlocking side chains (53). Amyloid structure can vary in terms of the numbers of wet and dry interfaces making up a fibril. The dry interface contains steric zippers and there are eight possible classes of steric zipper structures, five of which have been observed experimentally. These steric zippers differ by whether the β-strands are parallel or anti-parallel, are packed face-to-face or face-to-back, and are packed up-up (both sheet have the same 21 edge up) or up-down (the sheets have one edge up and one down) (Table 1;54). Amyloid fibrils can also vary by the number of protofibrils. There are four types of polymorph (multiple structure states) amyloids can adopt: packing, segmental, sidechain, and assembly. The polymorphs differ in which segment of the protein forms the amyloid fibril and the orientation of the side chains in the fibril. Multiple polymorphs can exist in a fibril. Table 1.1. The eight classes of steric zippers Example protein Class β-strand 1 parallel up-up face-to-face tau, sup35, RNase 2 parallel up-up face-to-back prion 3 parallel up-down face-to-face not yet observed 4 parallel up-down face-to-back amyloid-β, sup35 5 anti-parallel up=down face-to-face not yet observed 6 anti-parallel up=down face-to-back not yet observed 7 anti-parallel up-up face=back insulin 8 anti-parallel up-down face=back amyloid-β Packing Structure 22 1.4.1 Amyloid Properties There are functional as well as toxic amyloids. Disease caused by toxic amyloids is described by the term amyloidosis. For the purpose of this discussion, the focus will be on toxic amyloids. Since the cross-β fold in amyloid can be as simple as having four residues of β-strand that assemble repetitively, any protein that can form βstrands can eventually form amyloids. In fact, many proteins aggregate into amyloids or amyloid-like states under abnormal conditions such as high protein concentration, heat denaturation, extreme pH and non-aqueous solvents. However, even under normal conditions, some properties of the polypeptide itself can lead to amyloid formation. The primary mechanism for amyloid formation is destabilization of the folded (presursor amyloid protein) structure, leading to misfolding and aggregation (55). Studies of single point mutations in amyloidogenic proteins, such as human lysozyme, transthyretin and human immunoglobulin light chain, indicate that the mutants that form amyloid are significantly less stable than the wild-type counterparts (8-10). Hierarchical analysis of 2351 point mutations in 44 globular proteins using empirical tools indicates that the magnitude of destabilization in a mutant protein may correlate with amyloid formation (56). It has been found that, in general, mutants that are more stable than the wild-type counterparts do not form amyloids (10). Besides the mutations that result in a globally unstable structure, amyloids can form from mutations that change a specific physical property of the protein. Analysis of protein sequences using the Tango algorithm, a computer algorithm that predicts 23 aggregating regions in unfolded polypeptide chains, found proline (P), glycine (G), lysine (K), arginine (R), glutamate (E), and aspartate (D) residues play a protective role in flanking aggregating stretches within proteins (55). Mutations that reduce the number of P, G, K, R, E, and D residues increase the propensity for aggregation. Proline’s structural constraints make it difficult to form β-sheets; whereas glycine is small, has no β-carbon atom, and is highly flexible, requiring a high entropic cost to form it into β-sheets (55). Meanwhile, K, R, E, and D residues have the lowest hydrophobicity and are charged, also making it hard to form β-sheets. Properties that increase a protein’s propensity to aggregate into amyloids are long stretches of five or more hydrophobic residues, increased hydrophobicity in the protein core, decreased overall net charge, increased clustering of hydrophobic residues in a region, and high sequence identity between adjacent domains. These properties all make β-sheet formation in proteins easier. 1.5 Apolipoproteins and Amyloid Currently, over 25 proteins have been shown to form amyloids, contributing to diseases such as Alzheimer’s, Parkinson’s, the prion diseases and adult onset diabetes (57). Proteins in the apolipoprotein family also form amyloids that lead to disease. Understanding how apolipoproteins as a family form amyloids will aid in understanding how apo A-I forms amyloid because of the shared characteristics among these proteins. Apolipoproteins contain a high proportion of class A and Y amphipathic α-helices (57). Their function involves binding to lipids to form lipoprotein particles. Despite their 24 generally high helical content, which results in stable structures (57), they are prone to amyloid formation. For example, one of the three apo E isoforms is associated with enhancing amyloid formation of Aβ peptide (Alzheimer’s disease), while apo C-II spontaneously forms amyloids in lipid-free conditions and was found in atherosclerotic plaques. Mutations in apo A-II are responsible for familial systemic amyloidosis, and mutations in apo A-I lead to atherosclerosis and tissue specific amyloid deposition throughout the body. It is not clear why apolipoproteins form amyloids, but studies of apo E isoforms and apo C-II indicate that low conformational stability may be the main factor contributing to amyloid formation. Studies of apo E revealed that apo E fragments form helical fibrils and full-length apo E forms amyloid-like fibrils. The rate of apo E forming amyloid-like fibrils is related to the conformational stability of the N-terminal domain. Further investigation indicates that mutations that lower the conformational stability of the N-terminal portion of the protein increase the rate of apo E fibrillation (58). Apo C-II, unlike apo E, has no ordered structure (helix or β-strands) in lipid-free conditions and spontaneously forms amyloid (59). In a high micellar lipid environment, apo C-II forms structures with stable conformations that do not form amyloid (57). The disordered structure of apo C-II in lipid-free condition is less stable than its ordered structure in the lipid environment. These studies with both apo E and apo C-II indicate a decrease in conformational stability leads to fibril formation. Thus, low conformational stability may also be a factor in amyloid formation in apo A-I. 25 1.6 Apo A-I Amyloid There are 19 mutations in apo A-I that lead to amyloid formation (Table 2). Of the 19 mutations, 12 are single amino acid mutations that cluster in two regions, spanning residues 26-90 and 170-178 (Figure 1.12). Both clusters are within the Nterminal helix bundle (residues 1-189). Mutations near the N-terminus (“inside” mutations) lead to hereditary systemic amyloidosis, while mutations nearer to the Cterminus (“outside” mutations) lead to specific amyloid deposits in the larynx, cardiac tissue, kidneys, and cutaneous tissues (60). The mechanism of tissue-specific deposition of these amyloid variants is unknown. Mutations in both clusters result in fibrils in vivo consisting mainly of an N-terminal peptide of 80-100 residues in length, commonly 193 (61). Two full-length fibril-forming proteins, each with a mutation in either the Nterminal region (G26R) or the C-terminal region (L178H), have been studied by our laboratory in conjunction with Dr. John Voss at University of California, Davis, and Dr. Jens Lagerstedt at Lund University in Sweden. These are the only studies to date on full-length fibril-forming mutants of apo A-I. We found that the G26R mutation, which causes neuropathy, nephropathy and visceral amyloidosis (5), leads to the formation of amyloid fibrils with increased β-strand structure (62). In contrast, the L178H mutation, which deposits fibrils in the larynx and in cardiac tissue (5), results in helical fibril formation (63). We hypothesize that while lowered stability may underlie fibril formation in mutants from both clusters, fibril type may be determined by mutation 26 Table 1.2. Apo A-I mutations associated with amyloidosis Apo A-I Variant Clinical Features Renal impairment, Hepatomegaly, Peripheral neuropathy, Nephropathy, Gly26Arg Peptic ulcer disease, Visceral amyloidosis, Glu34Lys Renal impairment Renal impairment, Hepatomegaly, Trp50Arg Nephropathy, Visceral Amyloid Renal impairment, Hepatomegaly, Leu60Arg Cardiomyopathy, Nephropathy, Visceral Amyloid Leu64Pro Renal impairment Phe71Tyr Palatal mass, (normal organ function) Leu60Hepatomegaly, Visceral Amyloid Phe71delins6 particularly in liver Glu70_Trp72del Renal impairment Renal impairment, Gastrointestinal Asn74LysfsX10 amyloid, amyloid detected in uterus, 6 ovaries, pelvic lymph nodes Hepatomegaly, Renal impairment, Leu75Pro hepatic amyloid, gastrointestinal amyloid Skin lesions, cardiomyopathy, dysphonia, nephropathy, visceral Leu90Pro amyloid, dermal amyloid, laryngeal amyloid, Angina associated with amyloid in Lys107del aortic intima Ala154GlyfsX48 Renal impairment, polyneuropathy His155MetfsX46 Renal impairment, polyneuropathy Leu170Pro laryngeal amyloid Dermal amyloid, vocal cord amyloid, Arg173Pro visceral amyloid, cardiomyopathy, Infertility, cardiomyopathy, bilateral Leu174Ser carpal tunnel syndrome, polyneuropathy, systemic amyloid Laryngeal amyloid, Infertility, visceral Ala175Pro amyloid Dermal amyloid, laryngeal amyloid, Leu178His neuropathy, cardiomyopathy Cause of Death Refs Renal Failure 5, 6, 64, Renal Failure Renal Failure Renal Failure 6 5, 6, 64 5, 6, 64 Renal Failure Renal Failure 6, 64 6 5, 6, 64 6, 64 6, 64 Renal Failure 6, 64 Liver Failure Cardiomyopathy 5, 6, 64 6, 64 6, 64 6 6, 64 Cardiomyopathy 5, 6, 64 Cardiomyopathy 5, 6,64 Cardiomyopathy 5, 6,64 Cardiomyopathy 5, 6, 64 27 Figure 1.12. Clustering of single amino acid mutations in apo A-I. position. The first step in understanding the fibril-forming properties of these and eventually other mutants of apo A-I is to characterize their stabilities. 1.7 The Current Study The aim of this study was to compare the stability of apo A-I L178H, which has an “outside” mutation, to wild-type protein using equilibrium solvent denaturation. Denaturation of the protein was measured through changes in the intrinsic fluorescence of the protein’s four tryptophan residues. From intrinsic fluorescence measurements, the Gibbs free energy of the protein (ΔG⁰), the effectiveness of unfolding (m), and the midpoint of denaturation (D1/2) were obtained. These values tell us how stable L178H is compared to WT protein. Understanding the difference in stability caused by the mutation is key to understanding the mechanism of fibril formation in the protein. 28 Chapter 2 EXPERIMENTAL 2.1 Materials Ethanol (95%, ACS/USP grade) was obtained from Pharmco Products, Inc. Nanopure water was obtained by filtration through an EASYpure LF Compact Ultrapure Water System (Barnstead). His-bind (nickel chelating) resin was purchased from Novagen®. Ampicilin sodium salt (AMP), Triton-X-100 (electrophoresis grade), guanidine hydrochloride, ethylenediaminetetraacetic acid (EDTA) disodium salt (electrophoresis grade), sodium chloride, disodium hydrogen phosphate (Na2HPO4), monosodium phosphate (NaH2PO4), Luria broth components (LB; 10 g/L tryptone, 5 g/L yeast extract, 10 g/L sodium chloride), sodium dodecyl sulfate (SDS), glacial acetic acid, glycerol, Coomassie brilliant blue R-250, and methanol were purchased from Fisher Scientific®. Protease inhibitor cocktail P2714 (inhibitory components: AEBSF, Aprotinin, Bestatin, EDTA, E-64, Leupeptin) and P8849 (inhibitory components: AEBSF, Bestatin, E-64, Pepstatin A, Phosphoramidon) were purchased from SigmaAldrich Chemical Company®. Isopropyl-β-D-thiogalactopyranoside (IPTG) analytical grade was purchased from Affymetrix, Inc. Ultra pure urea and guanidine hydrochloride (Gn-HCl) were purchased from MP Biomedical, LLC. Imidazole was purchased from Janssen Chimica. Glycine was purchased from BDH Chemicals. Sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels were purchased from Jule, Inc. Acetone was purchased from Pharmco-Aaper. NiSO4•6H2O and 2-mercaptoethanol were purchased from Aldrich Chemical Co. Tris(hydroxymethyl)aminomethane 29 hydrochloride (Tris-HCl) was purchased from EMD Chemicals. Bromophenol blue was purchased from Sigma Chemical Co. Tris base was purchased from J.T.Baker®. 2.2 Methods 2.2.1 Protein Production 2.2.1.1 Clones and Cell Lines Wild type and mutant (L178H) apo A-I were donated from the laboratory of Dr. John Voss (UC Davis, CA) as inserts in pNFXex plasmid in Escherichia coli (E. coli) strain BL21(DE3) pLysS cells (Invitrogen) (65). The original pNFXex plasmid was constructed from pBluescript KS(+) vector (Stratagene, La Jolla, CA) by Ryan and coworkers to encode the human apo A-I gene (66). The (pNFXex) plasmid, constructed by Lagerstedt and co-workers, has increased ultility with the addition of 13 restriction endonuclease sites and improved expression due to optimizing codon for use in E. coli, compared to the original plasmid (65). 2.2.1.2 Induction of Expression and Cell Pellet Storage Fifity ml LB supplemented with 50 μl of 500 mg/ml AMP in a 125 ml Erlenmeyer flask was inoculated with E. coli containing the apo A-I plasmid either with one colony from an agar plate or from scraping the surface of a frozen (-80˚C) glycerol stock solution. The flask was then incubated in a New Brunswick Scientific Excella E24 Incubator Shaker Series at 37˚C with shaking (225 rmp) overnight (12-16 hours). Two 1 L flasks, each containing 500 ml LB broth and 500 μl of 50 mg/ml AMP, were 30 inoculated with 25 ml of the overnight culture until the OD at 600 nm reached 0.8 ± 0.05. IPTG was then added to a concentration of 0.5 mM, inducing expression of the target protein. After 3-5 hours, the cells were pelleted by centrifugation in a SuperLite® SLA-1500 rotor in a Sorvall® RC 5B PLUS centrifuge at 12,500 rpm for 15 minutes at 4˚C. The cell pellets were quickly frozen in a bath of dry ice and acetone and then stored at -80˚C. 2.2.1.3 Cell Lysis Ten ml of lysis buffer (phosphate saline buffer (PBS; pH = 7.4, 1.5 M sodium chloride, 0.2 M Na2HPO4, 0.2 M NaH2PO4 in nanopure water), 3 M Guanidine HCl, 0.1% Triton-X-100, 500 μl P2774 Sigma® protease inhibitor cocktail) was added to each frozen cell pellet and mixed by first inverting and then incubating on an InnOva 2300 Platform Shaker (30 rpm) for 20 minutes at room temperature. The solution then was kept in an ice bath and sonicated 3-5 times using a Fisher Sonic Dismembrator Model 150 from Artek Systems Corporation at 45 % burst power for 30 seconds with a 30 second cooling period. The solution was centrifuged in a Super-Lite® SLA-1500 rotor in a Sorvall® RC 5B PLUS centrifuge at 12,500 rpm for 15 minutes at 4˚C. The pellet was discarded and the supernatant was filtered through a syringe with a 0.45 µm Millex® Syringe Driven Filter Unit filter. 2.2.1.4 Protein Purification An econo chromatography column (Bio-Rad®) was loaded with 5 ml of Hisbind resin (binding capacity 5-10 mg protein/ml resin). The column was then washed 31 with 20 ml deionized water, charged with 25 ml charge buffer (400 mM NiSO4), and equilibrated with 25 ml binding buffer (PBS containing 40 mM imidazole). The cell extract was loaded onto the column with a flow rate of 0.5-1 ml/min. A 15 µl sample of the flow through was collected for SDS-PAGE analysis for the presence of protein (See section 2.2.1.5). The column was then washed with 70 ml binding buffer and the first seven 1.5 ml fractions were collected for SDS-PAGE analysis for the presence of protein. The target protein was then eluted with elution buffer (PBS containing 500 mM imidazole), and the first twelve 1.5 ml fractions were collected for SDS-PAGE analysis for the presence of protein. The column was then stripped with 30 ml strip buffer (0.5 M NaCl, 100 mM EDTA, 20 mM Tris-HCl, pH 8.0), to be re-used in the next round of protein purification. 2.2.1.5 Protein Analysis Sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE) was carried out with 15% Tris-Glycine gels (Jule, Inc.) to identify the protein-containing fractions collected from the column. Gels were loaded with 15 μl of column extract mixed with 5 μl of 4X sample buffer (240 mM Tris-HCl pH 6.8, 40% glycerol, 8% SDS, 0.04% bromophenol blue, 5% 2-mercaptoethanol), and then run at 4ºC for 1.5 hours at 120 V in running buffer (25 mM Tris base, 192 mM glycine, 0.1% SDS, pH 8.3). Gels were then stained with Coomassie blue (0.1% w/v Coomassie brilliant blue R-250, 10% glacial acetic acid, 45% methanol) for at least one hour, then de-stained overnight in 10% glacial acetic acid, 10% methanol. The fractions containing protein 32 were pooled and dialyzed in 1 L PBS for at least 9 hours, with a change of 1 liter of fresh PBS every 3 hours, to remove imidazole. 2.2.2 Determination of Protein Concentration The concentration of the pooled protein fractions were measured spectrophotometrically with an Ocean Optics, Inc. Chem 2000-UV-VIS spectrophotometer. Protein concentration was calculated from the OD at 280 nm using ε = 1.13 (cm2/mg). 2.2.3 Equilibrium Solvent Denaturation Apo A-I protein was diluted in PBS to a final concentration of 0.10 mg/ml and incubated at either 4˚C or 37˚C without stirring for 0 to 14 days. For 4˚C samples, protease inhibitor (P8849) was added at a 1:10,000 (v/v) ratio at day 3 of incubation. For 37˚C samples, P8849 was added at day 3 and again at day 10, both at a 1:10,000 (v/v) ratio, to prevent protein degradation. Aliquots of protein were removed periodically and combined with a denaturant (urea or Gn-HCl) in equilibrium denaturation experiments. Stock solutions of 8 M urea and Gn-HCl were prepared from their respective ultra pure solid crystals. Samples containing 0.050 mg/ml ApoA-I and varying concentrations of urea or Gn-HCl (0.0-4.0 M) were prepared. The final volume of each sample was 2.0 ml. The samples were incubated with denaturant for 24 hrs at 4°C prior to fluorescence measurements. 2.2.4. Intrinsic Fluorescence 33 Fluorescence measurements were conducted with a Shimadzu RF-5301 spectrofluorophotometer equipped with thermostat-controlled cell holders set to maintain the temperature constant at 25˚C using a circulating water bath. The experiments were performed using a 1 cm cell. The excitation wavelength for tryptophan was set at 295 nm in order to exclude the contribution of tyrosine residues (max absorbance at 280 nm) to the overall fluorescence emission. Emitted light was measured at 310-400 nm with a 5 nm excitation and 3 nm emission slit width, and corrected for background signal. The unfolding transition was monitored by recording the wavelength at maximum fluorescence intensity (WMF). When tryptophan is excited, the emission spectrum produces a curve with a WMF between 308-355 nm depending on the degree of solvent polarity around the tryptophan (67). During unfolding, the WMF increases and the fluorescence intensity decreases as the solvent polarity around the tryptophan increases (68). 2.2.5. Data Analysis Calculating D1/2, ΔGDº, and m Based on the unfolding of the protein monitored through the WMF, the stability of the protein can be calculated. A background scan of PBS was subtracted from each sample scan. The WMF of triplicate dilutions of each sample were averaged and plotted against the denaturant concentration. The resulting curves exhibited characteristics of a sigmoidal function and were fitted with the Hill equation (69), 𝑦 = 𝑦0 + 𝑎𝑥 𝑏 𝑐 𝑏+ 𝑥𝑏 [1] 34 where y is the WMF, y0 the minimum WMF (when the protein is folded), a is the difference between the minimum WMF and maximum WMF (when the protein is completely unfolded), b is the Hill coefficient, x is the denaturant concentration, and c is the midpoint of denaturation (D1/2). The midpoint occurs at a denaturant concentration where half of the protein population is folded and half is unfolded, which is determined by the spectroscopic change. Assuming a two-state model (folded and unfolded), the fraction of unfolded protein (fD) can be calculated using: 𝑓𝐷 = 𝑦− 𝑦0 𝑦𝑚𝑎𝑥 − 𝑦0 [2] where ymax is the maximum WMF. The equilibrium constant (KD) for the process can be calculated as: 𝐾𝐷 = 𝑓𝐷 1−𝑓𝐷 [3] and the free energy of denaturation (ΔGD) can be calculated as ∆GD = -RTlnKD [4] Plotting ΔG versus denaturant concentration results a linear plot with the equation: ∆GD = ∆GD0 – m [denaturant] [5] where ΔGDº is the free energy of protein folding in water (0 M denaturant), and m the effectiveness of denaturation. To obtain the D1/2, fD is set to 0.5 (in which half of the protein population is unfolded), resulting in KD = 1 and ΔGD = 0. The linear equation becomes 0 = ∆GD0 – m(D1/2). Thus, 35 𝐷1/2 = ° ∆𝐺𝐷 𝑚 [6] The goodness of fit for the sigmoidal function was determined by the correlation coefficient (R2) value. 2.2.6 Statistical Analysis 2.2.6.1 One Sample T-Test One sample t-tests are used to compare a mean value to a population mean. The variation in data (standard deviation) obtained in this project was due to instrument model, cell line, protein batch, reagents, and pipetting techniques, and one could assume the same sources of variation in other labs. In comparing this project’s data to a literature value, the standard deviations of both data are comparable using an unpaired ttest. However, this project’s data is compared to multiple literature values (five to six literature sources); and this project’s standard deviation is not comparable to the 5-6 standard deviations of the combined literature sources. Thus, the one sample t-test compares this project’s data to the literature data regardless of each literature data’s standard deviation. A one sample t-test uses the equation: 𝑡= |𝑥− 𝜇| 𝑠/√𝑛 [7] where x is the mean of the literature value, μ is the experimental mean value (from this study), s is the standard deviation of the literature values from their mean, and n is the number of literature values. The calculated t value is then used to obtain the p-value. 36 The data is tested at the 95% confidence level in which p > 0.05 indicates there is no difference between the two means, and p < 0.05 indicates the two means are different. 2.2.6.2 Unpaired T-Test Unpaired t-tests are used to compare two population means. Here, the unpaired t-test was used to compare the mean thermodynamic value between the two types of proteins (WT and L178H) in the same conditions. Due to the small sample sizes the variance may differ between each set of data, but if the sample size were to be infinite the true variance is assumed to be equal, due to the fact that the data were all obtained in the same way. Thus, this test assumes the two sets of data have equal variances. The test uses the equation: 𝑡= |𝑥̅ 1 −𝑥̅2 | 𝑠𝑝𝑜𝑜𝑙𝑒𝑑 𝑛1 𝑛2 [8] √𝑛 1 +𝑛2 where 𝑠𝑝𝑜𝑜𝑙𝑒𝑑 = √ 𝑠12 (𝑛1 −1)+𝑠22 (𝑛2 −1) 𝑛1 +𝑛2 −2 [9] s is the standard deviation, 𝑥 is the mean, and n is the number of samples. The calculated t value is then used to obtain the p-value. The data is tested at the 95% confidence level, in which p > 0.05 indicates there is no difference between the two means, and p < 0.05 indicates the two means are different. 2.2.6.3 Propagation of Error 37 Propagation of error is used to find the uncertainty associated with a function based on the uncertainty associated with the individual values. When comparing the average thermodynamic values of L178H to the average values of WT (meta-average), propagation of error is needed to find the uncertainty (standard deviation) of the metaaverage. To statistically compare the reduction in stability of L178H in urea to that in Gn-HCl at both 4⁰C and 37⁰C using an unpaired t-test, the standard deviation must be calculated for the reduced value. The equation for propagation of error was used to calculate the standard deviation. For a multiplying or dividing function (e.g. 𝑞 = 𝑥1 𝑥2 or 𝑞 = 𝑥1 ⁄𝑥2 ), the resulting standard deviation of q (stotal) can be calculated as: 𝑠𝑡𝑜𝑡𝑎𝑙 = 𝑥1 𝑥2 2 2 √(s1 ) + (s2 ) x x 1 2 where x is the mean and s is the standard deviation. [10] 38 Chapter 3 RESULTS AND DISCUSSION 3.1 Protein Purification The stability of L178H and WT apo A-I was determined through equilibrium solvent denaturation monitored by intrinsic fluorescence at two temperatures (4⁰C and 37⁰C) and in two denaturants (urea and Gn-HCl) over a 14 day period. This requires a large amount of protein (3.5-4.0 mg) for each experiment, so a good protein yield is important. In order to obtain protein, E. coli cells containing the apo A-I gene (WT or L178H in the pNFXex plasmid) (66) were first inoculated overnight in 50 ml of LB media. The next day, the cells were transferred to a new flask containing 500 ml of media and were induced to express the proteins. After induction, the cells were lysed, the protein was purified using nickel column chromatography, and the protein purity was confirmed using SDS-PAGE. The day the protein was extracted from the cell is considered day zero. The protein was then dialyzed for at least nine hours to eliminate imidazole salts. Finally, the protein’s concentration was determined and used in denaturation experiments beginning at day 1. The protein yield on average was five milligrams per 500 ml of culture, enough for one denaturation experiment (one expression of one protein at one temperature with one denaturant run as triplicate samples). In order to complete all the experiments, protein production was performed numerous times to obtain the required amount of protein. As the experiments were run 39 over a 14 day period, SDS-PAGE was used to periodically check for degradation because apo A-I is highly sensitive to proteolytic degradation (70); only intact protein was used in denaturation experiments. SDS-PAGE analysis confirmed both WT and L178H protein remained intact during the 14 days incubation at 4⁰C, as the band corresponding to intact protein persists through day 14 (Figure 3.1A). At 37ºC, both proteins were degraded by day 7, as the intact apo A-I band for day 7 is non-existent (Figure 3.1B). To prevent degradation, protease inhibitor cocktail (P8849 Sigma®) was added to the 37⁰C incubated protein at a 1:10,000 v/v dilution every seven days. With the addition of protease inhibitors, the proteins remained intact during the 14 days of incubation (Figure 3.1C). Figure 3.1. SDS-PAGE analysis of the integrity of WT and L178H protein at A) 4⁰C, B) 37⁰C, C) 37⁰C with protease inhibitor cocktail P8849. D1, D7, D14 corresponds to day 1, 7, and 14 respectively. The expected molecular weight of apo A-I is 28.1 kDa. 40 3.2 Equilibrium Solvent Denaturation Mechanism of Solvent Denaturation Equilibrium solvent denaturation is the unfolding of protein induced by increasing the denaturant (urea or guanidine hydrochloride) concentration. Urea is an uncharged molecule, whereas guanidine (a hydrochloride salt) is charged (Figure 3.2). The mechanism of denaturation for both denaturants is unclear, but research suggests that denaturation may be due to direct binding of denaturant to the protein, or to the ability of denaturants to change the properties of the water solvent (71), or to both. Urea has a carbonyl group and guanidine hydrochloride (Gn-HCl) has a charged guanidinium group that can interact with the protein backbone or side chains. The polarity of urea and Gn-HCl is different from that of the water solvent, which means their presence could change the bulk solvent properties in ways that promote protein unfolding. Figure 3.2. Structures of urea and guanidine hydrochloride. Because urea and Gn-HCl differ in their charges, they have different denaturing effects and yield different thermodynamic values (72). This was demonstrated by examination of the stability of a series of 4-helix coiled-coil analogs in which residues 41 involved in polar or electrostatic helix-helix interactions were systematically varied from attractive to repulsive charges. The residues contributing to nonpolar helix-helix interactions were the same in all four analogs. The stability of the four analogs was determined and yielded different thermodynamic values for each protein in urea but nearly identical values in Gn-HCl. This indicates that urea denaturation measures the contribution of both hydrophobic and electrostatic interactions that stabilize the protein, whereas Gn-HCl denaturation only measures the contribution of hydrophobic interaction (72). This is due to the fact that Gn-HCl is an ionic compound that masks electrostatic interactions. Thus, comparing proteins using two denaturants allows for a relative determination of stabilizing forces, electrostatic versus hydrophobic, between the proteins. To examine the stability of WT and L178H apoA-I, equilibrium solvent denaturation was conducted on both proteins using Gn-HCl and urea. Protein was diluted from a stock solution to 0.10 mg/ml and incubated at 4⁰C or 37⁰C for 24 hours to ensure the protein is in its monomeric form. The protein was then combined with denaturants (urea or Gn-HCl) in triplicate samples containing 0.05 mg/ml protein in varied concentrations of denaturants (0.0-4.0 M). The samples were incubated at 4⁰C for another 24 hours to allow the protein and denaturant mixture to reach equilibrium. Finally, the samples were measured using intrinsic fluorescence. 42 3.3 Intrinsic Fluorescence Measurements When tryptophan is excited, the emission spectrum produces a curve with a wavelength at maximum fluorescence intensity (WMF) between 308-355 nm depending upon the degree of solvent polarity around the tryptophan (67). During unfolding, the WMF increases and the fluorescence intensity decreases as the solvent polarity around the tryptophan increases (68). Typically, a WMF around 335 nm is characteristic of tryptophan residues buried inside the hydrophobic core of the protein, while a WMF of around 350 nm is characteristic of tryptophan residues exposed to aqueous solvent (67). Emission spectra typical for lipid-free WT apo A-I are shown in figure 3.3. Apo A-I contains four tryptophan residues (W8, 50, 72, 108) which are located in the N-terminal helix bundle (residues 1-189). Fluorescence studies indicate that all four of the tryptophans in lipid-free apo A-I exist in a non-polar environment and are protected from solvent (73). When the folded apo A-I is excited at 295 nm the buried tryptophan gives an emission spectrum with a wavelength at maximum fluorescence intensity (WMF) of around 338 nm (Curve A). As apo A-I unfolds due to denaturant, the tryptophan residues become exposed to solvent, and the solvent polarity around the tryptophan increases. This leads to an increase in WMF due to solvent relaxation of the excited fluorophore and a decrease in the fluorescence intensity due to increased quenching by the solvent. The WMF for fully unfolded apo A-I is approximately 352 nm (Curve D). 43 Figure 3.3. Emission spectra of WT apo A-I in varying denaturant concentrations. A) 0 M Gn-HCl, B) 1.5 M, C) 2.0 M, and D) 3.0 M. To determine stability of the protein, the WMF obtained from intrinsic fluorescence was plotted against denaturant concentration. This generated sigmoidal plots for both proteins (Figure 3.4A). The curves were then fitted using the Hill equation as described in Chapter 2. The sigmoidal plot shows the unfolding of apo A-I is a two state mechanism, transitioning from a native (N) state to denatured (D) state, without a substantially populated intermediate state. The cooperativity of the protein unfolding is dependent on the steepness of the transition from the N to D state. 44 Figure 3.4. A) Denaturation curve for WT apo A-I monitored by intrinsic fluorescence. B) Determination of ∆GH2O of WT apo A-I by linear extrapolation; ∆G is plotted as a function of denaturant concentration. Each point on both plots is an average of three samples from the same batch of protein. Parameters used to determine the stability of the protein are the midpoint of denaturation (D1/2), Gibbs free energy (∆G⁰) of unfolding, and protein stability in response to denaturant concentration (m). The m value has been shown to correlate with the change of accessible surface area (∆ASA) of the protein going from the native (N) to the denatured (D) state (74). The values were determined by plotting ∆G versus denaturant concentration (Figure 3.4B). The resulting linear function was analyzed using the equation ∆G = ∆G⁰ - m[denaturant] to obtain ∆G⁰, D1/2, and m. 3.4 Comparison of Experimental Data to Literature Values In the course of obtaining the stability data, there are many factors that could influence the data such as instruments, cell line, protein batch, reagents, pipetting technique, and fitting method. In order to prove the validity of our technique, data obtained for wild-type (WT) protein in urea and guanidine hydrochloride for day 1 at 4⁰C was compared to literature values, also obtained at 4⁰C (Table 3.1). There is no 45 literature value for WT at 37⁰C and no literature value for L178H at either temperature. The p-values from one-sample t-tests indicate that, statistically, there is no significant difference between the project values and literature values for either urea or Gn-HCl. This validates our technique and, by extension, the data obtained for WT and L178H protein at both temperatures. Table 3.1 Comparison of experimental values to literature values for WT protein. Urea* Gn-HCl** D1/2 D1/2 ∆G⁰ ∆G⁰ WT m m This Project 2.81 6.16 2.19 1.23 4.74 3.85 Literature 2.80±0.19 6.95±1.42 2.46±0.65 1.11±0.14 4.18±0.85 3.88±0.76 n=5 n=5 n=5 n=6 n=6 n=6 p-value p>0.05 p>0.05 p>0.05 p>0.05 p>0.05 p>0.05 (Diff?) (no) (no) (no) (no) (no) (no) * Literature value for urea is an average of data taken from ref. 75-79. **Literature value for Gn-HCl is an average of data taken from ref.78-83. Units for the thermodynamic parameters are as follow: D1/2 (M), ∆G⁰ (kcal•mol-1),m (kcal•M-1mol-1) 3.5 Stability of Initial Structure (Day 1) The stability of the initial structure (at day 1) of WT and L178H apo A-I was determined by incubating the proteins at 4⁰C and 37⁰C in both urea and Gn-HCl for one day. Protein incubated at 4⁰C was evaluated in order to compare the data to literature values, whereas the 37⁰C temperature was used to reflect relevant physiological (body) temperature. The data was collected from three batches of protein; each batch of protein resulted in a sigmoidal plot, as discussed above. Thus, the thermodynamic parameters are reported as an average from three batches of protein (three plots). 46 3.5.1 Stability at 4⁰C The urea denaturation curve at 4⁰C indicates L178H has reduced stability compared to WT, as shown by the shift of the unfolding curve towards lower urea concentration (Figure 3.5A). The denaturation curve for L178H is also less steep than that of WT indicating reduced cooperativity. The midpoint of denaturation (D1/2), conformational stability (∆G⁰), and the dependence of ∆G value on denaturant concentration (m) determined from linear plots of ∆G as a function of denaturant concentration are listed in table 3.2. Compared to WT, the stability of L178H is reduced by approximately 38%, 59%, and 33% corresponding to the D1/2, ∆G⁰, and m values, respectively. Similar to the results for urea denaturation, the values obtained for Gn-HCl denaturation indicate L178H is less stable and less cooperative than WT (Figure 3.5B). Compared to WT, the stability of L178H is reduced by approximately 12%, 59%, and 54% corresponding to the D1/2, ∆G⁰, and m values, respectively (Table 3.2). Figure 3.5. Urea induced unfolding of WT and L178H apo A-I at 4⁰C in A) urea and B) Gn-HCl. The solid line represents WT and the dashed line represents L178H. 47 Table 3.2. Comparison of L178H to WT at 4⁰C Urea Gn-HCl D1/2 D1/2 ∆G⁰ ∆G⁰ m m WT 2.81±0.05 6.16±0.19 2.19±0.11 1.23±0.02 4.74±0.38 3.85±0.34 L178H 1.74±0.03 2.55±0.07 1.46±0.07 1.08±0.07 1.93±0.14 1.76±0.06 Reduction of 38.2±1.1% 58.7±1.2% 33.3±3.1% 12.5±5.4% 59.2±3.0%, 54.2±1.5% L178H vs. WT p-value p<0.05 p<0.05 p<0.05 p<0.05 p<0.05 p<0.05 (Diff?) (yes) (yes) (yes) (yes) (yes) (yes) -1 Units for the thermodynamic parameters are as follow: D1/2 (M), ∆G⁰ (kcal•mol ), m (kcal•M-1mol-1) 3.5.2 Stability at 37⁰C Denaturation was then performed on protein incubated for one day at 37⁰C because it is the relevant body temperature and the temperature at which apo A-I normally functions. Similar to the results at 4⁰C, L178H incubated at 37⁰C has reduced stability and cooperativity (Figure 3.6). Urea denaturation shows the stability of L178H was reduced by approximately 37%, 59%, and 30% corresponding to the D1/2, ∆G⁰, and m values, respectively, compared to WT (Table 3.3). Gn-HCl denaturation at 37⁰C indicates L178H has reduced stability by approximately 25%, 59%, and 45% corresponding to the D1/2, ∆G⁰, and m values, respectively (Table 3.3). 48 Figure 3.6. Denaturation of WT and L178H apo A-I after incubation at 37⁰C in A) urea and B) Gn-HCl. The solid line represents WT and the dashed line represents L178H. Table 3.3. Comparison of L178H to WT after incubation at 37⁰C D1/2 Urea ∆G⁰ 5.39±0.33 2.34±0.12 m 1.91±0.08 1.34±0.05 D1/2 Gn-HCl ∆G⁰ 3.78±0.08 1.55±0.27 m 3.14±0.02 1.74±0.31 WT 2.82±0.07 1.20±0.03 L178H 1.78±0.12 0.90±0.02 Reduction of L178H vs. 36.9±4.1% 55.8±2.3% 29.7±2.5% 25.5±1.3% 58.8±7.5%, 44.6±9.8% WT p-value p<0.05 p<0.05 p<0.05 p<0.05 p<0.05 p<0.05 (Diff?) (yes) (yes) (yes) (yes) (yes) (yes) Units for the thermodynamic parameters are as follow: D1/2 (M), ∆G⁰ (kcal•mol-1), m (kcal•M-1mol-1) The results indicate that L178H is significantly less stable than WT protein, as indicated by the approximately 60% reduction in ∆G⁰ in both denaturants after one day incubation at either 4⁰C or 37⁰C. The D1/2 values for L178H in guanidine and urea are decreased at both temperatures compared to WT but the decrease is greater for urea than for guanidine. Since urea measures both electrostatic and hydrophobic contributions and Gn-HCl measures mainly hydrophobic contributions to stability (72), this indicates 49 L178H (relative to WT) has a larger reduction in electrostatic contributions than hydrophobic contributions. In other words, L178H has a larger proportion of hydrophobic interactions contributing to its stability. The m value of L178H is less than that of WT in both denaturants at both temperatures, but the reduction in m value in urea is less than that in Gn-HCl. The m value represents the dependence of ∆G⁰ on denaturant concentration (m = ∆G⁰/D1/2; Chapter 2, section 2.2.5, eq. 6). In general, the greater the m value, the more effective the denaturant because less denaturant is required to achieve a given ∆G⁰. Since the reduction in m value of L178H compared to WT is less in guanidine, this indicates guanidine is less effective than urea in denaturing L178H compared to WT. This result is consistent with the D1/2 results, since the more the protein is resistant to denaturation the less effective the denaturant. The m value is also proportional to the change in solvent accessible surface area (∆ASA) of the protein going from the native (N) to the denatured (D) state (74). The reduction in m values of L178H compared to WT indicates the change in surface area going from the N to D state for L178H is less than that of WT. WT is more stable and therefore its native state may be more ordered, so going from the N to D state would result in a large ∆ASA and thus a large m value. In contrast, L178H is less stable and therefore its native state may be more disordered, so the ∆ASA between the N and D state is smaller. Hence, the smaller m values for L178H. The more disordered N state of L178H is consistent with results obtained using ANS binding. ANS binds to accessible hydrophobic surfaces in proteins. The ANS binding of L178H is nearly 4 times that of WT indicating L178H has increased exposed hydrophobic surface area (63). 50 The lower stability of L178H compared to WT apo A-I is consistent with fibrilforming proteins being generally less stable than their WT counterparts. Recent analysis of the ∆(185-243) apo A-I crystal structure model by Gursky et al. reveals how the L178H mutation may destabilize the protein (36). In the crystal structure, all the amyloidogenic mutations in apo A-I are packed in the N-terminal helix bundle (Figure 3.7A; 44). The authors hypothesize that the mutations destabilize the bundle causing increased solvent exposure and leading to fibril formation. The L178H mutation is located in the hydrophobic cluster near the bottom of the helix bundle (Figure 3.7A). The mutation replaces a non-polar alkyl residue (L) with a polar imidazole ring (H), therefore disrupting the hydrophobic cluster and destabilizing the bundle. The disruption likely also changes salt bridge interactions involved in stabilizing the helix bundle. This could explain the reduced electrostatic contributions as indicated by the larger reduction of D1/2 in urea versus Gn-HCl for L178H at both temperatures, compared to WT. Since the crystal structure is that of a truncated apo A-I and not fulllength apo A-I, the position of L178H is also examined in the EPR model of full-length apo A-I. The EPR model of lipid-free apo A-I (37) also places position 178 in a helix bundle (Figure 3.7B) which would likely be destabilized by the L178H mutation. Thus, the lower stability of L178H can be explained by either structural model as being due to destabilization of the helix bundle. 51 Figure 3.7. Location of the L178H and R173C mutations in A) ∆(185-243) lipid-free apo A-I crystal structure and B) EPR model of lipid-free apo A-I. Whether lowered stability leads to fibril formation may depend on more subtle features of alterations caused by mutation. The apo A-I variant R173C is significantly less stable than WT protein; its conformational stability (∆G⁰) is reduced by approximately 50% in Gn-HCl and 70% in urea (Figure 3.8;78). Yet, the R173C mutation does not result in fibril formation. The location of R173C in both the X-RC and EPR models is also in the helix bundle, close to the location of L178H (Figure 3.7). The mutation replaces a charged residue (R) with a non-polar residue (C), potentially disrupting helix-helix interactions within the bundle (44). Therefore, since L178H and R173C destabilize the helix bundle, leading to similar reductions in global stability (Figure 3.8) while only L178H forms fibrils, it is unlikely that decreased stability per se is the only factor in fibril formation. An additional factor may be the balance of 52 electrostatic versus hydrophobic interactions that stabilize the mutant protein. R173C is less resistant to urea denaturation and slightly less resistant to guanidine denaturation compared to L178H, indicating that, relative to each other, R173C stability is dominated more by electrostatic forces while L178H stability is dominated more by hydrophobic forces. This suggests that the ability to maintain salt bridge interactions, a hallmark feature of helix-helix association in apolipoproteins (20,45), may help prevent fibril formation. Figure 3.8. Percent of reduction in thermodynamic value of L178H and R173C compared to WT at 4⁰C in A) urea and B) Gn-HCl. The white bar represents L178H and the gray bar represents R173C (78). 3.6 Stability Over 14 Day Period The magnitude of the initial stability difference of WT and its fibril forming mutants may be less important in fibril formation than the stability change over time. To examine this parameter, the thermodynamic values for WT and L178H were obtained on proteins incubated over a period of 14 days. One set of thermodynamic values for day 1-14 was determined from one batch of protein and the entire experiment was 53 repeated twice (i.e. three batches of each protein used). During the course of the thesis research, our research collaborators found that L178H incubated at 37⁰C converts to helical fibrillar structures over time with a t1/2 of 12 days, while WT fails to form fibrils (although small spherical aggregates were observed) (63). Since the distribution of monomer, aggregates, and fibrillar species for L178H and WT in solution over time is unknown, only the change in ∆G⁰ values over time were examined because ∆G⁰ measures the change in free energy of the whole system (all the protein species). The contribution of hydrophobic and electrostatic interactions of the system is much more complicated than that of a monomeric protein and will therefore not be discussed. 3.6.1 Stability changes at 4oC To compare the stability change over time, the ∆G⁰ values of day 7 and 14 were calculated as fractions of day 1. Both proteins experienced a decrease in stability when incubated at 4oC (Figure 3.9). The decrease in stability at day 7 is similar for the two proteins in both denaturants with the fraction of day 1 ∆G0 ranging from 0.79 to 0.84. However, the stability of WT protein to guanidine denaturation at day 14 is much lower than in urea, indicating a greater loss of hydrophobic interactions after two weeks. L178H also exhibits a bigger drop in stability in the second week of incubation but the change in the ∆G0 value is similar for the two denaturants for this protein (0.60 and 0.67 for urea and guanidine, respectively). These results indicate that both proteins undergo structural changes when incubated at 4oC and that the change is greater in the second week of incubation. The interactions involved in these changes are different for the two 54 proteins as WT experiences a greater loss of hydrophobic interactions than L178H (fractional decrease to 0.44 for WT versus 0.60 for L178H). Thus, even though both proteins are destabilized by incubation at 4oC, the resulting L178H structure is more stabilized by hydrophobic interactions than the WT structure. Figure 3.9. Change in free energy after incubation at 4⁰C. The ∆G⁰ values of day 7 and 14 were calculated as fractions of day 1 ∆G⁰ values. The standard deviations for the fractions were calculated using propagation of error. The white bar represents urea and the gray bar represents Gn-HCl. 3.6.2 Stability changes at 37oC Both proteins exhibited an overall increased stability when incubated at 37oC, but to different extents in the two denaturants (Figure 3.10). The WT stability to urea is similar to that of guanidine at both day 7 and day 14, as shown by the error bar (standard deviation), suggesting the increase in the stability of WT is due to an increase in both electrostatic and hydrophobic interactions. The stability at day 7 is similar to that of day 1 having a fraction of approximately 1, whereas at day 14 the stability increases 130% (fraction = 1.3) indicating a large change in stability occurs after day 7. 55 In contrast, L178H exhibits only a minor increase in stability to urea at both Day 7 and Day 14, whereas there is a large increase in stability of L178H to guanidine at Day 7 and again at Day 14. This indicates the increased stability of L178H is due to a substantial gain in hydrophobic interactions and a minor gain in electrostatic interactions. Figure 3.10. Change in free energy after incubation at 37⁰C. The ∆G⁰ values of day 7 and 14 were calculated as fractions of day 1 ∆G⁰ values. The standard deviations for the fractions were calculated using propagation of error. The white bar represents urea and the gray bar represents Gn-HCl. 3.6.3 Structural Changes in L178H versus WT It is clear from the denaturation data that while both proteins became less stable over time when incubated at 4oC and more stable over time when incubated at 37oC, the interactions involved in the stability changes are different for the two proteins. While at 4oC, both proteins experienced the greatest decrease in ∆G0 in guanidine at Day 14, the decrease was larger for WT, implying that L178H remains more stabilized by 56 hydrophobic interactions. The same trend toward greater hydrophobic interactions in L178H is given by the 37oC incubation data which shows L178H becomes much more stable to guanidine denaturation than WT. The basis for the decrease in stability at 4oC is unclear, but it may be due to cold denaturation. At low temperatures, the Gibbs free energy of hydration of nonpolar residues increases, leading to proteins unfolding and exposing nonpolar amino acids to water (84). Although this normally occurs in proteins at temperatures much below freezing, it may occur at 4oC in apolipoproteins due to their increased flexibility and increased hydrophobic surface exposure compared to globular proteins. The increase in stability at 37oC is most likely due to increased aggregation of the protein, with L178H experiencing greater aggregation. An increase in aggregation is supported by two types of evidence. First, as previously noted, L178H forms fibrils after one month of incubation at 37oC whereas WT forms smaller, spherical aggregates (63). Second, we have observed that light scattering in both WT and L178H is increased upon incubation for two weeks at 37oC, but the increase is much higher in L178H (Baal, Y., Nguyen, H., Roberts, L., unpublished observations). While the stability of aggregate protein species versus the native protein is unknown in this mixture, it has been shown that amyloid fibrils generally have higher stabilities than their normally folded counterparts (85,86). Thus, the increased stability of WT over time at 37⁰C is due to some form of aggregation. Presumably, the effect of the L178H 57 mutation is to both lower the stability of the protein and promote its rate of aggregation over time through increased hydrophobic interactions. 58 Chapter 4 CONCLUSIONS AND FUTURE WORK Research indicates that fibril-forming proteins are generally less stable than their wild-type counterparts, and that less stable proteins are more prone to misfolding and forming fibrils with increased stability (8-10). The goal of this thesis was to determine whether stability is a factor in fibril formation in apo A-I by comparing the stability of the fibrillogenic apo A-I mutant L178H to wild-type protein using equilibrium solvent denaturation. The stability data obtained in this work indicates L178H is significantly less stable than WT protein and that its stability is dominated more by hydrophobic than electrostatic interactions. Furthermore, the greater influence of hydrophobic interactions in L178H compared to WT persists through two weeks of incubation at either 4oC or 37oC. The increased stability of both proteins upon incubation at 37⁰C is most likely due to a greater tendency to form aggregated structure, with L178H experiencing more aggregation than WT protein. The formation of fibrils or other aggregated structures does not depend solely on conformational destabilization, as other apo A-I mutants, most notably R173C, have lower stability without forming fibrils (78). There are two clusters of single site mutations in apo A-I that lead to tissue-specific fibrillar deposits (6). L178H, from the 173-178 cluster, is the first full-length fibril-forming apoA-I mutant to be characterized in terms of its stability. Therefore, it is not yet known whether mutations from both clusters lead to lowered protein stability. Furthermore, the relationship between 59 stability and direction of fibril formation has yet to be established. The G26R protein forms amyloid (β-sheet) fibrils (7) whereas the L178H forms helical fibrils (63). Thus, even if it is found that G26R and other amyloidogenic apoA-I mutants have lowered stability, additional factors must dictate the type of fibril that forms. To understand the relationships between stability, fibril type, and mutation position, it will be necessary to characterize both the structures and stability of more mutants from both clusters. Fibril formation further depends on the protein's change in hydrophobic versus electrostatic interactions over time. There is almost no research at present on timedependent changes in stability and structure of apoA-I mutants, either amyloidogenic or non-amyloidogenic. To fully understand the development of fibrillar protein in disease will require these kinds of analyses for both types of mutants. 60 Appendix A List of Abbreviations ∆ASA: change of accessible surface area (∆ASA) of the protein going from the native (N) to the denatured (D) state ∆G: free energy of denaturation (protein unfolding) ∆G°: Gibbs free energy of protein. It represents the conformational stability of the protein, and is the energy required to unfold the protein in water (0 M denaturant). ABCA1: ATP-binding cassette transporter A1 Amino acid abbreviations: Symbol 3-Letter 1-Letter Alanine Ala A Arginine Arg R Asparagine Asn N Aspartic acid Asp D Cysteine Cys C Glutamic Acid Glu E Glutamine Gln Q Glycine Gly G Histidine His H Isoleucine Ile I Name Symbol 3-Letter 1-Letter Leucine Leu L Lysine Lys K Methionine Met M Phenylalanine Phe F Proline Pro P Serine Ser S Threonine Thr T Tryptophan Trp W Tyrosine Tyr Y Valine Val V Name ANS: 8-Anilinonaphthalene-1-sulfonate, ANS binds to accessible hydrophobic surfaces in proteins Apo A-I: apolipoprotein A-I 61 ATP: Adenosine-5'-triphosphate CD: circular dichroism CETP: Cholesteryl ester transport protein CHD: coronary heart disease CVD: cardiovascular disease D1/2: the midpoint of denaturation. It represent the stability against denaturant and is the concentration in which half the population of protein is unfolded. DMPC: dimyristoylphosphatidylcholine Electrostatic interaction: the interaction of charged amino acids along a helix. For example, a positively charge amino acid nearing a negatively charge amino acid results in electrostatic attraction and amino acids with the same charge (negative or positive) nearing each other result in electrostatic repulsion. EPR: electron paramagnetic resonance Equilibrium solvent denaturation: the unfolding of protein induced by increasing the denaturant (urea or guanidine hydrochloride) concentration. FRET: Förster Resonance Energy Transfer FTIR: Fourier transform infrared spectroscopy 62 G26R: a mutation at position 26 (in apo A-I), in which a glycine (G) residue mutates to an arginine (R) Gn-HCl: guanidine hydrochloride HDL: high density lipoprotein HX-MS: hydrogen exchange and mass spectrometry Hydrophobic interaction: the contribution of hydrophobic interaction depends on the number of buried hydrophobic amino acid side chain in the folded protein (87). These hydrophobic side chains are shielded from solvent. The more shielded hydrophobic side chains in a protein generally results in greater stability. L178H: a mutation at position 178 (in apo A-I), in which a leucine (L) residue mutates to a histidine (H) LB: luria broth LCAT: lecithin-cholesterol acyltransferase LDL: low density lipoprotein m: effectiveness of denaturation MIM: mixture of indirect method, which includes circular dichroism (CD), sedimentation velocity, ANS binding, fluorescence quenching and limited proteolysis of full-length and truncated protein (35) 63 NMR: nuclear magnetic resonance PBS: phosphate buffered saline PLTP: phospholipid transfer protein POPC: palmitoyloleoylphosphatidylcholine R173C: a mutation at position 173 (in apo A-I), in which an arginine (R) residue mutates to a cysteine (C) RCT: reverse cholesterol transport rHDL: reconstituted HDL SDS: sodium dodecyl sulfate SDS-PAGE: sodium dodecyl sulfate polyacrylamide gel electrophoresis SR-BI: scavenger receptor class BI WMF: wavelength at maximum fluorescence intensity WT: Wild-type X-RC: X-ray crystallography 64 Appendix B Experimental Data For all the data listed, units for the thermodynamic parameters are as follow: D1/2 (M), ∆G⁰ (kcal•mol-1), m (kcal•M-1mol-1). Initial stability data (day 1) for WT and L178H. Urea D1/2 ∆G⁰ Gn-HCl m D1/2 ∆G⁰ m WT 4°C 2.81±0.05 6.16±0.19 2.19±0.11 1.23±0.02 4.74±0.38 3.85±0.34 L178H 4°C 1.08±0.07 1.93±0.14 1.76±0.06 1.74±0.03 2.55±0.07 1.46±0.07 WT 37°C 2.82±0.07 5.39±0.33 1.91±0.08 1.20± 0.03 3.78 ±0.08 3.14± 0.02 L178H 37°C 1.78±0.12 2.34±0.12 1.34±0.05 0.90±0.02 1.55±0.27 1.74±0.31 * All thermodynamic values are an average of values from three batches of protein (n=3); except for L178H at 37°C in urea, which the values are an average of four batches of protein (n=4) Time course data at 4°C Urea Gn-HCl D1/2 D1/2 Day ∆G⁰ ∆G⁰ m m 1 2.76±0.04 6.46±0.72 2.34±0.23 1.21±0.02 4.95±0.15 4.11±0.16 7 WT 2.82±0.06 5.09±0.20 1.81±0.11 1.18±0.03 4.14±0.12 3.51±0.15 14 2.68±0.02 4.80±0.12 1.79±0.03 1.13±0.03 2.20±0.10 1.94±0.13 1 1.70±0.04 2.63±0.10 1.55±0.05 1.02±0.05 1.85±0.09 1.83±0.06 7 L178H 1.57±0.15 2.18±0.25 1.39±0.07 0.96±0.03 1.56±0.08 1.63±0.08 14 1.51±0.04 1.75±0.13 1.17±0.12 0.87±0.04 1.11±0.08 1.27±0.04 *All thermodynamic values are an average of three values (n=3) from triplicate samples. The data were obtained from the same batch of protein for each conditions (e.g. for WT in urea, the same batch of protein is used for day 1, 7, and 14 ). 65 Time course data at 37°C Urea Gn-HCl D1/2 D Day ∆G⁰ m ∆G⁰ m 1/2 1 2.86±0.08 5.52±0.73 1.92±0.20 1.22±0.03 3.85±0.19 3.16±0.09 7 WT 2.81±0.08 5.39±0.55 1.93±0.25 1.30±0.09 4.27±0.55 3.30±0.52 14 2.56±0.08 7.38±1.36 2.88±0.46 1.22±0.02 5.09±0.23 4.17±0.13 1 1.85±0.01 2.49±0.21 1.34±0.11 0.90±0.11 1.34±0.27 1.48±0.17 7 L178H 1.81±0.17 2.58±0.36 1.44±0.29 0.82±0.06 1.75±0.12 2.15±0.14 14 1.63±0.10 2.74±0.27 1.68±0.07 0.77±0.05 2.13±0.13 2.76±0.22 *All thermodynamic values are an average of three values (n=3) from triplicate samples. The data were obtained from the same batch of protein for each conditions (e.g. for WT in urea, the same batch of protein is used for day 1, 7, and 14 ). 66 REFERENCES 1. Andrikoula, M.; McDowell, I. F.W. The contribution of ApoB and ApoA1 measurements to cardiovascular risk assessment. Diabetes, Obesity and Metabolism. 2008, 10, 271–278. 2. Florvall, G.; Basu, S.; Larsson, A. Apolipoprotein A1 is a stronger prognostic marker than are HDL and LDL cholesterol for cardiovascular disease and mortality in elderly men. J Gerontol A Biol Sci Med Sci. 2006, 61, 1262–1266. 3. Ramella, N.A.; Rimoldi, O.J.; Prieto, E.D.; Schinella, G.R.; Sanchez, S.A.; Jaureguiberry, M.S.; Vela, M.E.; Ferreira, S.T.; Tricerri, M.A. Human Apolipoprotein A-I Derived Amyloid: Its Association with Atherosclerosis. PloS ONE. 2011, 6 (7), e22532. 4. Mucchiano, G.I.; Haggqvist, B.; Sletten, K.; Westermerk, P. Apolipoprotein A-1derived amyloid in atherosclerotic plaques of the human aorta. J Pathol. 2001 Feb, 193(2), 270-275. 5. Joy, T.; Wang, J.; Hahn, A.; Hegele, R.A. APOA1 related amyloidosis: a case report and literature review, Clin Biochem. 2003, 36, 641–645. 6. Mutations in Hereditary Amyloidosis. ApoA1. http://amygene.com/aapoai.html (Accessed October 11, 2011). 7. Lagerstedt, J.O.; Cavigiolio, G.; Roberts, L.M.; Hong, H.S.; Jin, L.W.; Fitzgerald, P.G.; Oda, M.N.; Voss, J.C. Mapping the structural transition in an amyloidogenic apolipoprotein A-I. Biochemistry. 2007 Aug 28, 46(34), 9693-9. Epub 2007 Aug 1. 8. Canet, D.; Last, A.M.; Tito, P.; Sunde, M.; Spencer, A.; Archer, D.B.; Redfield, C.; Robinson, C.V.; Dobson, C.M. Local cooperativity in the unfolding of an amyloidogenic variant of human lysozyme. Nat Struct Biol. 2002, 9, 308–315. 9. Hammarstrom, P.; Jiang, X.; Hurshman, A.R.; Powers, E.T.; Kelly, J.W. Sequencedependent denaturation energetics: a major determinant in amyloid disease diversity. Proc Natl Acad Sci USA. 2002, 99, 16427–16432. 10. Raffen, R.; Dieckman, L.J.; Szpunar, M.; Wunschl, C.; Pokkuluri P.R.; Dave, P.; Stevens P.W.; Cai, X.; Schiffer, M.; Stevens, F.J. Physicochemical consequences of amino acid variations that contribute to fibril formation by immunoglobulin light chains. Protein Sci. 1999, 8, 509–517. 67 11. “Atherosclerosis.” Pubmed Health. http://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0001224/ (Acessed 27 July 2011). 12. “Coronary Heart Disease.” National Heart Lung and Blood Institute: Diseases and Conditions Index. Last modified May 2011. http://www.nhlbi.nih.gov/health/dci/Diseases/Cad/CAD_WhatIs.html (Accessed 27 July 2011) 13. Roger, V.; Go, A.; Lloyd-Jones, D.; et al. Heart disease and stroke statistics—2011 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2011, 123, e1-e192. 14. Health Care Tips. Last modified 1 May 2010.http://healthcaretips-factguide.blogspot.com/2010/05/atherosclerosis-causes-risk-factors.html (Accessed 25 Nov. 2011). 15. Assmann, G.; Gotto, A.M. Jr. HDL Cholesterol and Protective Factors in Atherosclerosis. Circulation. 2004, 109, III-8-III-14. 16. Toth, Peter P. The “Good Cholesterol” High Density Lipoprotein. Circulation. 2005, 111, e89-e9. 17. Rader, Daniel J. Molecular regulation of HDL metabolism and function: implications for novel therapies. Journal of Clinical Investigation. 2006, 116, 30903100. 18. Movva, R.; Rader, D.J. Laboratory assessment of HDL heterogeneity and function. Clin Chem. 2008, 54(5), 788-800. 19. Ma, J.; Liao, X.L.; Lou, B.; Wu, M.P. Role of apolipoprotein A-I in protecting against endotoxin toxicity. Acta Biochim Biophys Sin (Shanghai). 2004, 36, 419-424. 20. Davidson, S.; Thompson, T. The Structure of Apolipoprotein A-I in High Density Lipoproteins. J of Bio Chemistry. 2007 Aug 3, 282(31), 22249-22253. 21. Law, S.W.; Brewer, H.B. Jr. Nucleotide sequence and the encoded amino acids of human apolipoprotein A-I mRNA. Proc Natl Acad Sci USA. 1984 Jan, 81(1), 66-70. 22. Segrest, J.P.; Jones, M.K.; Loof, H.D.; Brouillette, C.G.; Venkatachalapathi, Y.V.; Anantharamaiah, G.M. The amphiphathic helix in the exchangeable apolipoproteins: a review of secondary structure and function. J Lipid Res. 1992 Feb, 33(2), 141-66. 68 23. Mishra, V.K.; Palgunachari, M.N.; Datta, G.; Phillips, M.C.; Lund-Katz, S.; Adeyeye, S.O.; Segrest, J.P.; Anantharamaiah, G.M. Studies of synthetic peptides of human apolipoprotein A-I containing tandem amphipathic alpha-helixes. Biochemistry. 1998 Jul 14, 37(28), 10313-24. 24. Frank, P.G.; Marcel, Y.L. Apolipoprotein A-I: structure-function relationships. J Lipid Res. 2000 Jun, 41(6), 853-72. 25. Wang, Guangshun. How the lipid-free structure of the N-terminal truncated human apoA-I converts to the lipid-bound form: new insights from NMR and X-ray structural comparison. FEBS Letters. 2002, 529(2), 157-161. 26. Chetty, P.S.; Mayne, L.; Lund-Katz, S.; Stranz, D.; Englander, S.W.; Phillips, M.C. Helical structure and stability in human apolipoprotein A-I by hydrogen exchange and mass spectrometry. Proc Natl Acad Sci U S A. 2009 Nov 10, 106(45), 19005-10. Epub 2009 Oct 22. 27. Nolte, R.T.; Atkinson, D. Conformational analysis of apolipoprotein A-I and E-3 based on primary sequence and circular dichroism. Biophys J. 1992 Nov, 63(5),122139. 28. Chen, H.; Sanyal, G. Direct biophysical characterization of human apolipoprotein A-1 in ISCOMs. J Pharm Sci. 1999 Nov, 88(11), 1122-6. 29. Brubaker, G.; Peng, D.Q.; Somerlot, B.; Abdollahian, D.J.; Smith, J.D. Apolipoprotein A-I lysine modification: effects on helical content, lipid binding and cholesterol acceptor activity. Biochim Biophys Acta. 2006 Jan, 1761(1), 64-72. Epub 2006 Feb 10. 30. Silva, R.A.; Hilliard, G.M.; Fang, J.; Macha, S.; Davidson, W.S. A threedimensional molecular model of lipid-free apolipoprotein A-I determined by crosslinking/mass spectrometry and sequence threading. Biochemistry. 2005 Mar 1, 44(8), 2759-69. 31. McLachlan, A.D. Repeated helical pattern in apolipoprotein-A-I. Nature, 1977 Jun 2, 267, 465-466. 32. Suurkuusk, M.; Hallén, D. Denaturation of apolipoprotein A-I and the monomer form of apolipoprotein A-I (Milano). Eur J Biochem. 1999 Oct 1, 265(1), 346-52. 33. Jonas, A.; Wald, J.H.; Toohill, K.L.; Krul, E.S.; Kezdy, K.E. Apolipoprotein A-I structure and lipid properties in homogeneous, reconstituted spherical and discoidal high density lipoproteins. J. Biol. Chem. 1990, 265, 22123–22129. 69 34. Gillotte, K.L.; Davidson, W.S.; Lund-Katz, S.; Rothblat, G.H.; Phillips, M.C. Apolipoprotein A-I structural modification and the functionality of reconstituted high density lipoprotein particles in cellular cholesterol efflux. J Biol Chem. 1996 Sep 27, 271(39), 23792-8. 35. Rogers, D.P.; Roberts, L.M.; Lebowitz, J.; Engler, J.A.; Brouillette, C.G. Structural analysis of apolipoprotein A-I: effects of amino- and carboxy-terminal deletions on the lipid-free structure. Biochemistry. 1998 Jan 20, 37(3), 945-55. 36. Mei, X.; Atkinson, D. Crystal structure of C-terminal truncated apolipoprotein A-I reveals the assembly of high density lipoprotein (HDL) by dimerization. J Biol Chem. 2011 Nov 4, 286(44), 38570-82. Epub 2011 Sep 13. 37. Lagerstedt, J.O.; Budamagunta, M.S.; Liu, G.S.; Devalle, N.C.; Voss, J.C.; Oda, M.N. The "beta-clasp" model of apolipoprotein A-I - A lipid-free solution structure determined by electron paramagnetic resonance spectroscopy. Biochim Biophys Acta. 2012 Mar, 1821(3), 448-55. Epub 2012 Jan 8. 38. Borhani, D.W.; Rogers, D.P.; Engler, J.A.; Brouillette, C.G. Crystal structure of truncated human apolipoprotein A-I suggests a lipid-bound conformation. Proc Natl Acad Sci U S A. 1997 Nov 11, 94(23), 12291-6. 39. Okon, M.; Frank, P.G.; Marcel, Y.L.; Cushley, R.J. Heteronuclear NMR studies of human serum apolipoprotein A-I. Part I. Secondary structure in lipid-mimetic solution. FEBS Lett. 2002 Apr 24, 517(1-3), 139-43. 40. Jayaraman, S.; Abe-Dohmae, S.; Yokoyama, S.; Cavigiolio, G. Impact of selfassociation on function of apolipoprotein A-I. J Biol Chem. 2011 Oct 14, 286(41), 35610-23. Epub 2011 Aug 11. 41. Lund-Katz, S.; Phillips, M.C. High density lipoprotein structure-function and role in reverse cholesterol transport. Subcell Biochem. 2010, 51, 183-227. 42. Saito, H.; Dhanasekaran, P.; Nguyen, D.; Holvoet, P.; Lund-Katz, S.; Phillips, M.C. Domain structure and lipid interaction in human apolipoproteins A-I and E: A general model. J Biol Chem. 2003, 278, 23227–23232. 43. Brouillette, C.G.; Dong, W.J.; Yang, Z.W.; Ray, M.J.; Protasevich, I.I.; Cheung, H.C.; Engler, J.A. Forster resonance energy transfer measurements are consistent with a helical bundle model for lipid-free apolipoprotein A-I. Biochemistry. 2005, 44, 16413– 16425. 70 44. Gursky, O.; Mei, X.; Atkinson, D. The crystal structure of the C-terminal truncated apolipoprotein A-I sheds new light on amyloid formation by the N-terminal fragment. Biochemistry. 2012 Jan 10, 51(1), 10-8. Epub 2011 Dec 29. 45. Silva, R.A.; Huang, R.; Morris, J.; Fang, J.; Gracheva, E.O.; Ren, G.; Kontush, A.; Jerome, W.G.; Rye, K.A.; Davidson, W.S. Structure of apolipoprotein A-I in spherical high density lipoproteins of different sizes. Proc Natl Acad Sci U S A. 2008 Aug 26, 105(34), 12176-81. Epub 2008 Aug 21. 46. Li, H.; Lyles, D.S.; Thomas, M.J.; Pan, W.; Sorci-Thomas, M.G. Structural determination of lipid-bound ApoA-I using fluorescence resonance energy transfer. J Biol Chem. 2000 Nov 24, 275(47), 37048-54. 47. Panagotopulos, S.E.; Horace, E.M.; Maiorano, J.N.; Davidson, W.S. Apolipoprotein A-I adopts a belt-like orientation in reconstituted high density lipoproteins. J Biol Chem. 2001 Nov 16, 276(46), 42965-70. Epub 2001 Sep 13. 48. Gu, F.; Jones, M.K.; Chen, J.; Patterson, J.C.; Catte, A.; Jerome, W.G.; Li, L.; Segrest, J.P. Structures of discoidal high density lipoproteins: a combined computational-experimental approach. J Biol Chem. 2010 Feb 12, 285(7), 4652-65. Epub 2009 Nov 30. 49. Segrest, J.P.; Harvey, S.C.; Zannis, V. Detailed molecular model of apolipoprotein A-I on the surface of high-density lipoproteins and its functional implications. Trends Cardiovasc Med. 2000 Aug, 10(6), 246-52. 50. Dumoulin, M.; Bader, R. A Short Historical Survey of Developments in Amyloid Research. Protein & Peptide Letter. 2006, 13, 213-217. 51. Greenwald, J.; Riek, R. Biology of Amyloid: Structure, Function, and Regulation. Structure. 2010, 1244-1260. 52. Talaga Group. Research. Amyloid aggregation mechanism. http://talaga.rutgers.edu/research/index.php (Acessed 5 May 2012) 53. Nelson, R.; Sawaya, M.R.; Balbirnie, M.; Madsen, A.; Riekel, C.; Grothe, R.; Eisenberg, D. Structure of the cross-beta spine of amyloid-like fibrils. Nature. 2005 Jun 9, 435(7043), 773-8. 54. Sawaya, M.R.; Sambashivan, S.; Nelson, R.; Ivanova, M.I.; Sievers, S.A.; Apostol, M.I.; Thompson, M.J.; Balbirnie, M.; Wiltzius, J.J.; McFarlane, H.T.; Madsen, A.; Riekel, C.; Eisenberg, D. Atomic structures of amyloid cross-beta spines reveal varied steric zippers. Nature. 2007 May 24, 447(7143), 453-7. Epub 2007 Apr 29. 71 55. Monsellier, E.; Chiti, F. Prevention of amyloid-like aggregation as a driving force of protein evolution. EMBO Reports. 2007, 8(8), 737-742. 56. Sanchez, I.E.; Tejero, J.; Gomez-Moreno, C.; Medina, M.; Serrano, L. Point mutations in protein globular domains: contributions from function, stability and misfolding. J Mol Biol. 2006, 363, 422–432 57. Teoh, C.; Griffin, M.; Howlett, G. Apolipoproteins and amyloid fibril formation in atherosclerosis. Protein Cell. 2011, 2(2), 116-127. 58. Hatters, D.M.; Zhong, N.; Rutenber, E.; Weisgraber, K.H. Amino-terminal domain stability mediates apolipoprotein E aggregation into neurotoxic fibrils. J Mol Biol. 2006 Sep 1, 361(5), 932-44. Epub 2006 Aug 7. 59. Hatters, D.M.; MacPhee, C.E.; Lawrence, L.J.; Sawyer, W.H.; Howlett, G.J. Human apolipoprotein C-II forms twisted amyloid ribbons and closed loops. Biochemistry. 2000 Jul 18, 39(28), 8276-83. 60. Raimondi, S.; Guglielmi, F.; Giorgetti, S.; Di Gaetano, S.; Arciello, A.; Monti, D.M.; Relini, A.; Nichino, D.; Doglia, S.M.; Natalello, A.; Pucci, P.; Mangione, P.; Obici, L.; Merlini, G.; Stoppini, M.; Robustelli, P.; Tartaglia, G.G.; Vendruscolo, M.; Dobson, C.M.; Piccoli, R.; Bellotti, V. Effects of the known pathogenic mutations on the aggregation pathway of the amyloidogenic peptide of apolipoprotein A-I. J Mol Biol. 2011 Apr 1, 407(3), 465-76. Epub 2011 Feb 4. 61. Wong, Y.Q.; Binger, K.J.; Howlett, G.J.; Griffin, M.D. Methionine oxidation induces amyloid fibril formation by full-length apolipoprotein A-I. Proc Natl Acad Sci U S A. 2010 Feb 2, 107(5), 1977-82. Epub 2010 Jan 19. 62. Lagerstedt, J.O.; Cavigiolio, G.; Roberts, L.M.; Hong, H.S.; Jin, L.W.; Fitzgerald, P.G.; Oda, M.N.; Voss, J.C. Mapping the structural transition in an amyloidogenic apolipoprotein A-I. Biochemistry. 2007 Aug 28, 46(34), 9693-9. Epub 2007 Aug 1. 63. Petrlova, J.; Duong, T.; Cochran, M.C.; Axelsson, A.; Mörgelin, M.; Roberts, L.M.; Lagerstedt, J.O. The fibrillogenic L178H variant of apolipoprotein A-I forms helical fibrils. J Lipid Res. 2012 Mar, 53(3), 390-8. Epub 2011 Dec 19. 64. Eriksson, M.; Schonland, S.; Yumlu, S.; Hegenbart, U.; Hutten, H.V; Gioeva, Z.; Lohse, P.; Buttner, J.; Schmidt,H.; Rocken, C. Hereditary Apolipoprotein AIAssociated Amyloidosis in Surgical Pathology Specimens. Journal of Molecular Diagnostic. 2009, 11 (3), 257-262. 72 65. Lagerstedt, J. O.; Budamagunta, M.S.; Oda, M.N.; Voss, J.C. Electron paramagnetic resonance spectroscopy of site-directed spin labels reveals the structural heterogeneity in the N-terminal domain of apoA-I in solution. J Biol Chem. 2007, 282, 9143-9149. 66. Ryan, R.O.; Forte, T.M.; Oda, M.N. Optimized bacterial expression of human apolipoprotein A-I. Protein Expression & Purification. 2003 Jan, 27(1), 98-103. 67. Lakowicz, Joseph. Principles of Fluorescence Spectroscopy. Plenum Press. New York and London. 1986. 68. Mocz, Gabor. The Use of Fluorescence in Research into Protein Structures. Intrinsic Fluorescence of Proteins and Peptides. http://dwb.unl.edu/Teacher/NSF/C08/C08Links/pps99.cryst.bbk.ac.uk/projects/gmocz/i ndex.htm. (Accessed 6 June 2012) 69. Kurtin, W.; Lee, J. The free energy of denaturation of lysozyme. Biochem. Mol. Biol. Educ. 2002, 30, 244–247. 70. Roberts LM, Ray MJ, Shih TW, Hayden E, Reader MM, Brouillette CG. Structural analysis of apolipoprotein A-I: limited proteolysis of methionine-reduced and -oxidized lipid-free and lipid-bound human apo A-I. Biochemistry. 1997 Jun 17;36(24):7615-24. 71. England, J.L.; Haran, G. Role of solvation effects in protein denaturation: from thermodynamics to single molecules and back. Annu Rev Phys Chem. 2011, 62, 257-77. 72. Monera, O.D.; Kay, C.M.; Hodges, R.S. Protein denaturation with guanidine hydrochloride or urea provides a different estimate of stability depending on the contributions of electrostatic interactions. Protein Sci. 1994, 3, 1984–1991. 73. Davidson, W.S.; Arnvig-McGuire, K.; Kennedy, A.; Kosman, J.; Hazlett, T.L.; Jonas, A. Structural organization of the N-terminal domain of apolipoprotein A-I: studies of tryptophan mutants. Biochemistry. 1999 Oct 26, 38(43), 14387-95. 74. Myers, J.K.; Pace, C.N.; Scholtz, J.M. Denaturant m values and heat capacity changes: relation to changes in accessible surface areas of protein unfolding. Protein Sci. 1995 Oct, 4(10), 2138-48. 75. Rogers, D.P.; Roberts, L.M.; Lebowitz, J.; Datta, G.; Anantharamaiah, G.M.; Engler, J.A.; Brouillette, C.G. The lipid-free structure of apolipoprotein A-I: effects of amino-terminal deletions. Biochemistry. 1998 Aug 25, 37(34), 11714-25. 73 76. Koyama, M.; Tanaka, M.; Dhanasekaran, P.; Lund-Katz, S.; Phillips, M.C.; Saito, H. Interaction between the N- and C-terminal domains modulates the stability and lipid binding of apolipoprotein A-I. Biochemistry. 2009 Mar 24, 48(11), 2529-37. 77. Rogers, D.P.; Brouillette, C.G.; Engler, J.A.; Tendian, S.W.; Roberts, L.; Mishra, V.K.; Anantharamaiah, G.M.; Lund-Katz, S.; Phillips, M.C.; Ray, M.J. Truncation of the amino terminus of human apolipoprotein A-I substantially alters only the lipid-free conformation. Biochemistry. 1997 Jan 14, 36(2), 288-300. 78. Alexander, E.T.; Tanaka, M.; Kono, M.; Saito, H.; Rader, D.J.; Phillips, M.C. Structural and functional consequences of the Milano mutation (R173C) in human apolipoprotein A-I. J Lipid Res. 2009 Jul, 50(7), 1409-19. Epub 2009 Mar 24. 79. Lyssenko, N.N.; Hata, M.; Dhanasekaran, P.; Nickel, M.; Nguyen, D.; Chetty, P.S.; Saito, H.; Lund-Katz, S.; Phillips, M.C. Influence of C-terminal α-helix hydrophobicity and aromatic amino acid content on apolipoprotein A-I functionality. Biochim Biophys Acta. 2011 Aug 5. [Epub ahead of print] 80. Tanaka et al. Influence of N-terminal helix bundle stability on the lipid-binding properties of human apolipoprotein A-I. Biochimica et Biophysica Acta. 2011, 1811, 25–30 81. Tanaka, M.; Koyama, M.; Dhanasekaran, P.; Nguyen, D.; Nickel, M.; Lund-Katz, S.; Saito, H.; Phillips, M.C. Influence of tertiary structure domain properties on the functionality of apolipoprotein A-I. Biochemistry. 2008 Feb 19, 47(7), 2172-80. Epub 2008 Jan 19. 82. Saito, H.; Dhanasekaran, P.; Nguyen, D.; Holvoet, P.; Lund-Katz, S.; Phillips, M.C. Domain structure and lipid interaction in human apolipoproteins A-I and E, a general model. J Biol Chem. 2003 Jun 27, 278(26), 23227-32. Epub 2003 Apr 22. 83. Gross, E.; Peng, D.Q.; Hazen, S.L.; Smith, J.D. A novel folding intermediate state for apolipoprotein A-I: role of the amino and carboxy termini. Biophys J. 2006 Feb 15, 90(4), 1362-70. Epub 2005 Dec 2. 84. Privalov, P.L. Cold denaturation of proteins. Crit Rev Biochem Mol Biol. 1990, 25(4), 281-305. 85. Meersman, F.; Dobson, C.M. Probing the pressure-temperature stability of amyloid fibrils provides new insights into their molecular properties. Biochim Biophys Acta. 2006 Mar, 1764(3), 452-60. Epub 2005 Nov 16. 74 86. Morel, B.; Varela, L.; Conejero-Lara, F. The thermodynamic stability of amyloid fibrils studied by differential scanning calorimetry. J Phys Chem B. 2010 Mar 25, 114(11), 4010-9. 87. Matthews, Brian W. Hydrophobic Interactions in Proteins. Encyclopedia of Life Sciences. 2001. Nature Publishing Group. www.els.net 88. Koyama, M.; Tanaka, M.; Dhanasekaran, P.; Lund-Katz, S.; Phillips, M.C.; Saito, H. Interaction between the N- and C-terminal domains modulates the stability and lipid binding of apolipoprotein A-I. Biochemistry. 2009 Mar 24, 48(11), 2529-37. 89. Rogers, D.P.; Brouillette, C.G.; Engler, J.A.; Tendian, S.W.; Roberts, L.; Mishra, V.K.; Anantharamaiah, G.M.; Lund-Katz, S.; Phillips, M.C.; Ray, M.J. Truncation of the amino terminus of human apolipoprotein A-I substantially alters only the lipid-free conformation. Biochemistry. 1997 Jan 14, 36(2), 288-300. 90. Alexander, E.T.; Tanaka, M.; Kono, M.; Saito, H.; Rader, D.J.; Phillips, M.C. Structural and functional consequences of the Milano mutation (R173C) in human apolipoprotein A-I. J Lipid Res. 2009 Jul, 50(7), 1409-19. Epub 2009 Mar 24.