Alfred H. Caspary Professor of Physics and Biological Physics in the Department of Physics
Professor of Physics and Health Sciences and Technology in the
Harvard-MIT Division of Health Sciences and Technology
Study of kinetics of beta-Amyloid Fibrillogenesis
In this project we study the nucleation and growth of amyloid beta (Abeta) protein fibrils. The deposition of Abeta fibrils in plaques in the human brain is the hallmark pathology of Alzheimer's disease. The goal of the project is to develop a reliable method of monitoring the fibrillogenesis using quasielastic light scattering. We then apply this method to elucidate the molecular mechanism of fibril formation and, eventually, to find drugs capable of blocking Abeta fibril formation. The applications of quasielastic light scattering for monitoring protein assembly are discussed in a review article published in Methods of Enzymology, vol. 309, 429-459 (1999).
Characterization of the Effect of Inhibitors on the Oligomerization of Ab
Personnel: Aleksey Lomakin, Gal Bitan, David Teplow, and George Benedek
Fibrils formed from Amyloid beta-protein (Ab) are believed to be the key pathological factor in Alzheimer's disease. We undertook studies of C-terminal fragments (CTF) of Ab which might serve as inhibitors of Ab fibrillogenesis. Three CTF's were studied: Ab [34-40], Ab[34-42] and Ab [28-42]. We first examined aggregative properties of CTF themselves. We prepared samples by first dissolving the peptide in NaOH solution, pH 10.5 and then adding 1/19 part by volume of 0.2M phosphate buffer to bring solution condition to 10mM phosphate buffer pH 7.4. We found that Ab34-40] dissolves well, filters easily through a 20 nm filter and does not aggregate for 3 days. Ab[34-42] partially passed the filter and the filtrate quickly aggregated. In the case of Ab [28-42] only ~1% passed through the filter. No subsequent aggregation was observed in this very dilute (3.7mM) filtrate. We then prepared A?/CTF mixtures by dissolving lyophilized Ab and CTF mixed together, filtered these solutions through a 20 nm filter and compared the subsequent pattern of fibrillogenesis with that of pure Ab40 and Ab42. We saw no significant effect of Ab [34-40] on Ab40 or 42. As in its pure state, Ab [28-42] in mixtures with both Ab40 and Ab42, did not pass the filter. The eventual concentration of this CTF in the filtrate (determined a posteriori by AAA) was less then 3mM and had no effect on the subsequent fibrillogenesis. Ab [34-42] in mixture with Ab40 aggregated immediately after filtration as in its pure state. However, this CTF produced a very interesting result in a mixture with Ab42.
Normally, upon dissolution at physiological conditions pure Ab42 typically shows two fractions of scattering particles: an oligomer with hydrodynamic radius in the range of 6-10nm and much larger aggregates of about 40-60nm. This distribution remains unchanged over several days. During this time the presence of few very large particles becomes more and more prominent. Similar oligomer and aggregate fractions were observed in a mixture of 6.4 ?M Ab42 with 24.5 mM of Ab[34-42]. However, in contrast to the pure Ab42 case, oligomers and their aggregates significantly grew in size with time. This result clearly indicates that Ab[34-42] CTF interacts with Ab42 molecules and is incorporated in Ab42 oligomers. Further studies of concentration dependence of this effect are clearly warranted.
We reviewed the application of Quasielastic Light scattering to studies of Ab fibrillogenesis in: Aleksey Lomakin and David B. Teplow "Quasielastic light scattering study of amyloid beta-protein fibril formation", Protein Pept. Lett. 13, 247-254 (2006.)
Intensity Fluctuations Produced by Large Ab Fibrils Diffusing through the Scattering Volume.
Personnel: Aleksey Lomakin, Jennifer J. McManus, David Teplow, and George Benedek
We have established that clusters of multiple sharp spikes often observed during fibrillogenesis of Ab are produced by orientational diffusion of a single large fibril diffusing within the scattering volume. When the axis of such a fibril is oriented perpendicular to the scattering vector q, the entire fiber scatters coherently producing large scattering intensity. When the fiber turns out of the plane perpendicular to the scattering vector (plane of reflection) by an angle on the order of 1/qL where L is the fibril length, the coherence is destroyed and intensity is dramatically reduced. Thus, while a fibril is very close to the plane of reflection, it produces groups of short spikes of intensity. These groups of spikes are interleaved with longer periods of little scattering while the fibril is significantly out of the plane of reflection. The interval between the groups of pulses is the time needed for significant, ~ ? change for fibril orientation. This time is about 1/?, where ? is rotational diffusion coefficient. For example, an A? fibril ~500nm long has a rotational diffusion coefficient of about 10 sec-1. Therefore it produces about 10 groups of spikes per second on average. Since group of spikes are independent we expect that there will be fluctuations in this number equal to about 101/2 ~3. Since the light intensity averaged over an interval is proportional to the number of spikes, we expect that the reorientation of a 500 nm fibril will produce a 30 % random fluctuation in the scattering intensity averaged over a one second interval. Since ? scales roughly as 1/L3, and the fluctuations scale as a square root of the number of groups, the fluctuation in intensity will grow with fibril length as L3/2. The sharp fluctuations of intensity due to the orientational dynamics of fibrils overlays slower fluctuations in intensity due to the translation motion of the fibril through the scattering volume. Thus, we have discovered a tool with which we can not only determine the size of the particle producing intensity spike, but we also can discriminate between fibrillar and non-fibrillar aggregates. We are currently developing software for automated evaluation of the intensity spikes and accumulation of the relevant statistical data.
Investigation of the Structure of Ab Fibrils by Small Angle Neutron Scattering
Personnel: Aleksey Lomakin, Boris Khaykovich, Noel Laso, David Teplow, and George Benedek
We studied 40 amino acid long variety of Ab peptide that has the following sequence: DAEFRHDSGY10EVHHQKLVFF20AEDVGSNKGA30II-GLMVGGVV40. The area 29-40(42) of Ab monomer is highly hydrophobic, while the rest contains polar and charged amino acids. The solid state NMR data strongly suggest that Ab monomers make hairpin turn at positions 25-26 and form pair units connected by the hydrophobic regions of Ab that stack on the top of each other forming a protofiber. It is believed that a mature Ab fibril consist of 5-6 protofibers twisted together. To establish the structural organization of protofibrils within a mature we employed Small Angle Neutron Scattering using contrast variation method. Dr. Teplow's lab has synthesized A? monomers with deuterated central hydrophobic cluster LVFF20A which is known to play a pivotal role in Abfibrillogenesis. (Many familial forms of Alzheimer's disease are associated with mutations in this cluster.) Fibrils were prepared from these deuterated monomers and contrast variation experiments were conducted at NG3 SANS instrument at NIST. These results are now being compared with computed structure factors for several plausible molecular models of Ab fibrils.
Manuscript in preparation: Aleksey Lomakin, Noel Laso, Boris Khaykovich, Boualem Haummouda, David Teplow, George Benedek (2006). "Determination of the location of the central hydrophobic cluster in amyloid b-protein fibril using contrast variation small angle neutron scattering."