35 that unlike Aβ 10-35

Transcription

1 ph dependent self assembly of E amyloid(1035) and Eamyloid(1035) PEG3000 P Thiyagarajan, a* T.S. Burkoth, b V. Urban, ac S. Seifert, d T.L.S. Benzinger, e D.M. Morgan, b D. Gordon, e S.C. Meredith e and D.G. Lynn b a Intense Pulsed Neutron Source, Argonne National Laboratory, 9700 South Cass Avenue, Argonne 60439, USA b Department of Chemistry, University of Chicago, Chicago, IL 60637, USA c European Synchrotron Radiation Facility, Grenoble, France d Chemistry Division, Argonne National Laboratory, 9700 South Cass Avenue, Argonne 60439, USA e Department of Pathology, University of Chicago, Chicago, IL 60637, USA thiyaga@anl.gov Small angle neutron and xray scattering (SANS/SAXS) studies were conducted on the structure of the aggregates formed from both the truncated model peptide βamyloid(1035) ( ) and a block copolymer βamyloid(1035)peg3000 ( PEG) in D at phs from 3.0 to 7.0. These studies indicate that aggregates into rodlike particles (fibril) and their radii are strongly dependent on the ph of the solution. The fibrilfibril association in solutions is less at ph < 5.6, but becomes larger at higher ph. PEG also assembles into rodlike particles whose radius is larger by about 30 Å than that for fibril at ph 4., while it is about 3 Å larger at higher ph. Contrast matching SAXS/SANS experiments that eliminate the coherent scattering from PEG reveal that PEG moiety is located at the periphery of the fibril. Also the mass per unit length of the peptide portion is similar for both PEG fibrils at ph 5.6. The mass per unit length of the rods from SANS provides key information on the packing of peptides in the fibril. 1. Introduction Alzheimer s Disease, the third leading cause of death in the U.S., is characterized by amyloid deposits in the brain, the primary component of which is a peptide (Aβ) consisting of 3943 amino acid residues with a molecular weight of 5000 Da. Aβ peptides selfassemble into fibrils which in turn associate into insoluble precipitates both in vivo and in vitro. Understanding the structure of these fibrils and the detailed mechanism of their formation may reveal new methods for altering or preventing the process of fibrillogenesis. The single letter amino acid designation for is shown below with 10 YEVHHQKLVFFAEDVGSNKGAIIGL 35 M the hydrophobic residues marked in boldface. The full length Aβ 14 peptide is less soluble and precipitates from aqueous solution as a mixture of ordered fibrils and disordered amorphous precipitate. We have argued that maintains the essential elements of the pathogenic agent and have developed solidstate NMR methods for its initial characterization (Benzinger et al., 1998; Gregory et al., 1998). The system has been further modified by the covalent addition of a PEG3000 moiety to the hydrophobic Cterminal residue of, forming a block copolymer ( PEG). This modification prevented interfibril association, rendering fibrillogenesis completely reversible and amenable to thermodynamic investigations. Indeed, circular dichroism studies reveal that the secondary structure of the peptides in and Aβ 10 PEG are similar (Burkoth et al., 1998), and furthermore, showed 35 that unlike the aggregation of PEG was reversed by decreasing the ph or the concentration of peptide. Contrastmatching SANS measurements on PEG in 16% D solutions at ph 7.0 showed that PEG is at the periphery of the fibrils (Burkoth et al., 1999) and the organization of the peptide in both PEG fibrils are similar. ur ultimate goal is to develop methodologies to learn about both the structure of the fibrils and their kinetics of formation for Aβ 14 under relevant physiological conditions. We plan to accomplish this through studying appropriate model peptides that are amenable to both thermodynamic and structural analysis. The present study focuses on the structure of aggregates formed by the selfassembly of PEG in solution by using SANS and SAXS at a synchrotron source. The structure of the aggregates was investigated at 4 different phs. Contrastmatching SANS and SAXS experiments were conducted on PEG in order to study the organization of the peptides in both and PEG fibrils. Timeresolved SAXS studies were conducted on the kinetics of aggregation of PEG and the results from those studies will be presented elsewhere.. Experimental, PEG were synthesized using an ABI431A peptide synthesizer and the amino acids from Applied Biosystems as detailed in Burkoth et al. (1999). Cleavage by TFA and deprotection yielded a linear PEG 3000 covalently bound to the carboxyl terminus of the peptide. Molecular masses of all peptides were verified by MALDI TF mass spectroscopy. The peptide was solubilized in 500 ml of with 0.1% TFA(v/v) and lyophilized. SANS experiments were performed at the timeofflight smallangle neutron diffractometer (SAND) at the Intense Pulsed Neutron Source at Argonne National Laboratory (Thiyagarajan et al., 1998). SAND uses pulsed neutrons and timeofflight techniques for sorting out the wavelength of the neutrons in each pulse. The scattered neutrons are measured by using a 3 He gas filled position sensitive 40 cm x 40 cm area detector binned into 18 x18 arrays, while cold neutrons with wavelengths in the range of 0.5 to 14 Å in each pulse are binned into 68 different constant t/t = 0.05 wavelength channels. The simultaneous usage of neutrons with a wide range of wavelengths enables SAND to provide data in a wide Q region (Q = 4πsin(θ)/λ, where θ is the scattering angle and λ is neutron wavelength) of Å 1 in a single measurement. SANS data for each sample were corrected for the backgrounds from the instrument, suprasil cell and the solvent. The corrected azimuthally averaged data were placed on an absolute scale by following routine procedures (Thiyagarajan et al., 1997). Table 1 presents the list of samples considered for the present study. SANS measurements were made on 5.8 mg/ml or 11.5 J. Appl. Cryst. (000). 33, 535±539 # 000 International Union of Crystallography Printed in Great Britain ± all rights reserved 535

2 Table 1 List of Samples Studied using SANS and SAXS SANS Aβ PEG 100% D Aging Time (hrs.) ph Conc. g/l Crosssectional I c (0) (Å 1. cm 1 ) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Aβ PEG 16% D > ± 4 1.3E±5.4E4 > ± E3±5.E4 Aβ 100% D ± 0.0 ± ± ± ± ± ± ±.0E ± ±.0E ± ± 6.0E4 Low Q 70 ± ± 7.0E ± ± 9.0E ± ± Low Q 51 ± ± ± ± 1.0E03 Aβ 16% D ± E3±5.4E4 SAXS Aβ ± 1 AβPEG ± 1 50:50 Aβ+Aβ PEG in ± 1 mg/ml PEG in 0.1M phosphate buffer in 99% D in cylindrical suprasil cells with a path length of 5 mm at o C. The evolution of the scattering signal over several hours was investigated by alternating the samples and taking scans of 1 hour for each measurement. In order to determine the structural organization of the peptide moiety in the PEG aggregate it is necessary to contrastmatch the scattering from the PEG moiety by dispersing the copolymer in buffers containing 16% D (Thiyagarajan et al., 1995). These solutions were measured in 1 mm path length suprasil cells in order to reduce the multiple scattering effects from. SAXS experiments were conducted at the instrument on the undulator beam line of BESSRCCAT at APS (Seifert et al., 1999). For these experiments, we used a 18 cm x 18 cm position sensitive gas filled area detector and xrays with energies in the range of 8 to 13.5 kev. The sample to detector distance was 3.75 m, providing data in a Q region of to 0.09 Å 1 at 8 kev. Typical measurement time is 30 seconds with this detector because the beam flux has to be attenuated by three orders of magnitude due to the count rate limitation of the gas detector. The data were corrected for the backgrounds from the instrument, solvent and the quartz capillary. For SAXS measurements the samples in 0.1M phosphate buffer at similar concentrations were filled in 1 mm diameter quartz capillary tubes. Under these conditions PEG is invisible to xrays due to the similar electron densities of PEG and the buffer (Thiyagarajan et al., 1995). Hence SAXS provides easy access to obtain the speciesspecific structural organization of the peptides in the aggregates in PEG solutions. In this paper we present only a few SAXS data sets to illustrate this fact. EM studies (Burkoth et al., 1998; Burkoth et al., 1999) showed that PEG form long fibrils. In addition, the low Q region of the SANS data for PEG solutions exhibited a powerlaw scattering that varied as Q 1.0. Hence we interpreted the SANS/SAXS data for the above systems by using modifiedguinier analysis for rodlike forms, where Ln[I(Q) Q] is plotted versus Q (Porod, 198). Rodlike particles give rise to a linear curve at Q max.r c 1.1 in the modified Guinier plot. From the slope of the straight line in the above region, the crosssectional radius of gyration, R c, can be derived from the relation R c / = slope. The average crosssectional radius of the rod is then given by R = R c. Since the SANS data are on an absolute scale, the yintercept of the linear fit I c (0), can be used to determine the mass per unit length, m, in Da/Å: 1000Ic(0) d NA m = (1) πc ( ρp ρs) where C is the concentration in g/l and d is the mass density in g/cm 3 of the particle. N A is Avogadro s number. ρ p and ρ s are the scattering length densities in cm of the particle and the solvent. In principle, SAXS data can also be used to determine the mass per unit length, provided the data are on an absolute scale. In that case the scattering length densities in the above equation will correspond to the electron densities of the particle and the solvent. Since the data were not placed on an absolute scale, we could not use SAXS to obtain mass per unit length. 536 P. Thiyagarajan et al. J. Appl. Cryst. (000). 33, 535±539

3 a) Figure 1 SANS of 11.5 mg/ml PEG in 99% D in 0.1M phosphate buffer at different ph conditions. b) 3. Results and Discussion The differential neutron scattering cross section I(Q) for PEG samples in 0.1M phosphate buffer are shown in Fig. 1. The samples were aged in 99% D for ~35 hours at ph between 3.0 to 7.0. The observation that the scattering signal monotonically increases with increasing ph indicates that the particles are larger at higher ph, and the growth kinetics of the particles must be faster at higher ph. Modified Guinier plots for rodlike form are presented for the SANS data of PEG in Fig.. Figs. a, b and c are samples in 0.1M phosphate buffer in 99% D at ph 4., 5.6 and 7.0 respectively, and the results from the modified Guinier analysis are presented in Table 1. The aggregates in these solutions are rodlike as evidenced by the linear regions at Q.R c <1.1. Table 1 also includes SANS of PEG at ph 3.0 in D buffer and provides further evidence for the presence of rodlike particles. The aggregates of PEG at ph 4. have radii of 9±1 Å. An interpretation of the solidstate NMR analyses of (Benzinger et al., 1998) suggests that the peptide backbone is extended and organized perpendicular to the rod axis. In a fully extended βsheet conformation, the length of the peptide is ca. 90 Å and these measurements appear to be too small to account for a rod defined by the extended peptide. However, at ph 5.6 we observe two linear regions for, with radii of 4±5 Å and 70±8 Å. The smaller radius corresponds to a fully extended peptide while the larger radius would be consistent with the dimeric fibrilfibril bundles characteristic of this peptide (Burkoth et al., 1998). At ph 7.0 there seems to be a population of rods with a radius around 43± Å, but the radius of the rodlike particles increased enormously, reaching a value of 51±16 Å. Since the fibrils are large and polydisperse, the condition required for modified Guinier approximation did not apply in the determination of the larger radius. However, the larger radius indicates that at this ph fibrilfibril association was dominant leading to complex network of structures. This sample was also quite unstable and formed small precipitates during the course of the experiment. In contrast, the particles in PEG at ph 4. have a radius of 60± Å with the PEG moiety increasing the radius of the rods by 31± Å. At higher phs a single, larger radius is observed with a c) Figure Modified Guinier plots of the SANS data of PEG at a) ph 4., b) ph 5.6, c) ph 7.0. J. Appl. Cryst. (000). 33, 535±539 P. Thiyagarajan et al. 537

4 greatly attenuated ph effect (Table 1), consistent with the PEG moiety preventing the fibrilfibril association (Burkoth et al., 1998). At all phs, the PEG moiety increases the radius of the rods and clearly show that fibrilfibril association is virtually eliminated. The aggregates of PEG seem to be shorter than those of, as evidenced by the bend over of the data in the low Q region for samples at phs 5.6 and 7.0. Fig. 3 depicts the crosssectional radius and I c (0) values obtained from the analysis of the SANS data. Clearly, the radius of the rods increases rapidly with increasing ph, reaching a maximum around ph 5.6, but the I c (0) values exhibit a linear increase with ph and do not saturate. We believe that this increase in I c (0) is due to the increase in the number density of rods rather than real increase in the mass per unit length (see equation above). To investigate the organization of the PEG moiety in the particles formed by PEG, SANS studies were performed on both PEG solutions aged over several days at ph 5.6 in 16% D. At this D level, PEG becomes invisible to the neutrons and hence the measured coherent scattering must come solely from the peptide. Fig. 4a shows the modified Guinier plot of the normalized SANS data for the PEG solutions. The slope and intercept for both solutions are similar, implying that the radius and the mass per unit length of the particles are similar. Thus the peptide moiety organizes into a fibril with a radius = 44±4Å and mass per unit length = 3680 ± 150 Da/Å. The following parameters were used for the determination of the mass per unit length: chemical formula for = C 133 H 199 D 5 N 34 36, concentration of in 11.5 g/l of PEG3000 = 5.67 g/l, ρ p = 1.995x10 10 cm, ρ s = 0.547x10 10 cm, density = 1.35 g/cc. peptide with a molecular weight of 860 Da arranged perpendicular to the fibril axis at every 5 Å in a ribbonlike structure would correspond to a mass per unit length of 57 Da/Å. The larger value for the mass per unit length from SANS implies that lamination of the extended βsheet structures must occur in the fibril. These samples took over 0 hours for each SANS measurement due to the smaller contrast for the peptides in 16% D and the quality of the data is still poor. Nevertheless, the contrastmatching SANS measurements provided key information on the packing of the peptides in both fibrils. a) Ic(0) I c (0) (cm 1.Å 1 ) ph Figure 3 Crosssectional radius and I c (0) values for the PEG samples as a function of ph. While the radius values plateau above ph 5.6 the I c (0) values show a monotonic increase with ph. b) Figure 4 Modified Guinier plot of a) SANS data for PEG solutions at ph 5.6 in 16% D, b) SAXS data for PEG solution and 50:50 mixture of PEG solution. Note that the radii of the particles are similar for all solutions as PEG is invisible in all cases. Since synchrotrons provide extremely high flux xray beams, high quality data can be obtained for samples even at low contrast. ur studies indicate that PEG (electron density = 9.46x10 10 cm ) is slightly visible in (electron density = 9.36x10 10 cm ). However, in 0.1M phosphate buffer, PEG is completely invisible to xrays. Fig. 4b depicts the modified Guinier plot of the SAXS data for PEG solution at ph 5.6. A radius of 45±1 Å was determined, confirming the radius from the contrastmatching SANS of PEG solution. In contrast to the SANS experiment, the SAXS experiment took only a few seconds and the quality of the data was high. Another SAXS experiment conducted on a solution containing 50:50 mixture by weight of PEG at ph 5.6 aged for about P. Thiyagarajan et al. J. Appl. Cryst. (000). 33, 535±539

5 hours (See Table 1) produced data that resembled that for pure Aβ 10 PEG sample (Fig. 4b). This sample was measured by using kev xrays that increased the Q min to Å 1. The crosssectional radius of the rodlike particles in the mixture was identical to that determined for the PEG fibril, with again no evidence for the presence of the larger fibrilfibril associated particles seen in Fig. b for. 4. Conclusions SANS/SAXS studies on solutions of PEG were carried out to further understand the structure of the fibrils that evolve with time at different phs. The aggregates were determined to be rodlike particles with an average crosssectional radius that increases at higher ph for. At ph > 5.6, fibrilfibril associated arrays dominate in the structures at long incubation times. Most importantly, the fibril density measurements implied that the βsheet structures determined from solidstate NMR measurements were laminated into βhelical arrays. The degree of lamination seems to be ph dependent, possibly explaining the smaller radius obtained in acidic media. These studies therefore suggest a global structure for the entire fibril. Contrastmatching studies on PEG with both SANS and SAXS clearly indicate that PEG is at the periphery of the fibril. The cross sectional radius of the PEG fibrils also increases with ph, but unlike, plateaus around 65 Å, This data provides further conclusive evidence that the PEG modification dramatically reduces fibrilfibril associations. The packing of peptides seems to be similar in the aggregates of both PEG. Therefore, significant insight has been provided both as to the structure of this amyloid fibril and ways of controlling its selfassociation and precipitation. We thank ANC and the NIH (R1 RR 173, D.G.L.) for support and the NIH (5 T3 HL0737, T.S.B.; 5 T3 GM0781, T.L.S.B.), for fellowships. We thank Denis Wozniak at IPNS for his assistance in the SANS experiments. This work benefited by the use of IPNS and BESSRC at APS supported by US Department of Energy, BES Material and Chemical Sciences under the contract # W31109 ENG38 to U of Chicago. References Benzinger, T.L.S., Gregory, D.M., Burkoth, T.S., MillerAuer, H., Lynn, D.G., Botto, R.E. & Meredith, S.C. (1998). Proc. Nat. Acad. Sci. USA, 95, Burkoth, T.S. (1999). The Synthetic Control of Peptide Structure Apolipoprotein E (416) and βamyloid (1035), Dissertation submitted for the Degree of Doctor of Philosophy, Department of Chemistry, University of Chicago. Burkoth, T.S., Benzinger, T.L.S., Jones, D.N.M., Hallenga, K., Meredith, S.C. &. Lynn, D.G. (1998). J. Am. Chem. Soc., 10, Burkoth, T.S., Benzinger, T.L.S., Urban, V., Lynn, D.G., Meredith, S.C. & Thiyagarajan, P. (1999). J. Am. Chem. Soc., 11, Gregory, D.M., Benzinger, T.L.S., Burkoth, T.S., MillerAuer, H., Lynn, D.G., Meredith, S.C., & Botto, R.E. (1998). Solid State Nucl. Mag. Reson., 13, Porod, G.(198). Small Angle XRay Scattering, Chapter, edited by Glatter,. & Kratky,., pp New York, NY: Academic Press. Seifert, S., Winans, R. E., Tiede, D. M. & Thiyagarajan, P. (1999). J. Appl. Cryst.,(This Proceedings). Thiyagarajan, P., Chaiko, D.J. & Hjelm, R.P., (1995) Macromolecules, 8, Thiyagarajan, P., Epperson, J.E., Crawford, R.K., Carpenter, J.M., Klippert, T.E. & Wozniak, D.G., (1997). J. Appl. Cryst., 30, Thiyagarajan, P., Urban, V., Littrell, K., Ku, C., Wozniak, D.G., Belch, H., Vitt, R., Toeller, J., Leach, D., Haumann, J.R., strowski, G.E., Donley, L.I., Hammonds, J., Carpenter, J.M. & Crawford, R.K. (1998), ICANS XIV The Fourteenth Meeting of the International Collaboration on Advanced Neutron Sources, June 1419, 1998, Starved Rock Lodge, Utica, Illinois, edited by Carpenter, J.M. & Tobin, C.,Volume, Springfield, VA: National Technical Information Service. J. Appl. Cryst. (000). 33, 535±539 Received 1 June 1999 Accepted 5 November