Kinetic Intermediates in Amyloid Assembly

Supporting Information

Kinetic Intermediates in Amyloid Assembly Chen Liang‡, Rong Ni‡, Jillian E. Smith‡, W. Seth Childers†, Anil K. Mehta*, and David G. Lynn*

MATERIALS AND METHODS: Amidation of Glutamate side chain carboxylic acid: N-Fmoc-Glutamic acid t-Butyl ester (Anaspec, Inc.) (1 molar equiv.) was dissolved in dry CH2Cl2 containing dried molecular sieves (microwave), N,N-diisopropylethylamine (DIPEA, 2.6 equiv.), and 2-(1H-benzotriazol-1yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU, 1 equiv.). Either methylamine hydrochloride or dimethylamine hydrochloride (1.1 equiv.) were added sequentially with stirring at ambient temperature for 2 hr and monitored by thin layer chromatography (TLC: hexane / ethyl acetate (1:1, v/v), Rf = 0.25 and 0.3 for monomethyl and dimethyl products, respectively). When the starting material was no longer observed on the TLC plate, the reaction mixture was concentrated in vacuo and the residue purified with flash chromatography (hexane / ethylacetate (1:1, v/v)) to give the each product as colorless oils. Fmoc-Gln(NHMe)-OtBu: 1H NMR (400 MHz, CDCl3) δ 7.74 (d, J=7.6 Hz, 2H), 7.58 (dd, J=6.8, 4.8 Hz, 2H), 7.37 (t, J= 7.6, 7.2 Hz, 2H), 7.29 (ddd, J= 7.2, 1.6 Hz, 2H), 6.14 (s, 1H), 5.72 (d, J=6.0 Hz, 1H), 4.36 (dd, J= 7.2, 2.4 Hz, 2H), 4.19 (dd, J=7.2, 6.8 Hz, 2H), 2.77 (d, J= 4.4 Hz, + 3H), 2.44 (m, 2H), 2.21 (m, 3H), 1.91 (m, 1H), 1.44 (s, 9H). MS (ESI) m/z 439.6 (M+H) . Fmoc-Gln(NMe2)-OtBu: 1H NMR (400 MHz, CDCl3) δ 7.74 (d, J=7.6 Hz, 2H), 7.59 (dd, J= 8.0, 7.6 Hz, 2H), 7.37 (dd, J= 7.6, 7.2 Hz, 2H), 7.28 (dd, J= 7.6, 7.6 Hz, 2H), 5.84 (d, J=8.0 Hz, 1H), 4.34 (m, 2H), 4.21 (m, 2H), 2.94 (s, 3H), 2.92 (s, 3H), 2.50-2.28 (m, 2H), 2.18 (m, 1H), 2.02 (m, + 1H), 1.45 (s, 9H). MS (ESI) m/z 452.6 (M+H) Hydrolysis of t-Butyl esters: The above amidation products (0.3 mmol, 1 equiv.) were dissolved in dichloromethane (9.6 mmol, 32 equiv.) before trifluoroacetic acid TFA (0.3 mL, 3.9 mmol, 13 equiv.) and triethylsilane (0.12mL, 0.75 mmol, 2.5 equiv.) were added at room temperature with stirring. When the starting ester was completely consumed (~1 hr) as monitored by TLC (ethyl acetate, Rf = 0.25), the solvent was removed in vacuo. The resulting residues were triturated with cold diethyl ether, the white precipitate collected by centrifugation (16,100 xg for 5 min) and the pellet air-dried in fume hood. The white powder was dried by lyophilization overnight and stored for peptide solid-phase synthesis in a desiccator without further purification. Fmoc-Gln(NHMe)-OH: 1H NMR (400 MHz, CDCl3) δ 7.77 (d, J=7.2 Hz, 2H), 7.60 (t, J= 7.2, 6.4 Hz, 2H), 7.41 (t, J= 7.6 Hz, 2H), 7.33 (dd, J= 7.6, 7.2 Hz, 2H), 6.16 (m, 1H), 6.05 (d, J= 6.0 Hz, 1H), 4.38 (m, 3H), 4.23 (t, J= 6.8 Hz, 1H), 2.87 (d, J= 4.4 Hz, 3H), 2.60-2.40 (m, 3H), 2.20 + (m, 1H), 2.06 (s, 1H, OH). MS (ESI) m/z 383.5 (M+H) .

S1

Fmoc-Gln(NMe2)-OH: 1H NMR (400 MHz, CDCl3) δ 9.20 (brs, 1H, OH), 7.77 (d, J=7.2 Hz, 2H), 7.60 (dd, J= 7.2, 6.4 Hz, 2H), 7.41 (t, J= 7.6 Hz, 2H), 7.32 (dd, J= 8.0, 7.2 Hz, 2H), 6.16 (d, J= 6.0 Hz, 1H), 4.38 (m, 2H), 4.27 (m, 1H), 4.22 (dd, J= 7.6, 5.6 Hz, 1H), 3.04 (s, 3H), 3.00 (s, 3H), 2.88 (m, 1H), 2.54 (ddd, J= 16.4, 6.0 Hz, 1H), 2.29 (m, 1H), 2.04 (m, 1H). MS (ESI) m/z +

397.6 (M+H)

Peptide synthesis and purification: Aβ(16-22)E22QRX (R=XHCH3, (CH3)2) were synthesized using standard FMOC/HBTU chemistry with the FMOC rink-amide polystyrene resin (Anaspec, Inc.) on a Rainin Symphony Quartet peptide synthesizer with the first 5 residues (VFFAQ) single coupled for 2 hr and the last two residues (Lys16 and Leu17) double coupled for 4 hr total. The peptide N-terminus was capped manually on the resin by incubating overnight with acetic acid and HBTU/NMM. Ac-A(16-22)E22Q and [1-13C]F19 A(16-22)E22Q were synthesized using a Liberty CEM Microwave Automated Peptide Synthesizer (Matthews, NC, USA). Microwave assisted FMOC deprotection was completed using 20% piperdine in dimethylformamide at 45-55°C for 180 sec, and washed by 3 times with dimethylformamide. Each amino acid coupling step was performed using 0.1M FMOC protected amino acid and activated with 0.1 M 2-(1Hbenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU), and 0.2 M N,N – diisopropylethylamine (DIEA) in DMF. Coupling temperatures using microwave were maintained between 75-82°C for 330 sec, then rinsed with three aliquots of dimethylformamide. Final acetylation of the N-terminus was achieved by addition 20% acetic anhydride in dimethylformamide. Resin was filtered and washed with dichloromethane and dried in a vacuum desicator. Peptides were cleaved from dried resin by addition of cleavage cocktail (90 vol% TFA, 5 vol% thioanisole, 3 vol% ethanedithiol and 2 vol% anisole) for 4 hr at room temperature and the resulting filtrate was added drop wise to cold diethylether, centrigued and extracted in diethylether three more times. Peptides were purified by reverse-phase HPLC (Waters Delta 600) using a Waters Atlantis C-18 preparative column (19 x 250 mm) and employing a linear gradient at 20 mL/min starting at 20% acetonitrile and ending with 55% acetonitrile over 35 min. After removing acetonitrile in rotovap, the peptide fractions were frozen and lyophilized to yield a peptide powder that was stored at 4°C in a vacuum desiccator. Product mass was confirmed by MALDI-TOF on a Voyager-DETM STR Biospectrometry Workstation using -cyano-4hydroxycinnamic acid (CHCA) as matrix. Fibril Assembly: Aβ(16-22)E22Q powder was dissolved in 20% acetonitrile/water with vortexing, acidified with 0.1 vol% TFA, and incubated at room temperature for 1-3 weeks to allow for thermodynamic assembly. For samples where kinetics were followed, lyophilized peptide was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, Sigma) and sonicated for 30 min and then either dried in a vacuum desiccator for 48 hours or under a stream of dry N2 gas to a clear film. Film was dissolved in 20% acetonitrile/water with 0.1 vol% TFA, bath sonicated for 10 min and incubated at room temperature for assembly. Seeded Assembly: 0.8 mM [1-13C]F19 A(16-22)E22Q monomers were dissolved in 20% acetonitrile/water with 0.1% TFA and bath sonicated for 10 min. 1% mature [1-13C]F19 A(1622)E22Q fiber that had been confirmed as parallel by FT-IR were vortexed for 10 seconds to form seeds and added to the monomer solution,. S2

Transmission Electron Microscopy (TEM): A TEM copper grid with a 200 mesh carbon support (Electron Microscopy Sciences) was covered with 10 μL of a diluted peptide solution (0.05 mM to 0.1 mM) for 1 min before wicking the excess solution with filter paper. 10 μL of the staining solution, either (2% uranyl acetate, Sigma-Aldrich, for Figure 2 of main text or methylamine tungstate, (Ted Pella, Inc) was added and incubated for 2 min, excess solution was wicked away, and the grids were placed in desiccators to dry under vacuum overnight. Methylamine tungstate was used for early time point micrographs due to superior contrast of ribbons compared to uranyl acetate A Hitachi H-7500 transmission electron microscope was used to image the samples at 75 kV. Circular Dichroism spectroscopy: Jasco-810 circular dichroism (CD) spectropolarimeter was used to record CD spectra in a 20 μL cell with a 0.1 mm path length at room temperature. The reported spectra represent the average of three scans between 260 nm to 190 nm with a step size of 0.2 nm and a speed of 100 nm/s. Ellipticity ,, in mdeg was converted to Molar ellipticity [] with [] = /(10 x c x l), where c is the peptide concentration in moles/L and l is the pathlength in cm. Attenuated Total Reflectance Fourier Transform Infrared (AT-FTIR): Aliquots (10μL) of peptide solution were dried as thin films on an Pike GaldiATR (Madison, WI, USA) ATR diamond crystal. FT-IR spectra were acquired using a Jasco FT-IR 4100 (Easton, MD, USA) at room temperature and averaging 500 to 800 scans with 2 cm-1 resolution, using either an MCT or TGS detector, 5mm aperture and a scanning speed of 4mm/sec. Spectra were processed with zero-filling and a cosine apodization function. IE-IR spectra were normalized to the peak height of the 12C band. X-ray powder diffraction sample preparation: Mature fibrils were centrifuged at 16,100 xg for 10 min and the pellet was frozen and lyophilized to yield a dry powder and used directly for X-ray powder diffraction. The diffraction patterns were measured with Bruker APEX-II diffractometer with graphite monochromated Cu radiation K-alpha radiation, λ= 1.54184 Å, 40 kV and 35 mA, with a 0.5 pinhole collimator and with exposure times of 300 s per frame. The sample was loaded into a 0.2 mm mylar capillary. The data integration software XRD2SCAN[1] was used to convert the two dimensional data into a -2 scan. Solid-State NMR: NMR spectra were collected with a Bruker (Billerica, MA) Avance 600 spectrometer using a Bruker 4mm HCN BioSolids magic-angle spinning (MAS) probe. MAS frequency was actively controlled at 2 Hz with cooling and spinning air exit temperature maintained below -1 °C to ensure MAS and RF heating did not denature the samples. 13C (150.8 MHz) CP-MAS spectra before and after DQF-DRAWS and 13C{15N}REDOR experiments confirmed that the samples did not change during the experiment. Samples were centered in 4mm MAS rotors with boron nitride spacers. DQF-DRAWS experiments[2-3] were implemented with the addition of spin-temperature alternation of the initial 1H (600.3 MHz) 1.9µs /2 pulse to the pulse sequence and phase cycling previously described[4]. 1H cross-polarization RF fields were ramped from 50 to 70 kHz and the 13 C (150.8 MHz) cross-polarization RF field was kept constant at 50kHz. SPINAL-64[5] 1H decoupling at 128 kHz was used during both dipolar evolution and acquisition. For DRAWS S3

recoupling, a 41.23 kHz 13C RF field, measured by fitting a 13C nutation curve to a sine function with a decaying exponential, was used. The rotor period (206.2μs  r = 4.85 kHz) was set to 8.5 times the 13C  pulse length. T2DQ (Figure S7) was measured in separate experiments by placing a composite90x-90y-90x DQ coherence refocusing pulse between the two DRAWS evolution periods[6] which were fixed at 40-Tr, resulting in maximum DQ coherence excitation efficiency. Data points are the ratio of the sum of center- and sideband-integrated peak intensities for each evolution time to the 13C CP-MAS intensities. Error bars were calculated using the noise of each spectrum as the maximum peak height deviation. DQF-DRAWS curves were calculated using SIMPSON[7], where an array of 13C spins were approximated with a three spin “infiniteloop” model[6] and chemical shift tensor components 11=74.1ppm, 22=6.0ppm and 33=80.1ppm, which were measured from the 13C CP-MAS spectra. The infinite loop model consists of three spins with identical dipolar couplings but with the orientation of the CSA to dipolar tensors identical between spins 1-2 and spins 3-1. The effects of DQ-relaxation were approximated by multiplying the calculated SQ intensity with a decaying exponential[6] of the form ∗ . DRAWS curves were calculated from 3Å to 7Å and used to find a best fit to the experimental data points by minimizing the residual:

, where xi and wi are the experimental data and error respectively.

S4

Figure S1. FT-IR of 1mM Aβ(16-22)E22Q assembled at acidic pH (0.1% TFA) in 20% CH3CN. (Bottom) 1hr after dissolution, the strong 1625 cm-1 amide-I stretch is consistent with the growth of β-sheet structure. The broad transition centered around 1675 cm-1 is consistent with unassembled peptide. (Top) After 23 days, the 1625 cm-1 amide-I stretch indicates the peptides are in a -strand conformation and the absence of a 1690cm-1 band is consistent with anti-parallel orientation. The weak transition at 1676 cm-1 suggests ordered glutamine side-chains[8] consistent with a Q-track forming along parallel β-sheets.

S5

Figure S2. FT-IR spectra of (black) [1-13C]F19 Aβ(16-22)E22Q after one hour and (red) [113 C]F19 Aβ(16-22) nanotubes after 3 weeks of assembly. The 1 hr spectrum (black) has 12C and 13 C band positions similar to that of mature [1-13C]F19 Aβ(16-22) (red), which has been confirmed to be anti-parallel out-of-register by one residue by solid-state NMR distance measurements.[9]

S6

Figure S3. Fiber widths measured from TEM of Aβ(16-22)E22Q assembled under acidic pH conditions for 20 days. Red curve is the Gaussian fit with center and widths (w) to the measured fiber width histogram, Gaussian width is 2.3 nm .

S7

Figure S4. FT-IR of 1 mM [1-13C]F19Aβ(16-22)E22Q assembled at acidic pH (0.1% TFA) in 20% CH3CN and normalized to the 12C transition. Expansion highlights changes in 13C amide-I band with the positions of the 13C band indicated for parallel (p), anti-parallel in-register (i) and anti-parallel out-of-register (o) peptide orientations from this work and ref [9]. The intensity of the broad 1680 cm-1 and the 13C transition as well as the 12C-13C splitting decrease over time, consistent with peptide initially assembling as anti-parallel out-of-register -sheets and a transition to parallel -sheets.

S8

Figure S5. CD of 2.5mM Aβ(16-22)E22Q fibers assembled at acidic pH (0.1% TFA) in 20% CH3CN.The ellipticity minimum at 217 nm is consistent with peptides adopting a β-sheet conformation.

S9

Figure S6. X-ray powder diffraction of Aβ(16-22)E22Q fibers exhibit a cross-β conformation[10with reflections at 4.76 Å from hydrogen-bonded peptides and at 10.1 Å from -sheets that stack on top of each other. The sheet-stacking (lamination) distance of ~10 Å is similar to that observed for other amyloid assemblies[9, 12, 14-15] and is ~ 2 Å longer than that seen for polyglutamine peptides[16].

15]

S10

Figure S7. 13C Double-Quantum Filtered DRAWS calculated build-up curves comparing linear arrays of 13C spins to a 3-spin loop. The DQ build-up for the linear array of spins appears to be asymptotically approaching the 3-spin loop.

S11

Figure S8. 13C Double Quantum relaxation (T2DQ) of [1-13C]L17 Aβ(16-22)E22Q assembled at acidic pH (black). T2DQ was measured by inserting a composite 180° pulse between DRAWS pulses.[6] Solid line is best fit of a monotonically decaying exponential to the experimental data points.

S12

[1-13C]F19 A(16-22), pH2 anti-parallel in-register

[1-13C]F19 A(16-22), pH6 anti-parallel out-of-register

[1-13C]F19 A(16-22)E22Q parallel

Figure S9. Illustration of peptide backbone with [1- 13C] enriched F19 (indicated by red star) for A(16-22) congeners. The position and intensity of the 13C band and the magnitude of the 13C13 C splitting depends on both the peptide strand orientation and registry (Table S1), therefore parallel, anti-parallel out-of-register and anti-parallel in-register have distinct [1-13C]F19 IE-IR spectra.

S13

Figure S10 (A-C) TEM and (D) isotope-edited IR spectra of [1-13C]F19 enriched A(1622)E22Q, A(16-22)E22QNHCH3 and A(16-22)E22QN(CH3)2 peptides assembled for three weeks. (A) A(16-22)E22Q assembles into (D, black) parallel non-twisted fibers. (B) A(16S14

22)E22QNHCH3 assembles into twisted bundled fibers with (D, red) broad 12C and 13C IR bands, consistent with a mixture of anti-parallel out-of-register, anti-parallel in-register and parallel assemblies. The band at ~1677 cm-1 from ordered glutamine side chains suggests H-bonding between side chain N-methyl amide groups. (C) A(16-22)E22QN(CH3)2 peptides assemble into nanotubes and ribbons, similar to A(16-22) and A(16-22)E22L[9] assemblies and have (D, blue) IR spectra with a 12C-13C splitting of ~40cm-1 and a weak band at ~1695 cm-1 consistent with anti-parallel, out-of-register -sheets[9]. The lack of a 1677 cm-1 band suggests unordered glutamine side chain amide groups, presumably from the inability of these side chains to form Hbonds. IR signatures consistent with parallel and ordered Q-side chains are indicated with dashed lines. Assemblies were negatively stained with uranyl acetate.

S15

Figure S11. TEM micrographs of side-chain substituted-glutamine analog A(1622)E22QNHCH3, highlighting fibril bundles. Assemblies were negatively stained with uranyl acetate

S16

Figure S12. FT-IR spectra of [1-13C]F19 A(16-22)E22QNHCH3 as a function of time. Inset highlights changes in 13C amide-I band. Similar to A(16-22)E22Q (Fig S3, Fig 2F-G, main text), peptide assemblies start with a 12C-13C splitting of 40 cm-1 and over time the 13C band has both a red shift and a decrease in intensity. Unlike A(16-22)E22Q assemblies, the broadness of the 13C band (inset) at the final time point (28 days, red curve) is consistent with parallel (p), antiparallel in-register (i) and anti-parallel out-of-register (o) peptide orientations being present, as well as a band at 1677 cm-1 consistent with order glutamine side chains.

S17

Figure S13. (A-C) TEM and (D, E) FT-IR of [1-13C]F19 A(16-22)E22Q seeded with 1% parallel -sheet mature fibers. Addition of fiber seeds led to a dramatic decrease in the lag phase for the transition to parallel -sheet assemblies. (D) FT-IR spectra at 0hr (black line) show antiparallel out-of-register -sheet signatures, corresponding to ribbons in the TEM image (A). Dashed lines represent the positions of 12C and 13C bands for parallel -sheet assemblies. The transition to parallel -sheets is observed after three hours (D, 3hr. red line) and TEM (B) shows a mixture of non-twisted fibers and twisted ribbons. After eight hours, the transition to parallel sheet assemblies appears complete and (C) only non-twisted fibers are observed in the TEM micrographs. (E) Comparison of the 12C-13C splitting in the FT-IR spectra between seeded (red) and non-seeded (black) samples. For seeded assemblies, IR spectra were collected immediately, 0.5 hr,, 1.6 hr, 2.8 hr, 3.1 hr, 3.6 hr, 4.1 hr, 8.3 hr and 20hr after addition of mature E22Q assemblies. For non-seeded samples, data were collected at 1hr and multiples of 24 hours as indicated on the graph time-axis. Assemblies visualized by TEM were negatively stained with methylamine tungstate and scale bars are 200nm.

S18

Table S1. Isotope-edited IR bands for [1-13C]F19 Aβ(16-22) congeners Peptide sequence

Anti-parallel signature [cm-1]

Q-side chain [cm-1]

C [cm-1]

C [cm-1]

 [cm-1]

β-strand arrangement determined by NMR

Aβ(16-22) tube

~1693

N/A

1637

1597

40

anti-parallel out-of-register[9]

Aβ(16-22) fiber

~1693

N/A

1635

1607

28

anti-parallel in-register[9]

Aβ(16-22)E22Q

N/A

1677

1633.9

1601.4

32.5

parallel

Aβ(16-22)E22QN(CH3)2

~1695

N/A

1638.7

1597.9

40.8

12

13

Peptide identity, the presence of IR stretch above ~1690 cm-1 (consistent with anti-parallel sheet), the presence of glutamine side chain amide stretching frequency, position of 12C IR band, position of 13C IR band, the 12C-13C splitting ([cm-1]) and assignment of -strand registry based on backbone distance measurements by solid-state NMR ref [9] and this work. Band positions were determined from 2nd derivatives of IR spectra. Assembly conditions: Aβ(16-22)E22Q is assembled at acidic pH in 20% CH3CN/H2O; Aβ(16-22) tubes[9] assembled at acidic pH in 40% CH3CN/H2O and Aβ(16-22) fibers[9] assembled at neutral pH in 40% CH3CN/H2O.

S19

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

A. B. Rodriguez-Navarro, J Appl Crystallogr 2006, 39, 905-909. D. M. Gregory, M. A. Mehta, J. C. Shiels, G. P. Drobny, J. Chem. Phys. 1997, 107, 28-42. P. V. Bower, N. Oyler, M. A. Mehta, J. R. Long, P. S. Stayton, G. P. Drobny, J Am Chem Soc 1999, 121, 8373-8375. M. A. Mehta, M. T. Eddy, S. A. McNeill, F. D. Mills, J. R. Long, J Am Chem Soc 2008, 130, 2202-2212. B. M. Fung, A. K. Khitrin, K. Ermolaev, J Magn Reson 2000, 142, 97-101. D. M. Gregory, T. L. Benzinger, T. S. Burkoth, H. Miller-Auer, D. G. Lynn, S. C. Meredith, R. E. Botto, Solid State Nucl Magn Reson 1998, 13, 149-166. M. Bak, J. T. Rasmussen, N. C. Nielsen, J Magn Reson 2000, 147, 296-330. A. Barth, Prog Biophys Mol Biol 2000, 74, 141-173. A. K. Mehta, K. Lu, W. S. Childers, Y. Liang, S. N. Dublin, J. Dong, J. P. Snyder, S. V. Pingali, P. Thiyagarajan, D. G. Lynn, J Am Chem Soc 2008, 130, 9829-9835. W. T. Astbury, S. Dickinson, K. Bailey, Biochem J 1935, 29, 2351-2360. K. D. Parker, K. M. Rudall, Nature 1957, 179, 905-906. E. D. Eanes, G. G. Glenner, J Histochem Cytochem 1968, 16, 673-677. A. J. P. Geddes, K. D.; Atkins, E. D. T.; Beighton, E., J. Mol. Biol. 1968, 32, 343-358. H. Inouye, P. E. Fraser, D. A. Kirschner, Biophys J 1993, 64, 502-519. M. Sunde, C. C. Blake, Q Rev Biophys 1998, 31, 1-39. P. Sikorski, E. Atkins, Biomacromolecules 2005, 6, 425-432.

S20