Insights into the Molecular Architecture of a Peptide Nanotube Using FTIR and Solid-State NMR Spectroscopic Measurements on an Aligned Sample

Queuing up: Molecular orientation within macroscopically aligned nanotubes of the peptide AAAAAAK can be studied by solid-state NMR and IR spectroscopy. Line shape analysis of the NMR spectra indicates that the peptide N-H bonds are tilted 65-70° relative to the nanotube long axis. Re-evaluation of earlier X-ray fiber diffraction data suggests that the peptide molecules are hydrogen-bonded in a helical arrangement along the nanotube axis.


FTIR measurements
FTIR spectra were recorded using a Nexus-FTIR spectrometer equipped with a DTGS detector and a multiple reflection ATR system. Transmission FTIR measurements were performed using a CaF 2 plate, while grazing angle FTIR experiments were done using a ZnSe plate. For both transmission and grazing angle configurations, small aliquots of solutions (17 wt % A 6 K or 17 wt% A*A* 5 K) were deposited on the corresponding plate and allowed to dry, providing a solid film of dry peptide. Transmission FT-IR experiments were also performed on 17 wt % A 6 K or 17 wt% A*A* 5 K solutions in D 2 O, placing the samples between two CaF 2 windows (0.006 mm spacer). Spectra were scanned 64 times and 128 times for transmission and grazing angle geometries respectively.

Solid-state NMR measurements. Rotational resonance SSNMR measurements were
performed on the peptide [ 13 C 2 ]A6K, which incorporates [1][2][3][4][5][6][7][8][9][10][11][12][13] C]Ala at residue 2 (site C A ) and [2][3][4][5][6][7][8][9][10][11][12][13] C]Ala at residue 6 (site C B ). A 17 % w/v solution of the peptide in water was incubated at 25°C and self-assembled into nanotubes after 48 h, as confirmed by transmission electron microscopy ( Figure 1a). The hydrated gel was concentrated by centrifugation and a rotational resonance SSNMR method was used to obtain the inter-strand C A -C B distance restraint by detection of 13 C-13 C dipolar coupling. All experiments were performed using a Bruker Avance 400 spectrometer operating at a magnetic field of 9.3 Tesla. All experiments utilized an initial 4.0-s 1 H 90° excitation pulse length, 1-ms Hartmann-Hahn contact time at a matched 1 H field of 65 kHz, TPPM proton decoupling at a field of 85 kHz during signal acquisition and a 2-s recycle delay. RR experiments [2,3] were carried out by adjusting the sample spinning rate ( R ) to the exact difference between the resonance frequencies () of C A and C B (n = 1 rotational resonance) or to half the frequency difference (n = 2 RR). After cross-polarization, 13 C longitudinal difference polarization was created with a nonselective 4-s /2 pulse followed by a train of 18 DANTE pulses, representing an overall  pulse of 30 s, to invert the 13 C spin polarisation for C selectively. After a mixing period the 13 C magnetization was returned to the transverse plane by a second nonselective /2 pulse before digitization of the free-induction decay. A series of experiments using mixing periods of up to 34 ms was performed to measure the time dependence of difference polarization.
Curves representing exchange of Zeeman order were obtained from the difference in intensities of the Ala2 methyl carbon and Val5 carbonyl carbon peaks.
Simulation of RR exchange curves. Interatomic distances were determined by comparison of the data with numerically simulated curves. Curves were simulated for dipolar coupling constants d CC corresponding to fixed pairs of 13 C-13 C distances derived from molecular models of feasible -strand registrations. The 13 C-13 C interatomic distance r CC is related to d CC according to the equation: The zero quantum relaxation time ( ZQ T 2 ) also affects the shape of the curve. ZQ T 2 is generally not known, but can be estimated from the reciprocal sum of the NMR line widths. A series of curves was calculated by varying distance r between 3.0 Å and 7.0 Å. ZQ T 2 was varied between 1 ms and 5 ms, a conservatively broad range of values, and all other parameters were constant.

Nanotube modelling
Nanotube structural models were generated using a home-written C program (available free of charge by request to the authors). The atomic coordinates of a single A 6 K molecule were expressed in a reference frame with the z axis coincident with the principal axis of inertia (i.e., the long axis of the peptide -strand). The peptide coordinates were replicated radially using a series of translations and rotations to form a single molecule-thick, circular crosssection of the nanotube composed of antiparallel -strands, as summarised in Figure 3 of the main text. The nanotube was then elongated by propagating several such layers along the nanotube axis. The nanotube diameter was set to approximately 400 nm as determined by the electron microscopy, with molecules separated by 4.7 Å in the hydrogen bonding direction and 10 Å in the orthogonal packing direction, as determined by X-ray fibre diffraction. Peptides were rotated about their long axes to vary the mean peptide tilt angle  of the N-H bonds from 0° (parallel to the nanotube axis) to 90° (perpendicular to the nanotube axis). The peptide was also set to be either parallel or perpendicular to the nanotube axis. Each peptide orientation was varied randomly to produce for each molecular 3 position an ensemble of N-H bond angles in a normal distribution with mean  and upper and lower limits of ±n, the distribution angle. The intrinsic disorder within the nanotubes was thus increased by varying n from zero to 25°. Finally, the nanotube coordinates were rotated as a rigid body and expressed in a laboratory reference frame as being either parallel or perpendicular to the magnetic field. Hence complete control could be exercised over the peptide molecular orientation within the nanotube, and the orientation of the nanotube in the magnetic field. In total, chemical shifts were calculated for an ensemble of 2400 molecules in each nanotube model.

H spectra:
A similar procedure was used to simulate 2 H line shapes. In this case an axially symmetric quadrupolar interaction was assumed for a rapidly rotating CD 3 group, with the principal axis z coincident with the C-C bond orientation. Angle  defines the orientation of the quadrupolar tensor in the magnetic field, and the measured frequency for each orientation is given by: The term e 2 qQ/h is the static quadrupolar coupling constant. A quadupolar splitting of 37.5 kHz was measured from an unoriented nanotube sample ( Figure 2 of the main text).

Supplementary Results
-strand alignment. Rotational resonance SSNMR was used to determine whether the strands were hydrogen-bonded in a parallel or antiparallel configuration. For this purpose the peptide [ 13 C 2 ]A6K was prepared, which incorporates [1-13 C]Ala at residue 2 (site C A ) and [2][3][4][5][6][7][8][9][10][11][12][13] C]Ala at residue 6 (site C B ). A 17 % w/v solution of the peptide in water was incubated at 25°C and self-assembled into nanotubes after 48 h, as confirmed by transmission electron microscopy ( Figure S1a). The hydrated gel was concentrated by centrifugation and rotational resonance SSNMR was used to obtain the inter-strand C A -C B distance restraint by detection of 13 C-13 C dipolar coupling. The intermolecular C A -C B distance is strongly dependent on the orientation of the -strands, ~ 5 Å for antiparallel -strands ( Figure S1b). The measured difference polarization at n = 1 rotational resonance is seen to decay with time (Figure 1b), and comparison of the data with numerical simulations reveals that the C A -C B through-space distance is 5.1 Å ± 0.5 Ǻ, thereby confirming that the nanotubes are constructed from antiparallel -strands with adjacent peptides aligned with Ala1 in-register with Lys6. A parallel -strand alignment would result in a C A -C B through-space distance of ~16 Ǻ, which would not give rise to a detectable response in the rotational resonance measurement.