De Novo Design of Ln(III) Coiled Coils for Imaging Applications

A new peptide sequence (MB1) has been designed which, in the presence of a trivalent lanthanide ion, has been programmed to self-assemble to form a three stranded metallo-coiled coil, Ln(III)(MB1)3. The binding site has been incorporated into the hydrophobic core using natural amino acids, restricting water access to the lanthanide. The resulting terbium coiled coil displays luminescent properties consistent with a lack of first coordination sphere water molecules. Despite this the gadolinium coiled coil, the first to be reported, displays promising magnetic resonance contrast capabilities.


Materials and Methods
2. Figure S1 -HPLC and mass spectrum 3. Figure S2 -Ln(MB1) 3 mass spectra 4. Figure S3 -Sedimentation equilibrium experiments in the presence of Gd(III) 5. Table S1 -Models and data for fitting the sedimentation equilibrium experiments 6. Figure S4 -Plot showing percentage of Gd(III) bound as Gd(MB1) 3 7. Figure S5 -CD thermal unfolding 8. Figure S6 -Absorption and excitation spectra of Tb(MB1) 3 9. Figure S7 -CD spectra in the absence and presence of Ca(II) 10. Figure S8 -Tb(III) emission spectra in the absence and presence of Ca(II) 11. Figure

Materials and Methods
Peptide synthesis and purification.
All peptide synthesis reagents were purchased from AGTC bioproducts Ltd. Peptides were synthesized on a CEM Liberty 1 automated peptide synthesizer on Rink amide MBHA resin (0.25 mmol scale), using standard Fmoc-amino acid solid-phase peptide synthesis protocols 1 and purified and characterized as previously reported, 2 see Figure S1. Peptide concentrations were determined based on the tryptophan absorbance at 280 nm (ε 280 nm = 5690 M -1 cm -1 ) in 7 M aqueous solutions of guanidinium hydrochloride.

Circular Dichroism (CD) Spectroscopy.
CD spectra were recorded in 1 mm pathlength quartz cuvettes on a Jasco J-715 spectropolarimeter. The observed ellipticity in millidegrees was converted into molar ellipticity, (Θ), and is reported in units of deg dmol -1 cm 2 res -1 . The % folding was calculated based on the theoretical maximum ellipticity value -39054 deg dmol -1 cm 2 res -1 for our coiled coil, based on reports by Scholtz et al. 3 Aliquots of a 1 mM stock solution of GdCl 3 /TbCl 3 , and a 1 M stock solution of CaCl 2 were titrated into either a 100 or 30 μM solution of MB1 peptide monomer in 5 mM HEPES buffer pH 7.0 and the CD spectra recorded after 20 min equilibration. Sedimentation equilibrium experiments could be best fit for a simple monomer-to-trimer model in the presence of Gd(III) (vide infra) and so the CD titration data was fit to the following model, Where the fraction folded, f, is related to the observed molar ellipticity, θ obs, molar ellipticity of metal free peptide, θ free , and molar ellipticity of metal saturated peptide, θ saturated¸ all at 222 nm, see equation (3).  equilibrium constants were then determined using SEDPHAT. 4 The specific partial volume was taken to be 0.7571 and the buffer density was approximated to 1.00 using SEDNTERP software (Biomolecular Interaction Technologies Centre, New Hampshire, USA).

Results:
In the absence of Gd(III) the mass obtained from a simple fit assuming single species is slightly higher than would be expected (4733 as compared to monomer mass of 4004 amu). This is indicates a small amount of self-association in the system even in the absence of the Gd(III). To assess this further the data was fit to a monomer-to-trimer model.
This provides a fit that is at least as good as that for the single species fit, however, close examination of the K a for these fits shows the interaction to be very weak (10 -1.5 M -1 ), consistent with the requirement for Gd(III) to drive oligomerisation.
In the presence of Gd(III) ( Figure S3) an equivalent fit to a single species model produces a mass which is significantly above that expected for the dimer (9291 amu). These data suggest that the ligand bound complex is likely to contain a significant quantity of trimer. However the observation that the mass is not that of a trimer indicates that the solution contains an equilibrium between a number of species. To assess this, these data were best fit to a monomer-to-trimer equilibrium model with the major species now the trimer.
This model fits well to the data and provides an association constant of 67608297 M -2 . To assess whether other oligomerisation states could explain the behavior of the peptide in the experiment, other models were tested including those describing monomer-to-dimer and monomer-to-tetramer equilibriums. Neither of these models fit these data as well as that for the monomer-to-trimer equilibrium (see Table S1).
We are aware that the simple monomer-to-trimer model does not perfectly reflect the proposed assembly mechanism as it does not include the Gd(III). To address this, a more complex stepwise addition model was used to fit these data: This model also fits to these data, but with a chi 2 value that is slightly worse than that for the simple monomer-to-trimer model (9.4 for the stepwise model compared to 8.3 for the simple monomer-to-trimer). This stepwise model suggests an K a1 association constant for Gd(III) to the first MB1 of 47863 M -3 (see Table S1).
Overall these data confirm that the presence of Gd(III) mediates the formation of the Gd(MB1) 3 trimer. The data best fits to a single step trimerisation, although within the limits of the data the more complex sequential assembly cannot be ruled out.

Mass Spectrometry.
Peptides were diluted in 10 mM ammonium acetate to a concentration of 7.38 pmol/µl monomer (2.46 pmol/µl trimer) and 10 pmol/µl GdCl 3 or TbCl 3 . The pH of these solutions were identified to be between 6.5 and 7.0. 10 µl of the sample was collected and sprayed by the use of an Advion Biosciences TriVersa NanoMate electrospray source (Ithaca, NY, USA) into a Thermo Fisher LTQ Orbitrap Velos mass spectrometer (Bremen, Germany). Mass spectra were acquired in the orbitrap mass analyser at a resolution of 100,000 at m/z 400 and comprised of 30 scans each comprising of 4 co-added microscans.

Luminescence.
Emission spectra were recorded in a 1 cm pathlength quartz cuvette using an Edinburgh Instruments Fluorescence FLS920 system with a 450 W Xenon arc lamp and a Hamamatsu R928 photomultiplier tube. The emission monochromator was fitted with two interchangeable gratings blazed at 500 nm and 1200 nm and the data was collected using  equivalences of MB1, prepared in 10 mM HEPES pH 7.0. The higher concentrations and excess of peptide (6 monomers per Gd(III)) was used to ensure that >99% of Gd(III) was bound at the lowest concentration (50 µM), see Figure S4.
T 1 and T 2 maps of water protons in phantom samples prepared in 5 mm NMR tubes and shown in Figures 4 and S10, were acquired using a RARE (Rapid Acquisition with Relaxation Enhancement) 7 spin-echo imaging sequence. Horizontal images were acquired with a 1 mm slice thickness, using a 25 × 25 mm field-of-view and a 64 × 64 (T 1 map) or 128 × 128 (T 2 map) pixel matrix. A repetition time of T R = 5 s was used. T 1 relaxation maps were produced from a series of 9 spin echo images with varying T 1 inversion recovery delays from 6 -3500 ms and RARE factor of 1. T 2 relaxation maps were produced from 8 echo images with echo times from 103 -1516 ms and a RARE factor of 64.

Molecular Dynamics Simulations.
The Gd(III)-coiled coil model was built using Insight II (Biosym/MSI, San Diego, California), modifying amino acids from the previously published three-stranded coiled-coil structure 3H5F 8 and solvated with a water box at least 15 Å larger than the peptide in all dimensions. As the Gd(III)-coiled coil is neutral, no counterions were required. Minimization and molecular dynamic simulations were carried out using AMBER v8.0 9 with the ff03 10 forcefield, with additional parameters for Gd 3+ . Gd 3+ ions were treated as simple spheres with a charge of +3 and Van der Waals parameters (MOD4 RE format in AMBER) r = 1.7131 (radius) and e = 0.459789 (well depth). Prior to data-gathering dynamics, the peptide was subjected to 100000 cycles of conjugate gradient minimization and 500 ps equilibration dynamics at 300 K, with constant volume and periodic boundary conditions. Data gathering ran for 10.0 ns also at 300K, with constant volume and periodic boundary conditions, taking snapshots every 10 ps.   Table S1. The models within SEDPHAT used to fit the two datasets from the sedimentation equilibrium experiments. The minimum Chi 2 produced from each fit is provided alongside the constants produced from that fit.