Rules for the design of aza-glycine stabilized triple-helical collagen peptides†

The stability of the triple-helical structure of collagen is modulated by a delicate balance of effects including polypeptide backbone geometry, a buried hydrogen bond network, dispersive interfacial interactions, and subtle stereoelectronic effects. Although the different amino acid propensities for the Xaa and Yaa positions of collagen's repeating (Glycine–Xaa–Yaa) primary structure have been described, our understanding of the impact of incorporating aza-glycine (azGly) residues adjacent to varied Xaa and Yaa position residues has been limited to specific sequences. Here, we detail the impact of variation in the Xaa position adjacent to an azGly residue and compare these results to our study on the impact of the Yaa position. For the first time, we present a set of design rules for azGly-stabilized triple-helical collagen peptides, accounting for all canonical amino acids in the Xaa and Yaa positions adjacent to an azGly residue, and extend these rules using multiple azGly residues. To gain atomic level insight into these new rules we present two high-resolution crystal structures of collagen triple helices, with the first peptoid-containing collagen peptide structure. In conjunction with biophysical and computational data, we highlight the critical importance of preserving the triple helix geometry and protecting the hydrogen bonding network proximal to the azGly residue from solvent. Our results provide a set of design guidelines for azGly-stabilized triple-helical collagen peptides and fundamental insight into collagen structure and stability.


SYNTHESIS, PURIFICATION, SAMPLE PREPARATION, AND CD MEASUREMENTS a. Resin Preparation
All peptides were synthesized using manual SPPS. All peptides in this study were prepared on a 0.0125 mmol (1 equivalent) scale on Rink Amide MBHA resin LL (loading density: 0.36 mmol/g). The resin was swelled in DMF for 30 min. All Fmoc group deprotections were performed using 2 mL of a 1% HOBt (w/v), 2% DBU (v/v) in DMF solution and stirring for 1-2 min, repeated twice. The resin was then thoroughly washed with DMF. Finally, a coupling solution containing Fmoc-Pro-Hyp(tBu)-Gly-OH (described below) was added to the resin and stirred for 45 min at 60 ˚C. The solution was then drained, the resin was washed with DMF, the Fmoc group on the growing peptide chain was removed as previously described, and the resin was washed with DMF to give H-(PO(tBu)G) on resin. For all CMPs except 6c, 14c, and 20c this POG coupling was repeated two times to give H-(POG)3 on resin, and followed by standard amino acids couplings (described below) with Fmoc-Gly-OH and Fmoc-Hyp(tBu)-OH to give H-(OG)(POG)3 on resin. For CMPs 6c, 14c, and 20c only one POG coupling was performed.

b. Standard Amino Acid Couplings
Following Fmoc deprotection and washing, amino acid coupling solutions were added to the free amine on the growing peptide chain. The coupling solutions contained 5 equivalents of the appropriate Fmoc-amino acid (listed above), 5 equivalents of COMU, and 7.5 equivalents of TMP in 1 mL of DMF. Prior to addition to resin, the coupling solutions were allowed to sit for 5-10 min. All couplings were done over 45 min at 60 ˚C with the exception of His, Met, and Cys couplings which were done at RT. The coupling sln was then drained, and the resin was then washed with DMF. The Fmoc group was removed as described above and the resin was washed again with DMF.

c. Fmoc-Pro-Hyp(tBu)-Gly-OH Coupling
The amino acid trimer Fmoc-Pro-Hyp(tBu)-Gly-OH coupling solution was prepared in the same manner as the single amino acids. The coupling solution contained 5 equivalents of Fmoc-Pro-Hyp(tBu)-Gly-OH, 5 equivalents of COMU, and 7.5 equivalents of TMP in 1 mL of DMF. Prior to addition to resin, the coupling solution was allowed to sit for 5-10 min. The coupling solution was stirred with the resin for 45 min at 60 ˚C before being drained. The resin was then washed with DMF, the Fmoc group was removed using 1 mL of a 1% HOBt (w/v), 2% DBU (v/v) in DMF and stirring for 1 min (x3), and the resin was washed again with DMF. Whenever possible and appropriate, POG coupling solutions were used while synthesizing each CMP.

d. Aza-Glycine Couplings and Subsequent Amino Acid Couplings
Ten equivalents of 9-Fluorenylmethyl carbazate and ten equivalents of CDT were combined in 1 mL of DMF and activated at RT for 5-10 min before being added to the growing peptide chain on resin. The solution and resin were stirred for 24 hr before draining the coupling solution and washing the resin with DMF. The Fmoc group was removed using 1 mL of a 1% HOBt (w/v), 2% DBU (v/v) in DMF and stirring for 1 min (x3), and the resin was washed again with DMF. Following the azGly coupling, a Fmoc-Hyp(tBu)-OH coupling was performed followed by a Fmoc-Pro-OH coupling (as described above) to give H-(POazG)(XOG)(POG)3 on resin, where "X" represents any given amino acid.

e. Cleavage from Resin and Precipitation
Following all appropriate couplings, the final Fmoc group was removed using 1 mL of a 1% HOBt (w/v), 2% DBU (v/v) in DMF and stirring for 1 min (x3), and the resin was washed first with DMF and then with DCM. A 2 mL cleavage cocktail containing 95% TFA, 2.5% TIPS, and 2.5% H2O was added to the resin. The mixture was stirred for 2 hr before being collected into cold ether, causing the peptide to precipitate. The solid was collected by centrifugation, resuspended in cold ether, and collected by centrifugation again (3x). The final pellet was then dissolved in 1.5 mL of 50:50 H2O:CH3CN and stored at 4 ˚C prior to purification.

f. Purification
The peptide solutions were purified using semi-preparative reversed-phase HPLC using acetonitrile and 0.1% TFA in H2O. During purification, the column was heated at 60 ˚C to prevent triple helix formation and aid in separation. The absorbance at 215 nm was monitored to determine collection, collected fractions were analyzed using MALDI-TOF MS in positive ion mode. Appropriate fractions were combined and lyophilized to yield the desired peptide as a white solid. Purity was then checked using analytical HPLC, with the column heated at 60 ˚C. Please note that in many of the HPLC traces (below), there are peaks that appear around a 3-4 min retention time. These peaks are solvent artifacts resulting from differences in the sample solvent and the HPLC system solvent upon injection, as some of the samples were dissolved in 1X PBS prior to checking purity via analytical HPLC.

g. Sample Preparation
After obtaining the pure product, each peptide was then dissolved in a small amount of 1X pH 7.4 PBS.
Using UV-Vis spectrophotometry, the concentration of each sample was determined by measuring the absorbance at 214 nm and using an extinction coefficient of 60 mM -1 cm -1 as described by Engel et al. 3 . The stock solution was then diluted using PBS to a final concentration of 0.2 mM and stored at 4 ˚C for 24 hr before carrying out CD measurements.

h. CD Measurements and Tm Determination
For each peptide, approximately 200 µL of the 0.2 mM solution was placed into a 1 mm quartz cuvette. The ellipticity of these solutions was then measured from 260 to 190 nm while holding the temperature at 4 ˚C. Measurements were obtained in triplicate and then converted to mean residue ellipticity and averaged to generate the CD scan curves included for each peptide below. Following this, the solutions were then heated at a rate of 12 ˚C/hr starting at 5 ˚C and ending at 92 ˚C while monitoring the absorbance at 210, 215, 220, and 225 nm. These measurements, obtained in triplicate, were converted to mean residue ellipticity and averaged. The melting temperate for each peptide, Tm, was determined by using the program GraphPad Prism 7 by fitting the data to a two-state model to find the temperature at which 50% of starting ellipticity was lost, as described previously 1 .

CRYSTALLOGRAPHY a. Crystallization
Peptide stock concentrations were determined using UV-Vis. The absorption was measured at 214 nm and used an extinction coefficient of 60 mM -1 cm -1 was used. CMP 24. Peptide 24 was crystallized using sitting-vapor drop diffusion under conditions adapted from Okuyama et. al. 4 The peptide stock was prepared at a concentration of 12 mg/mL. Crystal trials were performed using 1 μL of the peptide solution and 1 μL of the reservoir solution of 0.

b. Data Collection, Refinement, and Analysis
Crystal data integration was performed with XDS 5 and iMOSFLM 6 (ver 7.2.2). Space group validation and data reduction was performed using Aimless 7 (ver 0.7.4) in the CCP4 8 (ver 7.0.078) software suite. Molecular replacement (MR) was performed using Phaser 9 . MR search models consisted of full length triple helical collagen structures and truncated triple helical collagen structures modified to be short (Ala-Ala-Gly)n sequences. Crystallographic restraints files for N-methylglycine, 4bromophenylalanine, and the C-terminal amidated glycine residues were generated using eLBOW 10 . Refinement was performed using Phenix 11 (ver 1.17.1-3660). Manual model building was performed using Coot 12 (ver 0.8.9.2).

c. PDB Analysis
The comprehensive list of all structures examined is as follows: 1A3I, 1A3J, 1BKV, 1CAG, 1CGD, 1EI8, 1G9W, 1ITT, 1K6F, 1Q7D, 1QSU, 1V4F, 1V6Q, 1V7H, 1X1K, 2CUO, 2D3F, 2D3H, 2DRT, 2DRX, 3A0A, 3A0M, 3A1H, 3A08, 3A19, 3ABN, 3ADM, 3AH9, 3B0S, 3B2C, 3DMW, 3P46, 3POD, 3PON, 3T4F, 3U29, 3WN8, 4AXY, 4DMT, 4GYX, 4OY5, 4Z1R, 5K86, 5Y46, 6A0C, 6HG7, 6JEC, and 6JKL. These structures were chosen as they contained only the collagen triple helix in their crystal structure and present a variety of primary sequences. Within each structure, the distance between the nearest water molecule to the Cα of every glycine residue and the backbone N of every Xaa position residue was measured using the UCSF Chimera software package. 13 Additionally, UCSF Chimera was used to measure the phi and psi angles for each structure.  Table S3. Measured phi and psi angles of CMP 24, containing the Xaa position Val residue. The first and last three terminal residues of the CMP have been excluded to account for the one-residue stagger of the collagen triple helix and helical fraying. The Xaa position residue, Val, is highlighted in blue and the adjacent Gly residue is highlighted in green.  Table S4. Measured phi and psi angles of CMP 25, containing the Xaa position NmetG residue. The first and last three terminal residues of the CMP have been excluded to account for the one-residue stagger of the collagen triple helix and helical fraying. The Xaa position residue, NmetG, is highlighted in blue and the adjacent Gly residue is highlighted in green.