Structural adaptability and surface activity of peptides derived from tardigrade proteins

Abstract Tardigrades are unique micro‐organisms with a high tolerance to desiccation. The protection of their cells against desiccation involves tardigrade‐specific proteins, which include the so‐called cytoplasmic abundant heat soluble (CAHS) proteins. As a first step towards the design of peptides capable of mimicking the cytoprotective properties of CAHS proteins, we have synthesized several model peptides with sequences selected from conserved CAHS motifs and investigated to what extent they exhibit the desiccation‐induced structural changes of the full‐length proteins. Using circular dichroism spectroscopy, two‐dimensional infrared spectroscopy, and molecular dynamics simulations, we have found that the CAHS model peptides are mostly disordered, but adopt a more α‐helical structure upon addition of 2,2,2‐trifluoroethanol, which mimics desiccation. This structural behavior is similar to that of full‐length CAHS proteins, which also adopt more ordered conformations upon desiccation. We also have investigated the surface activity of the peptides at the air/water interface, which also mimics partial desiccation. Interestingly, sum‐frequency generation spectroscopy shows that all model peptides are surface active and adopt a helical structure at the air/water interface. Our results suggest that amino acids with high helix‐forming propensities might contribute to the propensity of these peptides to adopt a helical structure when fully or partially dehydrated. Thus, the selected sequences retain part of the CAHS structural behavior upon desiccation, and might be used as a basis for the design of new synthetic peptide‐based cryoprotective materials.

F I G .S 2 Infrared spectra of CAHS11aa and control peptided at 0 and 50% TFE.We also report the IR spectrum of deuterated TFE (x 0.2), which also shows absorption bands in the amide I region.When subtracting the TFE IR spectrum to the peptide IR spectra, this overlap can cause problems leading to distortions in the background-corrected peptide spectra.These problems are overcome in 2D-IR since the TFE contribution to the 2D-IR spectrum is insignificant because 2D-IR signal scales as σ 2 , and the σ of the TFE absorption band is much lower than the ones of the amide I modes.

| METADYNAMICS ANALYSIS
To characterize the transitions between α and non-α structures, we calculated the free energy difference ∆G α,1−α between the two states: (1) where S 1 and S 2 run from N − ∆S to N in the integrals over α along the free energy profile A (S 1 , S 2 ) and over all the non-α states in the integrals over 1 − α.We extracted the optimal ∆S value, i.e. the deviations for which we sample conformations close to the initial extended helical conformation, from the ∆G α,1−α dependence on ∆S (Fig. 9).
For simplicity, we chose square regions in the conformation space, with S 1 = S 2 = ∆S .Briefly, the profiles rapidly decreases with increasing ∆S values, i.e. with less stringent boundaries on the consideration of α helical structures.
The decay levels off for CAHS11aa beyond ∆S = 1.5.A less evident plateau is observed for the control peptide, which further supports the findings that the non-α structure is not stable in solution.

A) B) D) C)
F I G .S 9 Free energy difference ∆G α,1−α (∆S ) between the α and non-α-structures as a function of ∆S calculated from eq. 1 for the peptides in hydrating and dessecating conditions.Larger values of ∆S correspond to bigger squares in the upper right corner of Fig. 4.
Table S1: Peak fitting parameters for SSP SFG spectra in Figure 4.A N R and ϕ N R are the amplitude and phase of the non-resonant signal, respectively.A n is the amplitude of the resonant signal, ω n is the resonant frequency, and Γ n is the width of transition.

D) C)
F I G .S 1 0 Average secondary structure analysis per residue in CAHS11aa and control peptide in desiccating conditions.Two sets of molecular dynamics simulations were set up, from helical (A,C) and random (B,D) conformations for both peptides.The error bars represent the standard error of the mean calculated as the standard deviation of the average values over the independent runs.
Comparison between 2D-IR diagonal slices of the bleach signals of the isotropic 2D-IR spectra of the 5 different peptides.

F
I G .S 3 Surface pressure of all peptides at 0.09 mg/ml in buffer solution and neutral pH.Note that the values were obtained before and after the SFG experiment was performed, which explains the missing values in this time range.Shown is the time series of the root mean square deviations of the control peptides from the extended helical conformation predicted by Alphafold2[1].Two sets of simulations are carried out for each system, i.e. starting from helical conformations (A,C) and randomly generated conformations (B,D).First the structural alignment of the individual snapshots saved along the MD simulations is carried out on the C α atoms of the peptides.Then for each MD snapshot the antibody C α root-mean-square deviation (RMSD) is calculated as1 N ab N ab i =1 (r i − r ref i ) 2, where r i and r ref i are the actual and reference coordinates, respectively.N ab is the number of residues in the antibody variable domain.The first 500 ns of the simulations are discarded as equilibration.The values in the plots represent the average RMSD and the error bars represent the standard error of the mean calculated as the standard deviation of the average values over the independent runs.F I G .S 5 Time series of the secondary structure per residue of the control peptide (A) and CAHS11aa (B) in a single run.The vertical dashed line indicates the separation between the equilibration and the production in the run.Average secondary structure analysis per residue in CAHS11aa and control peptide in hydrating conditions.Two sets of molecular dynamics simulations were set up, from helical (A,C) and random (B,D) conformations for both peptides.The error bars represent the standard error of the mean calculated as the standard deviation of the average values over the independent runs.

F
I G .S 7 Secondary structure analysis in 70 % TFE when averaged over the contributions of all residues.Two different sets of simulations are analysed: started from helical structures (A) and started from randomly generated conformations (B).The error bars represent the standard error of the mean calculated as the standard deviation of the average values over the independent runs.Shown is the time series of the root mean square deviations of the peptides from the extended helical conformation modified from the CAHS11aa peptide through mutations.Two sets of simulations are carried out for each system, i.e. starting from helical conformations (A,C) and from randomly generated conformations (B,D).First the structural alignment of the individual snapshots saved along the MD simulations is carried out on the C α atoms of the peptides.Then for each MD snapshot the antibody C α root-mean-square deviation (RMSD) is calculated as1 N ab N ab i =1 (r i − r ref i ) 2, where r i and r ref i are the actual and reference coordinates, respectively.N ab is the number of residues in the antibody variable domain.The first 500 ns of the simulations are discarded as equilibration.