Trpzip 1 has a primary structure composed of 12 amino acid residues: SWTWEGNKWTWK (Fig. 1). According to Cochran et al., its three-dimensional structure is characterized by a β-sheet fold, consisting of two antiparallel strands, joined by a β-turn [12]. This type of conformation is called β-hairpin [5].
The Trpzip 1 structure, obtained from the Protein Data Bank (PDB) [25] (PDB ID: 1LE0), was used to perform MD simulations. Once the frames were selected from the MD trajectory, they were used for the calculations of NMR ¹H chemical shifts. Table 1 shows experimental (δexp) [12] and calculated (δcalc) ¹H chemical shifts, obtained as an average of the values of each individual frame, as well as the Mean Absolute Deviation (MAD) and Root Mean Square Deviation (RMSD) values. The MAD and RMSD parameters were also calculated excluding amide hydrogens since these nuclei are usually very sensitive to experimental parameters, such as temperature and intra/intermolecular hydrogen bonds. Calculated ¹H chemical shift values for each individual frame are available in Electronic Supplementary Material (ESM; Table S1). A 2D structural model of the amino acid sequence, indicating the analyzed hydrogen atoms of each residue, is also available in ESM (Fig. S1).
Table 1
Calculated [GIAO-B3LYP/D95(d,p)//HF/3-21G] (δcalc) and experimental (δexp) ¹H chemical shifts, in ppm, obtained for Trpzip 1.
Residue | Nuclei | δcalc | δexp |
(1) Ser | Hα | 3.80 | 3.40 |
Hβ | 4.22 | 3.69 |
(2) Trp | HN | 7.68 | 8.81 |
Hα | 4.69 | 5.20 |
Hβ1 | 3.44 | 3.02 |
Hβ2 | 2.96 | 3.13 |
(3) Thr | HN | 7.32 | 9.56 |
Hα | 4.98 | 4.85 |
Hβ | 4.44 | 3.99 |
(4) Trp | HN | 10.73 | 8.92 |
Hα | 4.90 | 4.61 |
Hβ1 | 3.50 | 2.07 |
Hβ2 | 3.08 | 2.94 |
(5) Glu | HN | 6.63 | 8.36 |
Hα | 4.43 | 4.34 |
Hβ1 | 1.44 | 1.75 |
Hβ2 | 2.20 | 1.87 |
(6) Gly | HN | 5.11 | 8.21 |
Hα1 | 4.28 | 3.48 |
Hα2 | 5.17 | 3.77 |
(7) Asn | HN | 10.32 | 8.14 |
Hα | 5.37 | 3.93 |
Hβ1 | 2.56 | 2.74 |
Hβ2 | 3.48 | 2.79 |
(8) Lys | HN | 7.48 | 6.53 |
Hα | 4.88 | 4.16 |
Hβ1 | 1.89 | 1.66 |
Hβ2 | 2.46 | 1.72 |
(9) Trp | HN | 8.50 | 8.55 |
Hα | 4.48 | 5.17 |
Hβ1 | 3.44 | 2.95 |
Hβ2 | 2.70 | 3.27 |
(10) Thr | HN | 7.35 | 9.77 |
Hα | 5.12 | 4.86 |
Hβ | 3.76 | 4.00 |
(11) Trp | HN | 10.32 | 9.00 |
Hα | 4.04 | 4.26 |
Hβ1 | 3.19 | 2.01 |
Hβ2 | 3.92 | 2.76 |
(12) Lys | HN | 7.12 | 7.73 |
Hα | 4.71 | 4.16 |
Hβ1 | 1.91 | 1.37 |
Hβ2 | 1.69 | 1.50 |
MAD¹ | 0.81 |
RMSD¹ | 1.08 |
MAD² | 0.55 |
RMSD² | 0.67 |
¹MAD and RMSD values considering amide hydrogens. |
²MAD and RMSD values excluding amide hydrogens. |
Analysis of the differences between the calculated and experimental chemical shifts values, as well as the statistical parameters MAD and RMSD, indicate a satisfactory reproduction of the experimental data for Trpzip 1 [12]. This confirms that the level of theory employed in the calculations is adequate to be applied in structural studies of tryptophan zipper peptides.
Therefore, since the chemical shifts calculations for the Trpzip 1 peptide reproduced the experimental data accurately, this methodology can be used to evaluate the stability of the three-dimensional structure of tryptophan zippers against changes in the sequence. For this purpose, we theoretically proposed a new peptide with a mutation in the primary structure of Trpzip 1, particularly, in the region corresponding to the β-turn sequence. Therefore, we aimed to identify whether the exchange of an amino acid residue would influence the type of β-turn in the structure.
In the article published by Cochran et al., the synthesis of several tryptophan zipper peptides was reported. Three of the synthesized peptide structures differ from each other only by the change of two amino acid residues located in the β-turn region. As a result, the substitutions of the residues directly influenced the type of this secondary structure element. In Trpzip1, the β-turn is composed of EGNK amino acids and, according to the authors, is characterized as type II'. In Trpzip2, there is an exchange of position between residues N and G, resulting in the sequence ENGK, which generates a type I' β-turn. Finally, in Trpzip3, there is a substitution of the amino acid residue G for p, which leads to a sequence EpNK and a type II’ β-turn, the same as in Trpzip1 structure [12]. In this work, we proposed the replacement of one of the G residues by an N residue, in order to identify whether a second N amino acid would influence the type β-turn in the structure. Thus, the sequence of the mutant peptide is: SWTWENNKWTWK.
Table 2 shows the calculated (δcalc) ¹H chemical shift values, obtained as an average of the values of each individual frame. Calculated ¹H chemical shift values for each individual frame are available in ESM (Table S2). A 2D structural model of the amino acid sequence, indicating the analyzed hydrogen atoms of each residue, is also available in ESM (Fig. S2).
Table 2
Calculated [GIAO-B3LYP/D95(d,p)//HF/3-21G] (δcalc) ¹H chemical shifts, in ppm, obtained for the mutant peptide.
Residue | Nuclei | δcalc |
(1) Ser | Hα | 4.01 |
Hβ | 4.29 |
(2) Trp | HN | 8.33 |
Hα | 5.56 |
Hβ1 | 4.00 |
Hβ2 | 2.68 |
(3) Thr | HN | 7.80 |
Hα | 4.62 |
Hβ | 4.72 |
(4) Trp | HN | 10.05 |
Hα | 4.24 |
Hβ1 | 3.19 |
Hβ2 | 3.31 |
(5) Glu | HN | 7.54 |
Hα | 5.15 |
Hβ1 | 2.14 |
Hβ2 | 1.93 |
(6) Asn | HN | 9.71 |
Hα | 3.87 |
Hβ1 | 2.98 |
Hβ2 | 3.83 |
(7) Asn | HN | 6.95 |
Hα | 5.47 |
Hβ1 | 3.19 |
Hβ2 | 3.21 |
(8) Lys | HN | 7.27 |
Hα | 4.96 |
Hβ1 | 1.73 |
Hβ2 | 2.73 |
(9) Trp | HN | 7.72 |
Hα | 4.26 |
Hβ1 | 3.28 |
Hβ2 | 2.73 |
(10) Thr | HN | 7.52 |
Hα | 5.02 |
Hβ | 3.92 |
(11) Trp | HN | 9.20 |
Hα | 3.82 |
Hβ1 | 3.12 |
Hβ2 | 3.66 |
(12) Lys | HN | 7.72 |
Hα | 3.42 |
Hβ1 | 2.17 |
Hβ2 | 1.79 |
The amount of structural similarity between the folds of the mutant peptide and Trpzip 1 can be analyzed from the ¹H chemical shifts. Therefore, the differences in the ¹Hα chemical shifts of the mutant peptide and Trpzip 1 were obtained in relation to the random coil values, determined by Wishart et al. [27] (Fig. 2). For this calculation, experimental data from Trpzip 1 [12] and calculated chemical shift values for the mutant peptide were used.
The data from Fig. 2 show that the mutant and Trpzip 1 peptides have the same pattern of differences in chemical shifts in relation to random coil values, except for residues 5, 7 and 8, which comprise the β-turn region. With this, it can be inferred that, probably, the two peptides present similar folds in the β-strand regions, but they present different types of β-turns.
The β-turns are the most common type of turns and consist of four amino acid residues connecting two antiparallel β-strands [5], such as the one present in the Trpzip 1 structure. Recently, Shapovalov et al. identified the existence of 18 types of β-turns, which differ in the geometry of the peptide bonds in the 2nd and 3rd residues of the turn and in the distances between the Cα atoms of the 1st and 4th residues [28]. According to Cochran et al., in the case of Trpzip 1, the β-turn is constituted by EGNK residues, generating a pD-type turn [12, 28].
Once this probable change has been identified in the type of β-turn, the BetaTurnTool18 software [28] was used to identify the most likely type of β-turn present in the mutant peptide structure. For comparison reasons, the software was also used to indicate the type of β-turn present in the Trpzip 1 structure. For this, the mutant peptide structure, obtained after the MD simulations, and the Trpzip 1 structure obtained in the PDB [25] (PDB ID: 1LE0) were used (Table 3).
Table 3
Types of β-turns assigned to the calculated structure for the mutant peptide and for the experimental structure of Trpzip 1, confidence levels and values of angles Φ and Ψ.
Peptide | Angles Φ | Angles Ψ | Type of β-turn | Confidence level |
Residue 2 | Residue 3 | Residue 2 | Residue 3 |
Mutant | 167.59 | 179.97 | 69.79 | -121.68 | pD | 9 |
Trpzip 1 | 89.12 | -132.33 | -56.90 | -12.69 | dD | 5 |
dN | 4 |
The BetaTurnTool18 software, developed by Shapovalov et al., indicates the most likely types of β-turn to protein structures, based on factors such as the distance between the Cα atoms of the 1st and 4th residues of the β-turn; patterns of hydrogen bond formation between β-turn residues; and torsional angles Φ and Ψ relative to the 2nd and 3rd residues of the β-turn. When a type of β-turn is identified for a given structure, the program calculates the confidence level of that assignment. Confidence levels are given on a scale of 0 to 9, where 0 corresponds to a 0–10% confidence range and 9 corresponds to a 90%-100% confidence range. The authors consider that from level 7 (70%-80%), there is a high probability of a specific type of β-turn being assigned to the structure. However, there are cases where some β-turns resemble more than one type. In these cases, where there is more than one possibility of β-turn in the structure, the program marks the two most likely types, with confidence levels ranging from 0 to 6 [28].
As can be seen in Table 2, BetaTurnTool18 software assigned a pD-type β-turn for the mutant peptide, with a high level of confidence (9). However, for Trpzip 1, whose β-turn had been initially classified as pD-type when it was synthesized by Cochran et al., the software assigned two different types of turn, a dD and dN, also with a high level of confidence. Therefore, this is a clear indication that, possibly, the previous assignment of the β-turn for Trpzip 1 as pD-type might be inaccurate [12].
This result clearly illustrates how the applicability of the quantum calculation of chemical shifts can be useful in the assessment and identification of conformational changes in the three-dimensional structures of peptides and small proteins.
In this way, it becomes possible to obtain important structural information using a relatively fast and low-cost theoretical method, without the need to previously carry out syntheses and experiments, which would consume time and resources.