Poly-Gly Region Regulates the Accessibility of Metal Binding Sites in Snake Venom Peptides

It is supposed that the presence of poly-His regions in close proximity to poly-Gly domains in snake venoms is related to their biological activity; poly-His/poly-Gly (pHpG) peptides inhibit the activity of metalloproteinases during venom storage via the chelation metal ions, necessary for their proper functioning. This work shows that only the histidyl residues from the N-terminal VDHDHDH motif (but not from the poly-His tag) were the primary Zn(II) binding sites and that the poly-Gly domain situated in the proximity of a central proline residue may play a regulatory role in venom gland protection. The proline induces a kink of the peptide, resulting in steric hindrance, which may modulate the accessibility of potential metal binding sites in the poly-His domain and may, in turn, be one of the regulators of Zn(II) accessibility in the venom gland and therefore a modulator of metalloproteinase activity during venom storage.

Computational methods of theoretical chemistry have been used as useful tool to predict structure and stability of the ligands and complexes. 6,7,8,9,10 ,11 ,12 Molecular orbital studies on Cu(II) and Zn(II) cations 1:1 complex with pHpG-4 peptide have been done on the DFT level of theory with IEFPCM 13 solvent (water) model introduced upon potential energy surface investigation. The integral equation formalism for polarizable continuum model (IEFPCM) approximation describes a solvent as a homogeneous dielectric medium with electrical permeability (ε) equal to that of a pure solvent, and the cavity size is modeled for a solvent immersed molecule. The starting structure of the peptide for DFT calculations was generated on the basis of the amino acid sequence after 75 ps simulation at 300 K, without cutoffs using BIO+ implementation of CHARMM force field. DFT calculations were performed with Gaussian 09 C.01 14     wide range of pKa values assigned to His residues is also known, especially for peptides rich in histidyl residues. 16 Table S1. The distribution diagram for the obtained Cu(II)−pHpG-4 complex species is shown in Figure S4.    CuH7L is the first complex detected at acidic pH with maximum concentration at pH 4. In this complex, most probably two imidazole residues are involved in the copper(II) binding. The {2Nim} coordination mode is supported by (i) EPR parameters (A = 161.80 and g = 2.31) (Table S1, Figure S7 and Table   S2) 21  containing the poly-His motif in the sequence. 16,17 With the increase of pH and the occurrence of the next complex species, CuH5L, the maximum of absorption in the spectrum was shifted towards shorter wavelengths in the range of d-d transitions (608 nm → 577 nm, Figure S8A) and the new signal with a positive Cotton effect (Δε = 0.33 at 498 nm) in the CD spectrum occurred ( Figure S8B). These changes suggest the binding of the copper(II) ion by an additional nitrogen atom, possibly from an amide bond (negative band at 278 nm). 22 The {3N} coordination mode is much clearer at about pH 5.7, where the next complex species, CuH4L dominates in solution. Changes in the EPR parameters: i) the increase of the value of A parameter: 178.00 → 186.10 and ii) the decrease of the value of the g parameter: 2.27 → 2.25 strongly confirm the coordination of the copper ion by the three nitrogen donors (Table S1, Figure S6 and Table S2).
Additionally, a slight shift of the maximum absorption to shorter wavelengths with the intensity increasing at pH 5.7 in the UV-vis spectrum and also an increasing of the band intensity at 312 nm in the CD spectrum may suggest a change in the coordination mode from {2Nim, 1Nam} to {1Nim, 2 Nam}.
For the next complex species: CuH3L and CuH2L, no significant changes in spectroscopic parameters were observed. The pKa values of these complex species are close to those of the corresponding ligand species, which suggests the non-binding deprotonation of imidazole residues (pKa CuH3L = 5.95 (6.24), pKa for CuH2L = 6.55 (6.88)) and thus no changes in the coordination mode.   and ii) maximum absorbance at around 523 nm in the UV-vis spectra (Table S2, Figure S7, FigureS   8A). No significant changes in the far-UV CD spectra were observed with increasing pH value ( Figure   S5B) indicating no structural changes.
Based on a series of potentiometric titrations, the mode of zinc(II) coordination by the studied peptide was proposed. Initially, Zn(II) is bound to two His residues (in the ZnH6L complex species), and, with the increase of pH, subsequent imidazole residues become coordinated (ZnH5L -3Nim; ZnH4L -4Nim).
It is not excluded, that the {4Nim} coordinated species may be in equilibrium with other complexes, in which two different His residues coordinate the Zn(II) ion. Further complexes come from the deprotonation of unbound imidazole residues and the N-terminal amino group.

The Cu(II) -pHpG-4 complexes
Theoretical studies perfectly complement experimental work and often allow to explain the phenomenon of metal ion -peptide coordination. Studies of some of the Cu(II) complexes, compared with the Cu(II) -pHpG-4 system, were also supplemented by DFT calculations and allowed to accurately determine the interactions between metal-peptide. 16,17 ,23 In case of Cu(II)-pHpG-4 system, seven Cu(II) complexes have been found. CuH7L and CuH6L are 2N type and use H3 and H5 imidazole rings to bind the metal cation. The bond lengths are similar: 1.829 +/-0.03 Å (Table S4). Interestingly, short, 6 residue (19-24) unusual polyglycine helical fragments can be detected in CuH6L and CuH7L complexes (marked in red in Figure S10), similar to ones reported before. 24 The first 3N type complex is CuH5L; its binding pattern is similar to the previously mentioned complexes: two imidazole nitrogens from H3 and H5 bind Cu(II) with bond lengths 1.851 Å and 1.856 Å; the third nitrogen comes from the amide of H3, which is 2.101 Å away from the central Cu(II) ion.
CuH4L, CuH3L and CuH2L are 3N type complexes that share the same 3N binding pattern, in which D2 and H3 amide nitrogens and the H3 imidazole nitrogen are involved. The Cu(II) -amide nitrogen bond lengths are in the range 1.85-1.89 Å, and the Cu(II) -imidazole bond is moderately longer -2.06 Å.
CuHL is a typical albumine-like complex. The Cu(II) binding to amides of D2 and H3 (1.874 Å and 1.873 Å bond lengths, respectively) and to H3 imidazole nitrogen (2.167 Å bond length) is supported by the binding of the V1 amino nitrogen (bond length 2.241 Å).
All metal -ligand distances are shown in the Table S4.

The Zn(II) -(pHpG-4) complexes
Four Zn(II) complexes have been found ( Figure S11).  Table S5.  All Zn(II) complexes locate the metal at the D2-H7 region of the ligand. It is worth to note that in ZnH6L, there is a semi-linear fragment in the metal binding area, and surprisingly, proline is not involved in such build as one could expect (does not form the specific kink). Due to the presence of the semi-linear fragment in ZnH6L, the metal is relatively more exposed to further interactions in comparison to the rest of the investigated complexes.

Cu(II) complexes
Competition plot for Cu(II) complexes, which is a hypothetical simulation (based on the calculated stability constants) of a system in which equimolar amounts of metal ions and ligands are present and pH-dependent apparent affinity (pKD) values (at pH 7.4 and also at pH 5.4 -typical for snake venom glands) 25 may give us additional information about interaction specificity with the pHpG-4 peptide and other peptides containing the free N-terminus ( Figure S12A, Table S6), showed that below pH 7, pHpG-1Ath. sq (EDDHHHHHHHHHGVGGGGGGGGGG-NH2) is most effective in copper(II) binding, while the least effective peptide is pHpG-4 (VDHDHDHHHHHHPGSSVGGGGGGGGGGA-NH2). pHpG-1Ath. sq has a higher affinity towards Cu(II) than the pHpG-4 peptide containing the ATCUN motif, and most likely, this is due to the possibility of creating polymorphic states by pHpG-1 peptide and forming a helical structure in the presence of metal ions. 18 The plot shows that above pH 7, the situation changes and Cu(II) ions bind more effectively to the pHpG-4 peptide, which can be explained by metal coordination through the albumin-like binding, which begins to dominate at around pH 7 (CuHL species). The shorter analog of pHpG-4, without Val in the first position and poly-Gly domain in the sequence, N-DpH (DHDHDHHHHHHPGSSV-NH2), turns out to be the most effective in Cu(II) binding below pH 4.5 and the least effective above pH 7.5.
A comparison of the Cu(II) ion binding efficiency by pHpG-4 and C-and N-protected peptides containing poly-His and/or poly-Gly motifs ( Figure S12B) clearly show that about pH 8, almost 100% of the copper ions are bound to the peptide containing the ATCUN motif. At acidic pH values, the copper ion binding efficiency of the peptides decreases with the decreasing number of available His residues in the peptide sequences, which is in good agreement with our chemical expectations.  Comparison of copper ion complexes with the studied peptide from Echis ocellatus venom and other peptides containing the ATCUN motif found in the literature showed that the pHpG-4 peptide forms the most thermodynamically stable complexes with Cu(II) ions in the whole pH range ( Figure S13). The competition plot shows, i.a. that DAHQ is more efficient in Cu(II) binding than the DAH peptide, but less efficient than pHpG-4, mainly due to the number of histidine residues in the peptide sequences -the more His residues in the sequence, the stronger the binding, even despite the similar coordination pattern, as in the case of DAHQ and pHpG-4 peptides -up to two imidazole residues bind the Cu(II) ion at pH 4.5. However, it is worth to note that the aforementioned peptides start to form typical albuminlike binding at different pH values: in the case of DAHaround pH 4, in Cu(II)-DAHQat pH 5, and in the case of the Cu(II)-pHpG-4 complexalmost 3 units higher (at pH around 6.9) than in the first system. 26,27 Most likely, polymorphic binding sites present in the Cu(II)-pHpG-4 complex (and similar poly-His peptides) efficiently prevent amide binding at lower pH.
To conclude, the presence of i) proline, ii) poly-Gly, iii) ATCUN motif and number of His residues in the pHpG-4 sequence strongly stabilizes the Cu(II) complexes (above pH 7).  Table S7.

Zn(II) complexes
In the competition plot for Zn(II) complexes, the situation looks different when compared to Cu(II) systems. The Zn(II)-pHpG-4 complex forms the least stable complexes in comparison to other peptides ( Figure S14A, Table S8). Thus, it can be suggested that the presence of the Pro residue and the poly-Gly sequence in the peptide can significantly reduce the thermodynamic stability of zinc(II) complexes.
The competition plot in the Figure S14B shows an opposite trend in the affinity of metal ion coordination to the one observed in case of Cu(II) complexes with the same peptides which is probably related to the availability of metal binding sites. In case of pHpG-4, the 4Nim binding mode could be 'blocked' by the flexible poly-Gly sequence. In the N-DpH peptide, the poly-His motif is more accessible for Zn(II) ions ( Figure S14B, Table S8). Despite the availability of oxygen donors from the three aspartic acid residues in case of N-DpH peptide, which are also capable to bind zinc 30 , the more preferred Zn(II) binding sites are the imidazole nitrogen and the nitrogen from N-terminal amine. The potentiometric data of compared peptides are taken from 16,23,18,31