Dual Control of Peptide Conformation with Light and Metal-Coordination

Abstract

We describe the design of a stimuli-responsive peptide whose conformation is controlled by wavelength-specific light and metal coordination. The peptide adopts a defined tertiary structure whose conformation can be modulated between an α-helical coiled coil and β-sheet. The peptide is designed with a hydrophobic interface to induce coiled coil formation and is based on a recently described strategy to obtain switchable helix dimers. Here, we endowed the helix dimer with 8-hydroxyquinoline (HQ) groups to achieve metal coordination and a shift to a β-sheet structure. We find that the conformational shift only occurs upon introduction of Zn²⁺. The other metal ions (Cu²⁺, Fe³⁺, Co²⁺, Mg²⁺ and Ni²⁺) do not offer switching likely due to non-specific metal-peptide coordination. A control peptide lacking the metal-coordinating residues does not show conformational switching with Zn²⁺ supporting the role of this metal in stabilizing the β-sheet conformation in a defined manner.

Graphical Abstract

A coiled coil peptide programmed with 8-hydroxyquinoline (HQ) and imidazole groups exhibits a conformational shift to a β-sheet structure upon coordination of Zn²⁺, while the introduction of other metal ions (Cu²⁺, Fe³⁺, Co²⁺, Mg²⁺ and Ni²⁺) does not lead to a conformational change.

Conformational changes in peptide and protein structures in response to stimuli are implicated in a broad range of normal and pathogenic conditions.^[1] The biological importance of protein conformational changes has driven research efforts to understand and control these structural changes. Modulation of protein secondary and tertiary structure in response to external stimuli has been successfully demonstrated with pH changes,^[2] metal coordination,^[3] redox chemistry,^[4] and photochemistry.^[5] In the context of the latter, it was previously shown that photoisomerization of azobenzene units incorporated within peptides can lead to dramatic changes in the peptides conformation. These conformational changes include, among others, the enforcement or disruption of the helical structure of a coiled coil peptide resulting from a cis-to-trans photoisomerization of an incorporated azobenzene (Figure 3),^[5d] An α-helix-to-β-sheet photoswitch, however, has not been reported. Inspired by the design of a redox-triggered peptide that switches reversibly between an α helical monomer and a Cu(II)-bound β-sheet aggregate upon reduction and oxidation of copper,^[4] we wished to generate a coiled coil peptide that contains both azobenzene and metal-binding ligands and test its ability to change conformation upon a combination of both light irradiation and metal binding.

Figure 3. Circular dichroism spectroscopy reveals conformational changes in the designed metallopeptide.

(A) CD spectra of peptide T1 alone and in the presence of divalent metal salts in sodium phosphate buffer pH 7.25. The trans state (left) becomes more helical with metal added whereas the cis state (right) is negligibly affected. (B) CD spectra of peptide T2 alone and in the presence of divalent metal salts in sodium phosphate buffer pH 7.25, The trans state (left) shows a minor increase in helicity with Cu but a conversion from helix to an antiparallel sheet upon addition of Zn. The cis state, in contrast, only experiences helicity changes with both metals.

We hypothesized that appropriately designed metal coordination sites may allow us to trap discrete conformations such as a β-sheet and significantly alter the helical conformation in synthetic coiled coils. Based on tropomyosin, a cytoskeletal coiled coil protein,^[6] we have designed photoswitchable coiled coils containing an azobenzene dye linking two peptide strands in which the cis and trans azobenzene states can be reversibly accessed by irradiation with an appropriate wavelength of light.^[7] To create metal coordination sites that induce a coiled-coil to β-sheet switch, we chose one of these peptides, T1, as a control peptide and replaced two glutamic acid residues with 8-hydroxyquinoline (HQ) amides and two arginine residues with histidine to generate T2 (Figure 1A). The Glu-Arg pairs in T1 were positioned to form interhelical salt-bridging and we anticipated that their replacement with metal-coordinating groups in T2 would have similar bridging properties but in a metal-dependent manner.^[8] Both HQ-HQ and HQ-imidazole pairs have previously been employed to generate metal-responsive peptides and other oligomers.^[9]

Figure 1. Design of photoswitchable metallopeptide assemblies.

(A) Helical wheel diagrams of peptide constructs T1 and T2. Salt bridging residues Glu and Arg at e and g positions of the heptad are swapped with Glu(HQ) and His, respectively. (B) Proposed sheet structure formed by metal coordination bridging two Glu(HQ) residues. Evidence of this conformation is only observed in the presence of Zn(II). (C) Sequences of peptides T1 and T2. Coiled coil heptad positions are depicted.

We envisioned two different potential outcomes by combining photoswitching with metal coordination sites. The histidine-HQ pairs are placed at the e and g positions of a putative coiled coil heptad (Figure 1A, right); metal complexation of these residues would stabilize the helical tertiary configuration. Alternatively, two HQ may participate in metal chelation to induce a β-sheet conformation (Figure 1B). The sequences of T2 and its control T1 are compared in Figure 1C.

We began characterization of the T1 and T2 metal assemblies by analyzing the UV-Vis spectra of the peptides upon addition of 2 eq. Zn(ClO₄)₂ or Cu(ClO₄)₂ (Figure 2).

Figure 2. Impact of metal coordination on UV-Vis spectra of peptide sequences.

Curves of T1 (left) and T2 (right) peptides alone (14.78 μM, black), with 2 equiv. of copper(II) (blue), or zinc(II) (red) in sodium phosphate buffer pH 7.25 at are shown.

Metal-free peptides T1 and T2 in the trans-azobenzene mode exhibited two absorption bands each at micromolar concentrations in sodium phosphate buffer. The 244 and 367 nm absorptions can be assigned to the π-π* transition of azobenzene,^[10] and their positions remain unchanged after addition of metal ions. The sharp 248 nm band seen in the T2 spectrum can be assigned to the π-π* transition of HQ and reports on the metal coordination status of these groups.^[9c] Upon addition of Zn(ClO₄)₂ and Cu(ClO₄)₂ in individual experiments, the 248 nm peak shifted to 262 nm for Cu²⁺ and Zn²⁺, consistent with similar titration experiments described in the literature and implying complexation of metal by T2 through the His and HQ ligands (Figure 2).^{[9c, 11]} Similar metal-dependent shifts were not observed for T1. Mass spectrometry analysis supported metal complexation by T2 (Figure S8 and S9, Supporting Information).

We next probed if metal coordination enables a robust configuration switch in T2 in a manner dependent on whether azobenzene is in the cis or trans state. The trans-azobenzene state predominates in the dark and conversion to the cis state is achieved by irradiating the sample with 350 nm light. Successful photoisomerization was confirmed by comparing UV-Vis spectra before and after photoswitching and noting the disappearance of the absorbance band at 370 nm (Figure S3, Supporting Information). The conformations of peptides T1 and T2 were then analyzed by circular dichroism (CD) spectroscopy. CD spectra of peptide strands linked to azobenzene in both the trans and cis states were obtained at 400 μM peptide concentration (Figure S2, Supporting Information) – both alone and in the presence of various divalent metals in 50 mM sodium phosphate buffer pH 7.25 (Figure 3 and S4–6).

Addition of two equivalents Cu(II) and Zn(II) to T1 resulted in modest increases in helicity as determined by mean residue ellipticity (MRE) at 222 nm, which we attribute to the carboxylate coordination from Asp and Glu residues present at i, i+3 and i, i+4 positions (vide infra). Metal coordination of residues that lie on the same face of an α-helix is known to promote conformational rigidity,^[12] and the potential of Cu (II) to coordinate carboxylate residues and stabilize protein secondary structures has been studied.^[4] Predictably, the CD spectrum of T1 in the cis-azobenzene isomer was not perturbed by metal addition because this state is unstructured and does not provide alignment of the appropriate residues for coordination.

In the trans-azobenzene state, both the control peptide T1 and the metal-binding peptide T2 (Figure 3B) were moderately helical in the absence of metal ions. In our metal binding studies, we employed two equivalents of various metal ions for complexation because T2 presents two His-HQ metal coordination sites. The number of coordinated metal ions was validated by LC-MS analysis (vide infra). Addition of two equivalents Cu(II) slightly increased the helicity of T2 as expected, but addition of two equivalents Zn(II) resulted in a vastly different CD profile indicative of β-sheet formation.^[13] The CD minimum at 224 nm is slightly redshifted compared to minima of typical β-sheets at 218 nm; however, the CD spectra of β-sheets are known to be highly sensitive to aggregation due to aromatic interactions and peptide backbone twists.^{[13, 14]}

We further investigated the Zn complex with infrared spectroscopy (Figure S7, Supporting Information), which revealed stretching at 1621 and 1690 cm⁻¹ corresponding to the amide–carbonyl stretch expected of antiparallel β-sheets.^{[13, 14, 15]} LC-MS analysis of T1 and T2 with either Cu(II) or Zn(II) supports the formation of a bimetallic peptide complex LM₂ (peptide:M) (Figure S8–S11, Supporting Information). The solution obtained from the combination of T2 with Cu(II) was further analyzed by EPR (the solution was measured frozen). Simulation of the resultant spectrum provided the Hamiltonian parameters g_x = 2.41, g_y = 2.09 and g_z = 2.08 (Figure S12, Supporting Information), where g_x > g_y > g_z, suggesting the formation of a five coordinated square pyramidal complex.^[16] Without an unpaired electron we couldn’t determine the coordination geometry of the Zn(II) complex by EPR, but data from similar reported complexes suggest tetrahedral Zn center.^[17] We tried complexation with Cu(I) to determine if its tetrahedral geometry would induce conformational change as opposed to the square pyramidal geometry of Cu (II), but the EPR data suggests rapid oxidation to Cu(II) in buffer precluding our measurement. In the cis-azobenzene state, T2 was moderately helical and remained unperturbed upon addition of Cu(II). Unlike the trans-azobenzene state, Zn(II) addition did not result in significant deviation from its helical structure. We also tested the impact of Co(II), Fe (III), Mg(II) and Ni(II) on conformational switching and found that only Zn(II) causes an α to β switch in T2. Figure 3C schematically summarizes the CD results.

The conformational changes resulting from photoswitching of the peptide alone and in the presence of Cu(II) are reversible and remain sensitive to the azobenzene geometry; however, once the Zn(II)-induced antiparallel β-sheet assembly is formed, it becomes locked into that conformation (Figure S5, Supporting Information). We suspect that β-sheet formation led to higher order aggregates that were too stable to be overcome by azobenzene isomerization.

The observed differences caused by Zn(II) and Cu(II) complexation may be explained by the high affinity of Cu(II) for side chain carboxylates.^[4] As noted above and shown in Figure 4, complexation of the intrastrand i, i+3 and i, i+4 residues would stabilize helical conformations. We were surprised to observe that trans-T2 with Zn(II) formed the β-sheet structure rather than peptides linked to cis-azobenzene. Cis-azobenezene linked peptides have previously been shown to favor β-sheet conformation in non-metallic constructs. The trans-azobenzene conformation places the two individual chains in close proximity to each other to enable interhelical hydrophobic contacts.^[7] The designed β-sheet configuration also requires the two strands to interact with each other. We conjecture that the trans-azobenzene construct reconfigures to the β-sheet conformation upon addition of Zn(II) because the two strands are interacting with each other in the trans-azobenzene helical conformation but the helices are individually stabilized in the cis-azobenzene state. Selective binding to Zn²⁺ is advantageous in chemical biology as Zn²⁺ has a wide range of significant roles in several metalloenzymes where it acts as a structural co-factor by facilitating protein folding into a stable conformation - eventually enabling a specific function. ^[18] Further efforts are necessary to fully decode the conformational switching.

Figure 4. The helical conformation may be stabilized by inter or intrahelical metal coordination.

We envisioned a metal coordinated helix dimer (top left) but metals may also bind side chains in individual strands to stabilize the helical conformations. Potential coordination to existing aspartic and glutamic acid side chains on one face of the helix is depicted.

In conclusion, we have designed a peptide construct capable of significant change in tertiary structure (Figure 5). Using both photoswitching and metal coordination for dual control, the structure of T2 can be modulated to assume both α-helical coiled coil and β-sheet arrangements. The conversion to β-sheet structure is specifically dependent on Zn²⁺ while other divalent metals like Co²⁺, Cu²⁺, Fe³⁺, Mg²⁺, and Ni²⁺ resulted in maintenance of helicity. The likely difference between the coordination geometry of Zn²⁺ and these of the other metal ions may be a driving force for the conformational change. The β-sheet structure also depended on azobenzene being in the trans state whereas the cis state did not accommodate such a transition. We anticipate that this strategy may be useful in the development of stimuli-responsive biomaterials.

Figure 5. Model depicting potential T2 conformational states depending on photoisomerization of azobenzene and metal ion complexation. Black line connecting two components represent trans/cis states.

Trans-T2 maintains helicity upon addition of Cu(II) but reconfigures into a β-sheet upon addition of Zn(II). The exact nature of β-sheet aggregation is unknown, but the IR data supports an antiparallel assembly.