Cupryphans, metal-binding, redox-active, redesigned conopeptides

Design of the metal-binding site

Contryphan-Vn was used as a template for the design of a metal-binding, disulphide-constrained peptide. Analysis of the amino acid sequence of Contryphans reveals that amino acids at positions 2, 6, and 8 in the Contryphan-Vn sequence are not conserved in other Contryphans, despite a fairly good conservation of the overall molecular structure.16 Thus, residues Asp2, Lys6, and Trp8 were selected to be substituted by three histidine residues. Further, an additional His residue was added at the C-terminal end of the peptide, outside the disulphide ring, generating the new amino acid sequence shown in Figure 1.

The feasibility of the formation of a copper-binding site in the novel peptide was confirmed by modeling its three-dimensional structure based on Contryphan-Vn structure (PDB code 1NXN16). The side-chain conformation of the four histidine residues was adjusted to form a metal-binding site with tetragonal geometry, taking as a reference the copper-binding site of bovine superoxide dismutase (PDB code 2SOD18). Stereochemical compatibility of the metal-binding site formation was confirmed by energy minimization in explicit solvent using CHARMM.19 The final model is shown in Figure 2.

Cupryphan metal-binding ability probed by optical and fluorescence spectroscopy

Evidence of Cu²⁺ binding to the peptide was first obtained by fluorescence quenching experiments.21 Addition of copper quenches the fluorescence emitted by the unique Trp residue of the peptide (DTrp5), as shown in Figure 3.21 Fluorescence quenching is dependent on metal concentration and reaches a maximum at a [Cu²⁺]/[peptide] ratio higher than 1:1. All of these characteristics are consistent with an energy-transfer mechanism from the tryptophan residue to the copper bound at the binding site. Thus, the engineered peptide was named Cupryphan from the contraction of the words cupreous and Contryphan.

Fluorescence quenching because of copper binding to Cupryphan was used to determine the binding affinity of Cupryphan for Cu²⁺. The maxima of fluorescence intensity curves at 352 nm as a function of copper concentration were fitted with Scatchard transformation (see Fig. 3), obtaining a copper dissociation constant of 1.3 (± 0.2) × 10⁻⁷M. The good fit of the data with a linear function indicate that only one class of copper-binding site(s) is present in Cupryphan. In addition, fluorescence data indicated the presence of only one binding site per molecule. This is evident from the X-axis intercept value (2.06 × 10⁻⁵M) in the Scatchard transformation plot, which represents the total Cu²⁺ bound at infinite Cu²⁺ concentration (i.e., when [Cu²⁺]_bound/[Cu²⁺]_free = 0). In fact, this value correlates well with the total peptide concentration value (i.e., 2.1 × 10⁻⁵M).

Further evidence of copper–peptide binding was obtained by optical spectroscopy analysis. Addition of a stoichiometric amount of CuCl₂ to the peptide (1.0 mM final concentration) induces the appearance of absorption bands with maxima at 312 and 580 nm (see Fig. 4), the first indicative of a Cu²⁺-histidine charge-transfer,22 the second due to electronic transitions of copper d–d orbitals and typical of copper complexes with nitrogen ligands.22, 23 The maximum and the shape of the absorption band centered at 580 nm did not change with pH in the 6.0–9.5 range (data not shown), indicating a high pH stability of holo Cupryphan. The calculated molar extinction coefficient of Cupryphan at 580 nm is 50 M⁻¹ cm⁻¹. No absorption bands appear in the optical spectrum of Contryphan-Vn after addition of copper ions (data not shown), indicating that the natural peptide does not bind copper ions.

To investigate the possible competition of other metal ions in the binding to the peptide, tryptophan fluorescence quenching experiments by copper ions were carried out in the presence of approximately a fivefold excess of Zn²⁺ or Mg²⁺ (ion and peptide concentrations were 100 and 21 μM, respectively). The calculated copper dissociation constant in both cases is 1.4 (±0.1) × 10⁻⁷M, a value essentially identical to the one determined in the absence of Zn²⁺ or Mg²⁺. This result indicates that Zn²⁺ and Mg²⁺ do not appreciably outcompete copper for binding to Cupryphan. In this regard, it must be noted that addition of Zn²⁺ and Mg²⁺ does not lead to quenching of tryptophan fluorescence, thus a direct determination of the dissociation constant of these two ions for Cupryphan was not possible using fluorescence spectroscopy.

EPR spectroscopy characterization of Cupryphan

The EPR spectrum of Cupryphan, recorded at liquid nitrogen temperature (see Fig. 5), revealed a homogeneous signal, indicating the presence of only one copper species bound to the peptide. Coordination geometry appears to be axial, with values of the spectroscopic parameters (g_|| and A_||) typical of hexacoordinate complexes with an axial symmetry. Moreover, in the g_⟂ region (Fig. 5, panel B), five to seven superhyperfine lines are observed with position and coupling constant (≈12 G), typical of copper(II) complexes with at least two magnetically equivalent histidine residues on the coordination plane.

The presence of a single class of binding sites per molecule was further corroborated by saturation experiments in the 20–200 mW power range (not shown). In fact, the signal lineshape did not change upon increasing the power, suggesting that it likely arises from a single copper coordination environment.

NMR spectroscopy characterization of Cupryphan

Copper interaction with the native Contryphan-Vn peptide was first investigated. The assignment of the ¹H spectrum of Contryphan-Vn is reported elsewhere.16 The addition of CuCl₂ to a 3.1 × 10⁻⁴M solution of Contryphan-Vn (peptide/Cu ratio 1:0.27) caused a decrease of the intensity of all ¹H amino acid signals, whereas the corresponding linewidths were not significantly changed (see Fig. 6). Only the α-CH₂ signal of Gly at 3.78 ppm disappeared because of the large line broadening. The increase of the Cu²⁺ concentration resulted in a further decrease of amino acid signal intensities along with a partial signal broadening (see Fig. 6). Moreover, new signals, well evident at the 1:1.08 peptide/Cu²⁺ ratio, appeared together with the α-CH₂ signal of Gly at 3.94 ppm. The new signals probably belong to a Cu²⁺–peptide complex. Substantial signal broadening because of fast paramagnetic relaxation impeded the application of common 2D NMR experiments for spectral assignment and structure determination of this Cu²⁺/peptide complex. NMR data suggest that the interaction between native Contryphan-Vn and Cu²⁺ is characterized by low affinity as only a partial decrease of the native peptide signals was observed even at 1:1 peptide/Cu²⁺ ratio.

The copper interaction with Cupryphan was studied by the addition of different aliquots of CuCl₂ to a 2.7 × 10⁻⁴M solution of Cupryphan. To study this interaction, a complete NMR assignment of Cupryphan was performed. The ¹H and ¹³C NMR spectral assignment of Cupryphan, obtained by means of 2D experiments, is reported in Table I. As in the case of Contryphan-Vn,16 Pro4 residue in Cupryphan is in the cis conformation, whereas Pro7 is found in the trans conformation.

The first CuCl₂ addition (peptide/Cu²⁺ ratio 1:0.25) caused a drastic broadening of Hδ and Hε proton signals of the imidazole rings of all four histidine residues: the signal halfwidth increased by a factor of four or six depending on the signal (see Fig. 7). On the other hand, the signals of Trp and of other amino acid residues remained unchanged or only slightly broadened. Stepwise addition of CuCl₂ led to a further broadening of His signals and to a decrease of the intensity of other ¹H amino acid signals without any appreciable broadening, similarly to what observed for Contryphan-Vn titration (see Fig. 7).

These results suggest that at low Cu²⁺/Cupryphan ratios (<1), a peptide–Cu²⁺ complex is formed with a specific interaction between the metal and all the four His residues of the peptide, as indicated by the broadening of histidine signals.24 At high Cu²⁺/Cupryphan ratios (>1), when the specific binding is already saturated, a nonspecific interaction between Cu²⁺ and other amino acid residues is manifested by the stepwise intensity decrease of all the amino acid signals upon further addition of CuCl₂ to the peptide solution, as in the case of Contryphan-Vn.

The selectivity of Cupryphan for copper was further assessed by NMR titration of the apo form of Cupryphan with Zn²⁺. Stepwise additions of ZnCl₂ to a solution of Cupryphan (0.94 mM) give rise to a decrease of ¹H NMR signals of Cupryphan and a simultaneous increase of a new set of signals that can be ascribed to a Cupryphan–Zn²⁺ complex (see Fig. 8). Assignment of the ¹H and ¹³C resonances of the Cupryphan–Zn²⁺ complex was obtained by means of 2D NMR experiments (see Electronic Supp. Info.).

The ¹H NMR titration allowed an estimation of the K_d, obtained by a nonlinear fitting of the experimental data relative to Cupryphan/Zn–Cupryphan ratios (see Materials and Methods section) at different total Zn²⁺ concentrations. A K_d value of 9 (±4) × 10⁻⁶M was obtained, almost two orders of magnitude higher than that determined for the binding of copper to Cupryphan (see fluorescence quenching experiments), confirming the selectivity of Cupryphan for copper with respect to other divalent cations.

To obtain some structural information on the Cupryphan–Zn²⁺ complex, a ROESY experiment was performed. However, the NH protons signals of all amino acid residues are rather broad because of the relatively high exchange rate with H₂O/HDO at the experimental pH value and peptide concentration. Consequently, the ROESY correlations between NH and side chain protons of amino acids were not observed, making impossible a valid structural analysis of the peptide conformation based solely on NMR data.

Design and characterization of the Arg–Cupryphan variant

In the attempt to improve the metal-binding affinity of Cupryphan, a second variant was designed in which an Arg residue was inserted between Cys9 and His10 residues of Cupryphan (see Fig. 1). The rationale behind this choice was to confer higher flexibility to the C-terminal portion of the peptide allowing an easier access of His10 to the copper site while lowering at the same time the pK_a values of the His residues by a general electrostatic effect of the Arg guanidinium group. The three-dimensional structure of this second peptide was also modeled and the stability of the copper-binding site confirmed by energy minimization (see Fig. 2).

The novel peptide, named Arg–Cupryphan, was characterized by fluorescence and NMR spectroscopy. In detail, fluorescence quenching experiments confirmed binding of copper with a 1:1 stoichiometry and a K_d = 1.0 (±0.4) × 10⁻⁷M, a value slightly lower than that determined for Cupryphan. A slightly different copper coordination environment of Arg–Cupryphan with respect to Cupryphan was also confirmed by a 20 nm blue shift of the optical absorption peak in the visible region (centered at 560 nm when compared with 580 nm for Cupryphan) and by a twofold increase of the molar extinction coefficient (ε₅₆₀ = 100 M⁻¹ cm⁻¹ when compared with ε₅₈₀ = 50 M⁻¹ cm⁻¹ for Cupryphan).

The ¹H NMR experiments performed on Arg–Cupryphan (6.8 × 10⁻⁴M) upon addition of different aliquots of CuCl₂ showed the same changes observed for Cupryphan, that is, an initial significant broadening of Hδ and Hε signals of all four histidine residues and a progressive decrease of the intensity of the other ¹H amino acid signals at high (>1) Cu²⁺/peptide ratios (data not shown).

Determination of superoxide dismutase activity of Cupryphans

The possibility that the copper ion bound to Cupryphans could be reversibly reduced, a prerequisite for any copper-mediated catalytic activity, was tested by assaying the ability of Cupryphan to dismutate superoxide anions. Superoxide dismutase activity of Cupryphans was determined using the pyrogallol enzymatic assay.25 A solution of superoxide dismutase 0.1 μM in 20 mM Tris HCl buffer pH 8.2 was used as a standard, and a solution of Cupryphan or Arg–Cupryphan 6.8 × 10⁻⁴M in 50 mM sodium acetate buffer pH 6.5 was used to determine Cupryphans superoxide dismutase activity.

The rate of autoxidation of pyrogallol was recorded in the presence of superoxide dismutase, reaching a 50% inhibition of the reaction rate after addition of superoxide dismutase at a concentration of 2.15 × 10⁻⁹M.

The same experiment was performed in the presence of Cupryphan, reaching 50% inhibition at a peptide concentration of 8.5 × 10⁻⁵M. Taking as a reference the enzymatic activity of superoxide dismutase (3.92 × 10⁹M⁻¹ s⁻¹),26 and from the concentration of Cupryphan required to attain 50% inhibition of pyrogallol autoxidation, a superoxide dismutase activity rate for Cupryphan of ∼1 × 10⁵M⁻¹ s⁻¹ was calculated (Table II). Surprisingly, Arg–Cupryphan superoxide dismutase activity was significantly lower than that of Cupryphan (approximately one order of magnitude, Table II). In this regard, it must be noted that an analysis of the copper-binding proteins present in the Protein Data Bank reveals that no copper site is present with the sequence pattern His-X-Arg-His, which suggests that the poor superoxide dismutase activity of Arg–cupryphan could be due to a peculiar geometry of the metal site. An alternative explanation is that Arg10 exerts a “disturbing” effect on the electrostatic field generated by the copper ion, which leads to a lower productivity of the superoxide–copper encounter rate, a position-specific effect well known to occur in superoxide dismutase charge mutants.27, 28 Nonetheless, both peptides were found to be redox-active, highlighting the ability of the engineered copper sites to reversibly cycle between oxidized and reduced redox states.

Peptide synthesis

Contryphan-Vn, Cupryphan, and Arg–Cupryphan were synthesized by standard Fmoc chemistry on an automated Peptide Synthesizer (Pioneer, Applied Biosystems). The protected peptides were grown on a PAL-resin with a high level of modification, using the HATU/DIPEA amino acid activation, according to manufacturer's instructions. Side chain protection scheme was the following: Asp(Ot-Bu), Cys(Trt), Trp(Boc), and Lys(t-Boc). After synthesis, both peptides were released and deprotected by treatment with TFA/H₂O/triisopropylsilane (90:5:5 by volume) for 1 h at room temperature. The cleavage mixture was filtered under vacuum into t-butyl-methyl ether at −20°C. Peptides were collected by centrifugation at 2500 rpm for 20 min at 4°C and washed with t-butyl-methyl ether. The pellet was dissolved in 50% acetonitrile and lyophilized. Peptides were loaded on a Vydac C₁₈ semipreparative column (10 mm × 250 mm, 5 μm particle size, 300 Å pore size) and isolated from by-products with an HPLC apparatus (model K1001, Knauer GmbH, Berlin, Germany), using a linear gradient of 5–80% acetonitrile, containing 0.2% TFA, in 60 min. Elution was monitored by absorbance at 230 nm with an on-line UV detector (model K2501, Knauer GmbH, Berlin, Germany) and peptide fractions were manually collected. The major peptide-containing fractions were air oxidized at 0.01% concentration (w/w) by magnetic stirring in 0.1M NH₄HCO₃ for 24 h at room temperature. The cyclic peptides were finally purified from dimers/multimers by HPLC, as described for purification of synthetic products (see earlier), and characterized by infusion in an electrospray mass spectrometry (ES-IT, mod. LCQ, ThermoElectron, San Jose, CA).

Optical spectroscopy

UV–vis electronic absorption spectra were recorded on an Agilent spectrophotometer HP8452A. Protein concentration was determined by UV absorption at 280 nm, by using a molar extinction coefficient (ε₂₈₀) of 5699 M⁻¹ cm⁻¹.29

Fluorescence spectra of Cupryphan and Arg–Cupryphan were recorded at 25°C with a Perkin-Elmer LS50B spectrofluorimeter equipped with a thermostated cell holder. The excitation wavelength was 280 nm and the excitation and emission slits were set at 10 and 5 nm, respectively. Emission spectra were recorded in the 300–400 nm range. Peptides were dissolved at a concentration of 2.1 × 10⁻⁵M in 50 mM sodium acetate buffer (pH 6.5). For copper-binding studies, CuCl₂ (5.0 × 10⁻⁴M) was added in aliquots and fluorescence was measured after 5 min after addition. The dissociation constant values given in the text are the average of at least five independent experiments.

EPR spectroscopy

Low-temperature X-band EPR spectra were recorded on a Bruker ECS 106 spectrometer equipped with an ER4111VT temperature controller. In detail, spectra were recorded at 9.556 GHz and 115 K. Cupryphan concentration was 1.0 in 50 mM sodium acetate buffer, pH 6.5.

NMR spectroscopy

NMR spectra were recorded at 300K on a Bruker AVANCE AQS 600 spectrometer operating at 600.13 MHz and equipped with a Bruker multinuclear z-gradient probehead. Peptide solutions were prepared in D₂O or in H₂O with 10% of D₂O. pH values of Cupryphan and Arg–Cupryphan solutions were adjusted to 7.5–8.0 by the addition of 0.1M NaOH in D₂O to deprotonate the NH/ND group of imidazole ring of histidine residues. When only the deprotonated form is present, the chemical shifts of Hδ and Hε signals of the imidazole ring in the ¹H NMR spectrum are not changed by further addition of NaOH.

For Zn²⁺ titration experiments, the concentration of Cupryphan was 0.94 mM. Peptide titration was performed by the stepwise addition of aliquots of a 0.01M ZnCl₂ solution up to 1 equivalent of Zn²⁺ relative to Cupryphan. As the chemical shift of CH_α group of Trp (5.11 ppm) in Cupryphan was essentially identical to that of the Cupryphan-Zn²⁺ complex, this signal was used as a reference and the value of its integral was put to 10.0 in all integrations. The ratios of free Cupryphan to Cupryphan–Zn²⁺ complex were determined by the integration of the CH_β signal of Cys3 at 2.623 ppm (Cupryphan) and CH_β signal of Pro4 at 1.075 ppm (Cupryphan–Zn²⁺ complex). The sum of these two integrals was constant within the error of integration, (less than 10%) at all Cupryphan/Zn²⁺ ratios. One site binding model was used for fitting the experimental data, see Eqs. (1)–(3)

(1)

(2)

(3)

where [P], [P-Zn²⁺], and [Zn²⁺] are the concentrations of free Cupryphan, Cupryphan–Zn²⁺ complex, and Zn²⁺, respectively; P₀ and Zn_tot are total concentrations of Cupryphan and ZnCl_2. The Eq. (4) for fitting of the experimental data was derived from the Eqs. (1)–(3).

(4)

where y = Zn_tot, x = [P]/[P-Zn²⁺].

SigmaPlot 8.0 for Windows program was used for nonlinear regression and fitting of the experimental data to obtain K_d value.

In all of the ¹H spectra, a soft presaturation of the HOD residual signal was applied.30 ¹H and ¹³C assignments were obtained using total correlated spectroscopy (¹H-¹H TOCSY)31 and heteronuclear single quantum correlation (¹H-¹³C HSQC) experiments.31 All 2D experiments were carried out using 1024 data points in the f2 dimension and 512 data points in the f1 dimension; the recycle delay was 2 s. The TOCSY experiment was performed with a spin-lock duration of 80 ms. The HSQC experiment was performed using a coupling constant of 150 Hz. The ¹H-¹H TOCSY and the ¹H-¹³C HSQC experiments were processed in the phase-sensitive mode with (2048 × 2048) and (1024 × 1024) data points, respectively. Rotating frame nuclear Overhauser effect spectroscopy (¹H-¹H ROESY)32 with a spin-lock time of 150 ms was used to assign the sequence of amino acid residues.

¹H and ¹³C chemical shifts are reported in ppm with respect to sodium 3-(trimethylsilyl-2,2,3,3-tetradeutero)propionate, which gives a singlet signal at 0.0 ppm and was used as an internal standard.

Molecular modeling

Initial structural models of Cupryphan and Arg–Cupryphan were built using the average solution structure of Contryphan-Vn as a template (PDB code 1NXN16). Inserted His residues conformation was chosen using the standard Deep View rotamer library33 so as to form a tetragonal metal-binding site with geometry and ligand distances comparable to the copper-binding site of bovine Cu,Zn superoxide dismutase (PDB code 2SOD18). The holo forms of Cupryphan and Arg–Cupryphan were then equilibrated in water by energy minimization in explicit solvent using the CHARMM macromolecular mechanics package19 and the CHARMM27 parameters and force field.34 The three-site TIP3p model35 was used for water molecules. In detail, hydrogen atoms were added to the modeled peptides using the HBUILD routine of the CHARMM package. The minimized peptides were placed in a truncated octahedron, constructed from a cubic volume of water molecules of dimension 56.57 Å × 56.67 Å × 56.57 Å, and water molecules overlapping with protein atoms (cutoff = 2.8 Å) were removed. Solvated structures, containing ∼2900 solvent molecules, were further energy minimized to an energy gradient of 0.05 kcal/mol.

Superoxide dismutase activity assay

Superoxide dismutase activity of Cupryphan and Arg–Cupryphan was determined using the pyrogallol enzymatic assay.25 The assay was carried out at 30°C, in 20 mM Tris-HCl buffer pH 8.2, containing 1 mM EDTA and 200 μM pyrogallol, recording the absorption increase at 420 nm for 3 min. Bovine erythrocytes Cu,Zn superoxide dismutase (0.1 μM in 20 mM Tris HCl buffer pH 8.2) was used as a standard.