Since its discovery in transcription factor IIIA (TFIIIA) from Xenopuslaevis, zinc finger proteins (ZFPs) have been extensively researched due to their implications for eukaryotic protein-nucleic acid interactions.1-7 ZFPs are one of the most common DNA binding motifs and play important roles in gene regulation.3 In particular, Cys₂His₂ (C₂H₂) represent the most studied zinc finger motifs. ZFPs acquire DNA binding structures from their unstructured form via Zn²⁺ complexation, i.e., protein folding coupled with metal binding. The so-called zinc metallo-finger motif acquires a compact ββα form via the Cys-Zn²⁺ -Cys and His-Zn-His bridges.6 The ββα fold unit constitutes a single zinc finger domain that can recognize 3 to 4 DNA base pairs in a DNA major groove.7 DNA base recognition uses the interactions of amino acid side chains located in the recognition α-helix. Extended DNA recognition is also possible via the tandem repeating of zinc finger domains.

Mass spectrometry (MS) has been used widely to study metal ions interactions with oligonucleotides, peptides, and proteins.8-17 For zinc fingers, a variety of MS experiments have been used to study, for example, the stoichiometry for zinc ion transfer from metallothionein to zinc finger peptides, Zn²⁺ chelation by consensus zinc-finger arrays, the Zn²⁺ binding sites, and relative affinity for other transition metals.18-22

Recently, our group showed electrospray ionization-mass spectrometry (ESI-MS) can be used to study noncovalent complexes formed between a zinc-finger peptide and DNA double helix.23,24 In addition, it was found that the detected noncovalent complexes reflected the binding properties between these two interacting partners.

Because Zn²⁺ binding is crucially important to zinc fingers in inducing protein folding that ensures specific DNA base sequences are recognized, the Zn²⁺ binding (or dissociation) constants of zinc fingers are important information for studying zinc fingers. Zinc ion binding constants of zinc fingers are traditionally determined via UV spectroscopy. Despite UV method being widely used, it involves a back titration, which is time-consuming and requires significant sample amounts for analysis.

In contrast, ESI-MS, which is a fast analysis with low sample consumption, has been used extensively to determine the binding constants for various non-covalent complexes formed in solution.25-33 In this study, we proposed an ESI-MS based titration method for measuring the Zn²⁺ binding constants of various natural and artificial zinc finger proteins.

All of the peptides used in this study were custom synthesized via Fmoc-based solid-phase method (Peptron Inc., Daejeon, Korea) and purified using a Prominence LC-20AB RP-HPLC (Shimadzu, Kyoto, Japan). Table 1 shows the peptides and amino acid sequences being studied. These peptides are natural or designed with a C₂H₂ structure and generally contain a repeated amino acid sequence, i.e., the conserved sequence for Y/F-X-C-X(2 to 4)-C-XXX-Y/F-XXXXX-L-XX-H-X(3 to 5)-H with zinc ion-specific binding properties. The ammonium hydrogen carbonate (NH₄HCO₃) used to buffer the solution and zinc acetate (Zn(CH₃COO)₂) used as a zinc ion source were purchased from Sigma (Seoul, Korea) and used without further purification. The peptides and zinc acetate were prepared as a stock solution in 10 mM NH₃HCO₃ buffer solution (pH 7.5).

ESI mass spectrometry was performed using an ion trap mass spectrometer (LCQ Deca, Thermo, San Jose, CA, USA). The mass spectrometry parameters were set as follows: positive mode, +3.5 kV needle voltage, 110 ℃ capillary temperature (note, a low temperature was used to avoid thermally denaturing the zinc finger peptides), +10 V capillary voltage, +20 V tube lens offset voltage, AGC off, m/z 200~2,000 mass range. The ion injection time was carefully adjusted at approximately 5 ms so that the experiments were done below the space charge limit of the ion trap. The prepared sample solutions were infused directly with a flow rate of 3 µL/ min using a syringe pump built into the mass spectrometer.

We used the following titration model based on basic thermodynamic equations to determine the peptide-Zn²⁺ binding/dissociation constants for zinc-binding peptides.

where P represents a peptide.

For the peptide-Zn²⁺ binding reaction 1, the binding/ dissociation constant (K_b or K_d) can be calculated from Equation 2.

where [P] and [Zn²⁺] are the concentration of a free peptide and Zn²⁺, respectively, and [P:Zn²⁺] represents the Zn²⁺-bound peptide concentration. During the ESI-MS experiments, [P] and [P:Zn²⁺] were determined by measuring the ion intensities for the corresponding species. However, because the peptides and Zn²⁺-bound peptides exhibited different ionization efficiencies, the following equations were suggested.

where α and β are the ionization efficiency (response) parameters for peptide P and P:Zn²⁺ metal-bound complexes, respectively. The following mass balance equation was established for the Zn⁺ concentration.

Where [Zn²⁺]₀ is the initial Zn²⁺ concentration. The above equations yield Equation 6.

Futher,

The response factor, α, can be estimated from independent calibration experiments by measuring the ion intensities of P as a function of the peptide concentration. Furthermore, once α is determined, Equation 7 yields the ionization response factor, β, for the Zn²⁺ -bound peptides, i.e., P:Zn²⁺, for a reference zincfinger peptide with a known Zn²⁺-binding constant by measuring the ion intensities of its free and Zn²⁺-bound peptides (I_P and I_P:Zn²⁺) at different initial Zn²⁺ concentration, [Zn²⁺]₀. Once the α/β ratio for a ZFP with a known zinc ion binding onstant (K_b) is determined, the β value for ZFPs with unknown K_b values can be obtained by assuming the α/β ratio is more or less constant for all ZFPs, which is the only assumption used in this study. Because all ZFPs have similar compact ββα structural motifs, this assumption appears reasonable.

To determine the K_b or K_d values from the mass spectrometric results for zinc-binding peptides, the solvent pH and mass spectrometer cone voltage were optimized to favor folded ZFP conformations. For example, the ESI solution pH must be maintained at ~7.5, which is similar to the native environment.

Figure 1 shows the ESI mass spectra obtained using 20 ìM ZFP peptides alone (left panel) and with an equimolar peptide and Zn²⁺ mixture (right panel). The charge-state envelope generally shifted to lower charge states, when the peptide:Zn²⁺ complexes formed via Zn²⁺ addition. For example, ZiCo peptide (Figure 1a) under zinc free conditions exhibited +2 to +5 charges with the +4 charge state being most abundant. Adding zinc ions in an equimolar ratio caused the noncovalent peptide:Zn²⁺ complexes to dominate with a +3 charge state, while the +5 charge state was negligible. Similar charge-state shifts were observed for other zinc specific binding peptides: Figure 1b, Sp13; Figure 1c, Sp1HM1 [also including Sp11, CF2II4, and CF2II6 (spectra not shown)]. The same experiments were performed using other non-zinc ion binding peptides such as BBA1 and melittin with similar molecular masses, ~3 kDa, to the other ZFPs as a control (see Figure 1d and e). The charge-state shift after Zn²⁺ binding did not occur for these peptides. Furthermore, Zn²⁺ binding did occur to any insignificant degree compared to the ZFP peptides.

The above charge-state shift can be understood as the peptide folding after binding to Zn²⁺. Without Zn²⁺, the ZFPs tend to exhibit unfolded conformations. However, taking Zn²⁺ into a Zn²⁺ binding site formed the ZFP’s specific folded ββα structure, which lowered the overall charge-state distribution.

Here, we used ZiCo, whose Zn²⁺ dissociation constant was reported as (4±1)×10^-6 based on isothermal titration calorimetry (ITC)35, acted as a reference peptide for determining the response factor ratio, i.e., α/β. First, the MS peak intensities were measured at ZiCo concentrations from 10 to 50 µM without adding Zn²⁺ to determine α (see Figure 2a). All of the observed charge states were considered when calculating the MS peak intensities. The response factor, α, for ZiCo was estimated from this experiment as 0.17×10⁸/µM (R² = 0.99) using Equation 3. The α values for the other peptides were similarly determined (R² ≥ 0.99) and are shown in Table 2.

Second, the zinc-free and zinc-binding ZiCo ion peak intensities were measured by adding Zn²⁺ at 10~100 µM to the 20 µM ZiCo solution. The zinc-free and zinc-binding ZiCo ion peak intensities from the ESI full mass spectrum for each initial Zn²⁺ concentration, i.e., [Zn²⁺]₀, were measured to determine I_p and I_p:Zn²⁺, respectively. The response factor, β, for the zinc-binding ZiCo could be determined for each [Zn²⁺]₀ using the literature K_d value for ZiCo and Equation 7. The β value was generally constant over the measured [Zn²⁺]₀ range and averaged β = 0.044(±0.002)×10⁸/µM (see Figure 2b). These experimental results yielded an ionization response factor ratio, α/β, of 3.86.

Based on the assumption that all ZFPs have the same α/β ratio, the response factors, β, for other ZFPs could be determined as shown in Table 2.

Furthermore, to determine the Zn²⁺ binding/dissociation constants, the full ESI mass spectra were obtained for 20 µM equimolar Zn²⁺ :peptide mixed solution of the other ZFP peptides. The β value and Equation 6 could be used to calculate the Zn²⁺ binding/dissociation constants for a variety of ZFPs (see Table 3). In general, ZFPs exhibited dissociation constants from 10^-7 ~10^-9. Specifically, Sp13 exhibited the lowest dissociation constant, 7.2×10^-9, Sp1HM1/Sp1HM3 exhibited a moderate dissociation constant, ~10^-8, and CF2II4/CF2II6 exhibited the highest dissociation constant, ~10^-7. Non-specific peptides, such as BBA1 and melittin, exhibited large dissociation constants, 5.5×10^-6 and 4.3×10^-4, respectively, which is consistent with their Zn²⁺ non-specificity. Furthermore, the dissociation constant for Sp13 determined here is in the same order of magnitude of the literature value, i.e., (2.8±0.9)×10^-9.36

We propose an ESI-MS based method to determine the Zn²⁺ binding/dissociation constants for zinc finger peptides, which have conserved amino acid sequences and C₂H₂ zinc finger motif structures. By determining the ion efficiency parameters for zinc free peptides (α) and Zn²⁺ -peptide complexes (β) for a reference ZFP, i.e., ZiCo, we could readily determine the binding (or dissociation) constants for other ZFPs, for example, Sp11, Sp13, CF2II4, CF2II, Sp1HM1, and Sp1HM3. Our proposed method assumed the response factor ratio, α/β, was the same for all ZFPs with a ββα zinc-finger motif conformation. Although not described above, we performed a comparative study using the so-called De Pauw’s method,37 which agreed well the Zn²⁺ dissociation constants obtained via our proposed method. Whether this approach can be used for zincbinding peptides with different conformations is currently questionable. However, our model is useful for determining Zn²⁺ binding (or dissociation) constants for zinc-binding peptide analogues with similar structural motifs, and should be further applicable to other metal ion binding constant determinations.