Electrospray ionization mass spectrometry (ESI-MS) has been widely used to characterize specific noncovalent biological interactions in solution, including multi-protein assemblies, protein-/peptide-ligand, and DNA-/RNA-drug complexes.1-9 In practice, ESI-MS has been used to determine binding stoichiometry,5,10-12 and to measure relative13,14 and absolute15-17 association constants (binding affinities).

While ESI-MS is clearly a powerful bioanalytical tool, this technique does have certain limitations. One of its most notable limitations is related to the tendency of biological molecules to associate nonspecifically with other biomolecules, small molecules, or ions present in solution during the ESI process.5 The formation of nonspecific noncovalent adducts with neutral or ionic species derived from buffer or impurities has been previously discussed.18,19 The formation of nonspecific noncovalent complexes may lead to misinterpretation of ESI-MS results and obscure the binding stoichiometry of specific complexes. A variety of methods have been suggested to cope with the formation of nonspecific noncovalent complexes. In particular, dialysis,20 reversed-phase high-performance liquid chromatography (HPLC),21 and size exclusion chromatography22 have been used to resolve nonspecific metal binding issues. However, these methods are limited because desalting might interfere with the protein metalation process. As a result, protein metal binding properties may not be properly reflected.

A few different metal sources that contain different counter ions have been evaluated for use in minimizing nonspecific protein-metal ion adduct formation in ESI-MS experiments. For example, Pan et al. used chloride (Cl^？), acetate (CH₃COO^？), and tartrate (^？OOC-CH(OH)-CH(OH)- COO^？) to examine which counter ion is most suitable for studying protein-metal ion interactions.23 In particular, they studied Ca²⁺ and Zn²⁺ binding proteins. In their studies, tartrate, a weak chelator with a relatively high metal binding constant, was found to dramatically reduce nonspecific metalation. Tartrate was shown to be a particularly effective counter ion that produces metalation levels very close to those in bulk solution.

In the present study, we extend the previous study by Pan et al. to zinc finger peptides that have high Zn²⁺ binding affinity. Zinc fingers belong to an important protein family that has attracted extensive attention due to its intrinsic biological significance.24-26 Zinc fingers can recognize a specific DNA double-stranded sequence. Their unique specificity and DNA binding affinity give zinc fingers promising potential as DNA recognition motifs in artificial restriction enzymes. In our previous study, we showed that ESI-MS can be successfully used to examine the binding specificity of zinc finger peptides to DNA with a cognate sequence.26 In this study, we demonstrate that tartrate is most effective in reducing nonspecific Zn²⁺ adduction for zinc finger peptides, i.e., minimizing the formation of zinc finger？Zn²⁺ complexes with multiple zinc ions.

Zinc finger peptides were all custom-synthesized from Peptron Inc. (Daejeon, Korea) using Fmoc solid phase chemistry. Melittin was available from Sigma (Seoul, Korea). The sequences and monoisotopic masses of zinc finger peptides used in the present study are shown in Table 1.27-29 Zinc chloride (ZnCl₂), zinc acetate (Zn(CH₃ COO)₂), and ammonium tartrate were purchased from Sigma (Seoul, Korea). Zinc tartrate (Zn(OOC(CHOH)₂ COO)) was prepared by mixing equimolar amounts of ammonium tartrate and zinc acetate. All peptide samples were dissolved at a peptide concentration of 20 μM, and the pH was adjusted to pH 7.5 using aqueous NH₄OAc buffered solution.

All mass spectrometry measurements were made on an ion-trap mass spectrometer (LCQ Deca, Thermo Finnigan,

San Joes, CA) operated in the positive ion mode (needle voltage = +3.5 kV). The capillary was heated to 200 ℃ and applied potential was +10 V. The tube lens offset was maintained at +20 V. A solution of 100 mM aqueous NH4OAc buffer was infused at 3 μl min^？1. Full-scan mass spectra were recorded in the range of m/z 200？2000.

Three different zinc metal source compounds were added to electrospray solutions to evaluate the effects of various zinc-ion sources on the formation of zinc finger？Zn²⁺ complexes. Zinc chloride (ZnCl₂), zinc acetate (Zn(CH₃ COO)₂), and zinc tartrate (Zn(OOC(CHOH)₂COO)) were used as zinc metal sources.23

Figure 1 shows ESI mass spectra of Sp1-3, CF2II-6, and melittin obtained by co-spraying with zinc chloride, zinc acetate, and zinc tartrate, respectively. Figure 1 also shows the control ESI mass spectra obtained without any Zn²⁺ source. In Figure 1, only the m/z region corresponding to +3 ions is shown. Figure 1(a) and (e) show Sp1-3 and CF2II-6, respectively, when no Zn²⁺ source was added. Only protonated molecular ions, such as (M+3H)³⁺, were observed. When Zn²⁺ was added, Zn²⁺-containing molecular ions appeared. For example, when ZnCl₂ was added to the ESI solution for Sp1-3 at 200 μM, the following Zn²⁺- adduct molecular ions were observed: (M+H+Zn²⁺)³⁺ , (M？ H+2Zn²⁺)³⁺, (M？3H+3Zn²⁺)³⁺, (M？5H+4Zn²⁺)³⁺, and (M？ 7H+5Zn²⁺)³⁺ (see Figure 1(b)). Sp1-3 zinc finger peptide ions with a single Zn²⁺, i.e., (M+H+Zn²⁺)³⁺, reflect the native zinc finger metal binding stoichiometry. In addition to these ions, Zn²⁺-adducts with multiple Zn²⁺ ions were observed: (M？H+2Zn²⁺)³⁺, (M？3H+3Zn²⁺)³⁺, (M？5H+ 4Zn²⁺)³⁺, and (M？7H+5Zn²⁺)³⁺. Among many Zn²⁺ ions contained in [M+(3？2n)H+nZn²⁺]³⁺ molecular ions, one structural Zn²⁺ ion is very likely to coordinate in a nativelike metal ion binding site while the other Zn²⁺ ions are likely to be adducted. In other words, zinc finger peptide ions with multiple Zn²⁺ ions are likely to form nonspecific complexes, which should be avoided in this experiment. The addition of Zn(CH₃COO)₂ yielded very similar results. As shown in Figure 1(c), Zn²⁺-adducts with multiple Zn²⁺ ions were observed. In contrast, when Zn(tartrate) was added at 200 μM, Sp1-3 zinc finger peptide ions with a single Zn²⁺, i.e., (M+H+Zn²⁺)³⁺ , were dominant with a low abundance of (M？H+2Zn²⁺)³⁺ ions containing two Zn²⁺ ions (see Figure 1(d)). This result clearly indicates that when Zn(tartrate) was used, the formation of nonspecific complexes with multiple Zn²⁺ ions was significantly reduced, consistent with the previous study by Pan et al.23

For CF2II-6, the above-described Zn²⁺-adduct formation tendency was also observed. ESI solutions prepared with either ZnCl₂ or Zn(CH₃COO)₂ produced significant amounts of nonspecific zinc finger？Zn²⁺ complexes (see Figure 1(f) and (g)). Specifically, the ESI solution with ZnCl₂ showed ESI mass distribution of (M+3H)³⁺, (M+H+Zn²⁺)³⁺, (M？H+2Zn²⁺)³⁺, (M？3H+Zn²⁺)³⁺, and (M？5H+4Zn²⁺)³⁺. The ESI solution with Zn(CH₃COO)₂ yielded a distribution with (M+3H)³⁺ , (M+H+Zn²⁺)³⁺, and (M？H+2Zn²⁺)³⁺. As described above, zinc finger？Zn²⁺ complexes with multiple zinc ions are simple Zn²⁺ adducts that do not reflect the stoichiometry of native zinc finger peptides. ESI-MS results obtained with Zn(CH₃COO)₂ were very similar to those obtained with ZnCl₂. However, for Zn(CH₃COO)₂, the production of (M？3H+3Zn²⁺)³⁺ and (M？5H+4Zn²⁺)³⁺ was completely quenched. When Zn(tartrate) was used as a metal source, the generation of nonspecific zinc finger？Zn²⁺ complexes was not observed at all. As shown in Figure 1(h), only (M+3H)³⁺ and (M+H+Zn²⁺)³⁺ were observed, indicating that the formation of nonspecific complexes with multiple Zn²⁺ ions was completely inhibited. It is also noteworthy to note that when Zn²⁺ was added, Sp1-3 did not show any (M+3H)³⁺, while CF2II-6 showed abundant (M+3H)³⁺ peaks. This different complex formation behavior may be due to differences in the Zn²⁺ binding constants of the two peptides. In our unpublished data, the Zn²⁺ binding constant of Sp1-3 was higher than that of CF2II-6 by approximately two orders of magnitude.

To confirm whether Zn(tartrate) generally reduces the formation of nonspecific zinc finger？Zn²⁺ complexes, the same approach was repeated for other peptides, such as Sp1- 1 and CF2II-4. The formation of nonspecific complexes was significantly reduced (spectra not shown) for the other peptides. We also performed another control experiment with melittin, which has a mass similar to those of other zinc finger peptides but does not have any Zn²⁺ binding properties. Figure 1 (i)-(l) shows the resulting ESI mass spectra. As expected, the ESI solution with Zn(tartrate) did not produce any zinc finger？Zn²⁺ complex, while the addition of ZnCl₂ or Zn(CH₃COO)₂ caused the formation of zinc finger？Zn²⁺ complexes with multiple Zn²⁺ ion complexes. These control experiments clearly suggest that Zn(tartrate) is an appropriate zinc ion source that minimizes the formation of nonspecific zinc finger？Zn²⁺ complexes.

A detailed mechanistic study on why tartrate reduces the formation of nonspecific complexes between metal binding peptides and metal ions is still required.23 However, the relatively high Zn²⁺ binding affinity of tartrate was previously suggested to play a certain role in reducing nonspecific complexes. Acetate ion, which has a relatively higher Zn²⁺ binding ability than chloride ion, was shown to produce less nonspecific complexes than chloride ions did (see Figure 1).