Electrostatic Stabilization of a Native Protein Structure in the Gas Phase**

Recently, a general picture has been proposed of how long, and to what extent, native protein structure can be retained in the gas phase.1a In particular, molecular dynamics simulations suggest that salt bridges and ionic hydrogen bonds on the protein surface can transiently stabilize the global fold shortly after desolvation.1b However, the use of native mass spectrometry2 for studying protein solution structure is still controversial, mostly because site-specific experimental gas-phase data3 is scarce. Here we report electron capture dissociation (ECD)4 data on the gas-phase structures of the three-helix bundle protein KIX5 (Figure 1) that indicate substantial preservation of the native solution structure on a timescale of at least 4 s. We demonstrate that in the gas phase, the most stable regions are those stabilized by salt bridges and ionic hydrogen bonds.

shows site-specific yields of c and zC fragment ions [6] from ECD of (M + n H) n+ ions of KIX (see Figure S1 in the Supporting Information) formed by electrospray ionization (ESI). [7] For the 7 + ions, separated c and zC products were observed only from backbone cleavage near the termini (residues 1-13 and 89-91), but not from the threehelix bundle region, which forms a globular fold around a hydrophobic core (residues 16-88). [5] This observation is consistent with intramolecular interactions in the three-helix bundle region preventing separation of c and zC backbonecleavage products [3a-c] in the gaseous 7 + ions. Collisional activation of the 7 + ions (laboratory-frame energy: 28 eV) prior to ECD effected only marginal unfolding near the N terminus (see Figure S2 in the Supporting Information), revealing a notable stability of the three-helix bundle in the absence of solvent.
For the 8 + ions (Figure 2), the appearance of cleavage products from the N-terminal ends of helices a1 (residues 16-30) and a2 (residues 42-61) indicates partial unfolding, with helix a1 separating from the bundle, and helices a1 and a2 starting to unravel from their N-terminal ends. Unraveling of a1 and a2 continues in the 9 + ions, while helices a2 and a3 appear to largely retain their native antiparallel bundle structure. Separation of a2 and unraveling of a3 (residues 65-88), also from its N-terminal end, is evident from the fragmentation pattern observed for the 10 + ions. However, c-and zC-ion yields in the 65-88 region remained relatively small for the 10 + and 11 + ions, suggesting that partially intact a3 helix structure limits fragment ion separation. Further increasing the precursor ion charge gave increased c-and zC-ion yields and unfolding, similar to ECD data for Ubiquitin [3c] (see Figure S3 in the Supporting Information), with the fragmentation pattern of the 16 + KIX ions being largely unselective with respect to backbone cleavage site.
The data in Figure 2 provide substantial evidence for a correlation between the solution-and gas-phase structures of KIX. This supposition is corroborated by ECD of 12 + ions generated by nano-ESI from a solution (in H 2 O at pH 4.5) that better resembles the native protein environment, [8] which gave decreased c-and zC-ion yields in the a2 and a3 regions (see Figure S4 in the Supporting Information), along with a smaller total fragment ion yield (37 %) relative to that resulting from ECD of 12 + ions from ESI of solutions in H 2 O/CH 3 OH (80:20) at pH 4 (total fragment-ion yield: 49 %; see Figure S3 in the Supporting Information).
The temporal stability of nativelike KIX 7 + ions was studied by introducing a delay between ion trapping and structural probing by ECD. However, the ECD fragmentation patterns showed no significant differences for delay times of 1 ms and 2 s (see Figure S5 in the Supporting Information). To Figure 1. Structure of KIX in aqueous solution at pH 5.5 and 27 8C, as determined by NMR spectroscopic experiments (PDB entry: 2AGH, model 1). [5] [*] Dr. K expedite possible structural transitions, we next activated the gaseous 7 + ions by 28 eV collisions (see Figure S2 in the Supporting Information) prior to ion trapping. Despite the increase in ion internal energy, the fragmentation patterns from ECD with delays of 1 ms, 2 s, and 4 s ( Figure 3) are strikingly similar. [9] Apparently, the three-helix bundle structure of KIX is sufficiently stabilized by specific noncovalent interactions that outweigh the loss of hydrophobic bonding in the gas phase. Figure 4 a shows integrated c-and zC-ion yields for helix regions a1, a2, and a3 versus precursor ion charge. The data exhibit sigmoidal behavior, with transition charge values (at 50 % of the plateau value) of 9.2, 10.7, and 12.4 for a1, a2, and a3, respectively. This order of helix stability (a3 > a2 > a1) in the gas phase agrees with that in solution as determined by NMR spectroscopic experiments. [10] However, in solution, each helix unfolds cooperatively, [10] whereas the gasphase data ( Figure 1) show incremental unraveling from their N-terminal ends. This behavior is also reflected in the site-specific transition charge values from analysis of site-specific c-and zC-ion yields (see Figure S6 in the Supporting Information), which generally increase from the N to the C terminus (Figure 4 b). Transition charge values for cleavage sites between helix regions (31-41, 62-64) are similar to values for adjacent helix ends, indicating that helix separation does not precede helix unraveling.
Although the ECD data in Figures 2 and 3 demonstrate extensive preservation of the native solution structure in the 7 + ions, its stabilization in the gas phase must be based on interactions other than hydrophobic bonding. [3d,e] These include neutral [11] and ionic [1b, 12] hydrogen bonds, charge-dipole interactions, [13] and salt bridges. [1b, 14] Figure 5 shows helices a1, a2, and a3 with all basic (H, K, R) and acidic (D, E) residues highlighted in color. The density of charged residues is smallest for a1 (5 out of 15 residues, 0.33) and largest for a3 (14 out of 24 residues, 0.58); a2 exhibits an intermediate density of 0.4 (8 out of 20 residues). Importantly, the charge density values correlate (r = 0.9775) with transition charge values (as a measure of helix stability in the gas phase) for a1, a2, and a3 (Figure 6 a). This observation strongly suggests that interactions involving charged residues, that is, ionic hydrogen bonds and salt bridges, largely determine helix stability in the gas phase.
Close inspection of the native KIX structure revealed that one (D17/H21), three (R42/E45, K52/E55, K53/ D57), and six (R65/D66, E67/ H70, E74/K75, K78/E82, K81/ E84, E85/R88) intrahelix salt bridges can stabilize helices a1, a2, and a3, respectively ( Figure 5). The density of salt bridges correlates (r = 0.9999) with transition charge values (Figure 6 b) even better than the density of charged residues, suggesting that salt bridges are major determinants for protein structural stabilization in the gas phase. However, this conclusion does not exclude additional stabilization by ionic hydrogen bonds as well as charge-dipole interactions. In particular, interaction of the positive net charge at the C-terminal end of helix a3 ( Figure 5) with its electric dipole moment can further stabilize the a3 helix structure, [13] and is consistent with helix unraveling from the N-terminal end.
Stabilization of the global fold by interactions between the three helices probably involves helix dipole/dipole interactions; [15] the antiparallel helices a2 and a3 with larger dipole moments than that of the shorter helix a1 separate and unfold last. Additional stabilization of tertiary structure by ionic hydrogen bonding between charged residues and backbone amides [1b] is indicated by the scatter of site-specific transition charge values (Figure 4 b).
We show here that electrostatic interactions can compensate for the loss of hydrophobic bonding and stabilize the native three-helix bundle structure of KIX in the gas phase on a timescale of at least 4 s. Among these interactions, salt bridges were found to play a dominant role. However, a high number of surface-exposed charged residues alone does not guarantee protein stability in the gas phase: equine Cytochrome c has 24 basic and 12 acidic residues, [3a] with the number of salt bridges on the protein surface increasing from 6 in solution to an average value of 17.3 in the gas phase  within 10 ps after desolvation, [1b] yet its native fold disintegrates on a timescale of milliseconds. [3e, 16] The outstanding stability of gaseous KIX ions observed in this study must be attributed to the combination of favorable electrostatic interactions, including salt bridges, neutral and ionic hydrogen bonds, as well as charge-dipole interactions. Whether or not native mass spectrometry can reveal information about the solution structure of a protein critically depends on the timescale of the experiment [1a] and the extent of intramolecular stabilization by electrostatic interactions. KIX is the first protein for which site-specific ECD data indicate preservation of the solution structure in the gas phase. We propose KIX as a model protein for the evaluation of new and emerging methodology for the structural probing of gaseous proteins.

Experimental Section
KIX protein (91 residues, GSHMGVRKGW HEHVTQDLRS HLVHKLVQAI FPTPDPAALK DRRMENLVAY AKKVEGD-MYE SANSRDEYYH LLAEKIYKIQ KELEEKRRSR L) was expressed in Escherichia coli cells by using a plasmid that included the CBP KIX coding region [5] (residues 586-672; residue 586 corresponds to residue 5 in this study) and purified by Ni-affinity and size-exclusion chromatography. [10] The purified protein was desalted as described previously. [17] Solution pH was adjusted by addition of acetic acid. Experiments were performed on a 7 T Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer (Bruker) equipped with an ESI source (flow rate: 1.5 mL min À1 ) and a hollow dispenser cathode operated at 1.6 A for ECD. The desolvation gas temperature was 200 and 150 8C for 80:20 and 50:50 H 2 O/CH 3 OH solutions, respectively. Before ion trapping, precursor isolation (using radiofrequency waveforms), and irradiation with low-energy (< 1 eV) electrons for 17-50 ms in the FT-ICR cell, ions were accumulated in the hexapole ion cells for 0.3-2.0 s. Ion activation prior to ECD was realized in the second hexapole by energetic collisions with Ar gas. Between 250 and 500 scans were added for each ECD spectrum. ECD fragment ion yields were calculated as percentage values relative to all ECD products excluding aC/y ions, [6] considering that backbone dissociation of a parent ion gives a pair of complementary c and zC ions (100 % = 0.5 [c] + 0.5 [zC] + [other products], in which other products are reduced molecular ions and products from loss of small neutral species from the latter). [3c] Received: August 16, 2010 Published online: November 9, 2010 . Keywords: electron capture dissociation · electrostatic interactions · gas phase · native mass spectrometry · protein structure Figure 6. a) Density of charged residues (D CR , number of charged residues/number of residues) and b) density of salt bridges (D SB , number of salt bridges/number of residues) versus transition charge value for helices a1, a2, and a3 (linear-fit functions with Pearson correlation coefficients of r = 0.9775 (a) and r = 0.9999 (b) shown as dashed lines).