Introduction

Metal cations often play a variety of cellular roles [1], including stabilization of protein structural folds (Mg2+, Ca2+, and Zn2+) [2], oxidative catalysis via electron transfer (Mn2+, Fe2+, and Cu2+) [3], or cellular signaling and metabolism (Na+, K+, and Ca2+) [4]. It is estimated that a third of all proteins contain a metal cofactor, [1] with the divalent metal cations of Mg2+, Ca2+, Zn2+, and Mn2+ most often found in binding sites for metalloproteins [5]. Typically, the “hard” metal acids [6] of Mg2+, Ca2+, and Mn2+ prefer to bind to hard bases of oxygen-containing residue side chains or backbone carbonyls [6–8]. The transition metal cations of Co2+, Ni2+, Cu2+, and Zn2+ are classified as relatively softer “borderline” acids [5] and can coordinate with the softer bases of nitrogen- or sulfur-containing residue side chains of histidine or cysteine in addition to harder moieties such as carboxylic acids offered within the side chains of aspartic acid and glutamic acid [6–8].

One might expect that different metals substitute readily for each other in metalloprotein binding sites due to similarities in ionic radii and charge. This, however, is not normally the case. Rather, nature has engineered these systems to populate the correct metal into specific binding sites in metalloproteins [9]. This is often accomplished by elaborate biochemical mechanisms (e.g.. metal sensor proteins [10] and membrane-bound ion transporters [11]) to equilibrate the protein-cofactor demand with the metal supply. In some cases, the protein scaffold has evolved to orient specific side chains to interact around the preferred coordination geometry of the metal cofactor. These metal cations interact selectively with specific residue side chains [12, 13], where the interactions are largely determinant on the physical properties of the metal cation (i.e., charge, coordination number and geometry, and relative hardness) [14–18].

The innate flexibility and conformational dynamics of peptides offers an ideal model system to probe these metal interactions with preferred amino acid residues. Here, we use ion-mobility mass-spectrometry (IMS-MS) to generate a database of cross sections to investigate the general influence of the divalent transition metals (where X2+ = Mn2+, Co2+, Ni2+, Cu2+, or Zn2+) on peptide ion structure. With such a large number of cross sections, generalizations can be made relating the physical properties of X2+ with the size and range of structures that are observed for the peptide ions. Additionally, the peptide ion size can be normalized to account for increases that are dependent on the molecular mass. These normalized cross sections can then be correlated with the amino acid composition for each of the peptide sequences. This system of equations can then be used to derive intrinsic size parameters (ISP) [19–23] for the individual amino acid residues. As such, the ISP values describe the contribution to the cross section for each amino acid residue within specific peptide ion species that is independent of the influence from mass. Using ISP derivations from other charged species, these new derivations can be used to interpret the preferred binding interactions of X2+ with specific amino acid residues. Similar analyses have been performed with alkali cations [24] (954 cross sections coordinated with Li+, Na+, K+, or Cs+) and the alkaline earth cations [25] (1470 cross sections coordinated with Mg2+, Ca2+, or Ba2+) with general trends reported on the basis of increasing cation size and specific metal-residue interactions.

These gas-phase measurements provide insight into the intramolecular influence of the metal cations on the general structure(s) of peptide ions [26–46]. Recently, IMS-MS in tandem with soft ionization techniques [47] has demonstrated the ability to observe a solution dependence on the populations of gas-phase peptide ions [48, 49]. When a peptide is coordinated with a divalent metal cation, it is also possible to use the relative intensities of specific fragment ions produced via collision-induced dissociation to map back to solution-relevant binding interactions with the knowledge of the populations of the annealed ion mobility distributions [50]. As such, the assessment of the cross section distributions of these metalated peptide ions can yield information on the energetically favorable structures that these peptides can occupy from solution through the electrospray process into the gas phase. A comparison of our results with data mined from the Protein Data Bank (PDB) [7] further supports that the metal-peptide interactions observed here with IMS-MS are similar to those found in other physical states.

Experimental

Instrumentation

Nested IMS-MS measurements were recorded using a home-built instrument [51–53] where experimentation and theory have been described in detail previously [54–60]. Briefly, individual samples are nano-electrosprayed directly into an hourglass-shaped funnel [61, 62] using a Nanomate robotic autosampler (Advion Biosciences, Inc, Ithaca, NY, USA). Here, the ions are periodically pulsed into a ~1.8 m drift tube by lowering an electrostatic gate for 150 μs. Diffusing ions are guided through the drift tube (containing ~3.1 Torr of He buffer gas at 300 K) under the influence of a uniform electric field (~10 V·cm–1). Ions are radially focused with two additional funnels located at the middle and end of the drift tube. Following the drift experiment, ions are transferred through a differentially pumped region before extraction into an orthogonal time-of-flight mass spectrometer for mass analysis. The corresponding drift times and flight times in the mass analyzer for ions are recorded in a nested fashion [63].

Sample Preparation of Mixtures of Peptide Ions

The procedure to produce large mixtures of peptides via tryptic digestion of known proteins (and their subunits) has been detailed previously [24]. During this protocol, all cysteine residues are covalently modified with iodoacetamide to prevent disulfide bond formation via carboxyamidomethylation. Individual digest samples were suspended in a 50:50 water:methanol solution at a concentration of ~0.1 mg·mL–1. Transition-metal acetate (Aldrich or Sigma, St. Louis, MO) was added to a final salt concentration of ~0.5 mM to promote specific metal-peptide adduction.

Peptide Ion Assignments

We have described the experimental design and protocol for assigning peaks in detail previously [24, 25]. Briefly, distributions of peptide ions were produced by nanoelectrospray ionization and introduced into the IMS-MS instrument for the collection of nested mobility and m/z measurements. In-house software was used to pick all peaks of expected m/z values predicted by the ExPASy Proteomics Server PeptideMass [64] online tool. Drift time distributions were generated for each assigned peak by integrating all m/z bins around the mono-isotopic mass at each drift time. Drift times and peak maxima were recorded for all prominent features observed in the drift distributions. All distributions were cross checked against the nested IMS-MS datasets (absent metal acetate) to aid in the elimination of potential interfering peaks. When applicable, the charge state and expected isotopic distribution were used to minimize false positive assignments.

Determination of Experimental Collision Cross Sections

Individual drift times (t D ) can be converted to a collision cross section with Equation 1, [54],

$$ \varOmega =\frac{{\left(18\pi \right)}^{1/2}}{16}\frac{ze}{{\left({k}_b T\right)}^{{}^{1/2}}}{\left[\frac{1}{m_I}+\frac{1}{m_B}\right]}^{{}^{1/2}}\frac{t_D E}{L}\frac{760}{P}\frac{T}{273.2}\frac{1}{N} $$
(1)

where ze is the charge of the ion, k b is the Bolzmann’s constant, m I and m B correspond to the mass of the ion and the buffer gas, respectively, E is the electric field, L is the drift tube length, T and P are the temperature and pressure, and N is the neutral number density of the buffer gas at standard temperature and pressure. To account for the nonlinear fields of the two ion funnels (operated at ~12 V·cm–1), we apply a correction factor to E within Equation 1 using the drift time of specific ions of well-known systems (e.g., bradykinin) to traverse through the entire drift tube.

Results and Discussion

Nested IMS-MS Measurements

Figure 1 shows the nested tD (m/z) distributions for the protonated and transition metal-containing tryptic digestion of bovine beta-casein. The protonated digests are dominated by families of singly and doubly charged ions with triply charged ions also observed (to a lesser extent). Upon addition of the divalent metal acetate, the observed charge state families shift to favor doubly and triply charged ions. This shift in the observed charge state distribution is expected upon the introduction of the competing X2+ to H+.

Figure 1
figure 1

Nested t D (m/z) plots of the electrosprayed mixture of peptides for the tryptic digestion of bovine beta-casein (a) in the absence of acid or metal acetate or following the addition of (b) manganese(II) acetate, (c) cobalt(II) acetate, (d) nickel(II) acetate, (e) copper(II) acetate, or (f) zinc(II) acetate. The protein digest concentration was ~0.1 mg·mL–1 in a 50:50 water:methanol solution. Specific [Pep + X]2+ and [Pep + X + H]3+ sequences are labeled in white and green, respectively, to display drift and mass shifts upon association with X2+ (where X = 2H+, Mn2+, Co2+, Ni2+, Cu2+, and Zn2+)

Full size image

It is interesting to note specific cases where peptide ions are only identified with specific cation(s). For example, the peptide sequences INKK and INKKI are not favored for adduction by X2+, whereas the peptide GPFPIIV is observed with either a single X2+ or a single H+. Compressed mass spectra are supplied in Supplemental Figure 1 to showcase these differences in individual peptide ion abundances with differing cations available in solution.

Changes in the Cross Section Distribution Upon Metal-Coordination

Figure 2 displays the collision cross section distributions for doubly and triply charged ions of five peptide ions from the tryptic digestion of bovine beta-casein. It is clear that the metal cation is influencing the peptide structural landscape when comparing the similarly charged distributions of a given peptide sequence. For example, [EAMAPK + X]2+ and [EMPFPK + X]2+ ions are both smaller in cross section compared with [EAMAPK + 2H]2+ and [EMPFPK + 2H]2+ conformers. Conversely, the majority of all conformers observed for [AVPYPQR + X]2+ and [VLPVPQK + X]2+ are larger in cross section than the major feature of their respective doubly protonated species.

Figure 2
figure 2

Collision cross section distributions for [Pep + X]2+ ions (solid black) and [Pep + X + H]3+ ions (dashed gray) from the beta-casein datasets (where X = 2H+, Mn2+, Co2+, Ni2+, Cu2+, and Zn2+). These cross section distributions are obtained by integrating all bins at each drift time across the isotopic m/z range for each peptide ion of interest. The distributions are normalized to the total sum of each distribution and stacked for comparison

Full size image

It is often the case where similar cross sections are measured in the distributions irrespective of the coordinated transition metal. For example, the cross section distributions for [VLPVPQK + X]2+ ions appear to favor three conformers at cross sections of ~205 Å2, ~210 Å2, and ~216 Å2. The most compact conformer at ~205 Å2 is most abundant when coordinated with Mn2+. The most extended conformer at ~216 Å2 is most abundant when coordinated with Cu2+, Ni2+, or Co2+. When VLPVPQK is coordinated with Zn2+, however, the peptide ion primarily favors the intermediate conformer at ~210 Å2. It is important to note that in all distributions for [VLPVPQK + X]2+, conformers are observed at these approximate cross sections with the primary differences observed in the relative abundances of these conformers. Similar observations can be made with the other peptide ion cross section distributions provided in Figure 2 that are characteristic of the distributions that comprise the database.

Broader cross section distributions are typically observed with [Pep + X + H]3+ species. Examples of these are shown with VLPVPQK and AVPYPQR within Figure 2. These distributions generally contain multiple conformers of lower intensity (hereafter referred as minor features) that are often partially resolved from the major peak or observed at very low abundance at larger cross sections. These additional structures presumably arise from the counterbalance of repulsion and solvation of the H+ and X2+ charge carriers at the numerous sites offered by the specific peptide sequence.

Summary of the Transition - Metal-Coordinated Peptide Ion Database

Table 1 presents a subset of the cross sections for all observed conformers of doubly and triply charged peptide species from the tryptic digest of bovine beta-casein. Complete lists of cross sections for all observed doubly and triply charged peptide ions in all digests can be found in Supplemental Tables 1 and 2, respectively. These lists also provide information on the peptide sequence, number of residues, molecular mass, and the origin of the protein that produced the observed peptide sequence. Cross sections for protonated ions are reported here as the combination of the new measurements made here with those reported previously from a database of alkali-containing peptide ions [24]. The average error for the doubly protonated cross sections within these combined databases of measurements is found to be 0.73% ± 0.32%. The accuracy of these protonated cross sections was previously assessed [24].

Table 1 Cross Sections for Doubly- and Triply-Charged Peptides from Tryptic Digest of Bovine Beta-Casein a
Full size table

Here, we report cross sections for 214 doubly charged and 166 triply charged peptide ions coordinated by X2+. A total of 1253 cross sections are reported for [Pep + X]2+ ions, with 294 cross sections containing Mn2+, 299 with Co2+, 247 with Ni2+, 182 with Cu2+, and 231 with Zn2+. A total of 1035 cross sections are reported for [Pep + X + H]2+ ions, with 291 cross sections containing Mn2+, 245 with Co2+, 245 with Ni2+, 72 with Cu2+, and 182 with Zn2+.

Range of Cross-Sections for Species with Similar Masses

Cross sections generally become larger with increasing molecular mass for the systems studied here. Figure 3 displays this correlation for [Pep + Mn]2+ and [Pep + Mn + H]3+ species with comparisons to doubly and triply protonated species, respectively. For doubly charged species, the range of cross sections at a given molecular mass is relatively constant and suggests a high similarity in the X2+-coordinated peptide structure is present over this range of masses. The standard deviation for the [Pep + X]2+ species is reported to be ±3.4%, ±3.6%, ±3.7%, ±3.9%, and ±3.6% when X2+ = Mn2+, Co2+, Ni2+, Cu2+, and Zn2+, respectively.

Figure 3
figure 3

Cross sections as a function of molecular mass for all tryptic peptide ions that are protonated or coordinated with Mn2+. Comparisons of the range of cross sections at a given molecular mass are displayed for (a) doubly charged peptide ions ([Pep + Mn]2+ and [Pep + 2H]2+), and (b) triply charged peptide ions ([Pep + Mn + H]3+ and [Pep+3H]3+). Major features for protonated ions (gray diamonds) and manganese(II)-coordinated ions (pink diamonds) are displayed along with all minor features for the manganese(II)-coordinated species (green circles). The insets display an expanded view of the range of cross sections observed for a given molecular mass for the region displayed by the overlayed gray dashed box

Full size image

A larger range of cross sections is observed for the [Pep + Mn + H]3+ species with an increased number of observed minor features, indicative of added structural dissimilarity. A narrower range of cross sections is observed for the major features below m ~1000 and above m ~1500 with a broader range of structures between these mass regions. Considering the two charges in the [Pep + Mn + H]3+ system, it is expected that the smaller mass region (at m < ~1000) is principally influenced from Coulombic repulsion between the two cations where more extended conformations could result. We speculate that globular structures intrinsic to the peptide sequence alongside a lessened structural effect by metal ion solvation principally influences the range of structures in the larger mass region (at m > ~1500). The interplay between these forces likely determines the observed conformer distribution between m ~1000 and ~1500.

Plots relating the cross section to the peptide molecular mass for the remaining [Pep + X]2+ and [Pep + X + H]3+ species studied here (when X2+ = Co2+, Ni2+, Cu2+, and Zn2+) are supplied in Supplemental Figures 2, 3, 4, and 5, respectively. These plots highlight the similarity in the distribution of cross sections with respect to molecular mass for the X2+-coordinated species reported here.

Periodic Trends in Divalent Metal-Coordinated Cross Sections

Table 2 summarizes the general structural properties for divalent metal-coordinated peptide ions, inclusive of data previously reported for alkaline earth-coordinated peptide ions[25] (where M2+ = Mg2+, Ca2+, or Ba2+). For the alkaline earth [Pep + M]2+ series, the average change in cross section (compared with the respective [Pep + 2H]2+ species) increases with increasing cation size. Interestingly, each of the transition - metal-coordinated [Pep + X]2+ series show changes in average cross section that are independent of ionic radii yet comparable in magnitude to the [Pep + Mg]2+ data. This is unsurprising with the similarity in ionic radii (range of 0.57 Å–0.83 Å for X2+ compared with 0.72 Å for Mg2+).

Table 2 General Properties for Divalent-Metal-Coordinated Peptide Ion Species a
Full size table

The [Pep + M/X + H]3+ species generally show a decrease in the average change in cross section (compared with the respective [Pep + 3H]3+ species). As displayed by the [Pep + M+ H]3+ series, the average cross section of peptide decreases with increasing M2+ ionic radii that is presumably caused by the larger cations increasing peptide solvation by accommodating larger coordination numbers. The [Pep + X + H]3+ series display similar average decreases in cross section (compared with the respective [Pep + 3H]3+ species); however, the magnitude of these changes does not correlate with the ionic radii. Rather, these values seem to correlate with the average number of stable structures observed per peptide. We speculate that the ability of the metal cation to stabilize additional conformers is reflective of the inherent flexibility of each X2+ to orient itself in a variety of coordination numbers and geometries that is dependent on the local environment.

The overall similarity in the relationship of cross sections for a given molecular mass across all divalent - metal-coordinated peptide ions suggests that the peptide ion size is largely defined by the solvation of the divalent metal charge. The high similarities in the general properties between the X2+-coordinated and Mg2+-coordinated peptide species suggests that the ionic radii and noncovalent coordination sphere primarily determines the observed distribution of conformers. As such, it is reasonable to assume that differences noted in the cross section distributions between Mg2+-coordinated and the X2+-coordinated peptide ions primarily arise due to preferences of interaction with specific functional groups offered by the peptide sequence.

Derivation of Intrinsic Size Parameters for Individual Amino Acids

With such a large number of measurements, intrinsic size parameters can be used to assess the influence of specific amino acid residues on the cross section for specific peptide ion species. For each [Pep + X]2+ dataset, the cross sections are normalized with a second-order polynomial regression relating the cross sections to the molecular mass. These normalized values, termed reduced cross sections, can be equated to a matrix of the sum product of the amino acid intrinsic size parameters with the occurrence frequency for each amino acid residue for each identified peptide sequence, as defined by Equation 2: [19, 21].

$$ {\displaystyle {\sum}_{j=1}^n{X}_{i j}{p}_j={y}_i} $$
(2)

In this equation, the variables i and j correspond to each peptide sequence and the individual amino acid residues that comprise the dataset, respectively. There are 20 potential residues within these datasets that correspond to all naturally occurring amino acid residues except for cysteine, which is carboxyamidomethylated (hereafter referred to as Cys* or C*). X ij refers to the frequency of occurrence of each residue j within each peptide sequence i. The variable p j refers to the unknown intrinsic size parameter for each residue j. The sum product of these variables is equated to y i , which represents the reduced cross section for each peptide sequence i. This system of equations is solved for each p j using a linear least-squares regression [65]. The reported errors are derived as the square root of the variance, representative of one standard deviation from the population mean, and generally larger for lower frequency residues. These uncertainties presumably reflect the structural heterogeneity that exists within the datasets that is evidenced by the multiple conformation populations commonly observed within individual peptide ion assignments. Nevertheless, ion mobility predictions using ISPs have been shown to enhance peptide ion identifications [21, 22] and elucidate solvation preferences for alkali- [24] and alkaline earth-coordinated [25] peptide ions. Here, we aim to extend this analysis to explore the interaction preferences of the transition - metal-coordinated species.

Comparisons of ISP Values for Divalent Metal-Coordinated Peptide Ions

Table 3 lists these derived ISPs for protonated and M2+/X2+-coordinated species. The ISPs listed are generally similar among the separate residue classifications for the listed peptide ion species. For example, the nonpolar aliphatic residues are generally larger (values range from 0.91 to 1.17), whereas the polar aliphatic residues are generally smaller (0.77 to 1.00). Aromatic residues are intermediate in their contribution to size (0.88 to 1.07). These results highlight that the relative size of the peptide ion is heavily influenced by the ability of the peptide sequence to offer functional groups that can capably act to solvate the charge. The most noticeable difference between the alkaline - earth- and transition - metal-coordinated species is displayed with the softer side chains of His and Cys*. For both of these resides, the ISPs are generally smaller for the transition - metal-coordinated ions.

Table 3 Intrinsic Size Parameters of Individual Amino Acid Residues for Peptide Ion Species a
Full size table

Influence of Relative Hardness of Divalent Cations on Interactions with Specific Amino Acid Residues

The average selectivity of the transition metal interaction with specific amino acid residues can be evaluated in Figure 4 with a relative difference plot of ISPs upon the substitution of X2+ for Mg2+ (i.e., \( \frac{{\mathrm{ISP}}_{{\left[\mathrm{Pep}+\mathrm{X}\right]}^{2+}}\hbox{-} {\mathrm{ISP}}_{{\left[\mathrm{Pep}+\mathrm{Mg}\right]}^{2+}}}{{\mathrm{ISP}}_{{\left[\mathrm{Pep}+\mathrm{Mg}\right]}^{2+}}} \)). Considering the similarities in size between X2+ and Mg2+, this analysis can extrapolate the X2+-residue interactions that are dissimilar from those interactions associated with the relatively hard acid of Mg2+. It was previously interpreted that the alkaline earth M2+ binding occurs primarily with the peptide backbone with a lessened preference to interact specially with the residue side chains [25]. We assume that similar backbone solvation preferences are present when binding to divalent transition-metal cations. Thus, we consider the measurements of pentapeptide ions or larger in this analysis to reduce the nonspecific effects on the ISPs associated with backbone coordination.

Figure 4
figure 4

Relative difference plots of the ISP values for specific amino acid residues for the [Pep + X]2+ species compared with [Pep + Mg]2+ species (pentapeptides or larger). (a) ISP values used in the relative difference calculations are displayed for [Pep + Mg]2+ and [Pep + X]2+ species. Pink, purple, aqua, blue, and gray diamonds (with dashed lines) correspond to [Pep + X]2+ ISP values where X2+ = Mn2+, Co2+, Ni2+, Cu2+, and Zn2+, respectively. Black crosses (with solid lines) correspond to the ISP values for [Pep + Mg]2+ species. Error bars represent one standard deviation about the mean. (b) The relative difference plot upon substitution of X2+ for Mg2+ is displayed using the same color scheme for the [Pep + X]2+ ISPs detailed above. Uncertainties in the ISP relative difference plots (displayed as vertical error bars) are calculated as the square root of the product sum of the squared errors and their partial derivatives. These relative ISP plots are derived by equating the [Pep + X]2+ Lys and Arg ISP values to those derived from the [Pep + Mg]2+ ISPs [23]

Full size image

The residues of His, Cys*, Asp, and Met all generally show ISP values that are smaller than the Mg2+-coordinated system, suggesting that these residues form tight binding interactions with these metals. Of the transition metals studied here, Cu2+ displays the smallest ISP values for His, Cys*, and Met (by values of –28%, –19%, and –8%, respectively). It is interesting that some amino acid residues appear to interact favorably with different cations of differing relative hardness with the magnitude of the ISP relative differences reflecting the complex stabilities generalized by the Irving-Williams series. For example, larger relative ISP differences are noted for Asn, Gln, His, and Cys* when coordinated by Mn2+ or Zn2+ with successively smaller ISP differences when coordinated by Co2+, Ni2+, and especially Cu2+. The opposite trend is noted for several nonpolar residues, particularly with Val, Ile, Leu, and Gly. We assume this is reflective of the stronger affinity of the harder metal cations to interact preferably with backbone carbonyls, whereas the relatively softer metal cations can interact specially with softer side-chain bases offered by certain residue side chains (when available). Thus, we interpret the observation of decreased binding interactions of Mn2+ and Zn2+ with specific residue side chains as a preference for a harder backbone solvation similarly observed by Mg2+ coordination.

Figure 5 displays a comparison of 60 doubly charged peptide ion cross sections when measurements exist for sequences coordinated by either Cu2+ or Mg2+ [25]. Of these peptide ions, 30 sequences contain His, Cys*, Asp, or Met. Cross sections are significantly smaller for 20 of these 30 peptides (~66.7%) when coordinated by Cu2+ rather than Mg2+. For the 30 peptides that do not contain His, Cys*, Asp, or Met, only 11 of these peptides (~36.7%) are significantly smaller. Interestingly, nine of these 11 smaller peptides (that do not contain His, Cys*, Asp, or Met) have sequences that contain Gln or Thr. Figure 5 also shows the average percent change in cross section for each transition - metal-coordinated [Pep + X]2+ peptide ion compared with the [Pep + Mg]2+ species as a dependence on the inclusion of His, Cys*, Asp, or Met within the peptide sequence. A significant decrease in cross section is noted for those peptides containing His, Cys*, Asp, or Met when coordinated by Co2+, Ni2+, Cu2+, or Zn2+. Mn2+, however, shows less specificity of interaction with these residues. This is consistent with the relative hardness of Mn2+ with an insignificant change in percent cross section compared with Mg2+-coordinated ions.

Figure 5
figure 5

Comparisons in cross section for doubly charged peptide ions (pentapeptide or larger) when coordinated by either Cu2+ or Mg2+. (a) A comparison of specific cross section measurements for [Pep + Cu]2+ (x-axis) or [Pep + Mg]2+ (y-axis) peptide ions. The filled diamonds denote all peptide sequences containing Met, Asp, His, or Cys*, whereas the open diamonds denote that these specific amino acid residues are absent. (b) The average percent change in cross section for the transition-metal-coordinated [Pep + X]2+ species from the [Pep + Mg]2+ measurements is displayed using the same color scheme described above. The horizontal dashed line corresponds to the average percent error for the combined [Pep + 2H]2+ cross sections reported in Supplementary Table 1. Uncertainties for each [Pep + X]2+ are displayed as the standard deviation for the average percent error

Full size image

Comparison of Metalated ISPs with Metal-Binding Sites in Known Protein Systems

The frequency of interaction of biologically relevant transition metals with specific amino acid residues for all known metalloprotein structures has been characterized previously with a bioinformatic analysis of the Protein Data Bank (PDB) [7]. Overall, the data mining results from the PDB are consistent with our interpretation of the ISP values as an indicator for metal-residue binding preference. Our interpretation of the generally large ISP values for nonpolar aliphatic and aromatic residues is consistent with the findings from the PDB that indicate a limited specificity for metal binding. Similarly, our interpretation of the generally small ISP values for His, Cys*, and Asp agree strongly with the crystallography data that these residues interact specially with X2+. Our interpretation of specific metal-residue binding preferences also corroborates with the protein crystallography bioinformatics data where His has the highest frequency of binding, followed by Cys (Cys* in our data), Asp (specifically with Mn2+, Co2+, Ni2+, and Zn2+), and Met (specifically with Cu2+).

Some differences are also noted regarding the specificity of interaction. For example, the PDB data mining shows a larger frequency of interaction between Mn2+ and Co2+ with Glu than is noted with the ISP analysis. With the larger protein systems offered in the PDB, it is also rare to observe a transition metal interaction with a backbone amide. Differences in our ISP interpretation are attributed to the smaller peptide systems studied here, which offer a limited set of functional groups to effectively solvate the metal cation. Presumably, the smaller peptide system characterized here gives insight into the preference of interaction between X2+ and specific amino acid residues; these preferences are ultimately conserved by larger metalloprotein systems in the structuring of metal-binding pockets.

Summary and Conclusions

This work presents a database of 2288 transition - metal-coordinated peptide ion cross sections (containing Mn2+, Co2+, Ni2+, Cu2+, or Zn2+) for [Pep + X]2+ and [Pep + X + H]3+ species. Generalizations on the structural effect of a metal-peptide interaction can be made with such a large number of measurements. Cross sections are larger upon coordination with X2+ (compared with protonated ions) with sizes comparable to Mg2+-coordinated species (of similar ionic radii). Differences in the conformer distribution are largely attributed to the selectivity of the X2+ interactions that is largely determined by the offered peptide functional groups.

The subtle differences in conformer preference upon X2+-coordination were probed using ISP derivations to assess the preference of interaction for the transition metal cations with specific amino acid residues. The relative difference in ISPs upon the substitution of X2+ for H+ displayed a similarity of interaction to alkaline earth [Pep + M]2+ ions with an inferred preference to nonspecifically interact with the backbone. The relative difference in ISPs upon the substitution of X2+ for Mg2+ revealed the selective interaction of X2+ with certain residues (specifically His, Cys*, Asp, and Met). These metal-residue interactions are assessed with a comparison of cross sections for specific Mg2+- and X2+-coordinated sequences. Such metal-peptide binding preferences also offer great potential to complement proteomics analyses by (1) selectively enhancing peptide signals with high metal-binding affinity, and (2) improving peptide identification scoring methods via the application of the species-specific ISP set to the determined sequence.

Finally, it is prudent to stress the importance of the similarities between the interpretations offered from these gas-phase measurements and data mined of X-ray crystallography structures from the PDB [7]. It is well-established that X-ray crystallography provides a highly detailed perspective at atomic resolution of a stabilized protein molecular structure for an individual conformation favored in solution. Ion mobility provides a complementary assessment of the cross section distribution of stabilized structures. The abundance of secondary conformations implies that the metal cation can reorient the molecular structure, either by alterations in the noncovalent coordination sphere and/or interactions with other metal binding sites.