ARTICLE pubs.acs.org/ac
Identifying Specific Small-Molecule Interactions Using Electrospray Ionization Mass Spectrometry Elena N. Kitova, Naoto Soya, and John S. Klassen* Alberta Ingenuity Centre for Carbohydrate Science and Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
bS Supporting Information ABSTRACT: A simple method for establishing whether complexes composed of small molecules detected by electrospray ionization mass spectrometry (ES-MS) originate from specific interactions in solution or nonspecific binding during the ES process is described. The technique, referred to as the nonspecific probe method, exploits the tendency of small molecules to bind nonspecifically to macromolecules during the ES process to establish the presence of specific noncovalent interactions. To implement the method, a macromolecule probe (PNS), which does not bind specifically to any of the components present in solution, is added prior to ES-MS analysis. The existence of specific small-molecule complexes is determined from an analysis of the measured distributions of the small molecules bound nonspecifically to PNS. The principal assumption on which this methodology is based is that nonspecific binding of small molecules and their complexes to PNS during ES is a statistical (random) process. A mathematical framework for establishing the presence of specific heterocomplexes is presented. The reliability of the method for distinguishing specific from nonspecific small-molecule interactions is illustrated for peptideantibiotic and metal ionligand interactions in water.
E
lectrospray ionization mass spectrometry (ES-MS) has emerged as an important analytical technique for identifying and characterizing noncovalent interactions in vitro. The ESMS technique has been applied to a wide variety of complexes, including those composed of macromolecules, such as synthetic and biological polymers, organic molecules, and inorganic molecules and ions.13 In addition to being a rapid, sensitive, and versatile method for detecting specific noncovalent interactions in solution, ES-MS measurements can directly provide information on the binding stoichiometry and affinity (Ka), as well as kinetics of association/dissociation reactions.46 When combined with ion activation techniques (i.e., tandem MS), which can be used to dissociate (disassemble) complexes into their individual constituents, it is also possible to probe the composition and, in some cases, the topology of noncovalent complexes.2,710 Despite the widespread use of ES-MS to characterize noncovalent interactions in solution, there remain challenges to the general and routine application of this technique. Depending on the nature of the interactions investigated and the experimental and instrumental conditions used, ES mass spectra may not accurately reflect solution composition (e.g., abundance of noncovalent complexes). Typically, the major sources of error in the ES-MS measurements are nonuniform response factors,1114 in-source dissociation,1519 and nonspecific binding.2023 The abundance of the gas-phase ions corresponding to a given species (molecule or complex) present in solution, as measured by ES-MS, is related to its solution concentration by a response factor, which r 2011 American Chemical Society
collectively accounts for the ES ionization and detection efficiencies for each species. Although relative ionization efficiencies for different species in a mixture depend on many factors, they tend to be strongly influenced by surface activities,2426 whereas relative detection efficiencies are related to molecular weight of the species and charge states in the gas phase.27,28 In cases where response factors differ, the relative abundance of the gas-phase ions measured by ES-MS will not reflect the relative concentrations of the individual species in solution.1114,29,30 In the extreme case, where the response factors differ significantly, signal suppression of one or more species may occur, resulting in a false negative, i.e., no detectable signal.30 Thermal or collision-induced dissociation of gaseous complexes in the ion source of the mass spectrometer will reduce their relative abundance.1519 In extreme cases, in-source dissociation of a complex leads to a false negative whereby no gaseous ions of the complex are detected.17 The influence of in-source dissociation depends on the configuration of the ion source used and the gas-phase stability of the complex being investigated. Where gentle instrumental conditions do not eliminate in-source dissociation, the use of stabilizing solution or gasphase additives, such as imidazole or SF6, respectively, may aid in preserving complexes during ES-MS analysis.16,19 Another significant challenge to applying ES-MS to study noncovalent interactions is the possibility of nonspecific interactions Received: January 28, 2011 Accepted: May 27, 2011 Published: May 27, 2011 5160
dx.doi.org/10.1021/ac200244u | Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry forming between solution components during the ES process, which results in false positives. The phenomenon of nonspecific binding involving macromolecules (e.g., the nonspecific protein protein, proteincarbohydrate, or proteinmetal ion binding) during ES-MS analysis has been extensively investigated2023,31 and can be understood within the context of the charged residue model (CRM) of ES, which is the mechanism widely believed to be responsible for the formation of gaseous ions of large molecules and their specific complexes.3235 According to CRM, the initial (parent) ES droplets undergo solvent evaporation until they come close to Rayleigh limit, at which point they undergo fission, releasing several small multiply charged nanodroplets (often referred to as offspring or progeny droplets) containing no analyte or one or more molecules of analyte.34,35 Solvent evaporation from the nanodroplets ultimately yields gaseous ions. If a nanodroplet contains two or more analyte molecules, nonspecific intermolecular interactions can occur as the droplet evaporates to dryness, leading to the formation of nonspecific complexes. The probability of the nanodroplets containing more than one analyte molecule increases with increasing analyte concentration.20 Therefore, a general strategy for minimizing the occurrence of nonspecific binding involves limiting the concentrations of the analyte in solution. However, high analyte concentrations, typically >50 μM, are required to detect weak interactions. In such cases, nonspecific binding is often unavoidable in ES-MS analysis. Systematic investigations carried out by Klassen and co-workers into the nonspecific binding of carbohydrates to proteins in ES-MS have provided important insights into the phenomenon of nonspecific binding of small molecules to macromolecules during the ES process.2023,31 These studies revealed that the distribution of carbohydrates bound nonspecifically to protein ions typically resembles that of a Poisson process.20 Furthermore, in a given ES mass spectrum, the distribution of nonspecifically bound carbohydrates is independent of the nature (structure and size) of the protein.21 Similar observations have been made for other classes of molecules, including amino acids, as well as divalent metal ions.22,31 These findings indicate that nonspecific binding is a random process and that the distribution of molecules bound nonspecifically to proteins during the ES process reflects their statistical distribution among the nanodroplets. Several strategies have been developed to distinguish specific from nonspecific binding of small molecules to proteins during ES-MS analysis.16,21,23,3638 Among these is the reference protein method, which can be used to quantitatively correct ES mass spectra for nonspecific proteinligand binding.21 The method involves the addition of a reference protein (Pref), which does not bind specifically to any of the solution components, to the ES solution containing the protein and ligand of interest. The occurrence of nonspecific proteinligand binding is identified by the appearance of peaks in the mass spectrum corresponding to ions of nonspecific complexes composed of Pref and one or more ligand molecules. Additionally, the fraction of Pref undergoing nonspecific ligand binding provides a quantitative measure of the contribution of nonspecific ligand binding to the measured abundance of protein and specific proteinligand complex. Nonspecific complexes or clusters are also commonly observed in ES-MS analysis of solutions of carbohydrates,39 amino acids,4044 peptides,45 or salts.46,47 Several mechanisms have been suggested to account for their formation. Such nonspecific cluster formation is, perhaps, most easily rationalized on the basis of the CRM, whereby small molecules cluster together in the droplet as a result of solvent evaporation.46,47 It has also been
ARTICLE
reasonably suggested that nonspecific clusters can arise from the gas-phase fragmentation of larger clusters, which were formed initially by CRM.48 Alternatively, the ion evaporation mechanism (IEM),49 which is believed to be responsible for the formation of gaseous ions of small molecules and their specific complexes in ES, has been used to explain nonspecific cluster formation.40,46,50 According to IEM, solvent evaporation from the charged droplets promotes aggregation, which is followed by evaporation of the aggregate or cluster ions.40 In general, however, the exact mechanism(s) responsible for the formation of particular clusters (aggregates) ions cannot be definitively established.35 To our knowledge, there is no established method for determining whether gaseous ions corresponding to complexes or clusters of small molecules/ions observed by ES-MS originate from solution or whether they formed by nonspecific interactions during the ES process. Here, we describe a simple method for identifying the presence of specific interactions between small molecules or ions in solution using ES-MS. The technique, which we refer to as the nonspecific probe method, relies on the formation of nonspecific interactions between solution components and a macromolecular probe (PNS) during the ES process to probe solution composition. To implement the method a macromolecule, which does not bind specifically to any of the solution components, is added to the ES solution. The existence of specific interactions between small molecules in solution is established from an analysis of the measured distributions of the small molecules bound nonspecifically to PNS. It is important to emphasize that this method is distinct from the reference protein method in that specific interactions between small molecules are identified by considering only the distributions of small molecules bound to PNS. In contrast, the reference protein method is used to identify the presence of specific proteinligand interactions by comparing the distributions of small molecules bound to the protein of interest and to Pref. The reliability of the nonspecific probe method for distinguishing specific from nonspecific small-molecule interactions is illustrated for peptide antibiotic and metal ionligand interactions in water.
’ EXPERIMENTAL SECTION Materials. Lysozyme (Lyz, MW 14 311 Da), vancomycin (Van, MW 1448 Da), NR,Nε-diacetyl-L-Lys-D-Ala-D-Ala (Ac2KAA, MW 372 Da), and acetylSer-Gln-Asn-Tyr-Pro-Val-Valamide (AcSQNYPVVNH2, MW 847 Da), as well as EDTA and manganese acetate, were purchased from Sigma-Aldrich Canada (Oakville, Canada). Lyz was buffer-exchanged into 50 mM aqueous ammonium acetate (pH 7) and concentrated using an Amicon ultracentrifugal device (MWCO 10 kDa). All other compounds were used without further purification. Stock solutions of Lyz were prepared at a concentration of 150 μM in 50 mM aqueous ammonium acetate; stock solutions of the small molecules and salts were prepared at a concentration of 1 mM in water. The ES solutions were prepared from aqueous stock solutions of Lyz and small molecules or salt. Aqueous ammonium acetate (pH 7) was added to give a final concentration of 5 mM. Mass Spectrometry. All experiments were performed with an ApexII 9.4 T Fourier transform ion cyclotron resonance mass spectrometer (Bruker, Billerica, MA) equipped with an external nanoflow ES ion source. A description of the instrument and the experimental and instrumental parameters used in the ES-MS measurements is given elsewhere.51 5161
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry
ARTICLE
Implementation of the PNS Method: Theoretical Considerations. According to the nonspecific probe method, the origin
(whether they originate from specific interactions in solution or entirely or in part from nonspecific interactions during the ES process) of small-molecule complexes detected by ES-MS can be established from an analysis of the distributions of solution components, i.e., the small molecules and their complexes, bound nonspecifically to a macromolecule. In other words, the occurrence of nonspecific binding between small molecules during ES-MS analysis is established from the tendency of these same molecules to bind nonspecifically to a macromolecule. To implement the method, a macromolecular probe (PNS), which does not interact specifically with any of the solution components, is used. In principle, any “noninteracting” macromolecule that is readily detected by ES-MS may fulfill the role of PNS; in the present study the small protein Lyz was used. Below, a general mathematical framework for establishing the origin of heterocomplexes composed of small molecules is presented. When two noninteracting small molecules (A and B) and PNS are present in solution, nonspecific binding of both A and B to PNS may occur during the ES process, resulting in the appearance of gaseous ions corresponding to PNS bound to a variable number of A or B molecules, i.e., PNSAi and PNSBj species, as well as ions corresponding to PNS bound simultaneously to both A and B molecules, i.e., PNSAiBj species (Figure 1a). Because nonspecific binding during ES is a random process,20 the relative abundance (Abrel) of any PNSAiBj species will be given by the joint probability,52 p(i,j), eq 1: Abrel ðPNS A i Bj Þ ¼ pði, jÞ ¼ pi A pj B
ð1Þ
where piA and pjB are the individual probabilities of PNS binding nonspecifically to i molecules of A and j molecules of B, respectively. If the equality in eq 1 holds, such that for all i and j the measured or apparent Abrel, i.e., Abrel,app(PNSAiBj), is equal to piApjB, then it can be concluded that only individual A and B molecules are present in solution. On the other hand, if, for some i and j values, Abrel,app(PNSAiBj) is greater than piApjB, it can be concluded that specific interactions between A and B exist in solution. For example, if A and B interact to form the 1:1 AB complex, the mass spectrum will be influenced by the nonspecific binding of AB to PNS, which leads to formation of PNS(AB)k and PNSAiBj(AB)k species. Because the masses of the PNS(AB)k species are indistinguishable from those of the PNSAiBj species when i = j, i.e., PNS(AB)i=j, Abrel,app(PNS(AB)i=j) for these species will be increased compared to the situation where there is no AB complex present in solution (Figure 1b). The increase in relative abundance is related to the individual probability of PNS binding nonspecifically to k molecules of AB, i.e., pkAB, eq 2:
Figure 1. Simulated distribution of the nonspecific complexes of nonspecific probe protein (PNS) with small molecules A and B formed during the ES process. (a) Distribution in the case where there is no specific interaction between A and B in solution. (b) Distribution in the case where free A and B, as well as the 1:1 complex AB, are present in solution.
the masses of the PNSAiþkBjþk species are the same as those for the complexes resulting from the nonspecific binding of individual 6 j and A and B molecules to PNS, Abrel,app for all species with i ¼ i, j g k will also be increased by the presence of the AB complex in solution. It follows that the presence or absence of the 1:1 AB complex in solution can be established by comparing Abrel,app(PNSAB) with the joint probability, p1Ap1B. The magnitude of Abrel,app(PNSAB) can be calculated from the ES mass spectrum using eq 4: Abrel, app ðPNS ABÞ ¼
ð4Þ
p1 A ¼
AbðPNS AÞ AbðPNS A i Þ
ð5aÞ
AbðPNS BÞ AbðPNS Bj Þ
ð5bÞ
∑i
ð2Þ
where i = j = k. More generally, Abrel,app for any PNSAiþkBjþk species, including those for which i ¼ 6 j, is given by eq 3: Abrel, app ðPNS A iþk Bjþk Þ ¼ pði, j, kÞ ¼ pi A pj B pk AB
∑i, j
while p1A and p1B values can be calculated from the mass spectrum using eqs 5a and 5b, respectively:
Abrel, app ðPNS ðABÞi ¼ j Þ ¼ Abrel ðPNS A i Bj Þ þ Abrel ðPNS ðABÞk Þ ¼ pi A pj B þ pk AB
AbðPNS ABÞ AbðPNS A i Bj Þ
ð3Þ
where p(i,j,k) is the joint probability of PNS binding nonspecifically to i molecules of A, j molecules of B, and k molecules of AB. Since
p1 B ¼
∑j
where Ab(PNSAi), Ab(PNSBj), and Ab(PNSAiBj) represent the absolute abundance of the corresponding PNSAi, PNSBj, and PNSAiBj ions, respectively. A nonzero difference, i.e., Abrel,app(PNSAB) p1Ap1B > 0, indicates that the specific AB complex exists in solution. 5162
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry
ARTICLE
Figure 2. (a) ES mass spectrum obtained for an aqueous solution of 200 μM EDTA, 200 μM MnCl2, and 12 μM Lyz. The peak (m/z = 391) labeled as r corresponds to diisooctyl phthalate; peaks labeled by / correspond to electronic noise. (b) Normalized distribution of nonspecific (Lyz þ qMn(EDTA)) complexes determined from the ES mass spectrum (exp) and distribution expected for a single Poisson process with a λ of 0.28 (calc).
In the present study, the ratio (R) of Abrel,app(PNSAB) to p1Ap1B was used to test for the presence of the 1:1 AB complex in solution, eq 6: R¼
Abrel, app ðPNS ABÞ p1 A p1 B
ð6Þ
where an R > 1 indicates that AB is present in solution and an R ≈ 1 indicates the absence of specific AB complex. The mathematical treatment above deals with the situation where a specific 1:1 AB complex is present in solution. However, it should be noted that the nonspecific probe method is general and can be easily extended to situations where AiBj complexes with other stoichiometries exist in solution.
’ RESULTS AND DISCUSSION To demonstrate the reliability of the nonspecific probe method for establishing the presence or absence of specific interactions between small molecules in solution, a number of control experiments were performed. The high-affinity complex between EDTA and the divalent metal ion Mn2þ served as a positive control. At neutral pH, Mn2þ coordinates all four of the acidic groups of EDTA forming 1:1 specific complex (EDTA þ Mn) Mn(EDTA), for which the Ka is 7.8 1013 M1 at 25 °C.53 Measurements were also performed on a second model system involving the antibiotic glycopeptide vancomycin (Van), which binds specifically
to analogues of the bacterial cell wall peptides terminating in a D-alanyl-D-alanine sequence. The interactions between Van and the specific peptide ligand Ac2KAA, which binds with a Ka of 1.5 106 M1 at pH 7 and 25 °C,54 and the peptide, Ac SQNYPVVNH2, which does not bind specifically, served as positive and negative controls, respectively. Mn(EDTA) Complex. Shown in Figure 2a is a representative ES mass spectrum acquired for an aqueous solution containing equimolar (200 μM) MnCl2 and EDTA, as well as Lyz (12 μM), which served as the PNS. At equilibrium, the Mn(EDTA) complex is the only major species present in solution, with a concentration of ∼200 μM; the concentrations of free EDTA and Mn2þ are ∼80 nM. The free EDTA exists primarily in its triply deprotonated form, (EDTA 3H)3 EDTA3,55 whereas the Mn(EDTA) complex exists as the dianion, (EDTA 4H þ Mn)2 Mn(EDTA)2. As expected, the most abundant gas-phase ions containing EDTA that were detected by ES-MS originate from the Mn(EDTA)2 species, with the protonated and sodiated forms dominating, i.e., (EDTA H þ Mn)þ Mn(EDTA)þ, (EDTA 2H þ Mn þ Na)þ Mn(Na)(EDTA)þ, (EDTA 3H þ Mn þ 2Na)þ Mn(Na2)(EDTA)þ, and (EDTA 4H þ Mn þ 3Na)þ Mn(Na3)(EDTA)þ. Also detected were ions corresponding to the protonated and sodiated forms of the dimer of Mn(EDTA)2, i.e., (2EDTA 3H þ 2Mn)þ Mn2(EDTA2)þ, (2EDTA 4H þ 2Mn þ Na)þ Mn2(Na)(EDTA2)þ, (2EDTA 5H þ 2Mn þ 2Na)þ Mn2(Na2)(EDTA2)þ, 5163
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry
ARTICLE
Figure 3. ES mass spectrum obtained for an aqueous solution of 100 μM EDTA, 200 μM MnCl2, and 12 μM Lyz. þ
þ
(2EDTA 6H þ 2Mn þ 3Na) Mn2(Na3)(EDTA2) , and (2EDTA 7H þ 2Mn þ 4Na)þ Mn2(Na4)(EDTA2)þ. Ions corresponding to protonated Lyz, i.e., Lyznþ where n = 7 and 8, and Lyz bound nonspecifically to Mn(EDTA)2, i.e., (Lyz þ qMn(EDTA))nþ where q = 1 and 2, and n = 7 and 8, were also present. However, no ions corresponding to Lyz bound nonspecifically to only EDTA or only Mn2þ were detected. This finding is consistent with free EDTA and Mn2þ being present at very low concentrations in solution. Given the large relative abundance measured for the (Lyz þ qMn(EDTA))nþ ions, together with the negligibly small joint probability for the simultaneous attachment of EDTA and Mn2þ to PNS, i.e., p1EDTAp1Mn ≈ 0, it can be concluded that the Mn(EDTA) complex exists in solution and that the (Lyz þ Mn(EDTA))nþ and (Lyz þ 2Mn(EDTA))nþ ions originated exclusively, or nearly so, from the nonspecific binding of Mn(EDTA) complex to PNS during the ES process. Shown in Figure 3 is an ES mass spectrum acquired for a solution of EDTA (100 μM), MnCl2 (200 μM), and Lyz (12 μM). At equilibrium, both free Mn2þ and Mn(EDTA)2 ions are present at high concentration (∼100 μM). The ES mass spectrum is similar in appearance to the one shown in Figure 2a, with the exception that protonated (Lyz þ Mn)nþ ions were also detected. Although the probability of nonspecific binding of one Mn2þ to Lyz (p1Mn) is significant, 0.16, the probability of a single EDTA ion binding nonspecifically to Lyz (p1EDTA) is very small, ≈ 0, which is consistent with the low concentration of free EDTA in solution. As a result, the joint probability of EDTA and Mn2þ simultaneous binding to PNS is also negligibly small, p1EDTAp1Mn ≈ 0. Therefore, from this analysis it is correctly concluded that the Mn(EDTA)2 ion exists in solution at an appreciable concentration and that the (Lyz þ qMn(EDTA))nþ ions originate from the nonspecific binding of the intact Mn(EDTA)2 complex and not individual EDTA3 and Mn2þ ions. Shown in Figure 4 is an ES mass spectrum acquired for a solution containing MnCl2 (100 μM), EDTA (200 μM), and Lyz (12 μM). In this case, both the EDTA3 and Mn(EDTA)2 ions are present at high concentrations (∼100 μM), which leads to the formation of abundant protonated (Lyz þ Mn(EDTA))nþ and (Lyz þ EDTA)nþ ions. In contrast, the concentration of free Mn2þ in solution is very low and, as a result, no (Lyz þ Mn)nþ ions are detected. Therefore,
Figure 4. ES mass spectrum obtained for an aqueous solution of 200 μM EDTA, 100 μM MnCl2, and 12 μM Lyz.
although the probability of attaching one molecule of EDTA to Lyz is significant, 0.07, the joint probability of the simultaneous binding of EDTA and Mn2þ to PNS is negligibly small. Therefore, from the analysis of the nonspecific complexes of PNS, it can be concluded that the Mn(EDTA)2 ion exists in solution at an appreciable concentration. It is important to note that the distributions of (Lyz þ qMn(EDTA)) species measured in the ES mass spectra shown in Figures 24 are well-described by single Poisson functions. For example, the distribution determined from Figure 2a can be described, with >99% confidence (χ2 = 0.002), by a single Poisson function with a λ of 0.28 (Figure 2b).56 The distributions taken from Figures 3 and 4 can also be described by Poisson functions, although with λ values of 0.25 (χ2 = 0.0001) and 0.17 (χ2 = 0.00005), respectively (data not shown). The variation in λ values highlights the variability in the nonspecific binding process observed in different ES-MS measurements.20 Taken on its own, the observation that the formation of the (Lyz þ qMn(EDTA)) species is a Poisson process is sufficient to draw the conclusion that the Mn(EDTA)2 complex exists in solution. It can also be concluded that the dimer of the Mn(EDTA)2 complex does not exist in solution, at least not at a concentration high enough to lead to nonspecific binding to PNS. If the Mn2(EDTA2) species were present in solution at high concentration, the distribution of (Lyz þ qMn(EDTA)) species would no longer be adequately described by a single Poisson function. VancomycinPeptide Interactions. Control experiments were also performed on solutions containing Van and one of two peptides, the specific tripeptide Ac2KAA ( KAA) or AcSQNYPVVNH2 ( Pep), which served as a negative control. Shown in Figure 5a is a representative ES mass spectrum measured for an aqueous solution of Van (100 μM), Pep (10 μM), and Lyz (12 μM). As before, Lyz was used as PNS for these measurements. Abundant peaks corresponding to the protonated (Lyz þ iVan þ jPep)nþ ions, where n = 79 and i and j range from 0 to 2, were detected. The distribution of the nonspecific Lyz complexes is shown in Figure 5b. The magnitude of Abrel,app(PNSVanPep) determined from the mass spectrum is 0.022, while the probabilities of the nonspecific attachment of a 5164
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry
Figure 5. (a) ES mass spectrum obtained from an aqueous solution of 12 μM Lyz, and 100 μM Van and 10 μM Pep. (b) Normalized distributions of nonspecific complexes (Lyz þ iVan þ jPep).
single Van (p1Van) and a single Pep (p1Pep) to Lyz are found to be 0.10 and 0.18, respectively. It follows that the joint probability of concomitant nonspecific binding Pep and Van to Lyz (p1Vanp1Pep) is 0.018. The corresponding value of R, 1.22, is close to unity, which is consistent with an absence of a specific interaction between Van and Pep in solution. Six replicate measurements were performed on the same solution but using different nanoES tips. Importantly, the calculated R values exhibit a very narrow range of values, 0.981.24, with an average value of 1.09 ( 0.09. Shown in Figure 6a is a representative ES mass spectrum acquired for an aqueous solution of equimolar (100 μM) Van and KAA with Lyz (12 μM). At equilibrium, the specific (Van þ KAA) complex is present at a concentration of 92 μM, while the concentration of both free Van and KAA is 8 μM. Inspection of the ES mass spectrum reveals ion signal corresponding to the protonated KAA ions, i.e., KAAþ, as well as doubly charged, protonated ions of free and bound Van ions, i.e., Van2þ and (Van þ KAA)2þ, respectively. In addition, abundant signals corresponding to protonated Lyz ions, i.e., Lyznþ where n = 79, and protonated nonspecific complexes of Lyz with Van and KAA, i.e., (Lyz þ Van)nþ, (Lyz þ KAA)nþ, (Lyz þ Van þ KAA)nþ, and (Lyz þ 2Van þ KAA)nþ where n = 7, 8, were detected. The absence of nonspecific complexes at the þ9 charge state is intriguing and may be the result of preferential in-source dissociation of the nonspecific adducts at this charge state.16 The distribution of the nonspecific Lyz complexes, i.e., (Lyz þ iVan þ jKAA) where i = 02, j = 0 and 1, obtained from the mass spectrum, is shown in Figure 6b. The probabilities for the formation of the nonspecific (Lyz þ Van) and (Lyz þ KAA) complexes are 0.13 (p1Van) and 0.03 (p1KAA), respectively. It
ARTICLE
Figure 6. (a) ES mass spectrum obtained from aqueous solution of 12 μM Lyz, and 100 μM Van and 100 μM KAA. (b) Normalized distributions of nonspecific complexes (Lyz þ iVan þ jKAA).
follows that the joint probability for the simultaneous nonspecific attachment Van and KAA (p1Vanp1KAA) is 0.004. In contrast, the magnitude of Abrel,app(LyzVanKAA) is significantly larger, 0.12. Given that the value of R in this case is 30, it is correctly concluded that the (Van þ KAA) complex exists in solution at an appreciable concentration. Eight replicate measurements were carried out on the same solution but using different nanoES tips. In all cases, analysis of the mass spectra returned a value of R that was significantly larger than unity, although the magnitude of the value varied considerably, from 9 to 50 (data not shown). Measurements were also performed on solutions with different concentrations of Van and KAA. Shown in Figure S1 (Supporting Information) are representative mass spectra acquired at Van and KAA concentrations of 6 and 15, 13 and 25, 25 and 25, and 50 and 50 μM, respectively. Notably, the corresponding R values varied from 9 to 32, but no simple dependence on concentration was evident. Taken together, the results obtained for the positive controls involving the high-affinity interaction involving EDTA and the Mn2þ ion, as well as the positive and negative controls involving the antibiotic Van with an interacting and a noninteracting peptide, demonstrate the utility and reliability of the nonspecific probe method, combined with direct ES-MS measurements, for identifying small-molecule interactions in aqueous solution. Importantly, the nonspecific probe method is general and is expected to find applications in ES-MS studies of a wide variety of noncovalent interactions involving small molecules and ions. However, successful implementation of the method does require that the small molecules of interest and their noncovalent 5165
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry complexes bind nonspecifically to PNS during the ES process. Consequently, the method is most easily implemented when using experimental conditions that promote nonspecific binding of small molecules to PNS (e.g., high analyte concentrations, high solution flow rates).20
’ CONCLUSIONS In summary, a method for determining whether complexes composed of small molecules detected by ES-MS originate from specific interactions in solution or from nonspecific binding during the ES process is described for the first time. The technique, referred to as the nonspecific probe method, exploits the tendency of small molecules and ions to bind nonspecifically to macromolecules during the ES process and the statistical nature of nonspecific binding to gain insight into the solution composition. A series of control experiments were performed using peptideantibiotic and metal ionligand interactions as model systems to demonstrate the reliability and generality of the method. In all cases, the results were consistent with the known composition of the solutions investigated. Notably, this technique is general and is, in principle, applicable to any interactions involving small molecules and ions. Furthermore, the method is expected to be compatible with a wide range of ES-MS instruments. ’ ASSOCIATED CONTENT
bS
Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
’ ACKNOWLEDGMENT The authors acknowledge the Natural Sciences and Engineering Research Council of Canada and the Alberta Ingenuity Centre for Carbohydrate Science for funding. ’ REFERENCES (1) Daniel, J. M.; Friess, S. D.; Rajagopalan, S.; Wendt, S.; Zenobi, R. Int. J. Mass Spectrom. 2002, 216, 1–27. (2) Heck, A. J. R.; van den Heuvel, R. H. H. Mass Spectrom. Rev. 2004, 23, 368–389. (3) Schug, K. A.; Serrano, C.; Frycak, P. Mass Spectrom. Rev. 2010, 29, 806–829. (4) Simmons, D. A.; Wilson, D. J.; Lajoie, G. A.; Doherty-Kirby, A.; Konermann, L. Biochemistry 2004, 43, 14792–14801. (5) Natan, E.; Hirschberg, D.; Morgner, N.; Robinson, C. V.; Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14327–14332. (6) Wortmann, A.; Kistler-Momotova, A.; Zenobi, R.; Heine, M. C.; Wilhelm, O.; Pratsinis, S. E. J. Am. Soc. Mass Spectrom. 2007, 18, 385–393. (7) Sharon, M.; Robinson, C. V. Annu. Rev. Biochem. 2007, 76, 167–193. (8) Benesch, J. L. P.; Robinson, C. V. Curr. Opin. Struct. Biol. 2006, 16, 245–251. (9) Heck, A. J. R. Nat. Methods 2008, 5, 927–933. (10) Aquilina, J. A.; Benesch, J. L.; Bateman, O. A.; Slingsby, C.; Robinson, C. V. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 10611–10616. (11) Gabelica, V.; Galic, N.; Rosu, F.; Houssier, C.; De Pauw, E. J. Mass Spectrom. 2003, 38, 491–501. (12) Kempen, E. C.; Brodbelt, J. S. Anal. Chem. 2000, 72, 5411–5416.
ARTICLE
(13) Gabelica, V.; Rosu, F.; De Pauw, E. Anal. Chem. 2009, 81, 6708–6715. (14) Wortmann, A.; Rossi, F.; Lelais, G.; Zenobi, R. J. Mass Spectrom. 2005, 40, 777–784. (15) Van Dongen, W. D.; Heck, A. J. R. Analyst 2000, 125, 583–589. (16) Sun, J.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2007, 79, 416–425. (17) Clark, S. M.; Konermann, L. Anal. Chem. 2004, 76, 7077–7083. (18) Robinson, C. V.; Chung, E. W.; Kragelund, B. B.; Knudsen, J.; Aplin, R. T.; Poulsen, F. M.; Dobson, C. M. J. Am. Chem. Soc. 1996, 118, 8646–8653. (19) El-Hawiet, A.; Kitova, E. N.; Liu, L.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2010, 21, 1893–1899. (20) Wang, W.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2005, 77, 3060–3071. (21) Sun, J.; Kitova, E. N.; Wang, W.; Klassen, J. S. Anal. Chem. 2006, 78, 3010–3018. (22) Sun, N.; Soya, N.; Kitova, E. N.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2010, 21, 472–481. (23) Sun, N.; Sun, J.; Kitova, E. N.; Klassen, J. S. J. Am. Soc. Mass Spectrom. 2009, 20, 1242–1250. (24) Tang, L.; Kebarle, P. Anal. Chem. 1993, 65, 3654–3668. (25) Cech, N. B.; Enke, C. G. Anal. Chem. 2000, 72, 2717–2723. (26) P. Frycak, P.; Schug, K. A. Anal. Chem. 2008, 80, 1385–1393. (27) Page, J. S.; Kelly, R. T.; Tang, K.; Smith, R. D. J. Am. Soc. Mass Spectrom. 2007, 18, 1582–1590. (28) Pan, P.; McLuckey, S. A. Anal. Chem. 2003, 75, 1491–1499. (29) Sterner, J. L.; Johnston, M. V.; Nicol, G. R.; Ridge, D. P. J. Mass Spectrom. 2000, 35, 385–391. (30) Vaidyanathan, S.; Douglas, B.; Kell, D. B.; Goodacre, R. Anal. Chem. 2004, 76, 5024–5032. (31) Deng, L.; Sun, N.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2010, 82, 2170–2174. (32) Dole, M.; Mach, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L. P.; Alice, M. B. J. Chem. Phys. 1968, 49, 2240–2249. (33) de la Mora, J. F. Anal. Chim. Acta 2000, 406, 93–104. (34) Peschke, M.; Verkerk, U. H.; Kebarle, P. J. Am. Soc. Mass Spectrom. 2004, 15, 1424–1434. (35) Kebarle, P. J. Mass Spectrom. 2000, 35, 804–817. (36) Daubenfeld, T.; Bouin, A. P.; van der Rest, G. J. Am. Soc. Mass Spectrom. 2006, 17, 1239–1248. (37) Shimon, L.; Sharon, M.; Horovitz, A. Biophys. J. 2010, 99, 1645–1649. (38) Lane, L. A.; Ruotolo, B. T.; Robinson, C. V.; Favrin, G.; Benesch, J. L. P. Int. J. Mass Spectrom. 2009, 283, 169–177. (39) Gabelica, V.; Galic, N.; De Pauw, E. J. Am. Soc. Mass Spectrom. 2002, 13, 946–953. (40) Meng, C. K.; Fenn, J. B. Org. Mass Spectrom. 1991, 26, 542–549. (41) Nemes, P.; Schlosser, G.; Vekey, K. J. Mass Spectrom. 2005, 40, 43–49. (42) Cooks, R. G.; Zhang, D.; Koch, K. J.; Gozzo, F. C.; Eberlin, M. N. Anal. Chem. 2001, 73, 3646–3655. (43) Koch, K. J.; Gozzo, F. C.; Zhang, D.; Eberlin, M. N.; Cooks, R. G. Chem. Commun. 2001, 18, 1854–1855. (44) Koch, K. J.; Gozzo, F. C.; Nanita, S. C.; Takats, Z.; Eberlin, M. N.; Cooks, R. G. Angew. Chem., Int. Ed. 2002, 41, 1721–1724. (45) Counterman, A. E.; Hilderbrand, A. E.; Barnes, C. A. S.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 2001, 12, 1020–1035. (46) Camero-Castano, M.; de la Mora, J. F. Anal. Chim. Acta 2000, 406, 67–71. (47) Wang, G.; Cole, R. B. Anal. Chem. 1998, 70, 873–881. (48) Spenser, E. A. C.; Ly, T.; Julian, R. R. Int. J. Mass Spectrom. 2008, 270, 166–172. (49) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287–2294. (50) Wang, G.; Cole, R. B. Anal. Chim. Acta 2000, 406, 53–65. (51) Wang, W.; Kitova, E. N.; Klassen, J. S. Anal. Chem. 2003, 75, 4945–4955. 5166
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
Analytical Chemistry
ARTICLE
(52) Ibe, O. Fundamentals of Applied Probability and Random Processes; Elsevier Academic Press: Boston, MA, 2005. (53) (a) Ogino, H. Bull. Chem. Soc. Jpn. 1965, 38, 771–777. (b) Martell, A. E.; Smith, R. M.; Motekaitis, R. J. NIST Critically Selected Stability Constants of Metal Complexes; NIST Standard Reference Database 46; National Institute of Standards and Technology: Gaithersburg, MD, 2004. (54) Nieto, M.; Perkins, H. R. Biochem. J. 1971, 123, 789–803. (55) Patnaik, P. Dean’s Analytical Chemistry Handbook, 2nd ed.; McGraw-Hill: New York, 2004. (56) The value of λA can be found from the ratio of the relative abundance of any two PNSAi and PNSAiþ1 species using the general expression λA = (i þ 1)[Ab(PNSAiþ1)/Ab(PNSAi)], where Ab(PNSAi) and Ab(PNSAiþ1) represent the abundance of the PNSAi and PNSAiþ1 species, respectively.
5167
dx.doi.org/10.1021/ac200244u |Anal. Chem. 2011, 83, 5160–5167
