Formation and Dissociation Processes of Gas-Phase Detergent

Mansfield Road, OX1 3TA Oxford, United Kingdom. Langmuir , 2012, 28 (18), pp 7160–7167. DOI: 10.1021/la3002866. Publication Date (Web): April 18...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/Langmuir

Formation and Dissociation Processes of Gas-Phase Detergent Micelles Antoni J. Borysik and Carol V. Robinson* Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, OX1 3TA Oxford, United Kingdom S Supporting Information *

ABSTRACT: Growing interest in micelles to protect membrane complexes during the transition from solution to gas phase prompts a better understanding of their properties. We have used ion mobility mass spectrometry to separate and assign detergent clusters formed from the n-trimethylammonium bromide series of detergents. We show that cluster size is independent of detergent concentration in solution, increases with charge state, but surprisingly decreases with alkyl chain length. This relationship contradicts the thermodynamics of micelle formation in solution. However, the liquid drop model, which considers both the surface energy and charge, correlates extremely well with the experimental cluster size. To explore further the properties of gas-phase micelles, we have performed collisioninduced dissociation on them during tandem mass spectrometry. We observed both sequential asymmetric charge separation and neutral evaporation from the precursor ion cluster. Interestingly, however, we also found markedly different dissociation pathways for the longer alkyl chain detergents, with significantly fewer intermediate ions formed than for those with a shorter alkyl chain. These experiments provide an essential foundation for understanding the process of the gas-phase analysis of membrane protein complexes. Moreover they imply valuable mechanistic details of the protection afforded to protein complexes by detergent clusters during gas-phase activation processes.

■

INTRODUCTION The analysis of charged clusters is important for improving our understanding of the cohesive properties and dissociation mechanisms of condensed matter in the gas phase.1−3 Within this area, detergent clusters, or ‘gas-phase micelles’, are distinguished by their unique property of also possessing well-defined structures in solution.4−6 Despite this property, however, very little is known about gas-phase micelles compared with other clusters, including transition metals,7−10 peptides,11 van der Waals clusters,12 or nonartificial clusters such as noncovalent protein complexes.13−15 Furthermore, most studies on gas-phase micelles have focused on assessing the efficacy of electrospray ionization mass spectrometry (ESIMS) to determine critical micelle concentrations (CMC) or the aggregation numbers of detergents, both of which are solutionphase properties not gas-phase phenomena.16−18 While this work has been interspersed with several other studies aimed at improving the general understanding of gas-phase micelles, overall there has been very little activity in this area relative to that of other systems.19−24 Recent developments have, however, provoked more interest in the properties of gas-phase micelles by virtue of their direct involvement in the process of obtaining mass spectra for intact membrane protein complexes.25−29 In this approach, mass spectra are obtained directly from aqueous membrane protein solutions containing excess detergent. The detergent can then be removed from the protein complexes in the gas-phase by the use of collisional activation. One surprising aspect of this approach is that the detergents appear to protect the protein © 2012 American Chemical Society

complexes from the deleterious effects of collision voltages normally associated with the considerable loss of both tertiary and quaternary protein structure.30−33 Nevertheless, it is becoming increasingly apparent that the effective removal of detergents from protein complexes in the gas phase is a casedependent phenomenon. Thus, a renewed effort to fully understand the properties of gas-phase micelles is now clearly warranted. Here we focus on the factors involved in the distribution of different detergent cluster sizes observed during ESI-MS and also follow the dissociation of these clusters during collisioninduced dissociation (CID). The mass spectra of detergent clusters contain many overlapping charge state series rendering unambiguous assignment extremely challenging. We show that by using ion mobility mass spectrometry (IMMS) we can obtain a full and unequivocal assignment of highly polydisperse nC-trimethyl ammonium bromide (nCTAB) gas-phase micelle clusters (nC = alkyl chain length).34−37 Using this approach, we measure the detergent cluster sizes as a function of alkyl chain length. We show that an increase in the alkyl chain length leads to modest but significant reduction in the observed aggregation numbers of these clusters. Interestingly, therefore, these data clearly indicate that the relationship between aggregate number and alkyl chain length is opposed to that expected in solution.4,5 Thus, and in contrast with previous reports, we suggest that Received: January 21, 2012 Revised: April 17, 2012 Published: April 18, 2012 7160

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

a negative effect on the resolution of the peaks and was therefore omitted from all subsequent acquisitions. Mass spectra were processed using Masslynx software v 4.1 and Driftscope v 2.1 (Waters Corp. Manchester, UK). Irrespective of the detergent, all peaks in the mass spectra were found to obey the general formula [(nCTAB)a + (nCTA)b]b+ where nC = alkyl chain length in carbon atoms, a = any integer value and b = the charge state of the micelle. No other adducted ions, i.e., where H+ is replaced by Na+ or K+, were observed in the gas-phase micelles. The reproducibility of the observed cluster sizes was extremely good, with coefficients of variation no greater than 3% for charge states greater than 2+. Tandem mass spectrometry (MSMS) experiments were performed in argon without adjustment to the instrument parameters. Given the high degree of polydispersity in the mass spectra, we found that relatively few ions represented viable candidates for MSMS. This was because of the considerable limitations in finding m/z values that were unique for a particular aggregation number, or that were sufficiently resolved from neighboring ions. To reduce contamination from neighboring ions, quadrupole resolution settings were also set high and varied between 17.5 and 22.5, depending on the detergent. Note that in experiments to follow the dissociation of the ions, the 83n5+ ions of 16TAB could not be investigated because of the poor signal from this ion. Derivation of Critical Cluster Size. The critical sizes of the detergent clusters were calculated using the liquid drop model.1,38,39 The liquid drop model is a kinetic theory that permits the determination of the dimensionless ion fissibility parameter (X) for any gas-phase cluster, eq 1. As X approaches 1, the barrier height for charge separation is reduced with a concurrent increase in the rate of fission, until the maximal property X = 1 is reached, at which point barrierless fission ensues and the rate of this process is said to be spontaneous.

detergent aggregation numbers cannot be ascertained directly by ESI-MS.16−18 We next applied the liquid drop model to describe the observed distribution of micelle size obtained by ESI-MS.38,39 In contrast with the coefficients of micellation applicable to a condensed-phase environment, the liquid drop model also considers the repulsive (Coulomb) energy. This model therefore has been shown to have greater relevance to the behavior of charged gas-phase clusters.1−3 We used this model to calculate the critical sizes (ncrit) of the detergents at each charge state to explore how the experimental data would compare with a system that was constrained by electrostatics. Using this approach, we show a high degree of correlation between the average detergent cluster sizes and ncrit. This enables us to propose that this model is better suited to describe the sizes of gas-phase micelles obtained by ESI-MS than models typically used in solution. Having identified mass-to-charge ratios (m/z) that were unique for detergent clusters of a particular aggregation number, we selected these ions using tandem mass spectrometry (MSMS). We then monitored their dissociation during CID to obtain insights into the dissociation properties of these clusters during thermal agitation. We found that gasphase micelles dissociate gradually by CID in a process characterized by multiple rounds of asymmetric charge separation as well as the evaporation of neutrals from the clusters, typical of liquid drop behavior.1,11,12 Interestingly, however, our data also indicate the existence of a different dissociation pathway in which fewer ‘intermediate’ ions are formed. The experiments allude to potential mechanisms by which detergent clusters protect membrane protein complexes in the gas phase. They therefore serve as an essential foundation to understand the complex process of delivering intact membrane protein complexes into the gas phase.

■

X=

(z 2/n) (z 2/n)crit

(1)

The numerator in eq 1 (z2/n) is specific to the experimental conditions and relates to the charge state (z) and aggregation number (n) of a cluster. The term (z2/n)crit is material-specific and is defined in eq 2, where e is the elementary charge in cgs units, V is the molecular volume of the individual detergents in cm3, and γ is their surface tension in dyn/cm.

MATERIALS AND METHODS

Reagents and Sample Preparation. The detergents hexyltrimethylammonium bromide (6TAB), trimethyloctylammonium bromide (8TAB), decyltrimethylammonium bromide (10TAB), dodecyltrimethylammonium bromide (12TAB), myristyltrimethylammonium bromide (14TAB), and cetyltrimethylammonium bromide (16TAB) were all purchased from Sigma Aldrich Ltd. (Dorset, UK). All detergent solutions were prepared in deionized water without any other additions or treatment except for 16TAB, which required slight heating before use due to detergent crystallization. Mass Spectrometry. Mass spectra were acquired in positive ion mode on a second generation Synapt HDMS (Waters Corp. Manchester, UK) quadrupole-ion trap-IMMS instrument fitted with a nanoflow electrospray ionization source (nano-ESI). The instrument was calibrated externally with cesium iodide up to a mass range of 10 000 m/z. Nano-ESI needles were prepared in-house as described previously.40 Spectra were acquired at room temperature at detergent concentrations of 500 mM (6TAB), 300 mM (8TAB), 100 mM (10TAB), 30 mM (12TAB), 25 mM (14TAB), and 20 mM (16TAB) which were found to be the best concentrations for the quality of the spectra. Instrument parameters were carefully adjusted so as to maintain the interactions of the gas-phase micelles. Important settings were capillary voltage (0.8 kV), sampling, and extraction cone voltage (1 V). The trap, helium cell, and IMS pressures were 2.9 × 10−2, 1.4 × 10−3, and 2.3 mbar, respectively, and the backing pressure was maintained at between 5.0 and 6.0 mbar. The trap and transfer collision energies were 10 V and 5 V, respectively. IMMS separation of the detergent spectra was achieved with a fixed traveling wave height and velocity set to 30 V and 400 m/s, respectively. Pusher frequencies were optimized at an acquisition window of 50 to 9000 m/z. The effect of ammonium acetate was also investigated, but this was found to have

(z 2/n)crit =

12Vγ e2

(2)

Accordingly, taking the cohesive term (γ) of each detergent to be equivalent to its corresponding alkane at 20 °C and X = 1 (the Coulomb limit), eqs 1 and 2 can be combined and rearranged to produce eq 4 from which ncrit can be derived (see main text).

■

RESULTS AND DISCUSSION Gas-Phase Micelles Separated by Size and Charge. Preliminary experiments were focused on finding the optimum conditions for obtaining detergent clusters in the gas phase. Decyltrimethylammonium bromide (10TAB) was used initially at different concentrations. The optimum concentration, in terms of the quality of the mass spectra, was found to be 100 mM. Instrument parameters were then optimized so as to obtain conditions to best preserve the interactions of micelles in the gas phase, with emphasis also placed on the degree of resolution between the many peaks of the polydisperse spectra (see Materials and Methods). Important parameters for these experiments were capillary voltages (0.8 kV) and respective pressures of 2.9 × 10−2, 1.4 × 10−3, and 2.3 mbar for the trap, helium, and ion mobility cells (full parameters are given in Materials and Methods). The mass spectrum of 100 mM 10TAB in water is shown (Figure 1a). The spectrum features a broad distribution of 7161

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

demonstrate that the aggregation number increases with the charge state of the cluster. Gas-Phase Aggregation Number and Its Dependence on Alkyl Chain Length. Previous attempts to assess the aggregation numbers of detergents using ESI-MS have resulted in some encouraging results.16−18,20 However, the determination of the detergent aggregation number is often compromised by the ambiguity in the assignment of the detergent clusters. Thus, the ability to assign the polydisperse mass spectrum of 10TAB unambiguously prompted us to reevaluate the feasibility of using ESI-MS for the determination of detergent aggregation numbers. To perform this evaluation, we varied the detergent alkyl chain length to see how this affected the average cluster size of the polydisperse spectra of the gas-phase micelles. Mass spectra were obtained for the entire homologous series of the nCTAB family of detergents from hexyltrimethylammonium bromide (6TAB) to cetyltrimethylammonium bromide (16TAB). The spectra were acquired for each detergent in triplicate, and the average aggregation numbers of the different charge states were determined from the IMMS contour plots (see Materials and Methods). A plot of the average aggregation number against the alkyl chain length shows that the series from 6TAB to 16TAB leads to a decrease in the aggregation number (Figure 2a). Interestingly, therefore, this observation is in contrast with the established relationship in which increasing chain length leads to larger aggregation numbers in solution.4−6 To highlight this difference, we calculated the solution aggregation numbers of these detergents using their theoretical critical packing numbers nmax sp, assuming micelles of spherical geometry eq 3.5

Figure 1. Mass spectrum and IMMS contour plot of 100 mM 10TAB in water. (a) The ESI-MS spectrum shows a polydisperse distribution of ions between m/z values of 2000 and 6000, with 2-fold magnification applied. The peaks for the singly charged monomer and singly charged dimer ions are marked, and the structure of 10TAB is also shown. (b) Contour plot showing the presence of seven different charge states. The observed range of different aggregation numbers within each charge state are shown in brackets with the charge shown in italics. (b insert) An extracted spectrum of all ions with the 4+ charge state.

peaks with m/z values between 2000 and 6000. Peaks for singly charged monomeric ions are also apparent as well as singly charged dimers of the formula [10TA:10TAB]1+. Consistent with previous acquisitions of 10TAB in water, assignment of the mass spectrum is not trivial.17 However, it was found that the peaks could be assigned unambiguously by separation of the ions according to their unique drift times using IMMS (see Materials and Methods). Following separation of the ions by IMMS it can be seen that the mass spectrum of 10TAB is composed of seven different overlapping charge states (Figure 1b). Each charge state is composed of a Gaussian-like distribution of ions with successive ions separated from their neighbors by the mass of a single monomer [10TAB]. The use of IMMS not only allows unambiguous assignment but also allows us to extract all ions within a particular charge state series (see Figure 1b, insert). For example the 4+ charge state series shows the range of aggregate sizes formed from 46n to 84n with an average number of 65n. The highest aggregation number observed for this detergent is >150n for the 7+ charge state (see Table 1). Overall, these data show that we can assign mass spectra unambiguously with the aid of IMMS and also

ηmax sp =

mass (Da)

6TAB 8TAB 10TAB 12TAB 14TAB 16TAB

224.3 252.2 280.3 308.3 336.4 364.5

nmax sp

40 56 74 94

17 27 (40)b (55)b (70)b (90)b

(3)

where lc equals the maximum alkyl chain extension length (Å) and Vtail the volume of the respective chain (Å3) (calculated using the empirical relationships lc = 1.5 + 1.27nC and Vtail = 27.4 + 26.9nC).5,6 Comparing the cluster sizes determined experimentally with those predicted by nmax sp, it is clear that, in the gas-phase, shortening the alkyl chain length results in larger clusters (Figure 2a). Furthermore, because nmax sp does not depend on charge, this coefficient is also unable to predict the charge state dependency of detergent cluster sizes obtained by ESI-MS. This is particularly evident for the 7+ charge state of 6TAB where the average aggregation number exceeds its predicted condensed-phase value of 17n by >10-fold (see Table 1). Gas-Phase Micelles Rationalized by Their Critical Size. Given the apparent lack of correlation between the aggregation numbers of the detergents obtained by ESI-MS and their cognate values in solution, we investigated a different framework to describe the observed cluster sizes. A common framework used to describe the properties of charged clusters in the gas phase is the liquid drop model, first theorized by Lord Rayleigh.38 The liquid drop model not only considers the surface (cohesive) energy but also the repulsive (Coulomb) energy in terms of their relative contributions to the energy barrier for charge separation by fission. Thus, this model is wellsuited for describing the properties of charged gas-phase clusters, particularly under conditions where electrostatics become important. To make this comparison, we used the liquid drop model to calculate the critical size (ncrit) of the

Table 1. Some Properties of Detergents Used in This Studya detergent

lc3(4π /3) Vtail

n (gas phase)c 175n ± 2n 162n ± 5n 150n ± 1n 142n ± 1n 139n ± 1n 133n ± 2n

a

Molecular weights in Daltons represent detergent plus the bromide counterion. The detergent aggregation numbers (nmax sp) are theoretical values for the maximum aggregation number assuming an ordered spherical geometry (see eq 3).5 bThe aggregation numbers in parentheses for 10TAB−16TAB are experimentally derived values for the respective detergents and were taken from ref 43. cThe gas-phase aggregation numbers are the average values of the 7+ charge states ±1SD of at least three replicate acquisitions and are included as an example. 7162

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

Figure 2. A comparison of the average aggregation numbers of the various detergent cluster ions with their condensed-phase aggregation numbers (nmax sp) and their critical sizes (ncrit). The aggregation numbers determined from experiment are the average cluster size calculated for each charge state and represent the mean value of three separate acquisitions. In part a, the surface represents the theoretical aggregation numbers of these detergents in solution as defined by the parameter nmax sp. In part b, the surface shown represents the critical cluster sizes for these detergents (ncrit) calculated using the liquid drop model.

detergents at each charge state. ncrit defines the minimum cluster size at a given charge and therefore the Coulomb limit where electrostatics are so great that the process of charge separation, or fission, is instantaneous. ncrit was calculated for the series of detergents studied here, for all charge states observed, using eq 4, where z is the cluster charge, V is the molecular volume of the individual detergents in cm3, γ is a bulk cohesive term in dyn/cm, and e is the elementary charge in cgs units (see Materials and Methods).

⎡ z2 ⎤ ηcrit = ⎢ ⎥−1 ⎣ 12Vγ /e 2 ⎦

charge and aggregation numbers was kept to a minimum. The 83n5+ detergent clusters of 6TAB to 14TAB were selected in turn by MSMS and subjected to a stepwise increase in collision energy, up to 200 V (see Materials and Methods). The MSMS spectra produced from each of the precursor ions were monitored over the range of 50 to 9000 m/z, thus allowing the full dissociation patterns of the gas-phase micelles to be observed. MSMS spectra taken during the CID of the 83n5+ gas-phase micelles of 6TAB are shown in Figure 3. At 10 V, the MSMS spectrum contains peaks from two ions, the precursor ion (P) and also a dominant product ion formed from clusters that have lost a singly charged dimer of the formula [6TA:6TAB]1+ (shaded in red). At 40 V, the precursor ion cluster is no longer observed and the MSMS spectrum contains peaks from two different populations of ions that have lost one or two charges (shaded in red and blue, respectively). Peaks from singly charged monomer ions of the formula 6TA1+ are also seen. Interestingly, multiple peaks are now observed within each charge state with each peak separated from its neighbor by the mass of one 6TAB molecule. This indicates that further dissociation has occurred from the 83n5+ precursor ion by a process of neutral loss by evaporation, consistent with the loss of [6TAB]. From 50 to 100 V, the degree of evaporation from the precursor ions is seen to increase considerably such that at 100 V neutral losses up to ∼60n can be observed. Furthermore, at 60 V a third loss of charge is also seen (highlighted in green). At 150 V, four charges have been lost from the precursor ion, and the degree of evaporation is even more extensive. At 200 V, residual clusters that have originated from the precursor ion are still observed although these are only singly charged ions between 3n and 11n. Overall these spectra show that the 83n5+ gas-phase micelles of 6TAB dissociate gradually in a stepwise fashion during CID, leading to the formation of many intermediates that are stable on the time scale of these experiments. Furthermore, it can be seen that the dissociation of the 6TAB micelle is characterized by the sequential loss of charge and also extensive loss of neutrals by evaporation. 14TAB Micelles Populate Fewer Stable Intermediates during CID. The MSMS spectrum of the 83n5+ gas-phase micelles of 14TAB during CID is shown in Figure 4. At 10 V,

(4)

ncrit was projected as a surface and compared directly with the average detergent cluster sizes obtained by ESI-MS (Figure 2b). From this plot, we observe that ncrit decreases with increasing alkyl chain length in line with our experimental data. Moreover, because ncrit depends on charge, this value is also able to predict the charge state dependency of the detergent cluster sizes observed experimentally. It is clear, therefore, that ncrit is a significantly better descriptor of the experimental values than nmax sp. Interestingly, however, a notable discrepancy can be observed between the theoretical and experimental values at longer alkyl chain lengths for the ncrit plot. This could indicate that some of the gas-phase micelles are not charged to their maximum possible value or that their bulk cohesion value is overrepresented in eq 4. Nevertheless, it is clear that the liquid drop model better describes the detergent cluster sizes obtained by ESI-MS than detergent aggregation number in solution. 6TAB Micelles Populate Many Intermediates during CID. Having investigated the formation of detergent clusters by ESI-MS, we next investigated the dissociation of these clusters during thermal agitation by CID. This would allow some detailed insights into the dissociation processes of detergent clusters, under the same conditions used to remove them from protein complexes in the gas phase. For this purpose, m/z values equivalent to the 83n5+ gas-phase micelles for each detergent were selected. It is important to stress that the 83n5+ clusters were chosen on the sole grounds that their m/z values were found to have the greatest degree of separation from neighboring clusters for each detergent. Thus, in this way, the likelihood of introducing contaminating clusters with different 7163

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

precursor ion cluster for the 14TAB is seen to populate only a single intermediate species during CID corresponding to the loss of a singly charged dimer with the formula [14TA:14TAB]1+ (shaded in red) at very low relative intensity. Furthermore, singly charged monomers and dimers (m/d) are formed prior to the formation of this intermediate at a collision energy of 30 V. At 80 V, only singly charged monomer and dimer ions can be observed in the MSMS spectrum. This MSMS spectrum is relatively constant up to 200 V (data not shown). Consequently, the MSMS spectrum of the 83n5+ gasphase micelles of 14TAB suggests that these clusters do not form stable intermediates upon ejection of charged ions or neutral molecules unlike their 6TAB counterparts. The dissociation experiments of the 83n5+ gas-phase micelles of trimethyloctylammonium bromide (8TAB), dodecyltrimethylammonium bromide (10TAB), and dodecyltrimethylammonium bromide (12TAB) reveal a transition between these two extremes (see Figure S1 in Supporting Information). The 8TAB dissociation pattern is closely similar to that observed for 6TAB with ∼20n evaporation events and the sequential loss of 2 [8TA:8TAB]1+ ions up to 70 V. Both the 10TAB and 12TAB MSMS spectra show very low intensity populations of stable intermediates and intense peaks for monomers and dimers at low m/z. To determine whether or not this observation depends upon the charge state or aggregation number of a particular detergent cluster, we also carried out MSMS on the 53n4+, 31n3+, and 47n3+ ions of clusters of 6TAB and 16TAB. However, the same overall trend was observed as that described for the 83n5+ clusters described above in which decreasing alkyl chain length leads to an apparent increase in the stability of the intermediate clusters formed (see Figure S2 in Supporting Information).

Figure 3. MSMS spectra recorded during the CID of the 83n5+ ions of 6TAB. The precursor ion (P) is highlighted in gray The various intermediate populations with −1+, −2+, −3+, and −4+ charges are highlighted in red, blue, green, and orange, respectively, and the region for singly charged 6TAB monomers and dimers highlighted in light blue (m/d). −Xn represents the total number of detergents for the respective ion lost from the precursor 83n5+ ion cluster. The applied collision voltage is indicated. It is important to note that all of the peaks observed originate from a single population of detergent clusters with identical aggregation number and charge.

■

DISCUSSION One of the objectives of this study was to assess the utility of ESI-MS for the determination of detergent aggregation numbers. This was made possible by the use of IMMS which facilitated the full and unambiguous assignment of the mass spectra of a well-defined series of different detergent clusters. Using this approach we have shown that the well-known and predictable relationship between detergent alkyl chain length and micelle aggregation number in the condensed phase is not retained by ESI-MS. We considered that this phenomenon could be due to the different detergent concentrations used in this study. However, we found no change in the observed aggregation numbers of 12TAB gas-phase micelles between solution concentrations of 12.5 mM to 200 mM (Figure S3 in Supporting Information). Thus, these differences are not concentration-related. It is apparent, therefore, that while ESIMS is a sensitive probe for changes in detergent composition, the distributions of gas-micelle micelle ions obtained by this technique are not truly representative of detergent aggregation numbers in solution. We found that the charge-per-unit mass (zave) of the detergent clusters was significantly lower than the Rayleigh limit for droplet fission (zmax) or indeed that of protein complexes (zprot).41,42 The low zave of detergent clusters has been observed previously in gas-phase micelles of 16TAB, and this was reconciled by considering the reduced stability of these gas-phase micelles relative to proteins.21 This prompted us to describe the observed aggregation numbers of gas-phase micelles by their critical sizes (ncrit) rather than their aggregation numbers in solution using a model that would

Figure 4. MSMS spectra of the 83n5+ ions of 14TAB. The precursor ions are highlighted in gray and denoted by the letter P. The intermediate with −1 charge (−14TA:14TAB)+ is denoted red. Singly charged monomer and dimer ions (m/d) are highlighted in light blue. The collision voltage is indicated. The 6000 to 9000 m/z regions of the 60 and 80 V spectra have 5-fold and 10-fold magnification respectively.

the MSMS spectrum of these clusters is dominated by the precursor ion (P), the population of which then decreases with increasing collision energy up to 60 V. Interestingly, the 7164

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

Figure 5. A comparison of the observed mass distributions of different detergent clusters with the Rayleigh limit for droplet fission (zmax), the zave of proteins (zprot), and the critical detergent cluster sizes (ncrit) obtained with the liquid drop model. The horizontal bars in each plot are the observed masses of detergent clusters for each charge state for 6TAB, 8TAB, 10TAB, 12TAB, 14TAB, and 16TAB (a to f). The values of zmax and zprot were calculated using zmax = 0.0778 m0.5 and zprot = 0.0467m0.53 respectively (m = ion mass).41,42 ncrit is defined as in eq 4.

of [nCTAB + nCTA]1+. This is clearly shown in Figure 3 where at 10 V ions corresponding to the loss of a singly charged dimer dominate the MSMS spectrum. The loss of charged dimers was found to be the dominant fission pathway, irrespective of the detergent, cluster size, or applied collision energy. At higher collision energy, the dissociation of gas-phase micelles involves dual pathways of charge separation and evaporation, where the latter is characterized by the loss of [nCTAB]. This is evidenced best in the CID of gas-phase micelles formed from 6TAB where multiple peaks are observed within each charge state consistent with the loss of neutral detergent monomers from the precursor ion (see Figure 3). The extent of evaporation from these clusters is extremely high, and this would, therefore, further drive the process of charge separation due to the ensuing increase in Coulomb energy as the cluster size decreases. Thus, the CID of gas-phase micelles formed from 6TAB is complex and is characterized by the gradual disintegration of these clusters with increasing thermal agitation. In contrast with this, however, the 83n5+ gas-phase micelles of 14TAB populate very few intermediates during their CID (Figure 4). This also coincides with the observation that these ions populate singly charged monomers and dimers prior to the observation of any other intermediate species. This implies that these clusters dissociate readily without the formation of intermediate ions that are stable on the experimental time scale. We considered that this occurrence may be as a consequence of experimental discrimination factors with the 83n5+ clusters. However, the same overall trend was observed with various other clusters that differed in charge and aggregation number. The origins of this behavior are not known. Although, the structures and packing densities of clusters formed from this system do not vary with chain length (unpublished work). However, this behavior could be related to different fissibilities (proximities to ncrit) of the clusters across this detergent series, as suggested in Figure 5. In conclusion, this study has shown that the properties of gas-phase micelles are difficult to reconcile by their cognate properties in solution. A better understanding of the physical properties that govern the continued association of detergents

consider the Coulomb energy. A comparison of Figure 2, parts a and b, clearly shows that the experimental detergent clusters sizes correlate significantly better with ncrit than with their aggregation numbers in solution. A comparison of the observed distributions of the different gas-phase micelle cluster sizes with zmax and zprot is shown in Figure 5, where zmax = 0.0778m0.5 and zprot = 0.0467m0.53, respectively (m = ion mass). Here the low zave of the detergent clusters can be seen clearly with clusters of 40 kDa mass, for example, having charge states of between 6+ and 7+, as compared with proteins of the same mass where the predicted charge would be 13+. Thus, considerably more charge is available for transfer onto the detergent clusters during ionization than observed. The red line denoting ncrit suggests that the zave of the detergent clusters cannot approach that of proteins, however, as this would result in them populating unstable cluster sizes which would then undergo charge separation. Surprisingly, however, the zave of the detergent clusters is found to be relatively constant, regardless of alkyl chain length, with values of zave = 0.0062m0.665 and zave = 0.0066m0.645 for gas-phase micelles of 6TAB and 16TAB, respectively, taken from the central cluster within each charge state. This is surprising because at constant mass, the aggregation number of a 6TAB cluster will be significantly greater than its 16TAB counterpart and will therefore have potentially more ionizable sites. The consequences of this similarity in the zave are that gas-phase micelles formed from detergents with longer alkyl chains have zave values much lower than predicted by their critical size (compare Figure 5, parts a and f). This suggests that these clusters are not charged to their maximum value. Interestingly, however, the addition of supercharging agents had no effect of the zave of gas-phase micelles of 12TAB. We suggest, therefore, that the discrepancy between the zave and ncrit for gas-phase micelles formed from detergents with longer alkyl chains could also be due to an overestimation of the bulk cohesiveness of these clusters. Under conditions of low collision energy (10 V), gas-phase micelles can undergo dissociation by a Coulombically driven process that involves asymmetric charge separation and the loss 7165

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

Au(25)(SR)(18) Synthesis. J. Am. Chem. Soc. 2009, 131, 13604− 13605. (11) Jurchen, J. C.; Garcia, D. E.; Williams, E. R. Gas-phase dissociation pathways of multiply charged peptide clusters. J. Am. Soc. Mass Spectrom. 2003, 14, 1373−1386. (12) Echt, O.; Kreisle, D.; Recknagel, E.; Saenz, J. J.; Casero, R.; Soler, J. M. Dissociation channels of multiply charged van der Waals clusters. Phys. Rev. A 1988, 38, 3236−3248. (13) Jurchen, J. C.; Williams, E. R. Origin of asymmetric charge partitioning in the dissociation of gas-phase protein homodimers. J. Am. Chem. Soc. 2003, 125, 2817−2826. (14) Beardsley, R. L.; Jones, C. M.; Galhena, A. S.; Wysocki, V. H. Noncovalent protein tetramers and pentamers with “n” charges yield monomers with n/4 and n/5 charges. Anal. Chem. 2009, 81, 1347− 1356. (15) Boeri Erba, E.; Ruotolo, B. T.; Barsky, D.; Robinson, C. V. Ion mobility-mass spectrometry reveals the influence of subunit packing and charge on the dissociation of multiprotein complexes. Anal. Chem. 2010, 82, 9702−9710. (16) Nohara, D.; Ohkoshi, T.; Sakai, T. The possibility of the direct measurement of micelle weight by electrospray ionization mass spectrometry. Rapid Commun. Mass Spectrom. 1998, 12, 1933−1935. (17) Nohara, D.; Bitoh, M. Observation of micelle solution of decyltrimethylammonium bromide by electrospray ionization mass spectrometry. J. Mass Spectrom. 2000, 35, 1434−1437. (18) Nohara, D.; Kajiura, T.; Takeda, K. Determination of micelle mass by electrospray ionization mass spectrometry. J. Mass Spectrom. 2005, 40, 489−493. (19) Loo, R. R.; Dales, N.; Andrews, P. C. Surfactant effects on protein structure examined by electrospray ionization mass spectrometry. Protein Sci. 1994, 3, 1975−1983. (20) Siuzdak, G.; Bothner, B. Gas-Phase Micelles. Angew. Chem., Int. Ed. 1995, 34, 2053−2055. (21) Sharon, M.; Ilag, L. L.; Robinson, C. V. Evidence for micellar structure in the gas phase. J. Am. Chem. Soc. 2007, 129, 8740−8746. (22) Giorgi, G.; Ceraulo, L.; Liveri, V. T. Surfactant self-assembly in the gas phase: bis(2-ethylhexyl)sulfosuccinate-alkaline metal ion aggregates. J. Phys. Chem. B 2008, 112, 1376−1382. (23) Liu, J. B.; Fang, Y. G.; Bennett, A. Selective transport of amino acids into the gas phase: driving forces for amino acid solubilization in gas-phase reverse micelles. Phys. Chem. Chem. Phys. 2011, 13, 1466− 1478. (24) Friemann, R.; Larsson, D. S.; Wang, Y.; van der Spoel, D. Molecular dynamics simulations of a membrane protein−micelle complex in vacuo. J. Am. Chem. Soc. 2009, 131, 16606−16607. (25) Barrera, N. P.; Di Bartolo, N.; Booth, P. J.; Robinson, C. V. Micelles protect membrane complexes from solution to vacuum. Science 2008, 321, 243−246. (26) Lin, H. T.; Bavro, V. N.; Barrera, N. P.; Frankish, H. M.; Velamakanni, S.; van Veen, H. W.; Robinson, C. V.; Borges-Walmsley, M. I.; Walmsley, A. R. MacB ABC transporter is a dimer whose ATPase activity and macrolide-binding capacity are regulated by the membrane fusion protein MacA. J. Biol. Chem. 2009, 284, 1145−1154. (27) Velamakanni, S.; Lau, C. H.; Gutmann, D. A.; Venter, H.; Barrera, N. P.; Seeger, M. A.; Woebking, B.; Matak-Vinkovic, D.; Balakrishnan, L.; Yao, Y.; U, E. C.; Shilling, R. A.; Robinson, C. V.; Thorn, P.; van Veen, H. W. A multidrug ABC transporter with a taste for salt. PLoS One 2009, 4, e6137. (28) Barrera, N. P.; Isaacson, S. C.; Zhou, M.; Bavro, V. N.; Welch, A.; Schaedler, T. A.; Seeger, M. A.; Miguel, R. N.; Korkhov, V. M.; van Veen, H. W.; Venter, H.; Walmsley, A. R.; Tate, C. G.; Robinson, C. V. Mass spectrometry of membrane transporters reveals subunit stoichiometry and interactions. Nat. Methods 2009, 6, 585−587. (29) Zhou, M.; Morgner, N.; Barrera, N. P.; Politis, A.; Isaacson, S. C.; Matak-Vinkovic, D.; Murata, T.; Bernal, R. A.; Stock, D.; Robinson, C. V. Mass Spectrometry of Intact V-Type ATPases Reveals Bound Lipids and the Effects of Nucleotide Binding. Science 2011, 334, 380− 385.

in a solvent free environment could lead to better selection criteria for detergents best suited for the delivery of membrane protein complexes into the gas phase. Furthermore, by focusing on detergent clusters, we have gained valuable clues into the mechanism by which detergents protect membrane protein complexes during CID. One plausible protective mechanism that has emerged is evaporative cooling of the ions whereby the ejection of detergent neutrals during thermal agitation protects the protein structure. The exciting focus of future experiments will be to investigate these mechanisms further by characterizing the protein structure during the intermediate phases of gas-phase detergent release. Thus, this study represents an important initial step into an enhanced understanding of the gas-phase analysis of membrane protein complexes.

■

ASSOCIATED CONTENT

S Supporting Information *

Dissociation profiles of 83n5+ gas-phase micelles of 8TAB, 10TAB, and 12TAB and 53n4+, 31n3+, and 41n3+ clusters of 6TAB and 16TAB. Concentration dependence of observed cluster sizes of 12TAB gas-phase micelles. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 01865 275473. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Drs. Kevin Pagel, Mathew Bush, Todd Mize, and Profs. Mark Thachuk and Nelson Barrera for useful discussions. A.J.B. is a Waters Research Fellow and C.V.R. is a Royal Society Research Professor and is funded by an ERC advanced fellowship.

■

REFERENCES

(1) Saunders, W. A. Metal-Cluster Fission and the Liquid-Drop Model. Phys. Rev. A 1992, 46, 7028−7041. (2) Naher, U.; Bjornholm, S.; Frauendorf, S.; Garcias, F.; Guet, C. Fission of metal clusters. Phys. Rep. 1997, 285, 245−320. (3) Last, I.; Levy, Y.; Jortner, J. Beyond the Rayleigh instability limit for multicharged finite systems: from fission to Coulomb explosion. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 9107−9112. (4) Tanford, C. Thermodynamics of Micelle Formation - Prediction of Micelle Size and Size Distribution. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 1811−1815. (5) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. Theory of selfassembly of hydrocarbon amphilphiles into micelles and bilayers. J. Chem. Soc., Faraday Trans. 2 1976, 72, 1525−1568. (6) Lipfert, J.; Columbus, L.; Chu, V. B.; Lesley, S. A.; Doniach, S. Size and shape of detergent micelles determined by small-angle x-ray scattering. J. Phys. Chem. B 2007, 111, 12427−12438. (7) Brechignac, C.; Busch, H.; Cahuzac, P.; Leygnier, J. Dissociation Pathways and Binding-Energies of Lithium Clusters from Evaporation Experiments. J. Chem. Phys. 1994, 101, 6992−7002. (8) Saunders, W. A. Fission and Liquid-Drop Behavior of Charged Gold Clusters. Phys. Rev. Lett. 1990, 64, 3046−3049. (9) Kruckeberg, S.; Schweikhard, L.; Dietrich, G.; Lutzenkirchen, K.; Walther, C.; Ziegler, J. Decay pathway determination of even-size dicationic silver clusters: Ag-16(2+) and Ag-18(2+) revisited by preprecursor selection and sequential decay. Chem. Phys. 2000, 262, 105− 113. (10) Dass, A.; Dharmaratne, A. C.; Krick, T. Nanocluster Size Evolution Studied by Mass Spectrometry in Room Temperature 7166

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167

Langmuir

Article

(30) Benesch, J. L.; Aquilina, J. A.; Ruotolo, B. T.; Sobott, F.; Robinson, C. V. Tandem mass spectrometry reveals the quaternary organization of macromolecular assemblies. Chem. Biol. 2006, 13, 597−605. (31) Ruotolo, B. T.; Hyung, S. J.; Robinson, P. M.; Giles, K.; Bateman, R. H.; Robinson, C. V. Ion mobility-mass spectrometry reveals long-lived, unfolded intermediates in the dissociation of protein complexes. Angew. Chem., Int. Ed. 2007, 46, 8001−8004. (32) Pagel, K.; Hyung, S. J.; Ruotolo, B. T.; Robinson, C. V. Alternate dissociation pathways identified in charge-reduced protein complex ions. Anal. Chem. 2010, 82, 5363−5372. (33) Wang, S. C.; Politis, A.; Di Bartolo, N.; Bavro, V. N.; Tucker, S. J.; Booth, P. J.; Barrera, N. P.; Robinson, C. V. Ion mobility mass spectrometry of two tetrameric membrane protein complexes reveals compact structures and differences in stability and packing. J. Am. Chem. Soc. 2010, 132, 15468−15470. (34) Parent, D. C.; Bowers, M. T. Temperature-Dependence of Ion Mobilities - Experiment and Theory. Chem. Phys. 1981, 60, 257−275. (35) Dugourd, P.; Hudgins, R. R.; Clemmer, D. E.; Jarrold, M. F. High-resolution ion mobility measurements. Rev. Sci. Instrum. 1997, 68, 1122−1129. (36) Wyttenbach, T.; Bowers, M. T. Intermolecular interactions in biomolecular systems examined by mass spectrometry. Annu. Rev. Phys. Chem. 2007, 58, 511−533. (37) Ruotolo, B. T.; Benesch, J. L.; Sandercock, A. M.; Hyung, S. J.; Robinson, C. V. Ion mobility-mass spectrometry analysis of large protein complexes. Nat. Protoc. 2008, 3, 1139−1152. (38) Lord Rayleigh, L. On the equilibrium of liquid conducting masses charged with electricity. Philos. Mag. 1882, 14, 184−186. (39) Bohr, N.; Wheeler, J. A. The mechanism of nuclear fission. Phys. Rev. 1939, 56, 426−450. (40) Hernandez, H.; Robinson, C. V. Determining the stoichiometry and interactions of macromolecular assemblies from mass spectrometry. Nat. Protoc. 2007, 2, 715−726. (41) de la Mora, J. F. Electrospray ionization of large multiply charged species proceeds via Dole’s charged residue mechanism. Anal. Chim. Acta 2000, 406, 93−104. (42) Stengel, F.; Baldwin, A. J.; Painter, A. J.; Jaya, N.; Basha, E.; Kay, L. E.; Vierling, E.; Robinson, C. V.; Benesch, J. L. P. Quaternary dynamics and plasticity underlie small heat shock protein chaperone function. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 2007−2012. (43) Evans, D. F.; Allen, M.; Ninham, B. W.; Fouda, A. Critical Micelle Concentrations for Alkyltrimethylammonium Bromides in Water from 25-Degrees-C to 160-Degrees-C. J. Solution Chem. 1984, 13, 87−101.

7167

dx.doi.org/10.1021/la3002866 | Langmuir 2012, 28, 7160−7167