Addressing a Common Misconception: Ammonium Acetate as Neutral pH “Buffer” for Native Electrospray Mass Spectrometry
- 2.5k Downloads
Native ESI-MS involves the transfer of intact proteins and biomolecular complexes from solution into the gas phase. One potential pitfall is the occurrence of pH-induced changes that can affect the analyte while it is still surrounded by solvent. Most native ESI-MS studies employ neutral aqueous ammonium acetate solutions. It is a widely perpetuated misconception that ammonium acetate buffers the analyte solution at neutral pH. By definition, a buffer consists of a weak acid and its conjugate weak base. The buffering range covers the weak acid pKa ± 1 pH unit. NH4 + and CH3-COO− are not a conjugate acid/base pair, which means that they do not constitute a buffer at pH 7. Dissolution of ammonium acetate salt in water results in pH 7, but this pH is highly labile. Ammonium acetate does provide buffering around pH 4.75 (the pKa of acetic acid) and around pH 9.25 (the pKa of ammonium). This implies that neutral ammonium acetate solutions electrosprayed in positive ion mode will likely undergo acidification down to pH 4.75 ± 1 in the ESI plume. Ammonium acetate nonetheless remains a useful additive for native ESI-MS. It is a volatile electrolyte that can mimic the solvation properties experienced by proteins under physiological conditions. Also, a drop from pH 7 to around pH 4.75 is less dramatic than the acidification that would take place in pure water. It is hoped that the habit of referring to pH 7 solutions as ammonium acetate “buffer” will disappear from the literature. Ammonium acetate “solution” should be used instead.
KeywordsHenderson-Hasselbalch equation Dissociation equilibrium Acid/base chemistry Droplet evaporation Noncovalent complexes in the gas phase
Native electrospray ionization (ESI) mass spectrometry (MS) is a versatile tool for probing the structure and composition of proteins, protein–ligand complexes, all the way to MDa biomolecular assemblies [1, 2, 3, 4, 5, 6, 7, 8, 9]. This area has experienced significant growth in recent years, fuelled by the development of instruments that are capable of performing both m/z analyses and ion mobility measurements [10, 11, 12, 13, 14, 15]. For native ESI-MS, the analytes are dissolved in a nondenaturing solvent, usually aqueous ammonium acetate (NH4 + CH3-COO−) at neutral pH . Regular or nanoESI are used to spray the solution directly into the ion sampling interface. This article takes a critical look at the properties of ammonium acetate in a native ESI-MS context. In particular, we scrutinize the widely repeated claim that ammonium acetate acts as a “buffer” at neutral pH. For putting this issue into context, we will initially discuss aspects related to protein stability in solution and in the gas phase, as well as factors that can induce pH changes during the ESI process.
Proteins in Aqueous Solution
Proteins have evolved in an aqueous environment. Many of their biological functions involve noncovalent contacts with other proteins, nucleic acids, or membranes. Biomolecular structures and interactions are stabilized by intricate networks of hydrogen bonds, hydrophobic contacts, and electrostatic linkages (salt bridges, charge–dipole interactions, van der Waals forces, etc.) . The aqueous environment is an integral part of this interaction network. Unfavorable contacts between water and nonpolar sites cause proteins to form a hydrophobic core [18, 19]. The high dielectric constant of water (κe ≈ 80) results in screening effects, such that electrostatic interactions are much weaker than in the vacuum, where κe = 1 [20, 21]. This water-mediated screening is further amplified by dissolved electrolytes , keeping in mind that physiological ionic strength is around 150 mM .
The acid/base milieu is a critical factor for protein stability in solution. Intracellular pH values around 7 are maintained by various buffer systems, including CO2/H2CO3/HCO3 −, and by transmembrane ion transporters [24, 25]. A stable pH is essential because proteins possess titratable sites that alter their charge depending on pH [26, 27]. pH-induced changes of charge patterns can cause deactivation, unfolding, and precipitation [26, 27]. Often these effects are related to the release of metal cofactors, triggered by side chain protonation. Histidines are particularly prone to protonation due to their pKa values around 7, but Asp and Glu (pKa ~ 4) can also be involved . In an effort to mimic the intracellular environment, classic biochemical experiments are generally conducted in neutral aqueous solutions that contain background electrolytes, as well as a buffer system to stabilize pH (e.g., 150 mM NaCl with 25 mM sodium phosphate buffer at pH 7) [27, 29].
Transferring Native-Like Proteins from Solution into the Gas Phase
The development of ESI in the late 1980s for the first time allowed the transfer of proteins from solution into the gas phase, making them amenable for analyses by MS as intact [M + zH] z+ ions . Soon afterwards it became apparent that even protein–ligand and protein–protein complexes can be studied using this approach [31, 32, 33]. The preservation of noncovalent complexes sparked the idea that electrosprayed proteins can retain much of their solution structure. This concept was initially met with scepticism . However, it is now well established that gaseous protein ions can get kinetically trapped in solution-like structures, provided that the experimental conditions are properly optimized [7, 35, 36, 37].
In native ESI-MS, it is essential to minimize structural perturbations of the analyte during its journey from bulk solution into highly charged ESI droplets, and ultimately into the gas phase [16, 38]. One parameter that has to be carefully controlled is the extent of collisional heating in the ion sampling interface. Conditions that are too gentle cause inadequate desolvation, whereas excess activation disrupts noncovalent contacts and causes gas-phase unfolding [39, 40]. In addition to these gas-phase events, it is necessary to minimize structural perturbations that might take place while the protein is still surrounded by solvent. Examples of such events include thermal denaturation due to inadvertent heating [41, 42], as well as surface adsorption within the ESI capillary .
pH Changes During the ESI Process
Charge-balancing redox processes within the ESI capillary affect the solution chemistry . Positive ion mode (which is typically used in native ESI-MS) causes acidification due to oxidation reactions such as 2 H2O → 4 H+ + 4 e− + O2 . This effect was strikingly demonstrated by Van Berkel et al.  in experiments on nanoESI capillaries containing indicator dyes. For unbuffered solutions, those emitters showed pH changes from the near-neutral range to pH 3, under some conditions even down to pH 1.4. Similarly, the operation of an ESI source in negative ion mode increases pH because of reduction processes (e.g., 2 H2O + 2 e− → 2 OH− + H2) . The magnitude of these pH changes depends on the length of time during which the analyte solution is subject to solvent electrolysis . The high surface to volume ratio at the outlet of nanoESI emitters may cause additional pH modulations because of the effects of silanol groups (and/or other titratable sites) on the acid/base chemistry of the analyte solution .
Spectroscopic experiments revealed that pH changes can also take place within ESI droplets [53, 54]. Initial droplets generated at the Taylor cone have radii r around 150 to 1500 nm for nanoESI and ESI, respectively [16, 55]. Evaporation and fission events produce successively smaller droplets, down to r values of a few nm. Throughout this process the droplets stay close to the Rayleigh limit, where the number of excess charges is z R = 8π(ε 0 γr 3)1/2/e with ε 0 = vacuum permittivity, e = elementary charge, and γ = 0.072 N m−1 for water at 298 K [16, 56]. According to the charged residue model (CRM), analyte ions in native ESI are released as the final nanodroplets evaporate to dryness [9, 16, 57, 58].
Overall, the profiles of Figure 2a demonstrate that proteins within the ESI plume can experience significant pH-induced stress, with pH < 1 during the final stages of solvent evaporation (Figure 2b, c). The predicted pH is lowest when the droplets reach their minimum size, i.e., when the final layers of solvent evaporate and the protein is released. Small proteins will thus experience a more acidic pH than large proteins, illustrated in Figure 1a for ubiquitin (r ≈ 1 nm, final pH ≈ −0.2) and for pyruvate kinase (r ≈ 5 nm, final pH ≈1).
Admittedly, the pH estimates obtained for the magenta (H+) profile of Figure 2a represent a worst-case scenario. The actual evolution of droplet pH depends on many factors, including the possible presence of species that mitigate acidification. Also, the net droplet charge will usually be composed of different charge carriers, not just protons. When designing native ESI-MS experiments, one should nonetheless not dismiss the possibility that ESI droplets can undergo significant acidification. These effects are in addition to pH changes that may take place elsewhere in the ESI source (Figure 1).
Can Inadvertent pH Changes Affect the Outcome of ESI-MS Experiments?
It is undisputed that native ESI-MS is a powerful approach for probing protein structures and interactions, evidenced by the successful detection of numerous biomolecular systems in the gas phase [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. Nonetheless, a significant number of studies have reported discrepancies between solution and gas phase behavior. This issue seems to be particularly prevalent for complexes containing metal ions [62, 63, 64, 65, 66]. The underlying reasons are not always clear, but several authors have attributed effects of this kind to ESI-induced pH artifacts, consistent with the processes outlined in Figure 1 [62, 63, 65, 67, 68].
Whether or not ESI-induced pH artifacts affect the outcome of MS experiments depends not only on the thermodynamic stability of the analyte but also on kinetic aspects . Some binding equilibria respond to solvent changes quite slowly, such that pH changes during ESI are masked by the short (μs - ms) lifetimes of ESI droplets [70, 71]. Other systems react sufficiently fast to allow major pH-induced changes during the droplet stage [72, 73].
ESI charge state distributions are a widely used tool for monitoring protein conformational changes that take place in bulk solution. These studies rely on the fact that native proteins produce low charge states, whereas unfolded chains generate ions that are more extensively protonated [38, 74, 75, 76]. The addition of organic acids is the most common approach for inducing protein unfolding in bulk solution. Denaturation profiles obtained in this way are often in reasonable agreement with solution-phase spectroscopic data [74, 75, 76], seemingly indicating that ESI-induced pH changes are not a major issue. However, pH artifacts may be difficult to detect because unfolding midpoints measured by ESI-MS can be skewed by differences in the detection efficiency of folded versus unfolded proteins . In addition, the logarithmic nature of the pH axis, along with the fact that many proteins unfold below pH 3, tend to be somewhat deceiving. This can be illustrated by considering a bulk solution that contains 1 mM H+, corresponding to pH 3. Doubling the H+ concentration to 2 mM (e.g., as the result of water hydrolysis during ESI ) will lower the pH to 2.7. Such a relatively small pH shift may easily be overlooked. ESI-induced pH artifacts are more apparent for proteins that start to unfold around pH 5–6. In such cases, it is not uncommon to observe charge state distributions that are inconsistent with bulk solution data, implying that ESI-mediated acidification has affected the protein structure [77, 78]. In summary, there is clear evidence that inadvertent pH changes can take place during ESI, and that this phenomenon can influence the outcome of ESI-MS experiments in undesired ways.
For choosing an appropriate buffer, one has to select a system with an acid pK a that is close to the desired pH value. For physiological samples pKa ≈ 7 is desirable. Phosphate buffer is widely used for bulk solution studies because H2PO4 − has a pKa of 7.2. Its conjugate base is HPO4 2−. From Equation 2 it can be determined that a neutral 10 mM sodium phosphate buffer contains 3.9 mM Na2HPO4 and 6.1 mM NaH2PO4. When 1 mM H+ is added to this buffer (corresponding to water oxidation under typical nanoESI conditions ), the buffer will respond with a slight acidification to pH 6.8.
Next, we consider a buffer system that is not well suited for the neutral range. Acetic acid and acetate are a conjugate weak acid/base pair with pKa = 4.75. A 10 mM pH 7 acetate “buffer” contains 9.943 mM acetate and 0.057 mM acetic acid. The addition of 1 mM H+ decreases the pH to 5.8. This large drop reflects the fact that the initial pH is not within the buffering range of pKa ± 1 pH unit. For comparison, a drop down to pH 3 would take place in pure water.
Unfortunately, phosphate solutions and most other buffer systems commonly used in biochemistry laboratories are nonvolatile and therefore incompatible with ESI-MS . Major interferences arise from the fact that cationic and anionic components (such as Na+, H2PO4 −, and HPO4 2− form nonspecific adducts with analytes during ESI [16, 81]. The situation is different for ammonium and acetate, both of which are volatile. Like other electrolytes, they form nonspecific adducts with the analyte during the final stages of ESI. However, protonation of acetate in positive ion mode generates acetic acid that dissociates in the declustering region of the mass spectrometer. Similarly, NH4 + adducts release NH3 and leave behind a proton, thereby favoring the formation of clean [M + zH] z+ ions [16, 80]. This volatility makes ammonium acetate an electrospray-friendly additive that is widely used in native ESI-MS, often at concentrations around 10 mM [4, 12, 15, 16, 32, 82, 83, 84, 85, 86].
ESI-MS can readily deal with nonvolatile additives as long as on-line liquid chromatography or other sample clean-up steps are incorporated into the workflow. However, the current discussion focuses on native ESI-MS, which is much more unforgiving because analyte solution is sprayed directly into the ion sampling interface of the mass spectrometer.
Is Ammonium Acetate a Buffer at Neutral pH?
Numerous papers have repeated the claim that neutral ammonium acetate solutions are “buffered.” This habit is commonplace in ESI-MS publications, although it predates the inception of electrospray by several decades . The term “ammonium acetate buffer” implies a sense of security, suggesting that biomolecular analytes will be protected from pH-induced degradation throughout the ESI process (Figure 1). Sadly, the designation of ammonium acetate as a neutral buffer is misleading.
Ammonium is a weak acid, and acetate is a weak base. It is a quirk of nature that the acidity of NH4 + (pKa 9.25) exactly balances the basicity of acetate (pKb 9.25), such that dissolution of ammonium acetate salt in pure water produces a neutral pH solution [87, 88]. Perhaps it is this peculiar feature that has caused many practitioners to misjudge the properties of ammonium acetate.
Ammonium and acetate are not a conjugate acid/base pair. The two species are not related via a single (HA ↔ H+ + A− or HA+ ↔ H+ + A) equilibrium that would be required for the validity of Equation 2. A proper buffer has its pKa within ± 1 unit of the desired pH . Neither ammonium/ammonia (pKa 9.25) nor acetic acid/acetate (pKa 4.75) meet this criterion for pH 7. By definition, therefore, ammonium acetate is not a buffer at pH 7. This simple fact has been largely ignored in the ESI-MS literature, except for rare cautionary remarks that are easily overlooked .
Despite this negative verdict, the acid/base chemistry of ammonium acetate can be put to good use in native ESI-MS. Substoichiometric acidification converts a certain percentage of acetate to acetic acid. These conditions produce an acetate buffer that stabilizes the pH at 4.75 ± 1. The properties of acetate buffers have been discussed in the previous section, where we noted that addition of 1 mM H+ to a 10 mM neutral pH solution will decrease the pH to 5.8. This is less dramatic than the drop to pH 3 that would occur in pure water. Thus, ammonium acetate does have a pH buffering effect in the acidic range, but this effect is not felt until the solution is within roughly ± 1 pH unit of the acetic acid pKa (4.75). There are only a few instances in the ESI-MS literature where ammonium acetate was correctly referred to as a pH 5 buffer [33, 72].
Alkalinization of ammonium acetate solution generates NH3 via deprotonation of NH4 +, thereby generating a buffer around the pKa of ammonium (9.25). Thus, ammonium acetate also buffers pH in the basic range, but this effect is insignificant until the solution is within ±1 pH unit of the ammonium pKa.
Should Practitioners Continue to Use Ammonium Acetate?
Ammonium acetate has a number of attractive features that make it a useful additive in native ESI-MS, as long as practitioners are aware of the limitations. Although ammonium acetate is not a buffer at pH 7, it still exerts a moderating effect on pH alterations encountered during ESI. In positive ion mode there are several factors that tend to acidify the analyte solution (Figure 1). Ammonium acetate mitigates these effects by buffering the solution in the range of 4.75 ± 1. Hence, an initially neutral ammonium acetate solution should be expected to have a pH of 4.75 ± 1 at the point of analyte release in positive ion mode. This acidification is severe enough to cause the protonation of His (pKa ~7), but it will leave Asp and Glu carboxylates (pKa ~ 4) largely unaffected. Analogously, in negative ion ESI the electrolytic production of OH− tends to increase pH. Ammonium acetate will cap this alkalinization at pH 9.25 ± 1. Under these conditions, Lys (pKa ~ 11) and Arg (pKa ~12) will retain their canonical positive charge.
At any rate, the use of ammonium acetate solutions in native ESI-MS is preferable over pure water. Pure water may result in extreme acidification during the final stages of droplet shrinkage (Figure 2a, magenta). Acetate buffering will lessen this acidification by ensuring that the pH does not drop below 4.75 ± 1. This behavior is consistent with experimental observations, e.g., apomyoglobin in pure water produces high ESI charge states that indicate acid-induced unfolding. In ammonium acetate solution the spectrum shifts to lower charge states, implying the prevalence of compact structures (Figure 3 in ref. ).
Although many native ESI-MS studies have employed ammonium acetate at concentrations around 10 mM [4, 12, 15, 16, 32, 82, 83, 84, 85, 86], higher concentrations can often be tolerated [1, 5, 7, 8, 11, 40, 86]. These higher concentrations may help mitigate pH changes during the ESI process. As noted, the addition of 1 mM H+ (e.g., from water oxidation ) will lower the pH of a 10 mM ammonium acetate solution from pH 7 to 5.8. For a 100 mM solution the pH will only drop to 6.5 under otherwise identical conditions. However, the increased electrolyte concentration will enhance the rate of electrochemical processes within the ESI source, such that the H+ production is expected to rise when more ammonium acetate is added . Also, the volatile nature of ammonium acetate makes it difficult to predict in how far elevated concentrations would help suppress pH changes within ESI droplets (Figure 2a).
Ammonium acetate possesses additional attributes that contribute to its popularity in ESI-MS. Biomolecular structures and interactions in solution depend on the presence of dissolved salts [86, 90, 91, 92]. Classic electrolytes (NaCl, KCl, etc.) are nonvolatile and cause interferences [16, 81]. Owing to its volatile nature, ammonium acetate does not degrade the spectral quality, offering mass spectrometrists the opportunity to work in an ionic milieu that is acceptable to classically trained biochemists. As an added bonus, ammonium acetate can even suppress the formation of undesired salt adducts by competing with nonvolatile electrolytes for binding sites at the protein surface .
Other ESI-Friendly Solvent Additives and Buffers
The number of ESI-compatible (volatile) additives that buffer in the neutral range is limited. This may be the main reasons why ammonium acetate has become the de facto standard for native ESI-MS. Ammonium bicarbonate represents an interesting option, considering that HCO3 − has a pKa of 6.4, which is a reasonable match for pH 7 solutions . Unfortunately, ammonium bicarbonate tends to induce protein unfolding during ESI. This phenomenon has been attributed to protein interactions with bubbles formed by CO2 outgassing at the droplet stage , possibly in combination with other factors . Because of this destabilizing effect, it would be prudent to avoid ammonium bicarbonate in native ESI-MS . Ammonium formate is another volatile salt, but it is even less suitable for neutral pH buffering than ammonium acetate due to the pKa of formic acid/formate (3.75) . Various alkyl-ammonium acetate salts have been used for mechanistic investigations [81, 94], but they also do not provide buffering at neutral pH based on the pKa values involved.
The term “buffer” may also be used in different contexts. Redox buffers stabilize the interfacial potential within the ESI source. Examples of such substances include ascorbic acid and easily oxidizable metals [95, 96]. Redox buffers can reduce pH changes in the ESI source by suppressing water electrolysis. However, we will not discuss this aspect in greater detail because our focus is on conventional acid/base buffering. Readers interested in ESI-related redox chemistry can refer to recent reviews [49, 50].
Undergraduate chemistry students are taught that a buffer provides pH stabilization in a range of the weak acid pKa ± 1 pH unit . For pH 7, ammonium acetate does not match this criterion. Both of the relevant pKa values (4.75 for acetic acid and 9.25 for ammonia) are 2.25 units away from the desired neutral pH. It is true that dissolution of ammonium acetate salt in pure water produces a neutral pH solution. However, this solution is not buffered, and it will undergo significant pH changes upon addition of small amounts of H+ or OH−.
Aqueous ammonium acetate solution at pH 7 remains a useful solvent for native ESI-MS as long as practitioners are aware of its limitations. Protein stability is often reduced in the absence of dissolved salts, and ammonium acetate can serve as a stabilizing background electrolyte. Positive ion mode ESI tends to result in acidification of the analyte solution. In the presence of ammonium acetate the pH may drop to values as low as 4.75 ± 1 in the final ESI droplets, reflecting the pKa of acetate buffer. Conversely, experiments in negative ion mode likely result in final pH values around 9.25 ± 1 as a result of NH4 +/NH3 buffering.
It is hoped that this Critical Insight article will dispel some of the myths surrounding the properties of ammonium acetate. Practitioners should avoid the term “ammonium acetate buffer” when referring to neutral conditions, as this wording choice is highly misleading. The term “ammonium acetate buffer” is only justified if the pH is around 4.75 or 9.25. In all other cases (specifically around pH 7) “ammonium acetate solution” should be used. This change in terminology will help avoid false expectations regarding the acid/base environment of analytes under ESI conditions. Future work is expected to yield improved approaches for dealing with physiologically relevant buffer systems in ESI-MS. Interesting recent developments in this area include the use of nanoESI emitters with μm outlet diameter , as well as fused-droplet ESI .
Funding for this work was provided by the Natural Sciences and Engineering Research Council of Canada (Discovery Grant 217080–2013). Victor Yin and Yiming Xiao are acknowledged for critical reading of the manuscript.
- 5.Landreh, M., Marklund, E.G., Uzdavinys, P., Degiacomi, M.T., Coincon, M., Gault, J., Gupta, K., Liko, I., Benesch, J.L.P., Drew, D., Robinson, C.V.: Integrating mass spectrometry with MD simulations reveals the role of lipids in Na+/H+ antiporters. Nat. Commun. 8, (2017)Google Scholar
- 17.Fersht, A.R.: Structure and mechanism in protein science. W.H. Freeman and Co., New York (1999)Google Scholar
- 23.Creighton, T.E.: Proteins. W.H. Freeman & Co, New York (1993)Google Scholar
- 26.Nielsen, J.E.: Analyzing enzymatic ph activity profiles and protein titration curves using structure-based pKa calculations and titration curve fitting. In: Johnson, M.L., Brand, L. (eds.) Methods Enzymol, vol. 454, pp. 233–258. Elsevier Academic Press Inc, San Diego (2009)Google Scholar
- 44.Debaene, F., Wagner-Rousset, E., Colas, O., Ayoub, D., Corvaïa, N., Van Dorsselaer, A., Beck, A.H., Cianférani, S.: Time resolved native ion-mobility mass spectrometry to monitor dynamics of IgG4 Fab arm exchange and “bispecific” monoclonal antibody formation. Anal. Chem. 85, 9785–9792 (2013)CrossRefGoogle Scholar
- 50.Van Berkel, G.J., Kertesz, V.: Using the electrochemistry of the electrospray ion source. Anal. Chem. 79, 5511–5520 (2007)Google Scholar
- 55.Kebarle, P., Tang, L.: From ions in solution to ions in the gas phase: the mechanism of electrospray mass spectrometry. Anal. Chem. 65, 972A–986A (1993)Google Scholar
- 66.Jecklin, M.C., Touboul, D., Bovet, C., Wortmann, A., Zenobi, R.: Which electrospray-based ionization method best reflects protein–ligand interactions found in solution? A comparison of ESI, nanoESI, and ESSI for the determination of dissociation constants with mass spectrometry. J. Am. Soc. Mass Spectrom. 19, 332–343 (2008)CrossRefGoogle Scholar
- 68.Liang, Y., Du, F., Sanglier, S., Zhou, B.R., Xia, Y., Van Dorsselaer, A., Maechling, C., Kilhoffer, M.C., Haiech, J.: Unfolding of rabbit muscle creatine kinase induced by acid – a study using electrospray ionization mass spectrometry, isothermal titration calorimetry, and fluorescence spectroscopy. J. Biol. Chem. 278, 30098–30105 (2003)CrossRefGoogle Scholar
- 79.Masterton, W.L., Hurley, C.N.: Chemistry, principles and reactions,6th ed.; Brooks/Cole Cengage Learning: (2009)Google Scholar
- 88.Lide, D.R.: CRC Handbook of chemistry and physics, 82nd edn. CRC Press, Boca Raton, London, New York, Washington (2001)Google Scholar