Current perspectives on supercharging reagents in electrospray ionization mass spectrometry.

In electrospray ionization mass spectrometry (ESI-MS), analytes are introduced into the mass spectrometer in typically aqueous-organic solvent mixtures, including pH modifiers. One mechanism for improving the signal intensity and simultaneously increasing the generation of higher charge-state ions is the inclusion of small amounts (approx. <0.5% v/v mobile phase solution) of charge-inducing or supercharging reagents, such as m-nitrobenzyl alcohol, o-nitrobenzyl alcohol, m-nitrobenzonitrile, m-(trifluoromethyl)-benzyl alcohol and sulfolane. We explore the direct and indirect (colligative properties) that have been proposed as responsible for their modes of action during ESI. Of the many theorized mechanisms of ESI, we re-visit the three most popular and highlight how they are impacted by supercharging observations on small ions to large molecules including proteins. We then provide a comprehensive list of 34 supercharging reagents that have been demonstrated in ESI experiments. We include an additional 19 potential candidate isomers as supercharging reagents and comment on their broad physico-chemical properties. It is becoming increasingly obvious that advances in technology and improved ion source design, analyzers e.g. the use of ion mobility, ion trap, circular dichroism (CD) spectroscopy, together with computer modeling are increasing the knowledge base and, together with the untested isomers and yet-to-be unearthed ones, offer opportunities for further research and application in other areas of polymer research. This journal is © The Royal Society of Chemistry.

Introduction

Three main models seek to explain the mechanisms of electrospray ionization (ESI), on which early work began with John Fenn.[1] The generation of ions in the dynamic ESI environment which is under an electric field and at elevated temperatures but at ambient pressure is explored within the concept of these three popular models. Fundamentals such as solution chemistry and gas-phase interactions are summarised to explain the ESI processes. The role of supercharging reagents during ESI is discussed and key findings explored. Some of these findings are the direct and more complex interactions between these reagents and macromolecules, especially proteins in the ESI environment. The influence of colligative properties of both ESI solvents and supercharging reagents, including surface tension, polarizability, dipole moments, and acidity/basicity are highlighted. Konermann et al. offer a detailed and rigorous review, which includes experimentation and computational studies on the mechanisms of ESI, dwelling mainly on the interactions between supercharging reagents and proteins.[2]

In this brief review, we provide a comprehensive list of 34 supercharging reagents that have been used in studies and include an additional 19 potential candidate isomers commenting on their broad physico-chemical properties. These suggested isomers have the potential to further enhance our understanding of supercharging during ESI-MS. This report may be viewed as a first compendium or a ‘one-stop-shop’ of supercharging reagents for fellow ESI mass spectrometrists to peruse.

Supercharging during ESI

One of the unique advantages during ESI-MS is the formation of multiply charged ions, which enables the observation of molecules of high molecular weights at relatively low m/z values.[3,4] Thus, for instance, biological macromolecules including proteins (>100 KDa), nucleic acids, industrial polymers of high molecular weights can be analyzed in a mass spectrometer with a triple quadrupole (QqQ), several types of ion trap (IT) analysers (including orbitraps), and combinations, such as the quadrupole time-of-flight (QToF), quadrupole-ion trap (QIT), and with ion-mobility mass spectrometry.

In ESI, one of the main approaches for increasing the generation and detection of higher charge state in peptide ions (i.e. from [M + nH] to [M + (n + 1)H])+ and simultaneously improving the signal intensity (ESI response) is the inclusion of very small amounts (approx. <5% v/v of mobile phase solution) charge-inducing compounds or supercharging reagents. Supercharging reagents include glycerol,[5]m-nitrobenzyl alcohol (mNBA; 3-nitro(phenyl) methanol),[4,6,7]o-nitrobenzyl alcohol (2-nitro(phenyl)methanol) 8 and sulfolane.[8,9] Thus far, the supercharging reagents used in experiments are all small molecules (mol. wt. <180 Da (Table 1).

A list of supercharging reagents and some of their physico-chemical propertiesa

	Common name/[IUPAC name]α	Molecular structureα	Ave. mass,α Da	Boiling ptα °C at 760 mmHg	Vapour press.αβ mmHg at 25 °C	Surface tensionαβ mN m−1 at 25 °C	Densityαβ (specific gravity) 25 °C	Acidity pKa 25 °C	Dipole momentδ μ(D)/μD*	Ref.
3- to 4-membered & planar molecules
1	Dimethyl sulfoxide (DMSO) [(Methylsulfinyl)methane]		78.133	189	0.60	43.54	1.101 (ref. 65)	35.1	3.96/4.44	3 and 25
2	Ethylene glycol [1,2-ethandiol]		62.068	196–198	0.06	47.99	1.113	14.22	2.747/—	3
3	Glycerol [1,2,3-propanetriol]		92.094	182	1.68 × 10−4	72.6	1.26 (ref. 65)	14.4	2.68/—	5 and 8
4	Formamide		45.041	210	6.10 × 10−2 (ref. 66)	58.35 (ref. 67)	1.13	23.5 (in DMSO)	—	3
5	2-Methoxyethanol		76.094	124–125	6.2 (ref. 68)	30.84 (ref. 69)	0.97 (ref. 65 and 70)	14.8	—	3
6ε	Methoxypropanol [1-methoxypropan-2-ol]		90.122	120	10.9	2.64 × 10−3	0.926	14.49 ± 0.20 (ref. 70 and 71)	—
7	Crotononitrile[(2E)-2-butenenitrile]		67.089	120–121	31.95	—	—	—	4.3/—	15
Heterocyclic 4- & 5-membered ring structures, including organosulfur cpd (sulfones)
8	Dimethyl sulfone [methylsulfonylmethane]		94.133	238	—	—	1.45	31	4.5/—	15
9	N,N,N′N′-Tetraethylsulfamide (TES) [N-(diethylsulfamoyl)-N-ethylethanamine]		208.32	125	—	—	1.283	—	—	16
10	3-Chlorothiete-1,1-dioxide		140.59	328.9	—	—	1.64	—	—/3.38	9
11ε	3-Chloro-2H-thiete 1,1-dioxide		138.573	328.9 ± 41.0	3.2 ± 0.3	—	1.6 ± 0.1	—	—
12	2-Thiophenone [2H-thiophen-5-one]		100.139	197.4	0.4 ± 0.3	38.7 ± 3.0	1.24			49
13ε	4-Butyrothiolactone [thiolan-2-one]		102.16	76.5429	—	—	1.18	Poorly soluble in water	—
14	Sulfolane [2,3,4,5-tetrahydrothiophene-1,1-dioxide]		120.170	285	6.2 × 10−3 (27.6 °C)	35.5 (ref. 72)	1.26 (ref. 65)	12.9	4.35/5.68	7, 8, 52 and 53
15	Sulfolene [2,5-dihydrothiophene 1,1-dioxide]		118.154	64–65.5 °C mpt (ref. 73)	—	41.0	1.3 (ref. 68)	—	—/5.69	47
16	1,3-Propanesultone [1,2-oxathiolane 2,2-dioxide]		122.143	315.87	—	—	1.392	—	—	15
17	1,4-Butanesultone [1,2-oxathiane 2,2-dioxide]		136.169	304.047	—	—	1.335	—	—	15
18	3-Methyl-2-oxazolidone (MOZ) [3-methyl-1,3-oxazolidin-2-one]		101.1	88	—	—	1.17	—	—	16
Heterocyclic 4- & 5-membered ring structures (including the heterocyclic acetals)
19	Glycerol carbonate [4-hydroxymethyl-1,3-dioxolan-2-one]		118.088	353.9 ± 15	0.0 ± 1.8	41.1	1.375	—	—	48
20	Propylene carbonate [4-methyl-1,3-dioxol-2-one]		102.089	240	0.045 (ref. 66)	45	1.2 (ref. 69)	—	4.9	48 and 49
21	Ethylene carbonate [1,3-dioxolan-2-one]		88.062	244–245	—	54 (30 °C)	1.32	—	4.9/—	15
22	Butylene carbonate [4-ethyl-1,3-dioxolan-2-one]		116.115	281.18	—	—	1.141	—	—	15
23	Vinylethylene carbonate [4-vinyl-1,3-dioxolan-2-one]		114.099	238.72	—	—	1.188	—	—	15
Benzene-ring containing molecules
24	Nitrobenzene		123.109	210–211	4.2	0.245	1.201	1.204 slightly soluble in water	4.2/—	15
25	Benzyl alcohol [(hydroxymethyl)benzene]		108.138	205	0.094 (ref. 66)	39.0 (ref. 70)	1.0 (ref. 65)	15.4	1.71/1.79	8
26	m-Chlorophenol [3-chlorophenol]		128.556	214	0.125	44.7	1.3 (ref. 68)	9.12	1.03 ± 1.08	4
27ε	o-Chlorophenol [2-chlorophenol]		128.556	175–176	2.53 (ref. 66)	40.50	1.26 (ref. 65)	8.56	—
28ε	p-Chlorophenol [4-chlorophenol]		128.556	217–220	1 (ref. 74)	8.7 × 10−2 mm Hg	1.306 (ref. 65)	9.41	2.10/—
29	2-Nitrochlorobenzene [1-chloro-2-nitrobenzene]		157.555	245–246	—	48.4 ± 3.0	1.348	0.6	4.6/—	15
30ε	3-Nitrochlorobenzene [1-chloro-3-nitrobenzene]		157.555	235–236	0.097 (ref. 66)	4.37 × 10−2	1.3	—	—
31ε	4-Nitrochlorobenzene [1-chloro-4-nitrobenzene]		157.555	241.66	2.19 × 10−2 (ref. 66)	3.71 × 10−2	1.520	n/a	—
32	o-Nitrobenzyl alcohol (oNBA) [2-nitro(phenyl)methanol]		153.135	270 (solid at RT)	0.0 ± 0.5	—	1.3	—	—	8
33	m-Nitrobenzyl alcohol (mNBA) [3-nitro(phenyl) methanol]		153.135	349.8	0.0 ± 0.5	50 ± 5	1.29	—	—	4, 6, 7, 8, 48, 49 and 75
34	p-Nitrobenzyl alcohol (pNBA) [4-nitro(phenyl)methanol]		153.135	185 (solid at RT)	Negligible	—	—	7.15	—	8
35	m-Nitroacetophenone [1-(3-nitrophenyl)ethanone]		165.146	202 (solid at RT)	3.86 × 10−5	—	1.4	—	—	8 and 27
36	m-Nitrobenzonitrile [3-benzonitronitrile]		148.119	165	0.0 ± 0.5	—	1.3 ± 0.1	—	—	6, 8 and 27
37ε	o-Nitrobenzonitrile [2-benzonitronitrile]		148.119	321.78 165.0 °C (16.0 mmHg)	Insoluble in water	—	—	—	—
38ε	p-Nitrobenzonitrile [4-benzonitronitrile]		148.119	—	—	—	—	—	—
39	m-nitrophenyl ethanol [1-(3-nitrophenyl)ethanol]		167.162	281.3 ± 23.0	0.0 ± 0.6	—	1.3 ± 0.1	—	—	6 and 8
40ε	o-Nitrophenyl ethanol [1-(2-nitrophenyl)ethanol)		167.162	319.0 ± 17.0	0.0 ± 0.6	—	1.3 ± 0.1	—	—
41ε	p-Nitrophenyl ethanol (1-(4-nitrophenyl)ethanol]		167.162	290.68	0.0 ± 0.7 g	—	1.3 ± 0.1	—	—
42	o-Nitroanisole [1-methoxy-2-nitrobenzene]		153.135	272–273	—	48	1.254	—	5.0/—	15
43ε	m-Nitroanisole [1-methoxy-3-nitrobenzene]		153.135	256.3 ± 13.0	0.0 ± 0.5	—	1.2 ± 0.1	—	—
44	p-Nitroanisole [1-methoxy-4-nitrobenzene]		153.135	260	—	—	1.233	—	5.3/—	15
45	3-Nitrophenethyl alcohol [2-(3-nitrophenyl)ethanol]		167.162	341.7–349.0	—	—	—	—	—	8
46ε	2-Nitrophenethyl alcohol [2-(2-nitrophenyl)ethanol]		167.162	267	—	—	1.19	—	—
47ε	4-Nitrophenethyl alcohol [2-(4-nitrophenyl)ethanol]		167.162	337	—	—	—	—	—
48	m-(Trifluoromethyl)-benzyl alcohol [[3-(trifluoromethyl)phenylmethanol]		176.136	257–261	0.1 ± 0.4	—	1.3 ± 0.1	—	—	6
49ε	o-(Trifluoromethyl)-benzyl alcohol [2-(trifluoromethyl)phenylmethanol]		176.136	214–262	—	—	1.3 ± 0.1	—	—
50ε	p-(Trifluoromethyl)-benzyl alcohol [4-(trifluoromethyl)benzyl alcohol]		176.136	250	0.1 ± 0.4	—	1.28 (ref. 74)	—	—
51ε	o-Nitrobenzoic acid [2-nitrobenzoic acid]		167.12	295.67	0.0 ± 0.8 (ref. 75)	66.4 ± 3.0	1.575	2.17	4.07/—
52ε	m-Nitrobenzoic acid [3-nitrobenzoinc acid]		167.12	340.7 ± 25.0	3.71 × 10−5 (ref. 75)	66.4 ± 3.0	1.5 ± 0.1	3.47	4.03/—
53ε	p-Nitrobenzoic acid [4-nitrobenzoic acid]		167.12	It sublimes	—	—	1.58	3.41	4.05/—
Typical solvents used in ESI
	Common name/IUPAC name	Molecular structure	Ave. mass, Da	Boiling pt °C at 760 mmHg	Vap. press. mmHg at 25 °C (ref. 80)	Surface tension mN m−1 at 25 °C	Density (specific gravity)	Acididty pKa 25 °C	Dipole moment μ(D)/μD*	66 and 69
1	Water		18.02	100	17.44	72.8	1.00	14	1.85/2.16
2	Methanol		32.04	64.7	127	22.07	0.792	15.5	1.70/
3	Ethanol		46.07	78.2	59.3	21.97	0.789	15.9	1.69/
4	Acetonitrile		41.05	81.3–82.1	88.8	29.04	0.786	25	3.92/
5	Ethyl acetate		88.106	77.1	93.2	24	0.902	25	1.78/
6	Acetone		58.08	56.05	231	23.7	0.785	19.16	2.91/

Key: α chemspider: http://www.chemspider.com/, γ pKaε candidate isomers proposed supercharging reagents, β pubchem: https://pubchem.ncbi.nlm.nih.gov/, δ dipole moment μ(D)/μD*: determined by experiment/calculated*. We acknowledge that some values in Table 1 are incomplete.

Proposed modes of action of supercharging reagents: colligative properties

In ESI, supercharging reagents assist in surmounting deficiencies such as ion suppression and improve the linear dynamic range, that is, in the latter case, they can decrease the detection limit across a wide sample concentration linear range. Some colligative properties have been indicated as responsible for the supercharging phenomena. These are boiling points, surface tension, polarizability, dipole moments, and acidity/basicity.

Boiling points and surface tension

One of the fundamental properties of these reagents which promote higher ESI charging is that they all have much higher boiling points than both water and the organic solvents in the mobile phase mixture. They are therefore less volatile than the solvents. Although opposing arguments exist, the reason for their action has been attributed to two phenomena:

(1) Their ability to raise the solution surface tension.[10] Due to the low volatility of these compounds compared with that of common mobile-phase compositions (e.g. water/methanol or water/acetonitrile), their concentration increases as evaporation of the more volatile components from the droplets occurs. This enrichment of the supercharging reagents raises the surface tension which results in a requirement for a higher degree of surface charging to reach the Rayleigh limit for a spherical droplet. This, in turn, increases the boiling point of the supercharging reagent-enriched environment[10] However, Samalikova and Grandori[11] reported that their results were not easily reconciled by considering only charge availability of the precursor droplets at the Rayleigh limit as determined by surface tension.

(2) Protein denaturation: Williams and co-workers argued that protein unfolding induced by chemical and or thermal denaturation appears to be the primary origin of the enhanced charging observed in the protein complexes investigated.[7] In their experiments, the arrival time distributions obtained from traveling wave ion mobility spectrometry showed that the higher charge state ions that are formed with m-NBA and sulfolane are significantly more unfolded than lower charge state ions. Sterling et al.[12] argued that droplet heating, owing to the high boiling point of m-NBA, resulted in thermal denaturation. However, unlike Sterling et al.[12] Hogan et al.[13] reported that supercharging reagents do not cause structural protein modifications in solution and that the modest mobility decrease observed could be partly attributed to trapping of the sulfolane within the protein ions.

Polarizability and dipole moments

The addition of the supercharging reagent sulfolane on cytochrome introduced another dimension to the debate, in that, while supercharging was observed in the positive ionization mode, no change in the charge state distribution was observed in the negative mode.[9] This, therefore, eliminates polarity-independent factors such as conformational changes or surface tension effects. The authors also reported the formation of adducts between cytochrome and sulfolane thus supporting the report of Hogan et al.[13]

Further, the study by Douglass and Venter demonstrated that supercharging was shown to increase with increasing dipole moment for the supercharging reagents sulfolane and sulfolene.[9] Finally, if an increase in surface tension is the cause of supercharging, then it should occur in both the positive and negative modes since the effect of surface tension on the Rayleigh limit is independent of the polarity of the charge. Like many other identified supercharging reagents, sulfolane is highly polar with a dipole moment of 4.35 D, much higher than that of water (1.85 D) or methanol (1.70 D) (Table 1).

The interaction of supercharging reagents with charged basic sites would depend on the molecular structure, such as the type and location of functional groups. Thus, they exhibit low solution-phase basicities and relatively low gas-phase basicities.[8] For example, the most effective supercharging reagents have one or more carbonyl, sulfonyl, or nitro groups present in their structure (Table 1), which may be important for intermolecular interactions between the supercharging reagent and the charged basic site. Lomeli et al. demonstrated that supercharging occurs not only when liquid m-NBA is present, but also with solid o- or p-isomers. This provides some evidence against increased surface tension causing supercharging.[8]

Direct interaction

Adduct formation

Examination of the mass spectra generated from the addition of these reagents show adduct formation on the higher charges. This suggests that there is a direct interaction between these reagents and fragments of unfolding polypeptides and proteins.[8,9] Circular dichroism measurements while titrating myoglobin with 6 M guanidinium hydrochloride in the presence of 0%, 2.5%, 5.0%, or 7.5% sulfolane established that sulfolane destabilized the protein by ∼1.5 kcal mol−1 (moles L−1 of GuHCl)−1, which Sterling and co-workers[12] argued was evidence that chemical denaturation caused the charge increases observed when sulfolane is added to native-like solutions. However, they also found that without the added GuHCl, the helicity of myoglobin was lower, even at the highest sulfolane concentrations tested.

Chemical denaturation

Sterling's circular dichroism experiment indicates a chemical denaturation activity of sulfolane.[12] As supercharging reagents have low vapor pressures (higher surface tension), therefore it has been suggested that aqueous droplets are preferentially enriched in these reagents as evaporation occurs in the ESI environment. Sterling et al. reasoned that less evaporative cooling will occur after the droplets are substantially enriched in the low volatility supercharging reagent (i.e. the organic solvent is evaporated first), and the droplet temperature should be higher compared with when these reagents are not present, in the more aqueous environment.[12] Furthermore, Sterling et al. rationalized that protein unfolding induced by chemical and/or thermal denaturation in the electrospray droplet appeared to be the primary origin of the enhanced charging observed for noncovalent protein complexes formed from aqueous solutions that contain these supercharging reagents, although other factors most certainly influence the extent of charging. However, some studies are not in agreement with this. For example, Yao et al.[14] reported that supercharging of lysozyme by sulfolane was not related to protein unfolding during the ESI process.

A report evaluating the potential of low volatility of supercharging reagents demonstrated that there was not a strong correlation between the extent of analyte (protein) charging with either surface tension, dipole moment, or dielectric constant of the additives.[15] That is, there was no single physical property of an additive that significantly determined the extent of protein charging when formed from a solution mixture. What is clear, however, was that most of the supercharging reagents have higher boiling points, higher surface tension values, high dielectric constants, and higher dipole moments than the ESI solvents. Thus, a combination of these physicochemical properties together with other factors including adduct formation between analyte and supercharger, basicity at the solution, and gaseous phases, may all influence charging, although the extent of charging may depend on the experimental conditions.

A comprehensive list of 53 molecules of which 34 have been shown to have supercharging ability in experiments and 19 potential candidate isomers, together with some of their physico-chemical properties are given in Table 1. The molecules are grouped conveniently in terms of molecular structure and functional group given rise to four categories:

(1) Acyclic 3- to 4-membered and/or planar structures.

(2) Heterocyclic 4- to 5-membered molecules including the sulfones and S-containing ones. Dimethyl sulfone and N,N,N′N′-tetraethylsulfamide (TES) are acyclic molecules but are included in this category because they contain the sulfone functional group.

(3) Heterocyclic 4- to 5-membered molecules including the heterocyclic acetals.

(4) Molecules containing the benzene structure with at least two functional group attachments which may have opposite charges e.g. m-nitrobenzyl alcohol.

Recently, two new supercharging reagents N,N,N′,N′-tetraethylsulfamide (TES), and 3-methyl-2-oxazolidone (MOZ) were shown to counter the suppression activity of trifluoroacetic acid (TFA).[16] The study also demonstrated that the two molecules increased charge states of peptides and proteins and improved separation efficiency during reverse-phase LC-MS determinations.

We suggest that the following 19 additives (denoted by ε and in italics, Table 1) are potential supercharging reagents and could be evaluated:

(1) Based on the fact that some of their isomers have successfully been demonstrated; methoxypropanol, 3-chloro-2H-thiete 1,1-dioxide, 4-butyrothiolactone, o-chlorophenol, p-chlorophenol, 3-nitrochlorobenzene, 4-nitrochlorobenzene, o-nitrophenylethanol, p-nitrophenyl ethanol, m-nitroanisole, o-(trifluoromethyl)-benzyl alcohol, and p-(trifluoromethyl)-benzyl alcohol, and,

(2) Molecules having a benzene ring structure and with two or more attached functional groups having opposing charges (‘dual polarity’ isomers); e.g. the isomers o-, m-, and p- of nitrobenzoic acid, nitrobenzonitrile and nitrophenethyl alcohol.

ESI mobile phase solvents

Six solvents commonly used during ESI-MS are also included for comparison. The list in Table 1 includes the following physico-chemical properties: molecular structure, molecular weight (Da), boiling point (or melting point, °C), vapor pressure (mm Hg), surface tension (mN m−1), acidity (pKa)/basicity (pKb), dipole moment (D), and references of key publications in which the superchargers were employed. The sources of the properties are also indicated. We concede that some of the properties of the reagents are either not determined or could not be found in the literature.

The ESI mechanism

During ESI-MS operation and for samples extracted from a complex and an electrolyte-rich biological matrice, such as saliva, serum, cerebrospinal fluid, or microbial extract, resolution and detector signal response can be hindered. A clean-up procedure, for example, solid-phase extraction (SPE) or desalting steps is necessary to decrease the concentration of or remove these usually confounding electrolytes, detergents, or other compounds. In (positive) ESI-MS, analytes including peptides are introduced into the mass spectrometer in typically aqueous-organic solvent mixtures, together with other additives including pH modifiers like formic acid (≤0.3% v/v). Following the application of an electric charge at the capillary tip, there follows the simultaneous processes of solution-phase ion generation, solvent evaporation, a plume of charged ions in residual solvent, charge transfer, and/or concentration, and finally, gas phase ion generation (Fig. 1). These processes typically occur under an electric field and in a heated (approx. 90° to 120 °C) environment in the ion source compartment of the mass spectrometer. And finally, the generated highly charged gas-phase ions are then differentially drawn into the first stage of the mass spectrometer under negative pressure. The uncharged gaseous solutes, molecules, are also removed under negative pressure into a separate exhaust stream.

Fig. 1

A simple illustration of the ESI environment. For simplicity, droplets are presented with positive charges. According to the model presented by Loo et al.[27] opposing charges may also be present in the gas phase ion. Under an electric field, elevated temperature but at ambient pressure, the sprayed analyte solution undergoes the simultaneous processes of the droplet splitting, generation of highly charged droplets, repeated splitting and shrinkage of the droplets, solvent evaporation, and removal under a directed gas (typically N2) stream. These progeny droplets themselves undergo subsequent and repeated evaporation and splitting, and this process is repeated until only gas-phase ions are generated.[76] The ions are then drawn under negative pressure into the mass spectrometer.

Observations show that many parameters e.g. solvent composition,[3,4,17,18] analyte composition and concentration,[19-21] pH, flow rate,[18,21] denaturing solutions,[22,23] and non-denaturing solutions,[24-26] solution- & gas-phase basicity, solution-phase conformation,[26,27] instrument settings: source voltage,[28,29] sprayer orifice diameter,[30] gas pressures,[31] and, ion source type (laser ESI-LEMS vrs conventional ESI)[32] affect charge state distributions (CSD) especially, in large analytes such as proteins. Further, sub-ambient pressure ESI source combined with nano-flow significantly improves ion yield and sensitivity[33]

Several models have been put forward to explain the mechanisms of action of the processes of generating gas-phase ions from the liquid mobile phase. Among these, three are popular: the charge residue model (CRM),[34,35] the ion evaporation model (IEM)[36-40] and the third postulate, chain ejection model (CEM)[41,42] which was recently described for unfolded proteins.[2,43,44] These models are summarized in Table 2.

The ESI mechanism: gas-phase ion generation

	The three popular models that seek to explain the ESI mechanism
	Charge residue model (CRM)	Ion evaporation model (IEM)	Chain ejection model (CEM)
Background	The CRM was proposed based on the ability of the SCRs to raise the surface tension of the mobile phase solvent, a requirement for a higher degree of surface charging to reach the Rayleigh limit which, results in a coulombic fission event	Following the repeated splitting and shrinkage of droplets, the size of droplets reach a certain critical radius (10−6 cm = 10−8 m), the field strength at the surface of the droplet becomes large enough to assist the field desorption of solvated ions	Neither the CRM nor IEM mechanisms explain quantitatively the extent of macromolecular multiple charging, especially in unfolded proteins
Key points in the mechanism	Millions of small, highly charged progeny droplets containing the analyte leave the parent droplet with a disproportionately large fraction of the surface charge	Analyte ions desorb directly from charged nanodroplets, driven by the large electric field at the droplet surface	The inclusion of SCRs has opened up new ways to modulate protein charge states, thereby challenging the existing ESI models
	Progeny droplets containing single analytes are charged through a charge transfer process between the charge carriers on the surface of the droplet and the analyte molecule upon final droplet evaporation	The ion evaporation process then becomes operative for a highly electrified cloud of droplets at low solute concentrations	ESI of unfolded proteins yields [M + (z + 1)H]z+ ions that are much more highly charged than their folded counterparts
		That is, analyte ions desorb directly from charged nanodroplets, driven by the large electric field at the droplet surface	Solvent properties such as dipole moment and protein–SCR adduct formation have been shown to play prominent roles in supercharging in proteins
Analytes	Larger ions (e.g. from folded proteins under non-denaturing conditions) are formed by CRM	It has been demonstrated that small ions (i.e. Na+), generated from small molecules, are liberated into the gas phase through the ion evaporation mechanism	This proposed model accounts for the protein ESI behavior under such non-native conditions and recently has been proposed to apply to unfolded proteins
Illustration
References	34, 77 and 78	10, 36–40 and 79	2, 41–44, 52 and 53

The use of nano-ESI, conventionally a non-pneumatic operation with very low mobile flow rates (∼<1 μL min−1) and much small samples sizes which then provides improved desolvation, greater salt tolerance, and a higher ion yield[29,33] has added another dimension to the debate although the generated final ions are believed to be the same for both conventional and nano-ESI.[45]

Hogan et al. proposed that both the IEM and CRM are in play, in which macromolecules are charged residues but, in negative ion mode, carry less charge because highly-charged droplets field-emit anions. Why positively charged droplets would not also field-emit small ions has not yet been explained.[46,47] Excellent discussions are also given.[9,27,48,49]

The impact of supercharging reagents on ESI

After extensive investigations of ESI of macromolecules, neither model (CRM or IEM) adequately and quantitatively explain the extent of macromolecular multiple charging.[50,51] Quite recently, with increasing interest in using ESI-MS as a technology for top-down proteomics and for studying protein interactions, the expanding application of ion mobility mass spectrometry (IM-MS), and the use of supercharging reagents, most of the reported supercharging research has been on proteins in the native state, that is, with solution pH around neutral and the low voltage conditions to preserve protein structure. Thus, the role of additive properties such as dipole moment and adduct formation (between analyte and supercharger) have been shown to play prominent roles in supercharging in proteins. Therefore, a third model, the chain ejection model (CEM) has been proposed[42] where the ESI of unfolded proteins yields M + (z + 1)H] ions that are much more highly charged than their folded counterparts. This model, it has been suggested, accounts for the protein ESI behavior under such non-native conditions and has been proposed to apply for unfolded proteins.[2,43,44,52,53]

Proteins that are unfolded in solution produce higher charge states during ESI than their natively folded counterparts. Building on the work by Douglass and Venter,[9] who showed experimentally the formation of adducts between sulfolane and the most highly charged protein ions, Peters et al. recently, reported that the CEM involves the partitioning of mobile H+ (e.g. from formic acid) between the droplet and the departing protein.[44] Their results indicate that the supercharging of unfolded proteins is caused by residual sulfolane that stabilizes protonated sites on the protruding chains. This, therefore, promotes H+ retention on the protein. Their report suggested that charge stabilization on the sites of projecting chains is due to charge–dipole interactions which are mediated by the large dipole moment and the low volatility of sulfolane.

Denaturing ESI follows the CEM, where protein ions are gradually expelled from the droplet surface. Proton (H+) equilibration between the droplets and the protruding chains culminates in highly charged gaseous proteins.[2] The presence of internal disulfide (S–S) bonds on the extent of supercharging was determined in three proteins each containing multiple internal disulfide bonds; bovine serum albumin, β-lactoglobulin, and lysozyme.[43] Reduction of the disulfide bonds led to a marked increase in charge state following the addition of sulfolane without significantly altering folding in solution. This evidence supports a supercharging mechanism in which these proteins unfold before or during evaporation of the electrospray droplet and ionization would therefore, occur by the CEM.

Ogorzalek Loo et al. proposed a three-regime view of ESI, i.e. with solution, intermediate, and gas-phase regimes in the ESI environment.[27] The intermediate phase was introduced to explain the charge transfer and other accommodations that analytes especially protein ions undergo when exiting the bulk solution, this concept rationalizes observations that ion intensities vary little over pH 3–11, despite huge changes in solution compositions.[17,54,55] This kinetic model is unique in treating the evaporating droplets as charge-polarized and uses their decompositions and the variation in solution-to-gas phase properties to explain charge enrichment in ESI.

The 3-phase ESI environment could, in principle, be imagined for the IEM, CRM, and other models, particularly as distinctions between them blur in attempts to explain non-conforming observations.[29,56] Thus far, the IEM and the CRM appear to represent the foundations of any discussion related to the mechanism of ESI. Molecular dynamics (MD) simulations[52,57] suggest that small ions such as Na+ are ejected from the surface of an aqueous ESI droplet (IEM), while folded proteins in native ESI could be released by water evaporation to dryness (CRM).[43,52,53,58]

Other factors that increase the charge state during ESI are the diameter of the emitter opening: proteins (native and denatured) charge state distributions (CSDs) undergo a small, but reproducible shift to higher charge when delivered by nano-spray versus standard electrospray[59,60] and source voltage.[26,61] Ogorzalek Loo, et al.[27] noted that this relationship between the extent of charging and initial droplet diameter could not be observed for ions released as charge residues, thus, these observations argue against CRM being the primary source of native-like ions.

Improvements in ESI source design including sub-ambient pressure with nano-flow,[33,62,63] chemically-etched emitters with sheath gas capillaries and nano-flow[64] are reported to enhance signal intensity, increase desolvated ion transmission to the mass spectrometer proper and thus increasing overall sensitivity. The use of these improvements together with supercharging reagents should further enhance our understanding of the ESI mechanism and how it is impacted by the supercharging reagents.

Summary

The following are the main highlights on supercharging during ESI

(1) Konermann et al.[2] indicated that the ion evaporation model (IEM) and the charged residue model (CRM) represent the foundations underpinning the mechanism of ESI. Simulations of molecular dynamics (MD) supported by classical solution chemistry demonstrate that small ions such as Na+ or NH4+ are ejected from the surface of aqueous ESI droplets (IEM), while folded proteins in the native ESI are released by water evaporation to dryness (CRM).

(2) The application of ESI-MS in protein analysis and the inclusion of supercharging reagents (SCR) has been the main driver in the exploration of the mechanism of the ESI process and the dynamics governing proteins in the native and unfolded state.

(3) The Konermann team[2,44,52,53] and Donor et al.[43] indicated that the CEM largely explains the CSD in proteins as it accounts for the dipole-moment of solvents, charge distribution from the folded to the unfolded state in proteins, and also the presence of intramolecular di-sulfide bonds within proteins.

(4) SCRs have expanded and increased our knowledge of the ESI environment and protein unfolding. These molecules have boiling points and higher surface tension than the solvent mixtures used in ESI,[10] although studies by Samalikova and Grandori[11] reported a diminished role of surface tension. The vast majority of supercharging molecules have densities greater than that of water and other ESI solvents (Table 1). William's and Loo's teams observed that SCRs also have low solution-phase basicities (Brønsted bases weaker than H2O) and relatively low gas-phase basicities.[26-28] They also have higher dipole moments (except for acetonitrile (3.92 D)). In solution, SCRs are ionized, and highly polarized.

(5) Loo's team[8] suggested that for dual-polarity superchargers, the molecule should both be a weak Brønsted base and a weak Brønsted acid.

(6) Venter's team[9] demonstrated supercharging in cytochrome in the positive ionization mode, however, no change in the CSD was observed in the negative mode. This, therefore, diminishes the role of polarity-independent factors such as conformational changes or surface tension effects as key vehicles for supercharging.

(7) Venter's team[9] also demonstrated that when a SCR is added in concentration about equal to or greater than the proteins concentration in solution, during ESI, there is adduct formation between the proteins and SCR.

(8) Using SCRs such as nitrobenzyl alcohol, where there are two functional groups (hydroxyl and nitro), supercharging increases in the order of the para, meta, and ortho isomers. The same principle could perhaps be applied to the isomers of nitrochlorobenzene, nitroanisole, nitrobenzoic acid, nitrobenzonitrile, and nitrophenethyl alcohol.

(9) Improvements in ion source design to include operating at sub-ambient pressure combined with nano-flow increases desolvation of charged species and elevates the ion transmission rate into the analyser section.[33,62-64]

It should be pointed out that all these different supercharging attributes have been proposed to matter, but not all are of them are relevant. Experimental conditions would increase the prominence of one or several of them over the others depending on the aims of the researchers.

Conclusions

In summary, we presented a brief review of supercharging and the attributed factors which influence supercharging during ESI. We highlighted the three most popular models, IEM, CRM and CEM that seek to explain the ESI process. The use of supercharging reagents have unraveled and expanded on the ESI mechanisms. Advances in technology including improved source design, the applications of IM-MS, nano-ESI, detectors (e.g. ion traps), circular dichroism spectroscopy, and computer software to explore molecular dynamics coupled with the use of supercharging reagents have expanded the knowledge base of the ESI mechanism, especially as applied to proteins. The list of 19 untested candidate isomers and, many more to be unearthed, offers opportunities for further study. The use of extreme supercharging reagents should be very useful for maximizing MS and MS/MS performance. A combination of these tools offers avenues for further research and applications in other areas of polymer research.

Author contributions

DAA and BVN developed the original concept. All authors contributed to the writing of the original draft, and all authors contributed to the writing of and approved the final draft.

Conflicts of interest

There are no conflicts to declare.