1 Introduction

Almost since the discovery of multiple charging of protein molecules by electrospray ionization [1], mass spectrometrists have been trying to understand and manipulate the charge state distributions that one may obtain [2]. The shift in charge state distributions was shown to be dramatic when proteins passed from their native folded states to denatured extended conformations as solution pH was lowered [3]. However, the effect of pH on charge state distributions was much more subtle when small peptides that were not susceptible to “unfolding” were subjected to extreme pH variations [4]. Other solution additives such as nitrogen bases were shown to influence charge state distributions by creating a competition between basic sites bearing excess protons on an analyte protein and the small nitrogen base adducted to this charged site [5, 6]. For example, when introduced as a formate or acetate salt, the ammonium cation (NH4 +) will act largely as a protonating agent in the presence of highly basic amino acids. In addition to the above solution-phase considerations, in the early 1990s, gas-phase charge manipulation was achieved by reacting dimethylamine with multiply protonated cytochrome c [7], or 1,6-diaminohexane with multiply protonated myoglobin in an ion trap mass spectrometer [8]. These gas-phase studies resulted in a reduction in the overall charge of protein species as protons were transferred to the introduced bases.

In the beginning of the 21st century, meta-nitrobenzyl alcohol (m-NBA), previously used as a matrix to improve desorption/ionization processes in fast atom bombardment mass spectrometry [9–12], was proposed as a supercharging reagent in positive ion electrospray mass spectrometry (ESI-MS) by Williams and co-workers [13–16]. Through these studies, they showed not only the ability of m-NBA to enhance both the highest observed (zhigh) and average (zavg) charge states of multiply protonated proteins [13–16] and peptides [13], but also its ability to favor higher charge states of cationized synthetic organic polymers (e.g., PEG 1500, [14, 15] in ESI positive ion mode). They reported a limited influence of m-NBA on oxidized insulin chain A under negative ESI conditions [14]. Access to higher charge states allows analysis of larger biological systems when the m/z range of the mass spectrometer is limited; the achievable mass resolution may also increase for mass analyzers that exhibit decreasing resolving power at higher m/z values (e.g., Fourier transform ion cyclotron resonance. Electron capture efficiency may also be augmented at higher charge states [17–19].

From its introduction, many attempts have been made to better understand how the chemical characteristics of m-NBA lead to supercharging. Among the properties considered are surface tension [13–16, 20–22], vapor pressure (or boiling point) [14–16, 20–26], gas-phase basicity [14, 16, 20–25, 27–29], etc. According to the work of Kebarle et al. [6, 30, 31], these properties can be divided into two groups corresponding to the progressive steps of evaporation/desolvation in ESI. Vapor pressure and surface tension are important physico-chemical properties starting from the macroscopic early phases of the charged droplet’s lifetime. On the other hand, solvent-free (or sparsely solvated) charged molecular species appear only at the very end of the charged droplet’s lifetime, so Gibbs free energy change of protonation/deprotonation reactions [i.e., GB for positive ion production and ∆G°(acid) for negative ion production] intervene directly only at this stage of desolvation in the ESI process. Because the vapor pressure of m-NBA (bp = 177 °C at 3 Torr [32]) is lower than that of water (bp = 100 °C at 766 Torr [32]), a noticeable influence of m-NBA on ESI charge states is observed in aqueous solutions even at very low reagent concentrations [14–16, 20–28, 33]. This is because during the course of water molecule evaporation, the concentration of supercharging reagent increases until it becomes the most abundant solvent species present into the final droplets [14]. The high concentration of reactive aromatic molecules remaining after water evaporation could induce conformational changes of studied macromolecules, e.g. proteins [13–16, 20, 25] and protein noncovalent complexes [21–24, 26–28, 33]. Lower vapor pressure also implies a slower droplet cool down that could favor conformational changes of biomolecules due to their higher internal energy [22, 24].

During the last few years, protein possible conformational changes have been studied using denaturation conditions induced by solvent [29, 34, 35] or temperature variations [22, 24], but also by employing a reversed strategy (i.e., by chemically “locking” the protein tertiary structure with introduced covalent linkages [28]. As expected, the presence of m-NBA did not modify the existing correlation trend for biomolecular species (i.e., the more unfolded the protein is, the more charges it can carry [24, 27–29]. But it has not yet been clarified whether (1) the presence of additional charged sites causes increased Coulombic repulsion leading to protein unfolding or (2) protein unfolding induced by the presence of m-NBA reveals new accessible protonable sites, which lead to the observed charge state enhancement. Even if the GB value of m-NBA in positive ion mode has been implicated as a major thermochemical property responsible for these trends [23, 29, 33], conformational changes of the analyte (that may or may not occur in response to the presence of m-NBA) hinder the ability to unambiguously assign the underlying reasons for the success of m-NBA as a supercharging reagent.

In the current study, we examine the influence of m-NBA reagent in negative ESI ion mode on oligonucleotide charge state distributions [36]. The several phosphate groups present on the oligonucleotide backbone and, thus, the acidic character of single-stranded DNA (ssDNA) explains why these compounds have mostly been studied in the negative ion mode [37–43]. Positive ion studies of DNA are also possible [43–49] owing to the substantial GB values of nucleobases that are readily protonated in solution with counter-ions (e.g., deprotonated solvent molecules, nearby. Ammonium acetate, commonly used as buffer for ssDNA samples, does not only control the pH of the solution (e.g., preventing further duplex production), it also interacts with phosphate groups that are fully deprotonated in solution (pKa < 1) and, thus, induces a competition between NH4 + and alkali metals for the role of counter-ion. After complete evaporation of the solvent, and during the desolvation step in the microscopic state, intermolecular proton transfer(s) could occur between oligonucleotide anions and NH4 + cations leading to fewer deprotonated sites on the oligonucleotides [43]. Since ammonia (NH3) is not basic enough to form stable salt bridges with phosphate groups [50], such proton transfers are most likely to occur during the last stages of ESI desolvation and are guided by the respective GB of NH3 (819 kJ.mol–1, [51]) and the apparent Gibbs free energy change (∆G°app[acid]) of the deprotonation reaction of multiply deprotonated oligonucleotide species ([DNA–nH]n–). An excess of ammonium acetate can shift the charge state distribution of [DNA–nH]n– toward lower n values by favoring departures of NH3 molecules (leaving behind the proton) and resulting in a decrease in the number of negative charges located on the ssDNA backbone [52]; zavg of oligonucleotide anions are strongly dependent upon the ∆G°app(acid) of [DNA–nH]n– and, thus, upon oligonucleotide sequence, especially the density (i.e., number and consecutivity) of phosphor-thymidine units [53] (the most acidic nucleotide [50, 54–56]). Madsen and Brodbelt [57] have observed the charge enhancement of duplexes in negative ESI-MS induced by the presence of m-NBA, but the enhancement was accompanied by in-source dissociations into the corresponding ssDNA. Since no predefined conformations of single-stranded oligonucleotides have been established as being predominant for short sequences [58–60], the presence of m-NBA would not be expected to induce conformational changes that could potentially alter the number of apparent deprotonable sites of ssDNA. Because of the elimination of this conformational factor from the mechanistic interpretation of the influence of m-NBA, the use of oligonucleotides can perhaps give more precise information than proteins on m-NBA’s specific role in charging during ESI.

2 Experimental

2.1 Sample Preparation

ssDNA (12-mer: 5′-AAAAACCAAAAA-3′, 5′-TTGGTTAATTCC-3′, 5′-GGAATTAACCAA-3′, 5′-GGTTAATTCCTT-3′, 5′-CCAATTAAGGAA-3′, 5′-GGCCACAATTT-3′, 5′-AAATTGTGGCCC-3′, 5′-CCCTTGTGGAAA-3′, 5′-TTTCCACAAGGG-3′, 5′-TTTTTGGTTTGG-3′, and 6-mer: 5′-TTTTTT-3′, 5′-TTATTT-3′, 5′-TTCTTT-3′, 5′-GTTTTT-3′, 5′-GGGTTT-3′) with 5′ and 3′ hydroxyl terminal groups were obtained from EUROFINS MWG Biotech (Ebersberg, Germany) and were used without further purification. All other chemicals were purchased from Sigma-Aldrich Chemicals (St. Quentin Fallavier, France). All solutions were prepared using deionized water (10 MΩ) with 20 mM ammonium acetate, HPLC grade methanol (50:50, vol/vol) and with or without addition of 1 % volume of m-NBA. Oligonucleotide solutions at 20 μM were injected into the instrument at a flow rate of 5 μL.min-1.

2.2 Mass Spectrometry

Experiments were performed using an electrospray ionization source in negative ion mode combined with an LTQ-Orbitrap hybrid instrument (Thermo Scientific, San Jose, CA, USA) [61–63]. ESI conditions were as follows: accelerating voltage 3.5 kV; ion transfer tube temperature 280 °C; capillary voltage –50 V; and tube lens voltage –110 V. Nitrogen was used as sheath gas and auxiliary gas. High resolution mass spectra (m/z 200–2000) were acquired in the Orbitrap cell (theoretical mass resolving power of 60 000 at m/z 400), after external ion accumulation. The resolution power and mass accuracy allow the observation of the isotopic profile and facilitate the identification of molecular species and their resulting charge states. All data were acquired using external calibration. Relative standard deviations (RSD) were calculated from 10 repeated measurements and are presented in the respective tables.

3 Results

To gain insight into the influence of ∆G°app(acid) of [DNA – nH]n- on m-NBA’s action towards oligonucleotide charging processes, various DNA sequences and sizes (from 6- to 12-mer) were examined.

Figures 1 and 2 present negative mode ESI mass spectra of 5′-GTTTTT-3′ and 5′-GGAATTAACCAA-3′ solutions, respectively. Comparison of mass spectra obtained without (Figures 1a and 2a) or with (Figures 1b and 2b) addition of m-NBA, clearly shows the effect of m-NBA on the charge state distribution of oligonucleotide ions. The presence of m-NBA has broadened the charge state distribution of oligonucleotide ions, from 1 to 4 charge states for the 6-mer (Figure 1), and from 3 to 7 charge states for the 12-mer (Figure 2). In addition to this widening of the charge state distributions, the presence of m-NBA has dramatically increased the zhigh of smaller oligonucleotide ions from –2 to –5; the –5 charge state obtained with m-NBA corresponds to the deprotonation of all phosphate groups constituting the 6-mer ssDNA and, therefore, to the maximum potential charge state (zmax) (Figure 1). The same trend has been observed for longer oligonucleotides, shifting the zhigh from –4 to –8 corresponding to the deprotonation of 8 out of the 11 phosphate groups constituting the 12-mer ssDNA (Figure 2). For all tested oligonucleotides, the lowest observed charge state (zlow) stayed the same despite the presence of m-NBA. Note that “satellite” peaks present in Figure 1b correspond to alkali metal adducts, whereas no m-NBA adducts of oligonucleotides were observed.

Figure 1
figure 1

Mass spectra of 6-mer ssDNA ions (GTTTTT) in negative ESI mode, (a) without m-NBA and (b) with 1 % of m-NBA

Full size image
Figure 2
figure 2

Mass spectra of 12-mer ssDNA ions (GGAATTAACCAA) in negative ESI mode, (a) without m-NBA and (b) with 1 % of m-NBA

Full size image

Figure 3 presents positive ion mass spectra of 5′-TTTTTGGTTTGG-3′ solutions. Comparison of mass spectra obtained without (Figure 3a) or with (Figure 3b) addition of m-NBA shows a relatively minor effect on the charge state distribution of oligonucleotide ions observed in positive ESI mode.

Figure 3
figure 3

Mass spectra of 12-mer ssDNA ions (TTTTTGGTTTGG) in positive ESI mode, (a) without m-NBA and (b) with 1 % of m-NBA

Full size image

Both zhigh and zlow stayed the same while the ion abundance ratio zhigh/zlow has slightly decreased in the presence of m-NBA.

Tables 1 and 2 present negative mode ESI results obtained for all tested oligonucleotide solutions in the absence and presence of 1 % m-NBA. Total oligonucleotide ion abundances (Itot) have been calculated by summing the peak intensities of the entire natural isotopic cluster for each ion charge state. Two clear trends were deduced. First, a slight signal enhancement was observed for all tested oligonucleotides of any length or sequence. In order to quantify this signal enhancement that 1 % m-NBA induces on oligonucleotide ions, a “signal enhancement ratio” (SER1%) is defined as the quotient:

Table 1 Total Ion Abundance (Itot) Obtained Without or with Introduction of 1 % of m-NBA in Various ssDNA Solutions; Calculated Signal Enhancement Ratios (SER1%), Defined as the Quotient of Itot with 1 % of m-NBA Divided by Itot Without m-NBA. Relative Standard Deviations (RSD) were Calculated from 10 Different Measurements
Full size table
Table 2 Average Charge State (zavg), Highest Observed Charge States (zhigh) Obtained Without or with Introduction of 1 % of m-NBA in Various ssDNA Solutions; Calculated Charge Enhancement Coefficient (CEC1%), Defined as zavg in the Presence of 1 % of m-NBA Minus zavg in Absence of m-NBA. Relative Standard Deviations (RSD) were Calculated from 10 Different Measurements
Full size table
$$ {\mathrm{SER}}_{1\%}=\frac{I_{\mathrm{tot}}\mathrm{in}\ \mathrm{the}\ \mathrm{presence}\ \mathrm{of}\ m-\mathrm{NBA}}{I_{\mathrm{tot}}\mathrm{in}\ \mathrm{the}\ \mathrm{absence}\ \mathrm{of}\ m-\mathrm{NBA}} $$
(1)

Very importantly, in negative ESI mode, it was consistently found that all measured SER1% quotients were higher than 1.0, indicating that the presence of 1 % m-NBA invariably resulted in signal enhancement (Table 1).

The second clear trend relates to shifts in the charge state distribution (Table 2). To evaluate the latter shift, average charge states, zavg, were calculated using the following formula [64]:

$$ {\mathrm{z}}_{\mathrm{avg}}={\displaystyle {\sum}_i^n\frac{\left(i\kern0.5em {I}^{i\hbox{--}}\right)}{I_{\mathrm{tot}}}} $$
(2)

corresponding to the sum of each oligonucleotide ion abundance (Ii–) for a given charge state multiplied by its the charge state value (i), divided by the sum of all oligonucleotide ion abundances (Itot). A clear increase of zavg values was observed for all tested ssDNA after m-NBA addition, independently of the oligonucleotide length or sequence.

To better quantify the shift in charge states induced by 1 % m-NBA addition, a charge enhancement coefficient (CEC1%) is defined as:

$$ {\mathrm{CEC}}_{1\%}=\left({\mathrm{z}}_{\mathrm{avg}}\mathrm{in}\ \mathrm{presence}\ \mathrm{of}\ m-\mathrm{NBA}\right)-\left({\mathrm{z}}_{\mathrm{avg}}\mathrm{in}\ \mathrm{absence}\ \mathrm{of}\ m-\mathrm{NBA}\right) $$
(3)

The utility of this charge enhancement coefficient is that it allows a direct assessment of the influence of m-NBA addition (i.e., initial variability in charge state distributions is subtracted out. Very importantly, all calculated CEC1% values were found to be strictly positive, showing the ability of m-NBA to enhance zavg of [DNA–nH]n- ions. Additionally, CEC1% values obtained for larger oligonucleotides are shown to increase with the number of thymine units; moreover, small shifts are observed depending upon the thymine location.

4 Discussion

Through the last decade, m-NBA was extensively described in the literature for its ability to enhance, in positive ESI ion mode, the signal intensity [14], zhigh and zavg [13–16, 20–29, 33, 34] of peptides [13, 20, 25], proteins [13–16, 23, 25–29, 34], and noncovalent protein complexes [21, 22, 24, 33]. Early on, Williams and co-workers [15] proposed that the solvent surface tension is an important macroscopic factor involved in the supercharging mechanism. Indeed, the Rayleigh limit for a spherical droplet is classically given by:

$$ {\mathrm{z}}_r\mathrm{e}=8\pi {\left({\varepsilon}_0\ \gamma \kern0.5em {R}^3\right)}^{1/2} $$
(4)

where, z r is the unit charge limit, e is the elementary charge, ε 0 is the permittivity of the surrounding medium, γ is the surface tension, and R is the droplet radius. The overall surface tension modification induced by addition of m-NBA and the resulting Rayleigh charge limit are presumably independent of the studied polarity. Therefore, the enhancements should be observed regardless of the ion polarity when the same sample preparation and experimental conditions are employed within the limits of the analyte’s ability to undergo proton exchange processes. Attempts performed in negative ion ESI on amphoteric proteic ions, such as the oxidized A chain of insulin [14] or myoglobin noncovalent complexes [24], have only yielded slight increases in zavg.

However, for all tested oligonucleotides in positive ESI mode, SER1% values were measured to be less than 1.0, thereby revealing a decrease of [DNA + H]n+ ion abundance in the presence of m-NBA (Figure 3). Accordingly, CEC1% values were found to be strictly negative under these conditions, showing a decrease of zavg values in the presence of m-NBA (Figure 3). This contrast in results obtained for the two ion polarities shows that solvent surface tension is likely to not be the only factor involved in the supercharging mechanism of oligonucleotides. It is our view that the contradicting experimental results obtained in each polarity are likely indicative of the charging mechanism of DNA. No matter which ion polarity is used, the outcome of m-NBA addition on the oligonucleotide ionization resulted in a consistent decrease in the number of net positive charges (positive ion mode) or, similarly, an increased number of net negative charges (negative ion mode). This evidence supports the notion that m-NBA promotes a unidirectional intermolecular proton transfer ending in a higher number of deprotonated phosphate groups regardless of ion polarity.

For proteins, Loo and co-workers have shown that since (1) no protonated m-NBA reagent is observed in the low m/z region, and (2) m-NBA clustering onto proteins are lost as neutral m-NBA species (at higher transfer energies) while supercharging myoglobin ions in positive ESI mode, m-NBA has a lower GB than GBapp of [Myo + nH]n+ [21, 23]. From the analogous observations in negative ESI mode (no [m-NBA – H]– or negatively charged oligonucleotide adducts with m-NBA observed because of the harsh ESI conditions used), it could be concluded either that m-NBA appears to have a higher ∆G°(acid) value than ∆G°app(acid) of all observed [DNA – nH]n- species, or that m-NBA molecules do not interact directly with phosphate groups and need assistance to deprotonate ssDNA. Indeed, a signal loss has been observed when ammonium acetate buffer was not introduced during sample preparation. If a direct intermolecular proton exchange occurs between m-NBA and oligonucleotide ions during the last desolvation steps, the obtained charge enhancement should be observed even in the absence of NH4 + counter-ions. Effects induced by the presence of m-NBA could, therefore, be explained by the previously established conventional desolvation model for transfer of oligonucleotide ions from solution into the gas phase [43]. By definition, ∆G°(acid) values can be employed only when solvated ions (m-NBA/NH4 +/ssDNA negatively charged aggregates) are released into the gas phase from micro-droplets (mostly composed of m-NBA because of its low vapor pressure compared with that of water). During the last desolvation steps, when NH4 + ions have the opportunity to leave a proton to a [DNA – nH]n– ion, the presence of m-NBA seems to facilitate the other competitive desolvation pathway involving the departure of an NH4 + cation rather than a neutral NH3 molecule. This results in the creation of an additional negatively charged phosphate group on the oligonucleotide ion, and this process may occur more than once, until achieving a maximum level of deprotonation dictated by the ∆G°app(acid) of [DNA – nH]n-. When considering this type of consecutive deprotonation of an oligonucleotide, the ∆G°app(acid) value of each successive multi-deprotonated anion increases progressively until the last thermochemically allowed intermolecular proton transfer, at which point the highest observable charge state (zhigh) is reached. The width of the charge state distribution of oligonucleotides in the gas phase is thereby determined by the range of ∆G°app(acid) values of [DNA – nH]n– between the first and the last intermolecular proton transfers. The propensity for such intermolecular proton transfers correlates with ∆G°app(acid) values, which are dependent upon the specific DNA sequence and mostly upon the number of phospho-thymidine units (the most acidic nucleotide) present in the 12-mer ssDNA.

According to our model, the charge enhancement attributable to the presence of m-NBA occurs at the final stages of evaporation of the charged aggregated. The model can explain the difference in zavg obtained in the presence or absence of m-NBA for oligonucleotide sequences of the same length (12-mers) containing different numbers of phospho-thymidine units (i.e., the most acidic nucleotide). With or without the presence of m-NBA, it was observed that zavg values for oligonucleotide sequences containing three pairs of adjacent thymine units (i.e., three –TT- motifs) were higher than those obtained for sequences containing only one –TT- motif. For example, 5′-TTGGTTAATTCC-3′ and 5′-GGTTAATTCCTT-3′ oligonucleotide ions, respectively, exhibited zavg values of (1) 3.119 and 3.133 without introduction of m-NBA, and (2) 4.71 and 4.73 with 1 % of m-NBA, whereas 5′-GGAATTAACCAA-3′ and 5′-CCAATTAAGGAA-3′ oligonucleotide ions have only yielded respectively zavg values of: (1) 3.035 and 3.052 without m-NBA, and (2) 4.53 and 4.49 with 1 % of m-NBA. Additionally, CEC1% values of [DNA – nH]n- followed the same trend (i.e., 1.59 and 1.60 for these same T-rich sequences compared with 1.49 and 1.43 for the sequences poorest in thymine units. Indeed, the presence of additional thymine nucleobases increases the number of preferentially deprotonable sites and, therefore, facilitates production of higher charge states of [DNA – nH]n-. Thus, the first release of NH4 + occurs readily during desolvation (in preference to proton transfer to ssDNA), while further charging remains guided by the competition between NH3/NH4 + releases from other nucleotide units (e.g., G, C, and A) i.e., by the ∆G°app(acid) of [DNA – nH]n-. For example, 5′-AAAAACCAAAAA-3′ oligonucleotide ion (containing no thymine) exhibited zavg values of (1) 2.883 without introduction of m-NBA, and (2) 4.22 with 1 % of m-NBA; thus, a CEC1% value of 1.34, which is the lowest value of the 12-mer set.

It is not only the number of thymine nucleobases that influence the charge enhancement induced by m-NBA, but also their locations in the DNA sequence play a role. Indeed, CEC1% values obtained for 12-mer ssDNA sequences containing three consecutive thymine nucleobases (e.g., 1.38 and 1.41, respectively, for 5′-GGGCCACAATTT-3′ and 5′-TTTCCACAAGGG-3′ oligonucleotide ions) were found to be lower than those obtained for 12-mer ssDNA sequences containing three non-consecutive thymine nucleobases, (e.g., 1.52 and 1.54, respectively, for 5′-AAATTGTGGCCC-3′ and 5′-CCCTTGTGGAAA-3′ oligonucleotide ions). As noted above, intermolecular proton transfers between NH4 + and the ssDNA are guided by thermochemical properties (e.g., ∆G°app[acid] including electrostatic effects). Thus, they are also oriented by the proximity of other positive and negative charged sites, which will locally influence the activation energy required for a potential intermolecular proton exchange. For 12-mer ssDNA sequences containing consecutive thymine nucleobases (i.e., 5′-GGGCCACAATTT-3′ and 5′-TTTCCACAAGGG-3′), when one of the phosphate groups directly linked to a thymidine unit (most acidic site of the DNA sequence) already holds a negative charge, the energy required to stabilize another negative charge on an adjacent phosphate group directly linked to a second thymidine unit is raised, and this neighboring anion formation is thereby disfavored. Therefore, despite the relatively high acidity of phospho-thymidine units, the close proximity of several will disfavor the consecutive departures of NH4 + cations for a 12-mer and, instead, will achieve a balance wherein the departure of neutral NH3 molecules becomes more competitive, consistent with previously postulated ESI mechanisms describing desorption/ionization of ssDNA into the gas phase [43].

The above 12-mer example shows that local electrostatic effects can exert influence on the dominant orientation given by gas-phase acidity considerations on the supercharging phenomenon. The situation becomes even more complicated when considering the supercharging of 6-mers where the largest difference in the enhanced charging (CEC1% values) occurs for 5′-TTATTT-3′ (dispersed thymine nucleobases, lowest CEC1%) and 5′-GTTTTT-3′ (dense thymine nucleobases, highest CEC1%) even though each has an identical number of thymines. Thus, in the case of these T-rich 6-mer ssDNA, a high “density” of thymine actually leads to a higher CEC1%. The electrostatic effect that seems to come to the forefront in this case is that preferential protonation of the guanine nucleobase (highest GB) occurring in the aggregate apparently facilitates negative charge formation at the nearest phosphate group. During further desolvation, this proton can be lost as solvent molecules depart, thus leaving only negative charges on the ssDNA that now is a highly charged anion. The above 6-mer containing the adenine nucleobase (lower GB) apparently cannot exhibit the same level of electrostatic influence. Thus, this example shows that the deleterious effect of adjacent deprotonable sites can be overcome by other electrostatic effects that promote deprotonation.

The same rationale can explain why 5′-GTTTTT-3′ has a higher CEC1% value (1.36) than 5′-TTTTTT-3′ (1.22) even though the former has one fewer thymine nucleobase. Compared with these two, 5′-TTCTTT-3′ (CEC1% = 1.19) and TTATTT (CEC1% = 1.16) exhibit slightly lower CEC1% values, which is in agreement with our proposed model, but only 5′-GGGTTT-3′ (with the T-poorest sequence) gives a CEC1% = 1.23 that appears to be unusually high, again pointing toward a special role that guanine appears to play in the supercharging process.

5 Conclusion

m-NBA reagent was already known for favoring protonation of proteic biomolecules in positive ESI mode but it was shown here to promote deprotonation of various ssDNA sequences in negative mode during the ESI desolvation process. Using oligonucleotides as tools to scrutinize the underlying aspects of ESI fundamentals (e.g., transfer and desolvation of ions into the gas phase), m-NBA-induced charge enhancement of oligonucleotide ions in negative ESI mode is thus established. A model emphasizing the role of apparent gas-phase acidity has been proposed to rationalize the phenomenon of supercharging. Particular features examined are (1) the number of thymines, (2) the dispersion of thymines, and (3) the presence of protonable nucleobases (especially guanine). Each of these three components can apparently exert variable electrostatic effects on the overall tendency for charged site formation. The ability to produce more highly charged precursor oligonucleotide ions (induced by the presence of m-NBA) could provide a means to obtain more extensive ion fragmentations in tandem mass spectrometry experiments. This would thereby improve the relevance of fitness values obtained in automated tandem mass spectral sequencing of small oligonucleotides using, for example, the COMPAS, or other similar algorithms [65, 66]. Although no real consensus has yet emerged concerning the mechanism of supercharging in ESI, the current contribution is proposed as an additional brick in constructing a foundation to explain the mechanism of supercharging for both ESI ion modes.