B American Society for Mass Spectrometry, 2018 J. Am. Soc. Mass Spectrom. (2018) 29:1002Y1011DOI: 10.1007/s13361-018-1893-2

RESEARCH ARTICLE

Direct Analysis of Proteins from Solutions with High SaltConcentration Using Laser Electrospray Mass Spectrometry

Santosh Karki, Fengjian Shi, Jieutonne J. Archer, Habiballah Sistani, Robert J. LevisDepartment of Chemistry and Center for Advanced Photonics Research, Temple University, Philadelphia, PA 19122, USA

2040 2070 2100 2130 2040 2070 2100 2130

[L+7H]7+

ESI

[L+7H]7+

LEMS Abstract. The detection of lysozyme, or amixture of lysozyme, cytochrome c, and myo-globin, from solutions with varying salt con-centrations (0.1 to 250mMNaCl) is comparedusing laser electrospray mass spectrometry(LEMS) and electrospray ionization-massspectrometry (ESI-MS). Protonated proteinpeaks were observed up to a concentrationof 250 mM NaCl in the case of LEMS. In thecase of ESI-MS, a protein solution with salt

concentration > 0.5 mM resulted in predominantly salt-adducted features, with suppression of the proton-ated protein ions. The constituents in the mixture of proteins were assignable up to 250 mM NaCl forLEMS and were not assignable above a NaCl concentration of 0.5 mM for ESI. The average sodiumadducts (< n >) bound to the 7+ charge state of lysozyme for LEMS measurements from salt concentra-tions of 2.5, 25, 50, and 100 mM NaCl are 1.71, 5.23, 5.26, and 5.11, respectively. The conventionalelectrospray measurements for lysozyme solution containing salt concentrations of 0.1, 1, 2, and 5 mMNaCl resulted in < n > of 2.65, 6.44, 7.57, and 8.48, respectively. LEMS displays an approximately twoorders of magnitude higher salt tolerance in comparison with conventional ESI-MS. The non-equilibriumpartitioning of proteins on the surface of the charged droplets is proposed as the mechanism for the highsalt tolerance phenomena observed in the LEMS measurements.Keywords: Direct protein analysis, High salt concentration, Charge state distribution, Femtosecond, Laserablation

Received: 30 October 2017/Revised: 5 January 2018/Accepted: 17 January 2018/Published Online: 8 March 2018

Introduction

The analysis of biomolecules using electrospray ionization-mass spectrometry (ESI-MS) is often hindered by matrix

effects due to the presence of phosphate buffers, urea, orinorganic salts [1, 2]. ESI-MS measurements of proteins inthe presence of sodium chloride (NaCl) reveal extensive sodi-um ion adduction even for a low millimolar concentration [3].The presence of NaCl also results in a decrease in both thestability of the electrospray and the yield of protonated ions [4].

The adduction of sodium ions to protein molecules distributesthe signal over multiple mass-to-charge (m/z) peaks, reducingthe signal-to-noise ratio of each protein feature and broadeningthe mass spectral peaks, resulting in a decrease in accuracy formass measurement.

Since salt is ubiquitous in many biological samples, severalstrategies have been developed for desalting prior to ESI-MSanalysis including nanoparticle-based microextraction [5], mi-crodialysis [6], liquid chromatography [7], and ion-exchangechromatography [8, 9]. However, removal of salts can affectthe structure of protein complexes. For example, a sigma acti-vator protein (NtrC4) from Aquifex aeolicus requires millimolarconcentrations of Mg2+, BeF3−, and adenosine 5′-diphosphate(ADP) sodium salt to form an active hexamer [10, 11]. Desaltingprocesses that require multiple buffer exchange steps will dena-ture the hexamer and may also produce a spurious signal in theESI analysis due to inefficient removal of detergents [12].

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s13361-018-1893-2) contains supplementary material, whichis available to authorized users.

Correspondence to: Robert Levis; e-mail: rjlevis@temple.edu

Decoupling the sampling and electrospray ionization pro-cesses has enabled analysis of analytes from samples containinghigh salt concentrations. Fused-droplet electrospray ionization-mass spectrometry (FD-ESI-MS) involves ultrasonic nebuliza-tion of the sample solution to produce a fine aerosol which isthen combined with electrospray-generated charged droplets [3].Although this method showed a significantly higher salt toler-ance in comparison with conventional electrospray (1.70 versus0.17 M NaCl), interaction of protein molecules withelectrosprayed methanol droplets containing 1% acetic aciddenatures protein and shifts the charge state distribution (CSD)to higher charge in comparison with native state measurements.Probe electrospray ionization (PESI) uses a wire as both thesampling probe and electrospray emitter and has been used todetect biomolecules from solutions with high salt concentration[1]. The analysis of myoglobin using PESI displayed a salttolerance, or ability to detect protein without ion suppression,of up to 250 mM in comparison with nano-electrospray ioniza-tion (nano-ESI) which has a salt tolerance of 50 mM. The highsalt tolerance for PESI and nano-ESI is likely due to the forma-tion of higher charge states given the fact that sodium ionadduction occurs mostly for low charge states [4, 13, 14].Selective sampling of analytes rather than salt from a solutionhas been proposed to enhance detection in the PESI experimentin comparison with nano-ESI where the entire solution withinthe capillary is subjected to the nano-spray process resulting inanalyte signal suppression given the higher salt concentration[1]. The investigation of drug mixtures (cocaine anddiacetylmorphine) using desorption electrospray ionization(DESI) showed higher salt tolerance in comparison with ESI[15]. The salt tolerance in DESI was investigated by changingthe sample substrate such as paper, polytetrafluoroethylene(PTFE), and a variety of coated glasses. The salt tolerance wasdependent upon the surface being used, and PTFE was theoptimal substrate for reducing ion suppression effects at highersalt concentration. A raw urine sample spiked with atrazine and1,3,5-trinitroperhydro-1,3,5-triazine (RDX)was successfully ex-amined using extractive electrospray ionization (EESI), elimi-nating clean-up steps used to remove the matrix effects [16]. Thestable signal intensity in EESI during salt-rich biological sampleanalysis was attributed to the difference in surface partitioning ofpolar and non-polar analytes. The non-polar analytes tend tocompete for the surface while the polar analytes and salts arestabilized by increased solvation in the droplet interior.

Two-step ESI techniques decouple the vaporization and ion-ization process into separate processes to allow measurementsthat cannot be performed using either ESI or matrix-assistedlaser desorption ionization (MALDI) alone. These methodsinclude electrospray laser desorption ionization (ELDI) [17],matrix-assisted laser desorption electrospray ionization(MALDESI) [18, 19], and laser ablation electrospray ionization(LAESI) [20]. ELDI has been previously used to detect proteinsdissolved in various solutions containing buffers, organic andinorganic salts, and strong acids and bases without undergoingsolvent extraction, filtration, dialysis, or chromatographic sepa-ration [17]. The higher salt and detergent tolerance of ELDI in

comparison with ESI-MS enabled the direct analysis of proteinsfrom solution containing sodium dodecyl, octyl phenolethoxylate, tetrabutyl ammonium bromide, etc. [17, 21].

The use of laser electrospray mass spectrometry (LEMS)has enabled the quantitative analysis of complex mixtureswithout sample pre-processing at atmospheric pressure [22,23]. LEMS couples non-resonant femtosecond (fs) laser vapor-ization (1013W/cm2) with an electrospray ionization source forpostionization. The successful detection of 1,2-dihexanoyl-sn-glycero-3-phosphocholine (DHPC) spiked into whole bloodsample and the detection of lipids and proteins from reduced-fat milk sample demonstrate that LEMS is capable of detectinganalytes from complex mixtures with no pre-processing [24].In addition, LEMS measurements on small molecule mixtures[25] and multicomponent protein mixtures [26] have shownthat LEMS enables quantitative measurements up to ~ 2.5orders of magnitude and over 4 orders of magnitude higher inconcentration, respectively, in comparison with conventionalelectrospray where quantitative measurement of mixtures is notpossible. The non-equilibrium partitioning of analyte on thesurface of the charged droplet is likely responsible for thequantitative capability [25, 26]. This suggests that LEMSmay be utilized to significantly reduce ion suppression effectsin comparison with conventional ESI analysis, in particular forbiomolecules from solution containing high salt concentrationbecause the salt will partition into the droplet interior while theprotein remains on the surface where excess charge resides.

In this study, the mass spectral features of either a singleprotein (lysozyme) or a mixture of the proteins lysozyme, cyto-chrome c, and myoglobin as a function of salt concentration arecompared using LEMS and ESI-MS. The amount of protonatedprotein is measured as a function of salt concentration for bothLEMS and ESI, and the average number of sodium adducts(< n >) bound to the 7+ charge state of lysozyme is calculated.The ability to identify protein components in a mixture usingLEMS and ESI is compared as a function of salt concentration.Finally, a mechanism is presented for reduced salt adduction inthe LEMS measurement in comparison with ESI.

Materials and MethodsSample Preparation

Lysozyme, cytochrome c, myoglobin, and ammonium acetate(Sigma Aldrich, St. Louis, MO) were prepared in HPLC-gradewater (Fisher Scientific, Pittsburgh, PA) to yield the final con-centration of 1.0 × 10−3 M. A 1.0 M stock solution of sodiumchloride was prepared in HPLC-grade water (Fisher Scientific,Pittsburgh, PA). For ESI measurements, an aliquot of the stocksolution (single protein or protein mixtures) was diluted into10 mM aqueous ammonium acetate to yield a final proteinconcentration of 1.0 × 10−5Mwith the salt concentration rangingfrom 0.1 to 5.0 mM. For LEMS measurements, an aliquot of thestock solution (either individual or the protein mixtures) wasdiluted into water to yield a final protein concentration of 2.0 ×10−4 M with the salt concentration ranging from 2.5 to 250 mM.

S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS 1003

A 10 μL aliquot of the diluted protein solution was spotted onto astainless steel plate and then subjected to laser vaporization intothe electrospray (ES)-charged droplet stream consisting of aque-ous ammonium acetate.

Laser Vaporization

A Ti:sapphire laser oscillator (KM Laboratories, Inc., Boulder,CO) seeded a regenerative amplifier (Coherent, Inc., Santa Clara,CA) that delivered 75 fs, 0.6 mJ laser pulses centered at 800 nm.The laser, operated at 10 Hz to couple with the ES ion source,was focused to a spot size of ~ 250 μm in diameter with anincident angle of 45° with respect to the sample using a 16.9-cmfocal length lens, with an approximate intensity of 1 × 1013 W/cm2. The steel sample plate was biased to − 2.0 kV to compen-sate for the distortion of electric field between the capillary inletand the needle caused by the sample stage. Aqueous proteinsample (10 μL) deposited onto a steel substrate was vaporizedby the laser pulse allowing for capture and ionization by an ESplume traveling perpendicular to the vaporized material.

Mass Spectrometry and Data Analysis

The non-resonant femtosecond laser pulse transfers the analyteinto the gas phase for capture and ionization in an ES plume atatmospheric pressure. The flow rate for ES solvent was set at2 μL/min by a syringe pump (Harvard Apparatus, Holliston,MA). The ESI needle was 6.4 mm above and parallel to thesample stage and was approximately 6.4 mm in front of thecapillary entrance. The ES needle was maintained at groundwhile the inlet capillary was biased to − 4.5 kV to operate inpositive ion mode. The postionized analytes were dried beforeentering the inlet capillary by countercurrent nitrogen gas at200 °C flowing at 4 L/min. The charged sample was massanalyzed using a microTOF-Q II mass spectrometer (BrukerDaltonics, Billerica, MA).

Results and DiscussionDetection of Protein/Protein Mixturesfrom Solutions with Varying Salt Concentrations

The LEMS and ESI measurements for lysozyme are shown inFigure 1 for salt concentrations ranging from 0 to 100 mM and 0to 5 mM, respectively. Higher lysozyme concentration waschosen for LEMS (200 μM) in comparison with ESI (10 μM)analysis as only about 3.6% of the vaporized analytes are cap-tured in the electrospray plume, resulting in lower concentrationof the protein being analyzed. The LEMS measurement forlysozyme vaporized without NaCl, shown in Figure 1a, revealsa narrow range of low charge states ranging from 6+ to 9+. Thepredominance of the 8+ charge state in Figure 1b–e indicates thatthe native state of lysozyme was preserved when vaporized froma solution with salt concentration up to 100 mM. Salt adductionto the protein ions for LEMS increases with decreasing charge.The 9+ feature reveals no adduction in any of the spectra,minimal adduction in 8+, and significant adduction in the 7+

through 4+ species (see Figure 1b–e; also shown in Supplemen-tary Figure S1c–e, measurements taken at collision potential of70 eV). Previous ELDI measurement of cytochrome c fromsolution containing high salt concentration using a denaturingelectrospray solvent (acidic methanol solution) resulted in pro-tein CSD peaked at 15+ with minimal salt adduction [21]. Theincreased salt tolerance for ELDI measurement is likely due tothe formation of higher charge states where salt adduction to theprotein molecule ion is minimal. Salt adduction to lower charge(folded) states is likely due to the increased salt concentrationwithin the ES droplet during desolvation process in the chargeresidue mechanism [4]. The increasing salt concentration facil-itates non-specific pairing of Na+ and Cl− with the protein ions.Sodium adducts are evident at higher collision potential (~ 70 V)while a lower potential of ~ 5 V results in a series of n(Na −H)and m(Cl + H) peaks in the mass spectrum. In the case of highercharge states, Konermann has proposed that the unfolded proteinresides at the surface of the droplet, enabling charge ejectionduring the desolvation process when the salt concentration islower [4]. The increase in local kinetic energy for the 8+ chargestate in comparison with the 7+ and 6+ charge states (for a givencollision potential) may also remove salt from the protein mol-ecule through collisional processes resulting in the lower Na+

adduction for the 8+ charge state of lysozyme in comparisonwith the 7+ and 6+ charge states.

The ESI-MS measurements for lysozyme in aqueous ammo-nium acetate solution with NaCl concentrations ranging from 0to 5 mM are shown in Figure 1f–j. The ESI mass spectrum oflysozyme in ammonium acetate, without NaCl, reveals a narrowrange of low charge states ranging from 6+ to 9+ in Figure 1f.The spectra corresponding to ESI measurements in the presenceof salt (Figure 1g–j) again shows that increasing salt adductionoccurs for the 8+, 7+, and 6+ charge states with decreasingcharge. In the ESI measurement, the adduction is significantlyenhanced in comparison with LEMS for similar concentrationsof salt (e.g., Figure 1b, i). Peak broadening due to sodium ionadduction to the 6+ and 7+ charge states for salt concentration of≥ 1 mM makes the detection of protonated lysozyme ([L +7H]7+ and [L + 6H]6+) challenging. In particular, peak broaden-ing limited the detection of 6+ charge state of lysozyme for saltconcentration of ≥ 1 mM. This is in agreement with the previousnano-spray measurement of ubiquitin where peak broadening of[U + 6H]6+ and [U + 5H]5+ charge states hindered the detectionof protonated protein ions when analyzed from solution con-taining 1.0 mM of NaCl without the addition of solution addi-tives that facilitate desalting of proteins [2].

The LEMS and ESI measurements for an equimolar mixtureof lysozyme, cytochrome c, and myoglobin in salt concentrationranging from 0 to 250 mM are shown in Figure 2. The LEMSmeasurement of protein mixtures without NaCl vaporized intoan ES solvent consisting of aqueous ammonium acetate, shownin Figure 2a, reveals a narrow range of low charge states rangingfrom 7+ to 10+ for lysozyme, 6+ and 8+ for cytochrome c, and7+ to 9+ for myoglobin, respectively. The CSDs of lysozyme,cytochrome c, and myoglobin vaporized from concentrations ofNaCl ranging from 75 to 250 mM (Figure 2b–d) are similar to

1004 S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS

the 0 mM salt solution and correspond to the folded proteincharge state distribution. This suggests that the native states ofthe proteins were preserved when vaporized from solutions withhigh salt concentration. Previous LEMS measurement of aque-ous myoglobin into the electrospray solvent containing aqueousammonium acetate resulted in intact holo-myoglobin featuressuggesting that the non-covalently bounded protein complexescan be transferred intact into the electrospray plume [27]. This ispresumably due to shorter time scale at which the energy isdeposited into the system in comparison with the time requiredfor protein unfolding. The majority of salt adduction to theprotein mixture occurred for the 6+, 7+ and 8+ charge statesof lysozyme, for the 6+, 7+ and 8+ charge states of cytochromec, and for the 7+, 8+, and 9+ charge states of myoglobin (seeFigure 2b–d; also shown in Supplementary Figure S2b–d). Themajority of salt adduction occurred at the 6+ and 7+ chargestates in the case of lysozyme during single protein analysis (seeFigure 1b–e; also shown in Supplementary Figure S1b–e).

ESI measurements were performed for equimolar mixtures oflysozyme, cytochrome c, and myoglobin (10 μM of each) fromaqueous ammonium acetate solution containing NaCl concentra-tions ranging from 0 to 5 mM, as shown in Figure 2e–h. The ESImass spectra of protein mixtures with no added NaCl revealedcharge states ranging from 7+ to 10+ for lysozyme, 7+ and 8+ forcytochrome c, and 8+ and 9+ for myoglobin, as shown inFigure 2e. The ESI measurements for salt concentrations of 0.5

and 2mMNaCl (Figure 2f, g) show significant salt adduction forthe 7+ and 8+ charge states of lysozyme, the 7+ and 8+ chargestates of cytochrome c, and the 8+ and 9+ charge states formyoglobin. The ESI measurement of protein mixtures at saltconcentration of 5 mM resulted in charge states ranging from7+ to 14+ for cytochrome c and suppressed the detection of thelower charge states of myoglobin (8+ and 9+ charge states) andlysozyme (7+ and 8+ charge states) by significant salt adduction,as shown in Figure 2h. The observation of unfolded charge statedistribution for cytochrome c is likely due to the increase inelectrospray droplet temperature at higher salt concentrationdue to the non-volatile nature of the salt. The formation of highercharge states were also observed in a previous experiment whenthe electrospray solvent contained higher salt concentration [1].In that study, the addition of 50 mM NaCl into the 10 μMmyoglobin solution resulted in a CSD ranging from 10+ to 22+(peaked at 12+ and 20+), while the lower salt concentration(5 mMNaCl) resulted in a CSD ranging from 7+ to 18+ (peakedat 11+). Identification of individual proteins from multicompo-nent mixtures is difficult using ESI when salt is present in thesample solution at concentration > 5 mM due to ion suppressioneffects [28] and spectral congestion [29] arising from overlappingfeatures. Direct analysis of protein mixtures using ESI in thepresence of inorganic salts is complicated because of peak broad-ening due to adduction. This is not the case for LEMS where saltconcentrations up to 250 mM do not result in protein structural

Figure 1. Mass spectra representing laser-induced vaporization of 200 μM lysozyme with (a), no salt, (b) 2.5 mM NaCl, (c) 25 mMNaCl, (d) 50 mMNaCl, and (e) 100mMNaCl into the ES of 10mMaqueous ammonium acetate. Electrospray mass spectra of 10 μMlysozyme prepared in 10 mM aqueous ammonium acetate with (f) no salt, (g) 0.1 mM NaCl, (h) 1 mM NaCl, (i) 2 mM NaCl, and (j)5 mM NaCl. The measurements were performed at collision potential of 15 eV. A feature with an asterisk (*) represents solvent

S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS 1005

change or significant adduction, and thus, the identification ofindividual protein from the multicomponent mixture is possible.

Salt Adduction to Proteins: LEMS Versus ESI-MS

The adduction of sodium ion to protein was compared forLEMS and ESI-MS measurements to provide insight into theionization mechanisms of each experiment. Sodium adductionto the protein molecule can be expressed using Equation 1 [4],

Mþ zHþ þ n Naþ–Hþð Þ þ m Cl− þ Hþð Þ½ �zþ

þ 70 eV CID→ Mþ z–nð ÞHþ þ nNaþ½ �zþ þ mHCI ð1Þwhere M is the protein molecule, z is the number of protons(H+), n and m are the number of Na+ and Cl− adducts, respec-tively, z is the charge state of the protein, and z+ is the overallcharge on the protein molecule. The lower collision potential of15 eV used in this measurement resulted in a series of n(Na+ −H+) and m(Cl− +H+) features that complicate the mass spectra[4]. Increasing the collision potential to 70 eV resulted in thecomplete loss of HCl, and Na-adducted protein ions are pre-dominately formed as shown in Figures 3 and 4. The removalof HCl is mainly attributed to an in-source reaction [30, 31]where protein ions are subjected to collision-induced

dissociation (CID). The increase in collision potential alsoresulted in the loss of H2O from lysozyme as shown in Sup-plementary Figure S3.

To compare the Na+ ion adduction as a function of saltconcentration for both LEMS and ESI-MS measurements, wefocus on the 7+ charge state of lysozyme as shown in Figure 3.LEMS analysis of lysozyme with salt concentration of 25 mMNaCl revealed salt adducts ranging from 1 to 13 (Figure 3c).The dominant feature in the mass spectra collected with a CIDpotential of 70 eV is the protonated peak of lysozyme [L +7H]7+. Conversely, ESI analysis of lysozyme with 1 mMNaClresulted in the Na = 3 adduct dominating with sodium-adducted protein peaks ranging from 1 to 16 (Figure 3h). Thesodium adducts resulted in peak broadening for ESI-MS mea-surements (Figure 1h). The average number of sodium ionsadducts < n > can be calculated [4] as follows:

< n >¼ ∑Nn¼0 n*Inð Þ∑N

n¼0 In

where In is the peak intensity of the Na+ adducts and N is themaximum number of Na+ observed for each charge state. Theaverage number of sodium ions bound to the 7+ charge state oflysozyme (<n > ) for the LEMS measurement are 1.71, 5.23,5.26, and 5.11 for the salt concentration of 2.5, 25, 50, and

Figure 2. Mass spectra representing laser-induced vaporization of 200 μM protein mixtures (lysozyme (L), cytochrome c (C), andmyoglobin (M)) with (a) no salt, (b) 75 mM NaCl, (c) 125 mM NaCl, and (d) 250 mM NaCl into the ES of 10 mM aqueous ammoniumacetate. Electrospray mass spectra of 10 μM protein mixtures prepared in 10 mM aqueous ammonium acetate with (e) no salt, (f)0.5 mM NaCl, (g) 2 mM NaCl, and (h) 5 mM NaCl. The measurements were performed at collision potential of 15 eV

1006 S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS

100 mM NaCl, respectively. The Na+ adduction did not in-crease between 25 and 100 mM NaCl concentrations in theLEMS analysis. The average number of sodium ions bound tothe 7+ charge state of lysozyme for the ESI measurement forsalt concentrations of 0.1, 1, 2, and 5 mM NaCl are 2.65, 6.44,7.57, and 8.48, respectively. Conventional electrospray mea-surements showed increasing salt adduction with increasingsalt concentration with a corresponding decrease in intensityfor the protonated peak of the protein. The measurementsrevealed significantly reduced sodium adduction for LEMSmeasurements in comparison with ESI-MS for all charge states.

For the mass spectral analysis of multicomponent proteinmixtures, the collision potential was increased to 70 eV toeliminate HCl to form predominately Na-adducted protein ions[4]. To compare the sodium adduction to the protein moleculefor LEMS and ESI-MS measurements, the 7+ charge state ofcytochrome c (m/z 1766.5), 8+ charge state of lysozyme (m/z1789.1), and 8+ charge state of myoglobin (m/z 2122.4) werechosen. Salt adduction to the 7+ charge state of cytochrome c

overlaps with the protonated ([L + 8H]8+) lysozyme peak.LEMS measurements revealed well-resolved, protonated fea-tures for cytochrome c ([C + 7H]7+) and lysozyme ([L + 8H]8+)for salt concentration of 250 mMNaCl, as shown in Figure 4d.This suggests minimal adduction/overlap and limited ion sup-pression of the protonated protein ions even in the presence ofhigh salt concentration using laser vaporization to transfersample into the ES droplets. Conversely, in ESI-MS measure-ments, the mass spectra revealed congested features and sup-pression of protonated ions ((C + 7H)7+ and (L + 8H)8+) forNaCl concentrations ranging from 0.5 to 5 mM, as shown inFigure 4f–h. The ESI measurements for cytochrome c andlysozyme with 0.5 mM NaCl revealed that the Na = 2 andNa = 1 adducts were the dominant features with sodium ad-ducts ranging from 1 to 6 and 1 to 11, respectively (Figure 4f).At a salt concentration of 5 mM, unresolvable spectral featuresfor both cytochrome c and lysozyme are observed due toexcessive salt adduction. Adducted protein features were ob-served for 0 mMNaCl in both LEMS and ESI measurements of

Figure 3. High-resolution LEMSmass spectra of 200 μM lysozyme [L + 7H]7+ with (a) no salt, (b) 2.5 mM NaCl, (c) 25 mMNaCl, (d)50 mMNaCl, and (e) 100mMNaCl into the ES of 10mMaqueous ammonium acetate. High-resolution electrospray mass spectra of10 μM lysozyme [L + 7H+]7+ prepared in 10mMaqueous ammonium acetate with (f) no salt, (g) 0.1mMNaCl, (h) 1 mMNaCl, (i) 2 mMNaCl, and (j) 5mMNaCl. Themeasurements were performed at collision potential of 70 eV. Sodium adducts are indicated as n = 0, 1,2, etc

S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS 1007

cytochrome c and lysozyme, presumably due to impurities inthe solvents (Figure 4a, e).

The LEMSmeasurement for the 8+ charge state of myoglo-bin revealed salt adducts ranging from 1 to 14 for salt concen-tration of 75 mM NaCl (Supplementary Figure S4b). Thedominant feature in the mass spectra is the protonated peak ofapo-myoglobin [M + 8H]8+. The formation of apo-myoglobinis due to collision-induced unfolding of myoglobin upon in-creasing the collision energy to 70 eV. The ESI analysis ofmyoglobin with 0.5 mM NaCl resulted in the Na = 3 adductbeing the largest feature with sodium adduct features rangingfrom 1 to 15 (Supplementary Figure S4f). Again, the LEMSexperiment enabled measurements of protein molecules forsamples with higher salt concentration because of reducedadduct formation.

To test whether the higher protein concentration (200 μM)used in the LEMS measurements resulted in increased salt

tolerance and reduced adduct formation in comparison withESI-MS (where a protein concentration of 10 μM was used),we performed an experiment where the concentrations of lyso-zyme in both LEMS and ESI analyses were adjusted to be50 μM. The salt concentration was varied from 0 to 100 mMfor the LEMS measurements and 0 to 5 mM for the ESImeasurements. The mass spectra obtained are shown in Sup-plementary Figures S5 and S6. Similar to the analysis of200 μM lysozyme using LEMS, a protonated protein feature(Supplementary Figure S6d) was obtained up to the salt con-centration of 100 mM NaCl. Conversely, salt concentrationgreater than 0.1 mM NaCl resulted in salt-adducted featuresdominating the protonated feature of lysozyme (50 μM inconcentration, Supplementary Figure S6f) in electrospray anal-ysis. These results indicate that the adduct formation in LEMSand ESI measurements is independent of the proteinconcentration.

Figure 4. High-resolution LEMS mass spectra of 200 μM protein mixtures (cytochrome c [C + 7H]7+ and lysozyme [L + 8H]8+) with(a) no salt, (b) 75 mM NaCl, (c) 125 mM NaCl, and (d) 250 mM NaCl into the ES of 10 mM aqueous ammonium acetate. High-resolution electrospray mass spectra of 10 μM protein mixtures prepared in 10 mM aqueous ammonium acetate with (e) no salt, (f)0.5mMNaCl, (g) 2 mMNaCl, and (h) 5 mMNaCl. Themeasurements were performed at collision potential of 70 eV. Sodium adductsare indicated as n = 0, 1, 2, etc

1008 S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS

We also investigated the effect of an emitter size on salttolerance using a nano-spray emitter with the inner diameter of50 μm and a flow rate of 400 nL/min. The nano-spray massspectra of lysozyme prepared in 2.5 mM NaCl concentrationresulted in Na = 3 adduct being the dominant feature withsodium-adducted protein peaks ranging from 1 to 11 (inset ofSupplementary Figure S7e). The nano-spray analysis of lyso-zyme with 50 mM NaCl resulted in an unresolved mass spec-tral feature at the CID potential of 70 eV indicating that thepresence of high salt concentration results in ion suppressioneffects in nano-spray ionization under these experimental con-ditions. Conversely, the protonated lysozyme feature was ob-tained in the LEMS measurements up to the salt concentrationof 100 mM. Our preliminary studies using nano-spray emitterwith inner diameter of 50 μm did not show increased salttolerance in comparison with electrospray emitter with124 μm inner diameter. However, a recent study has shownthat the use of a nano-spray emitter with tip size of less than∼ 1.4 μm results in higher salt tolerance and reduced adductformation [32]. Future studies will compare the performance ofsuch emitters with LEMS measurements.

Thesemeasurements suggest that in the LEMS experiments,the Na+ ions are somehow excluded from interacting with theprotein molecules. The reduced Na+ ion adduction in LEMSmeasurements in comparison with ESI may be due to the non-equilibrium nature of laser-vaporized proteins interacting withthe ES droplets. We have previously proposed [25, 26, 33] thatthe partitioning of analytes into the ES droplets is limited by thetime that laser-vaporized analytes spend on the droplet surfaceprior to entering into the MS inlet capillary. The time forinteraction between the vaporized protein and the chargeddroplet is estimated to be < 100 ms [33], and this time willtherefore dictate the partitioning of analytes based on theirrespective sizes and solution enthalpy. Considering a simplediffusion process, the mixing time, t, of two analytes dependsupon their diffusion coefficient (D) and the distance separatingthe diffusing solutes (d), where t = d2/6D. Given the fact thatthe diffusion coefficient of analytes depend on their size andshape, we anticipate that Na+ and Cl− will diffuse into theinterior of the droplet much more quickly than the protein inthe LEMS process, thus limiting the opportunity for the chem-istry described by reaction 1 if the droplet lifetime is shortcompared to the protein diffusion time. The diffusion constantfor Na+ is 1.6 10−9 m2/s, and assuming d = 1 μm, the diffusiontime is (1 × 10−6 m)2/6 × 1.6 10−9 m2/s = 104.2 μs. This time isshort compared to the lifetime of the droplet ~ 10 ms, and thus,the salt ions can partition into the interior of the droplet. Thediffusion constant for hemoglobin is 6.9 × 10−11 m2/s, suggest-ing a 23× longer diffusion time. The proteins are also muchmore hydrophobic than salt ions and thus will tend to remain onthe surface of the charged droplet [34, 35]. This results in alower salt concentration on the surface of the newly formedelectrospray droplet from which a gaseous protein ion is re-leased. This also results in a higher concentration of protein onthe surface of the electrospray droplet where the charge resides.Thus, the non-equilibrium nature of the partitioning of salt and

protein between droplet surface and interior will reduce adduc-tion in the LEMS measurement.

A previous study investigating the ionization mechanism ofproteins in ESI-MS revealed a completely different salt adduc-tion pattern for folded and unfolded proteins [4]. The foldedprotein revealed higher salt adduction, which is in agreementwith the charge residue model (CRM) where the non-specificinteraction between salt and protein increases as the solventevaporates. The unfolded protein however revealed minimal tono salt adduction which is attributed to the difference in ioni-zation mechanism for folded and unfolded proteins in ESI.Unfolded proteins are likely expelled from the ES droplet viathe chain ejectionmodel (CEM) [36] prior to the increase in saltconcentration within the ES droplets resulting in minimal to nosalt adduction for higher charge states. The reduced salt adduc-tion for proteins in LEMS measurement suggests that proteinionization in LEMS likely occurs in a manner similar to thechain ejection model. However, the salt adduction in LEMS issignificantly lower than that for ESI, suggesting that the pro-teins do not equilibrate into the interior of the droplet. The polaranalytes (e.g., salt) equilibrating into the droplet interior due toincreased solvation will thus be isolated from proteins resultingin minimal salt adduction in LEMS measurements.

The preferential interaction of laser-vaporized protein withthe highly charged electrospray droplets in LEMS measure-ments has been proposed to enhance the detection of protonat-ed protein ions in comparison with conventional ESI measure-ments [37]. Protein residing in the droplet interior in the ESI-MS measurement has to partition onto the droplet exterior forionization. This enhances interaction of protein with salt in thecase of ESI resulting in an increased salt adduction.

ConclusionsThe analysis of lysozyme, and a mixture of lysozyme, cyto-chrome c, and myoglobin, in solution with varying salt con-centrations revealed a significantly higher salt tolerance forLEMS measurements in comparison with conventionalelectrospray with respect to suppression of the protonated ionsignal and the ability to distinguish mixture components. Theaverage sodium adducts bound to the 7+ charge state of lyso-zyme for LEMSmeasurements from salt concentrations of 2.5,25, 50, and 100 mM NaCl are 1.71, 5.23, 5.26, and 5.11,respectively, whereas conventional electrospray measurementsfor lysozyme from solutions containing salt concentrations of0.1, 1, 2, and 5 mM NaCl resulted in < n > of 2.65, 6.44, 7.57,and 8.48, respectively. The reduced salt adduction in the LEMSexperiment is likely due to the non-equilibrium nature ofpartitioning of laser-vaporized proteins between the surfaceand interior of the charged electrospray droplets. The reducedinteraction time of the laser-vaporized analytes with the ESdroplets presumably allows small ions like Na+ and Cl− topartition into the droplet interior on the time scale of transit tothe capillary leaving the much larger protein on the dropletsurface where excess charge resides. This results in a lower salt

S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS 1009

concentration on the surface of the ES droplets, which leads tothe formation of a higher fraction of protonated protein ions incomparison with conventional ESI measurements. Conversely,proteins in electrospray measurements have to partition fromthe droplet interior to the droplet’s surface for ionization whichresults in an increased interaction of protein with salt in the ESdroplet from which an ionized protein is released.

We note that the lower ion abundance of LEMS measure-ment in comparison with ESI-MS is due to the low captureefficiency (~ 3.6 ± 1.8%) of the laser-vaporized analyte by theelectrospray plume. The use of methods similar to previousstudies [38, 39] to increase the transfer efficiency of laser-vaporized analyte into the electrospray plume will enable anal-ysis of lower concentrations of samples. For instance, it hasbeen reported that the use of continuous flow solvent probe asthe plume capture source in LAESI measurement resulted inthe increase in transfer efficiency of the laser-vaporized mate-rial from 2 to 50% [38]. Similar designs will significantlyimprove the LOD of a two-step ionization technique likeLEMS that couples laser ablation with electrospraypostionization mass spectrometry.

AcknowledgementsThe authors gratefully acknowledge the funding support fromthe National Science Foundation (CHE-1362890, CHE-0957694).

References

1. Mandal,M.K., Chen, L.C., Hashimoto, Y., Yu, Z., Hiraoka, K.: Detectionof biomolecules from solutions with high concentration of salts usingprobe electrospray and nano-electrospray ionization mass spectrometry.Anal. Methods. 2, 1905–1912 (2010)

2. Flick, T.G., Cassou, C.A., Chang, T.M., Williams, E.R.: Solution addi-tives that desalt protein ions in native mass spectrometry. Anal. Chem. 84,7511–7517 (2012)

3. Chang, D.-Y., Lee, C.-C., Shiea, J.: Detecting large biomolecules fromhigh-salt solutions by fused-droplet electrospray ionization mass spec-trometry. Anal. Chem. 74, 2465–2469 (2002)

4. Yue, X., Vahidi, S., Konermann, L.: Insights into the mechanism ofprotein electrospray ionization from salt adduction measurements. J.Am. Soc. Mass Spectrom. 25, 1322–1331 (2014)

5. Shrivas, K., Wu, H.-F.: Modified silver nanoparticle as a hydrophobicaffinity probe for analysis of peptides and proteins in biological samplesby using liquid-liquid microextraction coupled to AP-MALDI-ion trapand MALDI-TOF mass spectrometry. Anal. Chem. 80, 2583–2589(2008)

6. Liu, C., Hofstadler, S.A., Bresson, J.A., Udseth, H.R., Tsukuda, T.,Smith, R.D., Snyder, A.P.: On-line dual microdialysis with ESI-MS fordirect analysis of complex biological samples and microorganism lysates.Anal. Chem. 70, 1797–1801 (1998)

7. Bauer, K.-H., Knepper, T.P., Maes, A., Schatz, V., Voihsel, M.: Analysisof polar organic micropollutants in water with ion chromatography–electrospray mass spectrometry. J. Chromatogr. A. 837, 117–128 (1999)

8. Huber, C.G., Buchmeiser, M.R.: On-line cation exchange for suppressionof adduct formation in negative-ion electrospray mass spectrometry ofnucleic acids. Anal. Chem. 70, 5288–5295 (1998)

9. Jiang, Y., Hofstadler, S.A.: A highly efficient and automated method ofpurifying and desalting PCR products for analysis by electrospray ioni-zation mass spectrometry. Anal. Biochem. 316, 50–57 (2003)

10. Batchelor, J.D., Doucleff, M., Lee, C.-J., Matsubara, K., De Carlo, S.,Heideker, J., Lamers, M.H., Pelton, J.G., Wemmer, D.E.: Structure and

regulatory mechanism of Aquifex aeolicus NtrC4: variability and evolu-tion in bacterial transcriptional regulation. J. Mol. Biol. 384, 1058–1075(2008)

11. Batchelor, J.D., Sterling, H.J., Hong, E., Williams, E.R., Wemmer, D.E.:Receiver domains control the active-state stoichiometry of Aquifexaeolicus σ 54 activator NtrC4, as revealed by electrospray ionization massspectrometry. J. Mol. Biol. 393, 634–643 (2009)

12. Nagaraj, N., Lu, A., Mann, M., Wiśniewski, J.R.: Detergent-based butgel-free method allows identification of several hundred membrane pro-teins in single LC-MS runs. J. Proteome Res. 7, 5028–5032 (2008)

13. Flick, T.G., Merenbloom, S.I., Williams, E.R.: Anion effects on sodiumion and acid molecule adduction to protein ions in electrospray ionizationmass spectrometry. J. Am. Soc. Mass Spectrom. 22, 1968–1977 (2011)

14. Pan, P., Gunawardena, H.P., Xia, Y., McLuckey, S.A.: Nanoelectrosprayionization of protein mixtures: solution pH and protein p I. Anal. Chem.76, 1165–1174 (2004)

15. Jackson, A.U., Talaty, N., Cooks, R.G., Van Berkel, G.J.: Salt toleranceof desorption electrospray ionization (DESI). J. Am. Soc.Mass Spectrom.18, 2218–2225 (2007)

16. Chen, H., Venter, A., Cooks, R.G.: Extractive electrospray ionization fordirect analysis of undiluted urine, milk and other complex mixtureswithout sample preparation. Chem. Commun. 19, 2042–2044 (2006)

17. Cho, Y.-T., Huang, M.-Z., Wu, S.-Y., Hou, M.-F., Li, J., Shiea, J.: Usingelectrospray laser desorption ionization mass spectrometry to rapidlyexamine the integrity of proteins stored in various solutions. Anal.Bioanal. Chem. 406, 577–586 (2014)

18. Sampson, J.S., Hawkridge, A.M., Muddiman, D.C.: Development andcharacterization of an ionization technique for analysis of biologicalmacromolecules: liquid matrix-assisted laser desorption electrospray ion-ization. Anal. Chem. 80, 6773–6778 (2008)

19. Ekelöf, M., McMurtrie, E.K., Nazari, M., Johanningsmeier, S.D.,Muddiman, D.C.: Direct analysis of triterpenes from high-salt fermentedcucumbers using infrared matrix-assisted laser desorption electrosprayionization (IR-MALDESI). J. Am. Soc. Mass Spectrom. 28, 370–375(2017)

20. Nemes, P., Vertes, A.: Laser ablation electrospray ionization for atmo-spheric pressure, in vivo, and imaging mass spectrometry. Anal. Chem.79, 8098–8106 (2007)

21. Huang, M.-Z., Yuan, C.-H., Cheng, S.-C., Cho, Y.-T., Shiea, J.: Ambientionization mass spectrometry. Annu. Rev. Anal. Chem. 3, 43–65 (2010)

22. Brady, J.J., Judge, E.J., Levis, R.J.: Mass spectrometry of intact neutralmacromolecules using intense non-resonant femtosecond laser vaporiza-tion with electrospray post-ionization. Rapid Commun. Mass Spectrom.23, 3151–3157 (2009)

23. Sistani, H., Karki, S., Archer, J.J., Shi, F., Levis, R.J.: Assessment ofreproducibility of laser electrospray mass spectrometry using electrospraydeposition of analyte. J. Am. Soc. Mass Spectrom. 28, 880–886 (2017)

24. Brady, J.J., Judge, E.J., Levis, R.J.: Analysis of amphiphilic lipids andhydrophobic proteins using nonresonant femtosecond laser vaporizationwith electrospray post-ionization. J. Am. Soc. Mass Spectrom. 22, 762–772 (2011)

25. Flanigan IV, P.M., Perez, J.J., Karki, S., Levis, R.J.: Quantitative mea-surements of small molecule mixtures using laser electrospray massspectrometry. Anal. Chem. 85, 3629–3637 (2013)

26. Perez, J.J., Flanigan IV, P.M., Karki, S., Levis, R.J.: Laser electrospraymass spectrometry minimizes ion suppression facilitating quantitativemass spectral response for multicomponent mixtures of proteins. Anal.Chem. 85, 6667–6673 (2013)

27. Karki, S., Sistani, H., Archer, J.J., Shi, F., Levis, R.J.: Isolating proteincharge state reduction in electrospray droplets using femtosecond laservaporization. J. Am. Soc. Mass Spectrom. 28, 470–478 (2017)

28. Wang, G., Cole, R.B.: Mechanistic interpretation of the dependence ofcharge state distributions on analyte concentrations in electrospray ioni-zation mass spectrometry. Anal. Chem. 67, 2892–2900 (1995)

29. Frey, B.L., Lin, Y., Westphall, M.S., Smith, L.M.: Controlling gas-phasereactions for efficient charge reduction electrospray mass spectrometry ofintact proteins. J. Am. Soc. Mass Spectrom. 16, 1876–1887 (2005)

30. Verkerk, U.H., Kebarle, P.: Ion-ion and ion-molecule reactions at thesurface of proteins produced by nanospray. Information on the number ofacidic residues and control of the number of ionized acidic and basicresidues. J. Am. Soc. Mass Spectrom. 16, 1325–1341 (2005)

31. Liu, J., Konermann, L.: Cation-induced stabilization of protein complexesin the gas phase: mechanistic insights from hemoglobin dissociationstudies. J. Am. Soc. Mass Spectrom. 25, 595–603 (2014)

1010 S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS

32. Susa, A.C., Xia, Z., Williams, E.R.: Small emitter tips for native massspectrometry of proteins and protein complexes from nonvolatile buffersthat mimic the intracellular environment. Anal. Chem. 89, 3116–3122(2017)

33. Brady, J.J., Judge, E.J., Levis, R.J.: Nonresonant femtosecond laservaporization of aqueous protein preserves folded structure. Proc. Natl.Acad. Sci. U. S. A. 108, 12217–12222 (2011)

34. Pace, C.N., Trevino, S., Prabhakaran, E., Scholtz, J.M.: Protein structure,stability and solubility in water and other solvents. Philos Trans R SocLondon B: Biol Sci. 359, 1225–1235 (2004)

35. Pinho, S.P., Macedo, E.A.: Solubility of NaCl, NaBr, and KCl in water,methanol, ethanol, and their mixed solvents. J. Chem. Eng. Data. 50, 29–32 (2005)

36. Konermann, L., Ahadi, E., Rodriguez, A.D., Vahidi, S.: Unraveling themechanism of electrospray ionization. Anal. Chem. 85, 2–9 (2012)

37. Karki, S., Flanigan IV, P.M., Perez, J.J., Archer, J.J., Levis, R.J.: Increas-ing protein charge state when using laser electrospray mass spectrometry.J. Am. Soc. Mass Spectrom. 26, 706–715 (2015)

38. O’Brien, J.T., Williams, E.R., Holman, H.-Y.N.: Ambient infrared laserablation mass spectrometry (AIRLAB-MS) of live plant tissue withplume capture by continuous flow solvent probe. Anal. Chem. 87,2631–2638 (2015)

39. Stolee, J.A., Vertes, A.: Toward single-cell analysis by plume collimationin laser ablation electrospray ionization mass spectrometry. Anal. Chem.85, 3592–3598 (2013)

S. Karki et al.: Direct Protein Analysis from High-Salt Solution Using LEMS 1011