Solvent Assisted Inlet Ionization: An Ultrasensitive New Liquid

Apr 29, 2011 - A new inlet ionization method requiring no voltage or laser, and using water, methanol, or water/organic solvent mixtures, is shown to ...
0 downloads 7 Views 1MB Size
LETTER pubs.acs.org/ac

Solvent Assisted Inlet Ionization: An Ultrasensitive New Liquid Introduction Ionization Method for Mass Spectrometry Vincent S. Pagnotti, Nicholas D. Chubatyi, and Charles N. McEwen* University of the Sciences, Philadelphia, Pennsylvania 19104, United States ABSTRACT: A new inlet ionization method requiring no voltage or laser, and using water, methanol, or water/organic solvent mixtures, is shown to produce mass spectra similar to those obtained with electrospray ionization (ESI) for small molecules, peptides, and proteins, at least as large as carbonic anhydrase, with sensitivity that surpasses ESI. With the use of wide mass range acquisitions at 100 000 mass resolution on an Orbitrap Exactive, detection limits below parts per trillion are obtained for small molecules such as arginine, ciprofloxacin, and acetaminophen. Low attomoles of bovine insulin consumed produced a multiply charged mass spectrum. Ions are generated, even using pure water as solvent, within the heated inlet tube linking atmospheric pressure with the first vacuum stage of the Orbitrap Exactive. The extremely high sensitivity observed at this early stage of solvent assisted inlet ionization (SAII) development suggests that inlet ionization may surpass nanoelectrospray in sensitivity but without the need for extremely low solvent flows.

he first inlet ionization method, introduced by Trimpin et al. and named laserspray ionization (LSI),1,2 produced multiply charged ions for mass spectrometry (MS) analysis similar to electrospray ionization (ESI)3 but using laser ablation conditions similar to matrix-assisted laser desorption/ionization (MALDI).4 This discovery is analytically useful because of its sensitivity (attomoles for peptides and low femtomoles for small proteins)2 and the ability to obtain multiply charged ions from surfaces with high spatial resolution allowing imaging of lipids in tissue5 and proteins and peptides to be identified directly from tissue using electron transfer dissociation fragmentation and high mass resolution and mass accuracy.6 The initial mechanistic expectation, similar to ESI,7,8 was that there are two ionization events in LSI.1 The first is initiated either by the energy imparted to the matrix by the laser (photoionization), as proposed for MALDI,9 or that the laser generated thermal expansion of the matrix produces matrix/analyte droplets that are charged during a fracturing process.10,11 Multiply charged ions were thought to be produced from the charged matrix/analyte droplets in a second process by a mechanism similar to the residue7 or ion evaporation8 models proposed for ESI.12 It has subsequently been discovered that the role of the laser is to transfer particles or droplets of the matrix/analyte sample into the heated atmospheric pressure (AP) to vacuum inlet transfer tube that normally links the ionization region to the mass analyzer.13 It is in this heated pressure drop inlet region where the initial ionization occurs and thus the term “inlet ionization”. Any physical means of transferring the matrix/analyte sample into the heated pressure drop region, including ultraviolet, visible, or infrared lasers, a sonifier, laser induced acoustic desorption, a center punch to produce an acoustic wave, and directly with, for example, a spatula also produces multiply charged ions.13 The preferred matrix materials are 2,5-dihydroxybenzoic acid

T

r 2011 American Chemical Society

(2,5-DHB) and 2,5-dihydroxyacetophenone (2,5-DHAP) which are also important matrix materials in MALDI. This matrix assisted inlet ionization (MAII) method is found to produce identical high sensitivity mass spectra as LSI. Thus, LSI was found to be a subset of MAII. Here, we report that inlet ionization is applicable to solution samples. Mass spectra of peptides, proteins, and small molecules similar to ESI, LSI, and MAII are obtained without use of voltage by introducing the solution into the inlet orifice of a heated transfer tube of a mass spectrometer. This method eliminates ion losses associated with transferring ions from atmospheric pressure to vacuum14 and has a sensitivity for those compounds studied (insulin and several randomly selected small molecules) that is much superior to the sensitivity achieved by ESI at the same flow rates. Flow rates between 2 and 50 μL min1 have been investigated. The characteristic strong dependence of ion abundance on the inlet capillary temperature found with LSI1,12 and MAII is also observed with solvent assisted inlet ionization (SAII). These results demonstrate that analyte introduced directly into the heated inlet transfer tube separating the AP region from the vacuum of the mass analyzer results in generation of multiply charged mass spectra of peptides and proteins and singly charged ions of small molecules consuming attomoles to zeptomoles of analyte without resorting to nanoliter flow rates associated with nanoelectrospray ionization.15

’ EXPERIMENTAL SECTION The mass spectrometer used for these studies is a Thermo Fisher Scientific (Bremen, Germany) Orbitrap Exactive. For the Received: March 2, 2011 Accepted: April 29, 2011 Published: April 29, 2011 3981

dx.doi.org/10.1021/ac200556z | Anal. Chem. 2011, 83, 3981–3985

Analytical Chemistry ESI and sonic spray ionization (SSI) studies, the heated ESI (HESI) probe was used. In SSI, a melting point tube held in place using an atmospheric solids analysis probe (ASAP), M&M Mass Spec, Hockessin, DE, was used as an obstruction to increase the ion abundance. The Ion Max ion source was removed and the interlock defeated to allow easy access to the inlet aperture of the ion transfer tube for SAII studies. Bovine insulin, arginine, acetonitrile, and Chromasolv HPLC water were purchased from Sigma Aldrich, St. Louis, MO. Ciprofloxacin HCl was obtained from retained chemicals at University of the Sciences and is of unknown origin. In all experiments reported here, fused silica was used to transport solution from polypropylene vials into the heated ion transfer tube linking the AP ion source region with the first vacuum stage. About 5 cm of polyimide coating was burned from the end of the tube that protruded into the inlet capillary. The internal diameter and length of the fused silica capillary as well as its temperature and the distance it protrudes inside the ion transfer tube determine the flow rate of the solution. For the studies here, flow rates of 1048 μL min1 were used as indicated, but flow rates as low as 2 μL min1 were successful. Two x,y-stages taken from old microscopes were combined to allow precise positioning of the fused silica tube within the inlet capillary. The fused silica was taped to one stage. For the high-sensitivity work, the position of the capillary was critical. Aligning the fused silica capillary against the wall of the inlet enhanced the ionization. The transfer tube linking AP and the first vacuum stage was heated to 425 °C for peptides and proteins and 375 °C for small molecules. The instrument tune files used were obtained from tune files using ESI conditions. The sensitivity test with ciprofloxacin used solutions prepared at 830 nM, 83 nM, 8.3 nM, 830 pM, 83 pM, 8.3 pM, and 830 fM and a blank. For the arginine sensitivity study, solutions were prepared at 6.8 ppm, 680 ppb, 68 ppb, 6.8 ppb, 680 ppt, 68 ppt, and 6.8 ppt and a blank. For the insulin sensitivity test using ESI, the instrument was tuned on the mass-to-charge (m/z) 1434 ion using a solution of 34.4 fmol μL1 in 1:1 acetonitrile/water with 0.1% formic acid (FA) at a flow rate of 10 μL min1. Afterward, a 1:1 solution of ACN/water, 0.1% FA, was infused for 5 h to achieve background runs with undetectable amounts of insulin. In order to match the flow rate of SAII, the ESI solution was infused at 10 μL min1 using a Chemyx (Thermo-Fisher scientific) syringe pump. In studies involving ion abundances of the singly and doubly charged ions of angiotensin II, mass spectra were acquired for every 50 °C increase of the ion transfer tube from 50 to 450 °C using SSI, LSI, ESI, and SAII methods. SSI required a solution concentration 2 orders of magnitude higher (15 pmol μL1) than used with ESI or SAII. SSI-like conditions were achieved by using the maximum gas flow allowed for the nebulizing gas of the HESI probe and a solvent flow rate of 10 μL min1. A melting point tube held in place by the ASAP probe obstructed the SSI flow of nebulized liquid droplets producing an increase in analyte ion abundance. With LSI, 2,5-DHAP was used as matrix with ∼150 fmol of angiotensin II per 1 μL spot. A Spectra Physics, Newport, CA, VSL 337ND-S nitrogen laser was used to ablate the matrix/analyte mixture. The ion transfer tube temperature steps were the same for all methods.

’ RESULTS AND DISCUSSION In AP MS, ionization takes place at AP in an ion source enclosure and the formed ions are transferred through an aperture

LETTER

Figure 1. Mass spectra of bovine insulin at 100 000 mass resolution and a flow rate of 10 μL min1. Single 1 s acquisition (a) using ESI at an insulin concentration of 3.44 fmol μL1 in 50% acetonitrile in water with 0.1% formic acid and (b) using SAII at the same insulin concentration in 70% acetonitrile in water.

(skimmer or capillary) to the vacuum of the mass analyzer. Typically, only a small fraction of the ions produced are transferred through the inlet aperture of the skimmer or capillary.16 Recently, inlet ionization methods were introduced in which ionization occurs in the inlet pressure drop region between AP and vacuum.1,2,13 Ions are not formed until the sample passes the entrance inlet aperture. In the inlet ionization methods, analyte introduced into the heated AP to vacuum inlet transfer tube of a mass spectrometer produces ions without an external means of ionization such as voltage, a laser beam, or a supersonic gas flow. Ions are formed within the inlet transfer tube and the abundances reflect a strong dependence on the inlet tube temperature.12 With the incorporation of the analyte into a matrix such as 2, 5-DHAP or 2,5-DHB, mass spectra nearly identical to those observed with ESI are obtained for proteins, peptides, carbohydrates, lipids, and small molecules.5,1719 Astonishing sensitivity has been obtained for a number of compound ranging from small molecules to proteins using pure water, water organic solvent mixtures, and pure organic solvents. Solid matrix materials are now found not to be required to obtain ESI-like mass spectra using inlet ionization. A scheme of the 3982

dx.doi.org/10.1021/ac200556z |Anal. Chem. 2011, 83, 3981–3985

Analytical Chemistry

LETTER

Figure 2. (a) SAII mass spectrum of a single 0.1 s acquisition from a solution of 830 nM ciprofloxacin HCl in water infused at 10 μL min1. (b) Plot of ion count vs concentration using solutions of 830 zmol μL1, 8.3 amol μL, 83 amol μL, 830 amol μL1, 8.3 fmol μL1, 83 fmol μL1, and 830 fmol μL1. The r2 value is 0.999.

basics of this process is provided in the TOC graphic. The ultra high sensitivity is attributed to the ability to introduce the solution into the inlet in a controlled manner without sample loss. Ionization occurs within the capillary, and in these studies no voltage was applied. Ion losses incurred in transferring ions from AP through the inlet aperture using ESI20 or other AP ionization methods are eliminated in SAII. This translates to a method that effectively ionizes compounds at exceedingly low concentrations. The higher flow rates increase the ruggedness of the method relative to nanoESI. Figure 1 is a comparison of the mass spectra of insulin obtained on the Orbitrap Exactive using ESI and SAII. The limit of detection for a single acquisition in ESI using 50% acetonitrile/ water at a flow rate of 10 μL min1 and one acquisition per second (100 000 mass resolution, full width half height (fwhh), m/z 200) is ∼3.44 fmol μL1 (Figure 1a). The instrument was tuned to give the best ESI results at a flow rate of 10 μL min1. SAII mass spectra were obtained at the same flow rate using the ESI tune conditions and the same solution concentration. The mass spectra of a single 1 s acquisition at the same mass resolution and covering the same mass range (m/z 1000  2000) used for ESI is shown for SAII in Figure 1b. The limit of detection for insulin using SAII is low attomoles using subpicomolar solution. A current drawback of SAII for proteins and peptides is sodium adduction. The degree of sodium adduction with small and medium sized proteins is greater than the number of charges suggesting that Naþ replaces some acidic protons. Addition of organic solvents and first flowing 10% acetic acid through the SAII inlet tube followed by a quick wash with water and then the sample solution is often sufficient to suppress sodium adduction for a few scans as is seen in the mass spectrum for insulin in

Figure 1. Proteins such as lysozyme (MW 14 300), myoglobin (MW 16 950), and carbonic anhydrase (MW 29 000) are also readily observed using SAII even though numerous isotope clusters containing differing amounts of sodium for each charge state are observed. Another issue is that salts or strong acids and bases strongly suppress ionization. Finally, this method is so sensitive that frequently low and even submicromolar concentration can result in signal suppression similar to that observed with ESI at high micromolar concentrations. Methods are being developed to alleviate these issues. Small molecules are also ionized by SAII and with exceptional sensitivity for compounds with basic functionality. Sodium cation adduction has not been an issue with the small molecules tested thus far. Only 14 fmol of ciprofloxacin produced the mass spectrum shown in Figure 2 giving 108 ion counts for the MHþ ion. A linear calibration curve (r2 = 0.999) was obtained for solutions of ciprofloxacin ranging from 830 fM (830 zmol μL1) to 830 nM (830 fmol μL1) in water on the Orbitrap Exactive acquired over a mass range from m/z 200500. The amino acid arginine produced a similar result with a limit of detection of a few parts per trillion (ppt) in water. Selected reaction monitoring is expected to achieve limits of detection in the sub zeptomole or sub ppt range for basic small molecules. In its simplest form, SAII requires no high voltage, laser, or other external ionization means such as a supersonic gas flow. Sample can be introduced to the heated inlet using the amount of solution entrapped in the eye of a sewing needle by simply placing the needle eye near the ion entrance aperture. Needles with larger eyes produce more abundant ionization. For continuous ionization, a fused silica capillary with the polyimide coating burned off the end can be inserted into the inlet aperture of the 3983

dx.doi.org/10.1021/ac200556z |Anal. Chem. 2011, 83, 3981–3985

Analytical Chemistry

LETTER

Figure 3. Plots of capillary inlet temperature of an Orbitrap Exactive mass spectrometer vs relative ion abundances of the singly ([) and doubly (9) charged ions for angiotensin II plotted over the temperature range of 50450 °C: sonic spray ionization (SSI), solvent assisted inlet ionization (SAII), laserspray ionization (LSI), and electrospray ionization (ESI).

heated transfer capillary. Simply dipping the AP end of the fused silica capillary into a solution will result in a flow of liquid through the fused silica and into the inlet tube driven by the pressure differential on the capillary ends. Exceptional sensitivity has been obtained with flow rates between 2 and 50 μL min1. This flow rate range will allow reasonably fast liquid chromatography without flow splitting. Ionization clearly requires the AP to vacuum inlet transfer capillary be heated similar to MAII and LSI. Figure 3 shows graphs of inlet temperature vs the ion abundances for the singly and doubly charged ions of angiotensin II obtained with 10 μL min1 flow rates with SSI, SAII, LSI, and ESI. In order to duplicate the SSI results obtained using the HESI probe with maximum nitrogen gas flow, ESI conditions used lower than normal desolvation. The increase in ion abundance with increasing temperature for SSI and ESI is primarily attributed to improved desolvation. Under high desolvation conditions, the heated ion transfer tube has little effect on the ion abundance with ESI.13 Additionally, 2 orders of magnitude more analyte was required to obtain the SSI results even though an obstruction was used with SSI which produced a 100 improvement in analyte ion abundance. In all cases, the singly charged ions increased with increasing inlet temperature up to the maximum of 450 °C but the doubly charged ions increased with temperature and then decreased. ESI and SSI had substantial ion current at the lowest inlet temperature and underwent a modest increase with increasing temperature (600 for the doubly charged ions and 2500 for the singly charged ions) with increasing temperature. The doubly charged ions with both SAII

and LSI reached a maximum at around 350 °C and then declined. The singly charged ions had little abundance until about 275 °C and then increased dramatically to the maximum temperature of the inlet capillary (450 °C). Inlet temperature studies are compound specific but, for any single compound, comparison of the results clearly show that SAII is more similar to LSI, and thus to MAII, than to either ESI or SSI.

’ CONCLUSIONS Inlet ionization using analyte in solution, similar to ESI, is highly sensitive competing in sensitivity with nanoelectrospray but without the difficulty of handling extremely small volumes and low flow rates of solution. The SAII method also operates from water solutions with and without organic modifiers. No voltage, laser, or even an ion source enclosure is necessary making the method easy to implement. For many small molecules, low and even subppt detection is achieved with full mass range acquisitions and without use of separation devices to provide a concentration benefit. SAII is expected to be even more useful when combined with liquid separation methods. Insulin can be detected at low picomolar concentrations. At this early stage of development, the simplicity and sensitivity of this method without use of ion funnels or special optics suggests that inlet ionization methods will be an important addition to future mass spectrometry development. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. 3984

dx.doi.org/10.1021/ac200556z |Anal. Chem. 2011, 83, 3981–3985

Analytical Chemistry

LETTER

’ ACKNOWLEDGMENT The authors wish to gratefully acknowledge Professor Sarah Trimpin, Wayne State University, for her original ideas that lead to inlet ionization and for discussions related to solvent assisted inlet ionization and especially the role of surfaces in inlet ionization methods. This work is supported by the Richard E. Houghton endowment to the University of the Sciences. ’ REFERENCES (1) Trimpin, S.; Inutan, E. D.; Herath, T. N.; McEwen, C. N. Mol. Cell. Proteomics 2010, 9, 362. (2) Trimpin, S.; Inutan, E. D.; Herath, T. N.; McEwen, C. N. Anal. Chem. 2010, 82, 11. (3) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64. (4) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299. (5) Richards, A. L.; Lietz, C. B.; Trimpin, S. Rapid Commun. Mass Spectrom. 2011, 25, 815. (6) Inutan, E. D.; Richards, A. L.; Wager-Miller, J.; Mackie, K.; McEwen, C. N.; Trimpin, S. Mol. Cell. Proteomics 2011, 10, M110.000760. (7) Dole, M.; Mack, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L. D.; Alice, M. B. J. Chem. Phys. 1968, 49, 2240. (8) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287. (9) Knochenmuss, R. Analyst 2006, 131, 966. (10) Karas, M.; Kruger, R. Chem. Rev. 2003, 103, 427. (11) Vestal, M. L. Mass Spectrom. Rev. 1983, 2, 447. (12) McEwen, C. N.; Trimpin, S. Int. J. Mass Spectrom. 2011, 300, 167. (13) McEwen, C. N.; Pagnotti, V. S.; Inutan, E. D.; Trimpin, S. Anal. Chem. 2010, 82, 9164. (14) Covey, T. R.; Thomson, B. A.; Schneider, B. B. Mass Spectrom. Rev. 2009, 28, 870. (15) Smith, R. D.; Shen, Y.; Tang, K. Acc. Chem. Res. 2004, 37, 269. (16) Page, J. S.; Tang, K.; Kelly, R. T.; Smith, R. D. Anal. Chem. 2008, 80, 1800. (17) McEwen, C. N.; Larsen, B. S.; Trimpin, S. Anal. Chem. 2010, 82, 4998. (18) Inutan, E. D.; Trimpin, S. J. Am. Soc. Mass Spectrom. 2010, 21, 1260. (19) Inutan, E. D.; Trimpin, S. J. Proteome Res. 2010, 9, 6077. (20) Page, J. S.; Marginean, I.; Baker, E. S.; Kelly, R. T.; Tang, K.; Smith, R. D. J. Am. Soc. Mass Spectrom. 2009, 20, 2265.

3985

dx.doi.org/10.1021/ac200556z |Anal. Chem. 2011, 83, 3981–3985