Anal. Chem. 2002, 74, 489-496
Accelerated Articles
MALDI-TOF-MS Analysis of Droplets Prepared in an Electrodynamic Balance: “Wall-less” Sample Preparation Michael J. Bogan and George R. Agnes*
Department of Chemistry, Simon Fraser University, Burnaby, BC, V5A 1S6, Canada
Methodology enabling mass spectral analysis of the composition of droplet(s) prepared in an electrodynamic balance (EDB) by matrix-assisted laser desorption/ionization (MALDI) is described. The dc field surrounding the electrodynamic balance was manipulated to eject single droplets at a time from the EDB thereby causing their deposition onto a MALDI sample plate precoated with matrix. When the laser was directed onto the droplet(s) and held stationary, marked gains in the signal-to-noise and signal-to-background ratios were realized with each subsequent mass spectrum due to the suppression of matrix cluster ion formation. Optical microscopy of the plate, after 1024 laser shots were fired at eight droplets that had been deposited one on top of the other, revealed a residual island of droplet matter (area ∼3.1 × 10-9 m2) inside the region where the crystalline matrix had been ablated away within the laser spot (area ∼1.6 × 10-8 m2). Removing the predried crystalline matrix layer and, instead, adding matrix into the starting solution was found to be a more effective means of suppressing matrix cluster ion formation. The chemical composition of the droplet(s) prepared in the EDB is discussed with regard to sample preconcentration, the images of the laser spot after MALDI, matrix cluster ion suppression, and the possibility for improved quantitation and detection limits by MALDI-TOF-MS.
Sensitivity improvements for matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOF-MS) have been reported through the use of ink-jet droplet generators.1-4 These microdispensing devices are capable of creating small * Corresponding author: (e-mail)
[email protected]; (phone) (604) 291-4387; (fax) (604) 291-3765. 10.1021/ac015638n CCC: $22.00 Published on Web 12/22/2001
© 2002 American Chemical Society
sample spots that reduce the need to search for hot spot areas within a large sample spot. Because small sample spot size is a necessity for realizing maximum sensitivity in MALDI-TOF-MS,5-7 efforts have been made to reduce the spread of sample material on a surface by confining the dispensed droplets into micromachined wells and increasing the rate of solvent evaporation by directing a stream of N2 over the wells or even heating the MALDI plate.4 The driving force behind such endeavors is that subattomole sensitivity and the ability to manipulate sample material before and after the separation steps remain key factors that are restricting the developmental pace of mass spectrometry-based methodologies for proteomics.8 A different approach would be to allow solvent evaporation in the wall-less environment of a levitated droplet of sample material to decrease the size of the droplet before deposition. This would directly preconcentrate the material in the droplet and create smaller sample spot sizes. Utilizing this evaporation step, which occurs in a matter of seconds, would save a great deal of time because it replaces the need to create micromachined surfaces or perform any postdeposition procedures. To achieve this, we inserted an electrodynamic balance between the droplet generator and the deposition plate to temporarily levitate droplets while their solvent evaporated. Features of this technology are its ability to (1) Ekstrom, S.; Ericsson, D.; Onnerfjord, P.; Bengtsson, M.; Nilsson, J.; MarkoVarga, G.; Laurell, T. Anal. Chem. 2001, 73, 214-219. (2) Ekstrom, S.; Onnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; MarkoVarga, G. Anal. Chem. 2000, 72, 286-293. (3) Little, D.; Cornish, T.; O’Donnell, M.; Braun, A.; Cotter, R.; Koster, H. Anal. Chem. 1997, 69, 4540-4546. (4) Onnerfjord, P.; Nilsson, J.; Wallman, L.; Laurell, T.; Marko-Varga, G. Anal. Chem. 1998, 70, 4755-4760. (5) Schuerenberg, M.; Luebbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.; Nordhoff, E. Anal. Chem. 2000, 72, 3436-3442. (6) Jespersen, S. Rapid Commun. Mass Spectrom. 1994, 8, 581-584. (7) Keller, B. O.; Li, L. J. Am. Soc. Mass Spectrom. 2001, 12, 1055-1063. (8) Mann, M.; Hendrickson, R.; Pandey, A. Annu. Rev. Biochem. 2001, 70, 437-473.
Analytical Chemistry, Vol. 74, No. 3, February 1, 2002 489
control the number of droplets levitated (sample loading) and the droplet levitation period (extent of desolvation), its ability to precisely deposit droplets onto a spatially well-defined position on a MALDI sample plate, and its glycerol-induced suppression of matrix cluster ions that facilitates analysis of molecules of 1 background matrix ions. Note that the most intense signal in Figure 6C is due to the sodiated adduct of acetone. This peak arose because the plate was washed with acetone. By simply washing with deionized water and air-drying, this peak as well as the peak assigned as [CH3COOR + Na+ + CH3COCH3] can be eliminated. These results suggest that the formation of background matrix cluster ions with two or more matrix molecules arises primarily from regions of crystallized matrix molecules. The signal intensity of such ions were dramatically reduced by adding glycerol and matrix to the starting solution, so that, in the deposited droplet, there was less chance for matrix crystallization. This observation has utility in the detection of small molecules by MALDI-TOFMS, because it effectively removes matrix cluster ions that otherwise dominate the background or cause chemical interference.43,47 The poor reproducibility of MALDI, due to the nonuniformity of the matrix/analyte crystallization process creating a highly variable fine structure in the target surface, is known as a major limiting factor in the use of MALDI in quantitative determinations.48 Upon laser irradiation, a thin crystalline layer of matrix suffers from localized depletion of matrix, rendering that area unable to further desorb analyte. Ring et al. have shown that, relative to a solid crystalline matrix layer, a matrix solution provides a more reproducible signal with laser shot number.49 By using the EDB to prepare droplets for analysis, the glycerolmatrix solution formed provides a much more uniform matrix from which to desorb. A total of 1087 laser shots were fired at the residue of the six droplets in Figure 6B before the S/N decayed below 10. The large number of microscans collected from the small amount of material in the collection of six droplets was a consequence of the fluid matrix present in the microspots. By analyzing liquid microspots, we achieved a sensitive and stable source of ions by MALDI. We speculate that, with further investigation into the effects of residual droplet size, matrix content, analyte, and reduction of the diameter of the laser spot, we can achieve lower absolute detection limits and improved quantitation with MALDI-TOF-MS by using the deposition of droplets from an EDB as the sample
preparation procedure. As an example of what may be expected from this technology, Figure 6B represents the average of consecutive laser shots 318-573 whereas the full spectrum in Figure 6C was the average of laser shots 610-865. The inset of Figure 6C shows that the sodiated ester, with S/N and S/B of 60 and 95 respectively, was still detectable after a large number of laser shots had been fired at the droplet residues. During the preparation of this paper, Keller and Li demonstrated the detection of 25 000 molecules of substance P by MALDI-TOF-MS.7 Their investigations into the fundamental detection limits in MALDI suggested that if the sample spot area is further decreased while maintaining g5 analyte molecules/µm2, the detection limit will be lowered. To do this using his nanoliter chemistry station,31,32 smaller capillary sizes can be used to create smaller droplet sizes. As they point out, handling of picoliter volumes becomes problematic in smaller inner diameter capillaries because of the higher surface-to-volume ratio that leads to stronger tension forces. We believe that the EDB provides a solution to this problem by offering a wall-less sample preparation procedure that is not limited by capillary tension forces. Work in progress in our laboratory has shown that systematic changes in the composition of the starting solution can lead to the formation of levitated droplets that have decreased size and also that solid particles of matrix and analyte can also be formed in the EDB with subsequent analysis by MALDI-TOF-MS. Detection of ∼10 amol of myoglobin in a single droplet (data not shown) has been achieved and further investigations of the application of wall-less sample preparation to protein analysis are ongoing.
(47) Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243-246. (48) Duncan, M. W.; Matanovic, G.; Cerpa-Poljak, A. Rapid Commun. Mass Spectrom. 1993, 7, 1090-1094. (49) Ring, S.; Rudich, Y. Rapid Commun. Mass Spectrom. 2000, 14, 515-519.
Received for review September 28, 2001. Accepted December 6, 2001.
496
Analytical Chemistry, Vol. 74, No. 3, February 1, 2002
CONCLUSIONS We have demonstrated controlled manipulations of droplets that were levitated in an EDB. Deposition of the droplets onto a spatially well-defined area enabled measurement of their solute composition by MALDI-TOF-MS. The addition of glycerol to the starting solution caused an increase in S/N and S/B with increasing laser shot numbers (up to at least 4000 laser shots) because background matrix cluster ion formation was suppressed by the transformation of crystalline matrix into a glycerol-matrix solution. Removal of the predried matrix layer on the MALDI plate and the addition of matrix to the starting solution proved a more effective means of suppressing matrix cluster ion formation at low laser shot numbers because the source of crystalline matrix was removed. Further investigation of the droplet composition on sensitivity and quantitation is underway. In terms of future applications, sample preparation using an EDB may find utility in interfacing separation methods to MALDI-TOF-MS because of the inherent sample preconcentration mechanism. ACKNOWLEDGMENT Financial support from NSERC is gratefully acknowledged. The authors thank Mario Pinto for the use of the MALDI-TOF-MS instrument, Zuo Ye for the use of the optical microscope, and Dev Sharma for the use of the digital camera.
AC015638N
