Cryogenic Frozen Solution Matrixes for Analysis of DNA by Time-of

Sukanta Dolai , Poulami Dutta , Barry B. Muhoberac , Charles D. Irving , and Rajesh Sardar. Chemistry of .... Hua Lin , Joanna M. Hunter , Christopher...
0 downloads 0 Views 171KB Size
Anal. Chem. 1997, 69, 3608-3612

Cryogenic Frozen Solution Matrixes for Analysis of DNA by Time-of-Flight Mass Spectrometry Joanna M. Hunter,* Hua Lin, and Christopher H. Becker

GeneTrace Systems Inc., Menlo Park, California 94025

We present an alternative matrix system for mass analysis of high molecular weight biomolecules by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS). Single-stranded DNA oligomers are embedded in thin films of frozen solutions containing photoabsorbing substituted phenols. Ultraviolet laser desorption from these cryogenically cooled matrixes generates internally colder ions than from the equivalent room-temperature crystalline matrixes containing no solvent, as evidenced by a substantial reduction in the extent of fragmentation in the corresponding mass spectra of oligonucleotides. Furthermore, the homogeneity of the frozen solution samples allows analyte signals to be obtained uniformly throughout the sample. Sensitivity and mass resolution approaching that obtained using the most effective room-temperature crystalline matrix for DNA (3-hydroxypicolinic acid) are consistently and reproducibly obtained in both positive and negative ion mass spectra. As fragmentation is a principle factor limiting the maximum attainable mass resolution in MALDI-TOF mass analysis of DNA, this improvement in generating intact DNA ions is advantageous. Because a time-of-flight mass spectrometer has an almost infinite mass range, it is in principle a particularly appropriate tool for measuring the mass of high molecular weight biomolecules. Such molecules can be efficiently volatilized using a technique called matrix-assisted laser desorption/ionization (MALDI), in which the analyte molecules are embedded in a large molar excess of photoabsorbing matrix, usually crystals of small aromatic organic molecules. Upon laser excitation, the matrix molecules are rapidly heated and ejected into the gas phase, carrying analyte molecules into the expansion plume of molecules and ions. One widely accepted model for ionization of the neutral analyte molecules is gas-phase ion-molecule reactive collisions in the near-surface region, often via proton transfer.1,2 The matrix thus functions as an energy- and charge-transfer agent. An additional requirement for the matrix is that it must not react or interact strongly with the analyte and that the analyte be soluble in the matrix crystals. As a result, different analytes sometimes require different matrix systems. Using MALDI, proteins of up to several hundred thousand daltons are routinely detected by time-of-flight (TOF) mass spectrometry. However, because of their relatively fragile structure, DNA oligomers are more difficult to volatilize intact, and the efficiency for detection of these molecules at high masses is (1) Bo¨ckelmann, V.; Spengler, B.; Kaufmann, R. Eur. Mass Spectrom. 1995, 1, 81. (2) Ehring, H.; Karas, M.; Hillenkamp, F. Org. Mass Spectrom. 1992, 27, 472.

3608 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

currently much lower than that for proteins. As the length of the oligonucleotide increases, the mass resolution is degraded by widening kinetic energy spreads, prompt fragmentation, delayed fragmentation (metastable decay), and the formation of matrix adducts. Ideally, these deleterious effects may be reduced by constructing a matrix system that will produce cooler ions by minimizing the internal energy of the desorbed analyte and by lowering the binding energy of the analyte with its surrounding molecules. Among the few matrix molecules that have been found to desorb/ionize intact DNA, 3-hydroxypicolinic acid (3-HPA) is probably the most widely used.3 Using a matrix mixture of 3-HPA with picolinic acid, oligonucleotides have been detected which are greater than 500 bases (up to ∼200 kDa) in length.4,5 There are several reasons that matrix systems consisting of frozen films of volatile (at room temperature) solutions might be effective and useful for laser desorption/ionization of DNA oligomers. The primary reason is that frozen solution matrixes are more spatially homogeneous than crystalline matrixes, which often contain “sweet spots” from which the strongest analyte ion signal is obtained. Sample preparation techniques such as the “crushed crystal” method6 produce more homogeneous samples and protein ion signals than standard crystallized matrixes but are not as widely utilized for DNA mass analysis in 3-HPA.7 Frozen solutions minimize the dependence of the ion signal on analyte/matrix crystallization. Another reason is that excess small solvent molecules in the desorption plume might enhance collisional cooling of the desorbed analytes and minimize their decomposition.8 Furthermore, MALDI would be a more versatile technique if biomolecules could be analyzed in a more natural physiological state in aqueous solution. Previous studies have shown that it is indeed possible to desorb proteins and DNA from ice matrixes. In the first case, visible laser photons (581 nm) were used to desorb frozen aqueous DNA solutions from an absorbing copper oxide substrate.9 Although reproducibility was very poor, positive ion mass spectra of oligonucleotides free of fragmentation and adducts were recorded at desorption wavelengths of 578 and 589 nm.10 A parallel study (3) Wu, K. J.; Steding, A.; Becker, C. H. Rapid Commun. Mass Spectrom. 1993, 7, 142. (4) Tang, K.; Taranenko, N. I.; Allman, S. L.; Chang, L. Y.; Chen, C. H. Rapid Commun. Mass Spectrom. 1994, 8, 727. (5) Liu, Y.-H.; Bai, J.; Liang, X.; Lubman, D. M.; Venta, P. J. Anal. Chem. 1995, 67, 3482. (6) Xiang, F.; Beavis, R. C. Rapid Commun. Mass Spectrom. 1994, 8, 199. (7) Dai, Y.; Whittal, R. M.; Li, L.; Weinberger, S. R. Rapid Commun. Mass Spectrom. 1996, 10, 1792. (8) Zhang, J.-Y.; Nagra, D. S.; Li, L. Anal. Chem. 1993, 65, 2812. (9) Nelson, R. W.; Rainbow, M. J.; Lohr, D. E.; Williams, P. Science 1989, 246, 1585. (10) Nelson, R. W.; Thomas, R. M.; Williams, P. Rapid Commun. Mass Spectrom. 1990, 4, 348. S0003-2700(97)00376-4 CCC: $14.00

© 1997 American Chemical Society

demonstrated that infrared laser ablation of frozen aqueous matrixes followed by ionization with a second laser is effective for analysis of small peptides.11 More recently, mass spectra have been obtained using pulsed infrared laser desorption (IR-MALDI). A sample consisting of hydrated lysozyme (MW 14 000) was found to be sufficient to desorb/ionize the intact protein,12 In this same work, signals obtained of proteins up to 30 kDa containing only water of hydration appear to be 1 order of magnitude less sensitive than conventional crystalline matrixes. Also using IR-MALDI, a matrix consisting of a suspension of activated carbon particles in frozen water was found to desorb intact non-covalently bound myoglobin.13 The present study was undertaken to investigate the effectiveness of volatile solvents in MALDI matrixes. The addition of some photoabsorbing substituted phenol molecules has been shown to significantly improve the quality of mass spectra of oligonucleotides in frozen solution matrix systems by reducing base-loss fragmentation.

obtained by this method are ∼3 mm in diameter and 100 µm thick. The temperature of the sample on the liquid nitrogen-cooled sample stage was measured by a chromel-alumel thermocouple attached with silver paste to the silicon surface. During data collection the sample temperature was maintained at ∼180 K. At this temperature, the thin layer of water that condensed on the surface of the cold sample during transfer through air into the vacuum was pumped away within several minutes before the samples were analyzed.

EXPERIMENTAL SECTION Some modifications have been implemented in the time-of-flight mass spectrometer previously described.14 The ion source was modified for pulsed delayed ion extraction. The sample stage was floated at 20 kV, and after some delay time (approximately several hundred nanoseconds, dependent on mass), ions were extracted by a 2.7 kV pulse (high-voltage switch from Directed Energy, Inc., Fort Collins, CO) and focused into a 2 m flight tube. Laser wavelengths of either 355 or 266 nm were employed for desorption/ionization. Either positive or negative ions were detected by switching the polarity of the potentials. The signal output from the dual microchannel plate detector was amplified and digitized. Spectra were collected using 5 ns time resolution and typically summed over 20 laser shots. Synthetic single-stranded mixed-base DNA was obtained from BioSource International (Menlo Park, CA). All matrix compounds were obtained from Aldrich (Milwaukee, WI) and used without further purification. Research grade solvents were used, and water was deionized and filtered. Solutions of the matrix additives were prepared at a concentration of 0.5 M or in saturated solution if the solubility of the additive was lower. To reduce alkali-metal adduct ion formation, 50 mM diammonium citrate was added.15 This buffers the sample solutions at pH ∼5. Aliquots of aqueous solutions of mixed-base single-stranded oligonucleotides were first evaporated in a vacuum evaporator to remove the water, and then the matrix solution was added to the dried DNA. The samples contained 0.25-2.5 pmol of DNA in 0.5 µL of liquid matrix solution. Samples were prepared by a quick-freezing method. A 0.5 µL of the sample solution was pipetted onto precooled (to ∼10 °C) silicon substrates mounted on a copper sample holder and then rapidly cooled by dipping into liquid nitrogen. The high rate of cooling minimized solvent evaporation. The frozen sample films

RESULTS AND DISCUSSION Several small, volatile heterocyclic nitrogen-containing ring molecules that are liquids at room temperature were initially examined as frozen matrixes for laser desorption/ionization of oligonucleotides.16 These molecules included the following: (1) pyridine, (2) pyrrole, (3) 3-pyrrolidinol, (4) pyrrolidinone, (5) piperidine, (6) thiazolidine, (7) pyrimidine, (8) pyrrolidine, and (9) pyridazine. 3-Nitrobenzyl alcohol was also tested because it has been previously used as a matrix for proteins.17 These molecules were investigated primarily because the nitrogen ring systems are structurally similar to nucleotide bases, and therefore they are perhaps compatible as matrixes. They were tested as frozen matrixes in both neat and mixed solutions with water (if miscible) to ensure DNA solubility, but no analyte signals were detected. Because several of these liquids are basic and do not have an available proton for DNA ionization, additional photoabsorbing charge-transfer species were added. Carboxylic acid compounds are unsuitable for this purpose because they form solid salts. Instead, substituted phenols were used. Although the ground-state acidity of the phenol compounds is typically much lower than the commonly used carboxylic acids, a characteristic that might reduce acid-catalyzed cleavage of a nucleobase during the sample preparation, the excited-state acidity is much higher.18 This may facilitate gas-phase proton-transfer ionization. Mass spectral analysis of frozen films of single-stranded DNA oligomers 36 bases in length (36-mer) in the liquid matrixes containing added phenols produced extremely weak positive and negative ion signals that were broadened by adducts and fragments of the molecular parent ion. In addition, strong featureless background ions were observed over a wide mass range, often prohibiting detection of analyte peaks. Lack of appreciable analyte ion signal from matrixes of frozen liquids) led to the investigation of a series of substituted phenols as additives to other volatile nonphotoabsorbing solvents such as acetonitrile, water, tetrahydrofuran (THF), ethanol, and methanol. Oligonucleotide signals were observed for several of the substitutedphenol additives frozen in a variety of these solvents. Table 1 lists the substituted-phenol molecules and tabulates relative signal intensities for a 36-mer in a 1:1 (volume) methanol/water (50 mM diammonium citrate). This solvent system was most often used because most of the phenols have adequate solubility in this solution and because this solution wets the substrate sufficiently

(11) Becker, C. H.; Jusinski, L. E.; Moro, L. Int. J. Mass Spectrom. Ion Processes 1990, 95, R1. (12) Berkenkamp, S.;Karas, M.; Hillenkamp, F. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 7003. (13) Alimpiev, S.; Chen, Y.-C.; Dratz, E.; Kraft, P.; Sunner, J. Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, May 12-16, 1996; p 641. (14) Wu, K. J.; Shaler, T. A.; Becker, C. H. Anal. Chem. 1994, 66, 1637. (15) Pieles, U.; Zu ¨ rcher, W.; Scha¨r, M; Moser, H. E. Nucleic Acids Res. 1993, 21, 3191.

(16) Hunter, J. M.; Lin, H.; Sannes, K. L.; Becker, C. H. In Ultrasensitive Biochemical Diagnostics; Cohn, G. E., Soper, S. A., Chen, C. H. W., Eds.; Proc. SPIE 1996, 2680, 384. (17) (a) Zhao, S.; Somayajula, K. V.; Sharkey, A. G.; Hercules, D. M.; Hillenkamp, F.; Karas, M.; Ingendoh, A. Anal. Chem. 1991, 63, 450. (b) Chan, T. W. D.; Thomas, I.; Colburn, A. W.; Derrick, P. J. Chem. Phys. Lett. 1993, 202, 93. (18) Ireland, J. F.; Wyatt, P. A. H. In Advances in Physical Organic Chemistry; Gold, V., Bethell, D., Eds.; Academic Press: New York, 1976; Vol. 12, p 131.

Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

3609

Table 1. Relative Intensity of 36-Mer Cation Signal in Frozen Solutions of 1:1 Methanol/Water with 50 mM Diammonium Citrate and 0.5 M Photoabsorbera photoabsorber 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

phenol hydroquinone catechol 4-nitrophenol 3-nitrophenol 2-nitrophenol 2-fluoro-4-nitrophenol 3-fluoro-4-nitrophenol 4-fluoro-2-nitrophenol 4-fluorophenol 4-iodophenol 4-cyanophenol 2-cyanophenol R,R,R-trifluorocresol 4-aminophenol 4-hydroxybenzamide 4-hydroxyacetophenone 2-hydroxyacetophenone 2,4-dihydroxyacetophenone 2,5-dihydroxyacetophenone 2,6-dihydroxyacetophenone 2,4,6-trihydroxyacetophenone 4-aminoacetophenone 4-hydroxy-3-nitrobenzaldehyde 2,5-dihydroxybenzoic acid 3-hydroxypicolinic acid 2-hydroxy-5-nitrobenzoic acid 4-hydroxybenzoic acid 3-hydroxybenzoic acid 2-hydroxybenzoic acid

266 nm

355 nm

++ + + + + +++ + ++

+++ ++ +++ ++ + + +

++

+++

+++ + + -

+++ ++ + -

a A saturated solution is used if the solubility is lower than 0.5 M. Key: -, no signal; +-, very weak; +, weak; ++, strong; +++, very intense.

to spread to a relatively thin (∼100 µm) final film thickness. Water is necessary to ensure DNA solubility, but the fraction can be varied over 20-80% (v) methanol with no significant change in the mass spectral ion signal. Several of these matrix solutions generated a uniformly intense oligonucleotide ion signal over the entire sample spot but are not useful matrixes for reasons such as too much adduct formation or fragmentation. In general, comparable positive and negative ion yields were obtained from these frozen liquid matrixes. Several of these compounds, including 4-nitrophenol,14,19 2,4,6trihydroxyacetophenone,15,20 2,6-dihydroxyacetophenone,21 and other hydroxyacetophenones,22 have previously been reported as room-temperature solid matrixes for oligonucleotides, but the DNA mass spectra obtained from frozen matrix solutions containing these molecules are significantly improved. An example of the dramatic difference between frozen solution and solid crystal matrixes is plotted in Figure 1. Shown in the figure are mass spectra of a mixed-base 36-mer comparing dried (crystalline) 2,4dihydroxyacetophenone matrix with 2,4-dihydroxyacetophenone frozen in a 1:1 (v) methanol/water matrix. The intensity of the (19) Gimon, M. E.; Preston, L. M.; Solouki, T.; White, M. A.; Russell, D. H. Org. Mass Spectrom. 1992, 27, 827. (20) Zhu, Y. F.; Chung, C. N.; Taranenko, N. I.; Allman, S. L.; Martin, S. A.; Haff, L.; Chen, C. H. Rapid Commun. Mass Spectrom. 1996, 7, 383. (21) Cohen, L. R. H.; Strupat, K.; Hillenkamp, F. Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, May 1216, 1996, p 1049. (22) Krause, J.; Stoeckli, M.; Schluneggar, U. P. Rapid Commun. Mass Spectrom. 1996, 10, 1927.

3610 Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

Figure 1. Positive ion time-of-flight mass spectra obtained from 2,4dihydroxyacetophenone matrixes containing 50 mM diammonium citrate: (a) saturated in 1:1 (v) methanol/water frozen solution and (b) room-temperature crystal from the same solution. The parent molecular ion is protonated single-stranded 36-mer [M + H]+. The doubly charged ion [M + 2H]2+ is also plotted. Total analyte loading is 2.5 pmol. Each spectrum is the sum of 20 laser pulses at 266 nm.

base-loss fragment ions in the spectrum obtained from the frozen matrix is significantly reduced relative to that in the spectrum obtained from the dried solid matrix. The relative ratio of fragment to parent peaks in the two cases persists from near to well above the laser pulse energy threshold for desorption/ ionization. When this same sample is warmed in vacuum to evaporate the solvent, the fragmentation reappears in the mass spectrum. This effect is extremely reproducible and is also observed for several other substituted-phenol additives to frozen solution matrixes, although it is not generally true for all of the compounds listed in Table 1. In MALDI, loss of nucleotide bases followed by phosphate backbone cleavage is thought to be the primary mechanism for dissociative cooling of the laser-heated DNA oligomers,23,24 and thus lower oligonucleotide internal energy would decrease the degree of fragmentation. The observed difference in the extent of fragmentation between frozen solution and crystalline matrixes is evidence that the solvent molecules contribute to dissipating excess desorption energy. There is no effect of sample stage temperature on mass spectra obtained from dried crystalline matrixes. Spectra obtained from crystalline samples contain the same level of DNA fragmentation when the sample stage temperature is varied between room temperature and 180 K, an observation that further supports the conclusion that the solvent (23) Nordhoff, E.; Karas, M.; Cramer, R.; Hahner, S.; Hillenkamp, F.; Kirpekar, F.; Lezius, A.; Muth, J.; Meier, C.; Engels, J. W. J. Mass Spectrom. 1995, 30, 99. (24) Zhu, L.; Parr, G. R.; Fitzgerald, M. C.; Nelson, C. M.; Smith, L. M. J. Am. Chem. Soc. 1995, 117, 6048.

molecules in frozen solution matrixes aid in cooling the analyte ions. There are several possible explanations for the observed decrease in fragmentation. First, the higher density of molecules desorbed from frozen matrixes, dominated by the light solvent molecules, likely increases the degree of adiabatic cooling in the desorption plume, thereby decreasing the fragmentation.25,26 Another possible explanation is that the desorption from volatile matrix solutions may require less energy per analyte molecule than more conventional low vapor pressure crystalline matrixes. In the desorption plume, this consequently means that the analyte internal energy would also be lower. The ratio of solvent/matrix/analyte in the prepared sample is typically about 107:105:1. Although the solvent slowly sublimes in the ∼10-7 Torr vacuum, mass spectra collected as a function of time after the sample is inserted into the vacuum chamber indicate that the ion signal is stable for more than 1 h, and no measurable increase in base-loss fragmentation is evident in this period of time. This observation suggests that there is a range of solvent/matrix/analyte ratios within which a high enough density of small molecules is desorbed to effect cooling of the analyte ions. Although collisional cooling is considered relatively inefficient, it appears that there is some cooling of the analyte ions in the presence of solvent molecules. A characteristic of other frozen solution matrixes containing hydroxyacetophenone compounds is relatively high mass resolution of the parent oligonucleotide ion (greater than 500 for a 36mer with a mass of ∼11 kDa), regardless of the extent of fragmentation or adduct formation. Typically, the resolution from these compounds is a factor of ∼2 times higher than that of the other phenol/solution matrixes. In many cases, even though hydroxyacetophenone frozen solution mass spectra exhibit less fragmentation than room-temperature crystalline samples, the oligonucleotide ions produced from these matrixes are still hot enough to undergo substantially more prompt fragmentation than in crystalline 3-HPA matrix. A drawback of the hydroxyacetophenone matrixes is that the analyte ion signal quickly disappears at a given spot, possibly resulting from photodegradation under UV laser irradiation. Although the sensitivity is not as high as that of crystalline 3-HPA matrix, negligible matrix and alkali metal adduct formation are characteristics that make hydroxyacetophenones appealing as frozen solution matrix additives for DNA analysis. Of the frozen phenol matrix additives tested, some of the most sensitive and versatile are the nitrophenols. Figure 2a illustrates a typical mass spectrum of a mixture of single-stranded 36-mer and 55-mer in a frozen 1:1 methanol/water solution containing 0.5 M 4-nitrophenol. Although there is some matrix adduct formation, there is almost no fragmentation, and the signal intensity is comparable to that obtained using a room-temperature 3-HPA crystalline matrix. In the spectrum obtained from roomtemperature crystalline 4-nitrophenol in the absence of solvent (Figure 2b), it can be seen that at a signal-to-noise level approximately equivalent to the frozen solution spectrum (Figure 2a), the base-loss fragments are at least 3 times greater and the peaks are sitting on a broad background. In frozen 4-nitrophenol matrix solutions, sensitivity to the 100 fmol level is easily attained for oligomers up to at least 65 bases in length. In addition, mass spectra of proteins up to 66 kDa are easily obtained from this (25) Vertes, A. Microbeam Anal. 1991, 1991, 25. (26) Zhang, J.-Y.; Davinder, S. N.; Li, L. Anal. Chem. 1993, 65, 2812.

Figure 2. Positive ion time-of-flight mass spectra of 2.5 pmol each of single-stranded 36-mer [36]+ and 21-mer [21]+ in (a) 0.5 M 4-nitrophenol in 1:1 (v) methanol/water frozen solution containing 50 mM diammonium citrate and (b) room-temperature crystalline 4-nitrophenol from the same solution. The parent molecular ions are protonated. Each spectrum is the sum of 20 laser pulses at 355 nm.

frozen solution matrix. However, the signal-to-noise ratios observed are not as good as those of conventional crystalline matrixes for proteins, e.g., 2,5-dihydroxybenzoic acid. One reason that these phenol compounds might be potentially useful matrixes is that many of them are much more acidic upon photoexcitation than in the ground state and thus might easily transfer a proton to the DNA. However, no strong correlation between phenol excited state acidity and analyte ion signal was discerned (at any threshold level), in agreement with the conclusions of previous studies.27,28 The behavior of 3-HPA and other commonly used roomtemperature carboxylic acid matrixes is much different from that of the phenols when these compounds are frozen in a solvent matrix. Frozen matrix solutions consisting of 0.5 M 3-HPA in 25% acetonitrile/water 50 mM diammonium citrate (as is used for room-temperature crystalline sample preparation) do not function to effectively desorb/ionize DNA. High levels of background ions typically dominate the mass spectra obtained from frozen solutions containing carboxylic acids. The ion signals obtained from 3-HPA frozen solution matrix are extremely weak and exhibit extensive addition of Na and K cations to the oligonucleotide. For oligonucleotides longer than ∼15 bases, the adducts degrade the parent ion signal to such an extent that the parent peak is not resolvable. It seems, therefore, that the carboxylic acid-type additives are generally less tolerant of impurities such as alkali metal salts than the phenol-type additives. (27) Fitzgerald, M. C.; Parr, G. R.; Smith, L. M. Anal. Chem. 1993, 65, 3204. (28) Nelson, C. M.; Zhu, L.; Tang, W.; Smith, L. M.; Crellin, K.; Berry, J.; Beauchamp, J. L. In Ultrasensitive Biochemical Diagnostics!; Cohn, G. E., Soper, S. A., Chen, C. H. W., Eds.; Proc. SPIE 1996, 2680, 247.

Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

3611

Although the microscopic physical structure of the frozen films is unknown and difficult to control and characterize, several observations lead to the conclusion that they do indeed contain a homogeneous distribution of analyte, at least in the surface region accessed by the laser pulse. It is well-known that when aqueous solutions undergo slow freezing, phase separation occurs. The pure water crystallizes first, forcing the remaining solution into areas of high analyte and matrix concentration which freeze at lower temperatures. On the other hand, a rapid rate of cooling will minimize the size of the pure ice microcrystals, resulting in greater film homogeneity.29 We therefore attempted to freeze the sample solutions as quickly as possible. Another way to inhibit the formation of large ice crystals during the freezing process is to add a small percentage of a “cryoprotector” such as glycerol.30 We observed that addition of a small percentage of glycerol (up to 5%) to the matrix solution does not significantly alter the analyte ion signal intensity or distribution obtained from the frozen film matrix. This result suggests that the uniformity of the samples without glycerol is likely similar to those with added cryoprotector. In the methanol/water frozen matrixes, it is possible that methanol functions as a cryoprotector. [As it is extremely difficult to reproducibly control the evaporation of volatile solvents before freezing the sample solution, the final volume fraction of methanol is certainly less than the initial volume fraction (50%).] Thus, we conclude that the rate of cooling attained by dipping samples into liquid nitrogen ensures minimal segregation of the matrix into crystallites and uniform dispersion of matrix and analyte on a lateral scale smaller than the laser spot size (less than ∼100 µm). The homogeneity of frozen film matrixes prepared in this manner is manifested in the uniformly intense analyte ion signal that can be obtained over a large area of the frozen solution samples. Such uniformity is a great advantage over crystalline solid matrixes, which typically have “sweet spots” where particularly strong analyte signal is collected, and pure ice matrixes, in which signal (29) Studer, D.; Michel, M.; Hunziker, E. B.; Buschmann, M. 13th Internation Congress on Electron Microscopy; Les Editions de Physique Les Ulis: Paris, 1994; p 719. (30) Boryak, O. A.; Kosevich, M. V.; Shelkovsky, V. S.; Blagoy, Y. P. Rapid Commun. Mass Spectrom. 1996, 10, 197.

3612

Analytical Chemistry, Vol. 69, No. 17, September 1, 1997

can only be obtained from the perimeter of the sample spot. CONCLUSIONS The reduction in the extent of oligonucleotide fragmentation in mass spectra of DNA desorbed from frozen solution matrixes containing substituted phenols is evidence that the DNA ions have lower internal energy than when desorbed from the corresponding room-temperature crystalline phenol matrixes. It is likely that solvent molecules co-desorbed from the frozen matrix contribute to dissipating excess energy acquired by the analyte molecules in the desorption/ionization process. The high density of small (solvent) molecules in the desorption plume increases the number of gas-phase collisions, possibly leading to enhanced jet cooling (adiabatic expansion). Alternatively, less energy may be required for desorption/ionization of DNA from frozen solutions. In addition, frozen thin film matrix samples are spatially homogeneous, unlike crystalline matrixes, and therefore generate highquality mass spectra of DNA oligomers throughout the sample area. These advantages demonstrate the potential utility for frozen solution matrixes in pulsed laser desorption. Reduction (or elimination) of oligonucleotide fragmentation may ultimately result in greater mass resolution and sensitivity for applications of MALDI-TOF mass spectrometry to DNA sequencing and sizing. ACKNOWLEDGMENT Financial support from the National Institutes of Health, National Human Genome Research Institute (Grant R01 HG00174) is gratefully acknowledged.

Received for review April 9, 1997. Accepted June 17, 1997.X AC9703764

X

Abstract published in Advance ACS Abstracts, August 1, 1997.