Anal. Chem. 1998, 70, 3757-3761
On-Probe Solid-Phase Extraction/MALDI-MS Using Ion-Pairing Interactions for the Cleanup of Peptides and Proteins Maria Esteban Warren, Adam H. Brockman,† and Ron Orlando*
Complex Carbohydrate Research Center and the Departments of Biochemistry & Molecular Biology and Chemistry, University of Georgia, 220 Riverbend Road, Athens, Georgia 30602-4712
Samples originating from biological sources often contain a complex mixture of inorganic salts, buffers, chaotropic agents, surfactants/detergents, preservatives, and other solubilizing agents. However, the presence of these contaminants virtually ensures the failure of any subsequent analysis of the sample by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). Sample cleanup procedures, therefore, must be performed prior to MALDI-MS analysis. This paper reports a probe-surface derivatization method that greatly simplifies this sample preparation process. MALDI probes possessing self-assembled monolayers (SAMs) terminated with ionic functional groups can rapidly extract peptides/ proteins via ionic interactions from e1-µL volumes of sample solutions placed directly on their surface. We have found that MALDI probes modified in this manner are a practical solution for analyzing very small volumes of peptide/protein solutions contaminated with high levels of inorganic salts, buffers, detergents, chaotropic agents, and other solubilizing agents. Contaminants from a complex mixture of inorganic salts, buffers, chaotropic agents, surfactants/detergents, preservatives, and other solubilizing agents are often found in samples that have been obtained from biological sources. Biologists employ these agents in preparing samples for many reasons, such as to maintain a nontoxic cellular environment, to stabilize solvated samples, and to preserve enzymatic or other biological activity. Although the presence of these agents is critical, they often cause problems for subsequent analyses, particularly within the domain of mass spectrometry. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) is one of the most sensitive MS approaches, is probably the MS procedure most compatible with biological buffers, and is now routinely used for biomolecular analysis.1-3 However, MALDI-MS is not immune to interferences from the complexity of the biological media and is often crippled by the †
Present address: Covance Laboratories Inc., 3301 Kinsman Blvd., Madison, WI 53704. (1) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshido, Y.; Yoshido, T. Rapid Commun. Mass Spectrom. 1988, 2, 151. (2) Hillenkamp, F.; Karas, M.; Beavis, R.; Chait, B. T. Anal. Chem. 1991, 63, 1193A. (3) Wang, R.; Chait, B. T. Curr. Opin. Biotechnol. 1994, 5, 77. S0003-2700(98)00210-8 CCC: $15.00 Published on Web 08/15/1998
© 1998 American Chemical Society
presence of inorganic salts, chaotropic agents, buffers, and detergents. For example, samples containing sodium at concentrations of 200 mM typically lead to signal suppression and the presence of multiple, intense sodium adducts.4 Since these adducts increase peak width and decrease resolution, desalting prior to MALDI analysis can increase mass accuracy. At significantly higher concentrations of sodium, no signal can be obtained from the sample. Similarly, in high concentrations of detergent, low signal yield and low resolution are also observed. It is currently thought that these agents prevent proper inclusion of the analyte into the matrix crystal lattice, prevent crystallization of the matrix totally, or compete for ionization and thereby lead to signal suppression. The primary obstacle encountered in obtaining a MALDI-MS spectrum of a biological sample is its cleanliness. Hence, the investigator is required to perform sample cleanup procedures before performing the analysis. Several sample preparation techniques have been designed specifically for MALDI-MS analysis of contaminated biological samples. One strategy relies upon a thin film of matrix deposited in acetone.5 The peptide sample is applied to the surface of the matrix film and desalted using aqueous organic acids. Another development involves a recrystallization method that incorporates more peptide molecules than the salt contaminants.6 The excluded salts can then be selectively washed away from the adhering crystals. Samples can also be freed of contaminants by depositing them onto polymeric membranes, such as porous polyethylene (PE)7,8 and poly(vinylidine difluoride) (PVDF).9,10 Again, the salts are washed from the sample, which itself is bound to the membrane. The “cleansed” sample and membrane are attached to the MALDI probe, layered with matrix solution, and the sample is analyzed. All of these methods have been incorporated into the arsenal of sample preparative methods available to the MALDI analyst and have become valuable tools to overcome the potential problems involved with MALDI analysis of unknown peptide samples. (4) Brockman, A. H.; Dodd, B. S.; Orlando, R. Anal. Chem. 1997, 69, 4716. (5) Vorm, O.; Mann, M. J. Am. Soc. Mass. Spectrom. 1994, 5, 955. (6) Xiang F.; Beavis, R. C. Rapid Commun. Mass Spectrom. 1994, 8, 199. (7) Blakledge, J. A.; Alexander, A. J. Anal. Chem. 1995, 67, 843. (8) Worrall, T. A.; Cotter, R. J.; Woods, A. S. Anal. Chem. 1998, 70, 750. (9) Vestling, M. M.; Fenselau, C. Anal. Chem. 1994, 66, 471. (10) Strupat, K.; Karas, M.; Hillenkamp, F.; Eckerskorn, C.; Lottspeich, F. Anal. Chem. 1994, 66, 464.
Analytical Chemistry, Vol. 70, No. 18, September 15, 1998 3757
Figure 1. Mechanistic illustration detailing the synthesis of ionpairing SPE/MALDI-MS probe.
We have developed an on-probe sample cleanup strategy based on chemically modifying the surface of a MALDI probe so that it is capable of extracting biological polymers from solution while having little or no affinity for the contaminants in the sample.4,11 We have named this approach SPE/MALDI-MS, since the probe derivatizations we use provide surfaces that are chemically similar to the media used in solid-phase extraction (SPE). For example, we have created MALDI probes with a hydrophobic self-assembled monolayer (SAM),4,11 a surface chemically similar to the C18 stationary phase commonly employed in SPE. These hydrophobic SPE/MALDI-MS probes are capable of binding peptides/ proteins via hydrophobic interactions and allow both inorganic salts and chaotropic agents to be removed with aqueous washes.4,11 A major limitation in hydrophobic SPE/MALDI-MS probes is the low degree of “wetability” of the surface. The hydrophobic nature of the surface combined with the aqueous solutions of the sample results in strong sample surface tension and the formation of a near-perfect sphere on the surface, all of which severely limit the amount of peptide that can be extracted.4 We, therefore, developed a better method to utilize these probes by immersing them in the sample solution for an extended period.4 The inherent limitation of the hydrophobic surfaces provided the impetus for the development of probes with “wettable” surfaces for SPE/MALDI-MS, since this would overcome the need to immerse the probe in the analyte solution. Here, we describe the use of MALDI probes possessing SAMs that are terminated with ionic functional groups. MALDI probes modified in this manner have excellent “wettability” character and can rapidly extract peptides/proteins via ionic interactions from e1-µL volumes of sample solutions placed directly on the modified probe. We have found that ion-pairing SPE/MALDI probes are practical for analyzing very small volumes of peptide/protein solutions contaminated with high levels of inorganic salts, buffers, detergents, chaotropic agents, and other solubilizing agents. EXPERIMENTAL SECTION Probe Preparation. The mechanistic approach for preparing ion-pairing SPE/MALDI-MS probes is shown in Figure 1. In general, “disposable” commercial stainless steel rings with eight (11) Brockman, A. H.; Orlando, R. Proceedings of the 44th ASMS Conference of Mass Spectrometry and Allied Topics, Portland, OR, May 12-16, 1996.
3758 Analytical Chemistry, Vol. 70, No. 18, September 15, 1998
“mesas” were manufactured to fit a MALDI-TOF probe tip (Hewlett-Packard, Palo Alto, CA). These were first etched with aqua regia [5.4 mL of 70% nitric acid (Aldrich, Milwaukee, WI) and 24.6 mL of 37% hydrochloric acid (Sigma, St. Louis, MO)]. The etching serves to increase the tip’s surface area and, thus, its binding capacity. While etching, the aqua regia is “stirred” through the probe tip using a glass pipet, followed by sputtering each ring with fine microcrystals of gold using a commercial sputtering instrument (Sputter Module Coater, West Chester, PA). SAMs were assembled in a similar fashion for each surface explored. The freshly sputtered gold tips were incubated for 30 min in an anhydrous ethanol solution saturated with the thiol of interest. Care was taken to stir the incubating tips using a shaker or rotator. After incubation, the tips were washed first in ethanol and then in water to remove unbound thiolates. In this manner, strong cation-pairing (3-mercapto-1-propanesulfonic acid and 2-mercaptoethanesulfonic acid), strong anion-pairing (2-aminoethanethiol hydrochloride), and weak cation-pairing (3,3′-dithiodipropionic acid) monolayers were all assembled to further explore the separation properties of each. The SAM-forming species 3-mercapto-1-propanesulfonic acid (MPSA or “sulfonate-terminated”) and 2-aminoethanethiol hydrochloride (referred to as “amine-terminated”) were purchased from Aldrich, and 3,3′-dithiodipropionic acid (“carboxylate-terminated”) was purchased from Fluka (Buchs, Switzerland). All reagents were used as purchased without further purification. Ethanol (Fisher, Fair Lawn, NJ), the thiol solvent, was dried over aluminate, filtered with a 0.22-µm filter, and then deaerated with helium for 30 min. Sample Preparation. Stock solutions of 1 mg/mL concentration of insulin, somatostatin, angiotensin, cytochrome c, human serum transferrin (all purchased from Sigma), and biotinylated BSA (Pierce, Rockford, IL) were prepared in ultrafiltered deionized water. Appropriate aliquots of each solution were used to give working solutions of 2 pmol/µL of each peptide/protein. Triton-X 100 (Aldrich), Tween-20, urea (Fluka), guanidinium hydrochloride, and sodium acetate (Aldrich) were used as representative nonionic detergents, chaotropes, and salts. Loading and Washing of Samples. A volume of 0.3 µL of the analyte is loaded onto the mesa of a derivatized probe tip. The analyte is allowed to dry under either ambient conditions or vacuum crystallization. At these high concentrations of salt and detergents, drying at room temperature requires at least 10 min for each sample. The sample is then washed several times with deionized water (regardless of the surface), dried, and 0.3 µL of matrix is applied. We also experimented with acidic (0.1%, 1%, and 5% aqueous trifluoroacetic acid), basic (11 mM sodium hydroxide), organic (70% aqueous acetonitrile and 70% methanol), and phosphate-buffered saline washes. MALDI-TOF Measurements. All mass spectra were obtained using a Hewlett-Packard G2025A MALDI-TOF mass spectrometer. The instrument was operated at an accelerating voltage of 28 kV, an extractor voltage of 7 kV, and a pressure of ∼1 × 10-6 Torr. The samples were desorbed/ionized from the probe tip using a nitrogen laser source with an output wavelength of 337 nm. Calibration was performed with mixtures of peptides/ proteins with known molecular masses. Sinapinic acid (Aldrich), dissolved in 70% aqueous acetonitrile (Baker, Phillipsburg, NJ)
Figure 2. Schematic illustration of on-probe sample cleanup using ion-pairing SPE/MALDI-MS.
Figure 3. Ion-pairing SPE-MALDI/MS spectrum of insulin (2 pmol/ µL) from a solution of the peptide saturated with sodium acetate. These spectra were acquired with (A) a strong cation-pairing (3mercapto-1-propanesulfonic acid) surface and (B) a strong anionpairing (2-aminoethanethiol hydrochloride) surface.
with 0.01% trifluoroacetic acid (Aldrich), was used as the matrix for all samples. “Conventional MALDI” refers to a typical MALDI preparation in which the analyst mixes a volume of matrix and a volume of sample and then dries the resulting mixture onto the probe surface. This method is the technique to which we wish to compare our method using SPE/MALDI-MS technologies. This procedure is referred to throughout the text as conventional MALDI. RESULTS AND DISCUSSION The basic steps of an SPE/MALDI-MS experiment are shown in Figure 2. In these experiments, a 0.3-µL aliquot of the analyte solution is first placed on the modified probe and allowed to stand for a few minutes. This is followed by vacuum-drying and washing with deionized water to remove any contaminants. Finally, a solution containing the matrix is deposited on the sample regions, vacuum-dried, attached to the probe holder, and inserted into the mass spectrometer. This entire process adds only a few minutes to sample preparation yet allows successful MALDI-MS analysis of samples whose concentrations of salts, chaotropic agents, and/ or detergents/surfactants would have otherwise required extensive sample purification prior to MALDI/MS analysis. Existing procedures used to eliminate these interferences can require several hours of sample preparation time. The effectiveness of various ion-pairing surfaces for on-probe sample purification was evaluated with a series of SPE/MALDI experiments performed on peptide and protein solutions containing a high concentration of salt. The on-probe desalting ability of this technique is clearly illustrated by the analyses of insulin (2 pmol/µL) and human serum transferrin (10 pmol/µL) dissolved in an aqueous solution saturated with sodium acetate (Figures 3 and 4, respectively). These spectra were obtained on modified probes possessing SAMs with either a negatively (Figures 3A and 4A) or a positively (Figures 3B and 4B) charged terminus. Under these same sample conditions, conventional MALDI analysis, as expected, did not provide any molecular weight information due to the high concentration of salt. An additional series of experiments was performed to test these ion-pairing surfaces for the analysis of peptide and protein
Figure 4. Ion-pairing SPE-MALDI/MS spectrum of human serum transferrin (10 pmol/µL) from a solution of the protein saturated with sodium acetate. These spectra were acquired with (A) a weak cationpairing (3,3′-dithiodipropionic acid) surface and (B) a strong anionpairing (2-aminoethanethiol hydrochloride) surface.
solutions containing either a chaotropic agent (8 M urea or guanidinium hydrochloride) or a detergent/surfactant [up to 20% (v/v) Triton X-100 and Tween-20]. In each instance, conventional MALDI failed to yield molecular weight information. This result was expected due to the high concentration of contamination (spectra not shown). SPE/MALDI, on the other hand, provided intense molecular ions that were free from all interferences from contaminants (spectra not included for brevity). We proceeded to analyze samples with multiple interfering agents, since isolation of an analyte from a solution containing high concentrations of multiple contaminants is a better demonstration of SPE/MALDI’s capabilities. To dramatically illustrate the sample cleanup prowess of ionpairing SPE/MALDI, each type of SAM was used to analyze a series of peptides and proteins, ranging in molecular mass from 500 to 80 000 Da, dissolved in a solution containing 20% (v/v) Triton X-100, 8 M urea, and saturated sodium acetate or sodium chloride. These samples are probably more contaminated than those most scientists would ever encounter, but they serve to Analytical Chemistry, Vol. 70, No. 18, September 15, 1998
3759
Figure 5. Ion-pairing SPE/MALDI-MS spectra of insulin (2 pmol/ µL) from a solution of the peptide containing 20% (v/v) Triton X-100, 8 M urea, and saturated sodium acetate. These spectra were acquired with (A) a strong cation-pairing (3-mercapto-1-propanesulfonic acid) surface, (B) a weak cation-pairing (3,3′-dithiodipropionic acid) surface, and (C) a strong anion-pairing (2-aminoethanethiol hydrochloride) surface.
Figure 6. Ion-pairing SPE/MALDI-MS spectra of somatostatin and angiotensin (at a concentration of 2 pmol/µL) in 20% (v/v) Triton X-100, 8 M urea, and saturated sodium acetate, placed onto adjacent sample regions of an amine-terminated SAM probe. This experiment demonstrates that samples do not migrate between sample regions during the deionized water wash.
demonstrate the level of interference that can be dealt with successfully by ion-pairing SPE/MALDI. In each of these experiments, probes modified with the carboxylate-, sulfonate-, and amine-terminated SAMs were successful. The quality of these spectra is demonstrated by the analysis of insulin from this solution (Figure 5). Conventional MALDI analysis of these samples, on the other hand, was unsuccessful in all of these instances, as expected, since the presence of any one of these contaminants at this concentration ensures the failure of conventional MALDI analysis. We implemented various controls to ensure that the probe surface modifications were the cause of our successful isolation of the peptides/proteins in the above solutions. We were unsuccessful in obtaining spectra from peptides/proteins in either the singly or multiply contaminated solutions when an underivatized probe was substituted for a probe with an ion-pairing surface. These results show that an underivatized surface does not share the binding characteristics of ion-pairing surfaces. Also, we were unable to acquire a spectrum from a peptide/protein dissolved in either the singly or multiply contaminated solutions by conventional MALDI, i.e., mixing a 1-µL aliquot of sample with 1 µL of matrix, depositing this mixture on the probe, and directly analyzing. The absence of signal in these experiments confirmed that conventional preparatory methods are not suitable for this level of contamination. At this point, we were principally concerned with the nature of the peptide-monolayer binding. Was binding truly due to ionpairing, or was it a result of other forces, such as hydrophobic interactions? To investigate the ion-pairing theory, we designed several experiments using a variety of wash solutions ranging from slightly to strongly acidic and washes with organic solvents. Mild acid washes (pH 5.2) had no effect on the performance of the tips. Both insulin and cytochrome c were retained on the surfaces after mild acid washes on the carboxylate, sulfonate, and amino tips. However, stronger acid washes (0.1% TFA, pH 2.3) removed both insulin and cytochrome c from the carboxylate surface but not on the other negatively charged monolayer, the sulfonate. This
result can be attributed to the pKa of the two negatively charged surfaces. Specifically, under these conditions the carboxylate surface is expected to be fully protonated and, hence, to have no charged groups available for ion-pairing, while the sulfonate surface should still possess a negative charge. Washes with organic solvents had no effect on signal, suggesting that hydrophobic interactions play little or no role in the interactions we are observing with these tips. These data are strong evidence that the peptides/proteins are, indeed, attracted and held to the surface via ion-pairing interactions. Although the monolayer allows the sample to bind, it is the washing that ultimately removes the contaminants from the probe, permitting mass analysis to occur with no interference from contaminants. Because the washing procedure used to remove contaminants may also remove any unbound peptide/protein, we were concerned with contamination of neighboring samples on the same probe tip during the wash. Theoretically, any unbound peptide removed in the wash could contaminate neighboring samples. Such migration, if unforeseen, could destroy the integrity of the samples involved. To study the possibility of such migration, we loaded two different peptides, somatostatin and angiotensin (at a concentration of 2 pmol/µL in the multiply contaminated solution described above), onto adjacent sample regions of an amine-terminated SAM probe. There was no apparent migration of either somatostatin or angiotensin after washing with deionized water (Figure 6). Similar results were obtained from the two negatively charged surfaces. Consequently, sample migration does not appear to occur at concentrations above our limit of detection. The ion-pairing SPE/MALDI spectra of the two peptides shown in Figure 6 also demonstrate the level at which interferences are removed with this approach. Specifically, the molecular mass of these peptides is similar to that of the detergent/surfactant. However, the only signal seen in this m/z range results from the peptides, even though the detergent/surfactant was present at 20% (v/v), while the peptide was present at only 2 pmol/µL.
3760 Analytical Chemistry, Vol. 70, No. 18, September 15, 1998
CONCLUSIONS We have used MALDI probe tips derivatized with selfassembled monolayers terminated with a variety of chargeable functional groups to separate both peptides and proteins from solutions containing salts, chaotropic agents, and nonionic detergents/surfactants. Both negatively charged and positively charged surfaces allowed separation of the sample, effectively providing conditions favorable for MALDI-TOF analysis. However, due to the nature of the ion-pairing interactions, samples containing ionic detergents, such as SDS, are not amenable to ion-pairing SPE/ MALDI at this time. The preparation of the monolayers and samples and the subsequent analyses of the solutions are relatively easy to perform and yield spectra that would have been difficult, if not impossible, to obtain using conventional MALDI techniques. The ion-pairing
nature of these surfaces may be useful for a broad range of applications such as the analysis of oligonucleotides and oligosaccharides. These surfaces also offer the potential for on-probe separations. ACKNOWLEDGMENT This work was supported by grants from the National Science Foundation (NSF Grant No. 9626835) and the National Institutes of Health (NIH Grant No. 2-P41-RR05351) which made this project possible.
Received for review February 23, 1998. Accepted June 2, 1998. AC980210I
Analytical Chemistry, Vol. 70, No. 18, September 15, 1998
3761
