Feature-Level MALDI-MS Characterization of in Situ-Synthesized

pubs.acs.org/Langmuir © 2009 American Chemical Society

Feature-Level MALDI-MS Characterization of in Situ-Synthesized Peptide Microarrays Matthew P. Greving,† Pallav Kumar,† Zhan-Gong Zhao, and Neal W. Woodbury* The Biodesign Institute at Arizona State University and the Department of Chemistry and Biochemistry at Arizona State University, Tempe, Arizona 85287. † These authors contributed equally to this work. Received September 17, 2009. Revised Manuscript Received November 10, 2009 Characterizing the chemical composition of microarray features is a difficult yet important task in the production of in situ-synthesized microarrays. Here, we describe a method to determine the chemical composition of microarray features, directly on the feature. This method utilizes nondiffusional chemical cleavage from the surface along with techniques from MALDI-MS tissue imaging, thereby making the chemical characterization of high-density microarray features simple, accurate, and amenable to high-throughput.

Introduction Microarrays are one of the leading platforms used in highthroughput experimentation and data acquisition. A primary example is the DNA microarray, which serves as an integral experimental platform in genomics.1 The information content and the throughput achieved with microarrays continue to advance, partly because of advances in the fabrication of in situ synthesized DNA microarrays.1-7 This has resulted in greatly increased feature densities. In situ-synthesized microarrays of peptides were described more than 15 years ago.2 However, non-nucleic acid-based microarrays have not received the level of attention afforded to DNA microarrays during this period. Recently, the advances in DNA microarray fabrication are beginning to be applied to the production of peptides and other non-nucleic acid microarrays.8-17 *To whom correspondence should be addressed. E-mail: nwoodbury@asu. edu. (1) Chee, M.; Yang, R.; Hubbell, E.; Berno, A.; Huang, X. C.; Stern, D.; Winkler, J.; Lockhart, D. J.; Morris, M. S.; Fodor, S. P. Science 1996, 274, 610– 614. (2) Fodor, S.; Read, J.; Pirrung, M.; Stryer, L.; Lu, A.; Solas, D. Science 1991, 251, 767–773. (3) Gao, X.; LeProust, E.; Zhang, H.; Srivannavit, O.; Gulari, E.; Yu, P.; Nishiguchi, C.; Xiang, Q.; Zhou, X. Nucleic Acids Res. 2001, 29, 4744–4750. (4) Hughes, T. R.; Mao, M.; Jones, A. R.; Burchard, J.; Marton, M. J.; Shannon, K. W.; Lefkowitz, S. M.; Ziman, M.; Schelter, J. M.; Meyer, M. R.; Kobayashi, S.; Davis, C.; Dai, H.; He, Y. D.; Stephaniants, S. B.; Cavet, G.; Walker, W. L.; West, A.; Coffey, E.; Shoemaker, D. D.; Stoughton, R.; Blanchard, A. P.; Friend, S. H.; Linsley, P. S. Nat. Biotechnol. 2001, 19, 342–347. (5) Lipshutz, R. J.; Fodor, S. P.; Gingeras, T. R.; Lockhart, D. J. Nat. Genet. 1999, 21, 20–24. (6) Pease, A. C.; Solas, D.; Sullivan, E. J.; Cronin, M. T.; Holmes, C. P.; Fodor, S. P. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5022–5026. (7) Singh-Gasson, S.; Green, R. D.; Yue, Y.; Nelson, C.; Blattner, F.; Sussman, M. R.; Cerrina, F. Nat. Biotechnol. 1999, 17, 974–978. (8) Beyer, M.; Nesterov, A.; Block, I.; K€onig, K.; Felgenhauer, T.; Fernandez, S.; Leibe, K.; Torralba, G.; Hausmann, M.; Trunk, U.; Lindenstruth, V.; Bischoff, F. R.; Stadler, V.; Breitling, F. Science 2007, 318, 1888. (9) Frank, R. J. Immunol. Methods 2002, 267, 13–26. (10) Kuruvilla, F. G.; Shamji, A. F.; Sternson, S. M.; Hergenrother, P. J.; Schreiber, S. L. Nature 2002, 416, 653–657. (11) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760–1763. (12) Maurer, K.; McShea, A.; Strathmann, M.; Dill, K. J. Comb. Chem. 2005, 7, 637–640. (13) Northen, T. R.; Greving, M. P.; Woodbury, N. W. Adv. Mater. 2008, 20, 4691–4697. (14) Park, S.; Shin, I. Angew. Chem., Int. Ed. 2002, 41, 3180–3182. (15) Pellois, J. P.; Zhou, X.; Srivannavit, O.; Zhou, T.; Gulari, E.; Gao, X. Nat. Biotechnol. 2002, 20, 922–926. (16) Pirrung, M. Chem. Rev. 1997, 97, 473–488. (17) Ramachandran, N.; Hainsworth, E.; Bhullar, B.; Eisenstein, S.; Rosen, B.; Lau, A. Y.; Walter, J. C.; Labaer, J. Science 2004, 305, 86–90.

1456 DOI: 10.1021/la903510y

Synthesis efficiency is essential for reliable microarray data; even minor inefficiencies result in cumulative in-situ stepwise synthesis errors.18 The characterization of the synthesis efficiency of in situsynthesized DNA microarrays is facilitated by the ability to hybridize a complementary strand of DNA and the sensitivity of hybridization to base pair mismatches.19-21 However, characterizing the synthetic fidelity of peptides and other non-nucleic acidbased microarrays is much more difficult given that the hybridization of a complementary strand is not possible. Typically, the characterization of non-nucleic acid-based microarrays is done using direct-label fluorescence or staining,13,15,22 antibody binding to a synthesized epitope,8,15 protein binding to a known ligand,12,23 or molecules are cleaved from the surface and then analyzed using traditional methods such as HPLC and mass spectrometry.9,24,25 Although these methods do provide some information about the success or failure of a particular microarray synthesis, much of the information about the chemical composition of the microarray spot remains unknown. For example, although directly labeling or staining the microarray spots provides information about the total number of molecules available at the spot, the chemical identity of these molecules remains unknown. Antibody binding to a known epitope or protein binding to a known ligand indicates if the epitope or ligand exists at the spot. However, these are typically very high affinity interactions, and even a small amount of epitope or ligand in the presence of large numbers of side products at the microarray spot will produce significant signal, particularly given the avidity and nonspecific binding effects that occur at surfaces.26-28 Finally, cleaving in situ-synthesized molecules from the microarray surface (18) Pirrung, M. Angew. Chem., Int. Ed. 2002, 41, 1276–1289. (19) Wallace, R. B.; Shaffer, J.; Murphy, R. F.; Bonner, J.; Hirose, T.; Itakura, K. Nucleic Acids Res. 1979, 6, 3543–3557. (20) Conner, B. J.; Reyes, A. A.; Morin, C.; Itakura, K.; Teplitz, R. L.; Wallace, R. B. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 278–282. (21) Maskos, U.; Southern, E. M. Nucleic Acids Res. 1992, 20, 1679–1684. (22) Hilpert, K.; Winkler, D. F. H.; Hancock, R. E. W. Nat. Protoc. 2007, 2, 1333–1349. (23) Michel, O.; Ravoo, B. J. Langmuir 2008, 24, 12116–12118. (24) Bhushan, K. R. Org. Biomol. Chem. 2006, 4, 1857–1859. (25) Northen, T. R.; Brune, D. C.; Woodbury, N. W. Biomacromolecules 2006, 7, 750–754. (26) Chen, S.; Phillips, M. F.; Cerrina, F.; Smith, L. M. Langmuir 2009, 25, 6570–6575. (27) Tapia, V.; Bongartz, J.; Schutkowski, M.; Bruni, N.; Weiser, A.; Ay, B.; Volkmer, R.; Or-Guil, M. Anal. Biochem. 2007, 363, 108–118. (28) Zichi, D.; Eaton, B.; Singer, B.; Gold, L. Curr. Opin. Chem. Biol. 2008, 12, 78–85.

Published on Web 12/22/2009

Langmuir 2010, 26(3), 1456–1459

Greving et al.

does provide a complete characterization of the synthesis, but this analysis is not compatible with arrayed chemical libraries because of diffusion and mixing at the surface upon cleavage and sample collection. Here we describe an approach to fully characterize, directly on the feature, the composition of arrayed peptide libraries synthesized in situ using mass spectrometry. Direct characterization of in situ-synthesized small-molecule microarrays has been previously described.29,30 These studies utilized TOF-SIMS and a custom-synthesized cleavable linker. In the characterization of microarrays, TOF-SIMS has significant spatial resolution advantages when compared to MALDI-TOF. In addition, TOFSIMS does not require the application of a matrix, thereby limiting the diffusion of analyte between microarray features. However, TOF-SIMS is generally limited to low-mass ions such as short peptides or small molecules when compared to the highmolecular-weight ions that can be detected with MALDI-TOF.31 The high-mass range of TOF-SIMS can be extended with the application of a matrix, similar to that used with MALDI-TOF,31 but with the application of a matrix, the advantage of limited diffusion offered by matrix-free TOF-SIMS is lost. In the approach described herein, microarrays containing different peptide sequences attached to the surface via commercially available cleavable linkers are synthesized in situ using both light-directed and electrochemically patterned synthesis. The resulting features are then cleaved and the MALDI matrix is applied, with limited diffusion. After applying the MALDI matrix, the chemical composition of the features is analyzed using MALDI-TOF mass spectrometry directly on the feature. In this work, MALDI-TOF was used to demonstrate the approach. However, with the application of the MALDI matrix, TOFSIMS could also presumably be used to obtain in situ mass spectra from the peptide array features and would likely offer higher spatial resolution than MALDI-TOF.32

Results and Discussion To demonstrate the generality of this approach in characterizing in situ-synthesized microarrays on different platforms, both photolithographically and electrochemically synthesized peptide microarrays were tested. Photolithographic in situ synthesis utilizes photolabile protecting groups and masks to deprotect features selectively on the microarray,2 whereas electrochemical in situ synthesis utilizes electrochemically generated acids to remove acid-labile protecting groups and addressable electrodes to produce distinct microarray features.12 In this work, microarray feature dimensions were limited to several hundred micrometers to demonstrate the approach. Direct MALDI-MS Characterization of Photolithographically Synthesized Peptide Microarrays. The light-directed synthesis of peptide microarrays was performed using the mixedbase-labile/photolabile (FMOC/NPPOC) protecting group approach previously described (Supporting Information, Figure 1).13 The advantage of this approach is that high-yield FMOC (fluorenylmethyloxycarbonyl) chemistry can be used for invariable peptide positions and photolabile NPPOC (30 nitrophenylpropyloxycarbonyl) chemistry can be used for variable positions to generate microarray diversity. Also, the mixed protecting group strategy allows for the use of commercially (29) Chen, C.; Nagy, G.; Walker, A. V.; Maurer, K.; McShea, A.; Moeller, K. D. J. Am. Chem. Soc. 2006, 128, 16020–16021. (30) Chen, C.; Lu, P.; Walker, A.; Maurer, K.; Moeller, K. D. Electrochem. Commun. 2008, 10, 973–976. (31) Heeren, R. M. Proteomics 2005, 5, 4316–4326. (32) MacAleese, L.; Stauber, J.; Heeren, R. M. Proteomics 2009, 9, 819–834.

Langmuir 2010, 26(3), 1456–1459

Letter

available FMOC-protected amino acids and avoids the synthesis and purification of NPPOC-protected amino acids. Using the mixed protecting group strategy, an array of four peptides; YERYGS, YGRYGS, YRRYGS, and YSRYGS;was synthesized with the second position varied and all other positions fixed (Supporting Information, Figure 2a). Each peptide had a PEG2 (polyethylene glycol) spacer followed by a Rink linker at the C-terminus to provide separation and an acid cleavage site between the C-terminus and the APTES (aminoproplytriethoxysilane) glass surface. In the case of light-directed synthesis, an acid-cleavable Rink linker was chosen because it is orthogonal to the base-labile FMOC and photolabile NPPOC protecting groups used in the synthesis of the peptide microarray. To cleave the peptides from the surface without diffusion between features, dry cleavage was carried out using TFA (trifluoracetic acid) vapor in a sealed glass chamber. Acid vapor cleavage of peptides on beads has been described.33 However, to the best of our knowledge, this is the first report of using acid vapor to cleave molecules from the surface of a microarray. Before cleaving peptides from a microarray, the TFA vapor cleavage protocol was tested with a peptide synthesized on beads and compared to the TFA solution cleavage of the same peptide on beads (Supporting Information, Figure 3). These results show that TFA vapor cleavage not only efficiently cleaves the Rink linker but also removes the acid-labile side-chain protecting groups from the peptide. After validation on beads, TFA vapor cleavage was performed on a light-directed array of four peptides. Following peptide cleavage, an R-cyano-4-hydroxycinnamic acid MALDI matrix was applied to the array surface as an aerosol using an airbrush, as has been done in MALDI-MS imaging of tissues to limit analyte diffusion.34 The glass slide containing the array was then mounted in a MALDI target plate designed to hold 75 mm  25 mm substrates (microscope slide dimensions). Initial spectra collected directly from the microarray features showed significant charging effects on the glass slide;35 this was apparent from the downfield peak shifts observed with additional shots of the MALDI laser.36 Charging effects were completely suppressed by applying conductive tape to all four edges of the glass slide, bridging the gap between the glass slide and the metal frame of the MALDI target plate. As a result, very good signal-noise MALDI-MS spectra were obtained from the four array spots (Figure 1). At each spot, the desired peptide is a significant peak in the spectrum (Figure 1). Interestingly, the YERYGS peptide peak (1003 Da) from array feature 1 is also visible in the spectra obtained from features 2-4. This contaminant peak appears to be due to the mixed FMOC/NPPOC protecting group strategy used to synthesize the array in which the FMOC group is removed after coupling the arginine in position three, with respect to the N-terminus, and then substituted with the subsequent coupling of NPPOC-chloride to the N terminus. The spectra obtained from this array indicate that the NPPOC substitution is not complete; therefore, the incoming amino acid for the first patterned feature (feature 1, glutamic acid) couples to free amines available at subsequent features (features 2-4). This is a good example of an array synthesis problem that could not have been detected by characterization using direct fluorescence, colorimetric assay, or antibody binding. (33) Brummel, C. L.; Lee, I. N.; Zhou, Y.; Benkovic, S. J.; Winograd, N. Science 1994, 264, 399–402. (34) Monroe, E. B.; Jurchen, J. C.; Koszczuk, B. A.; Losh, J. L.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2006, 78, 6826–6832. (35) Knochenmuss, R. Anal. Chem. 2004, 76, 3179–3184. (36) Iban~ez, A. J.; Muck, A.; Svatos, A. J. Mass Spectrom. 2007, 42, 634–640.

DOI: 10.1021/la903510y

1457

Letter

Greving et al.

Figure 1. MALDI-MS spectra of a four-peptide light-directed array. The template sequence Y(X)RYGS-Peg was used at each feature, with each of the following amino acids substituted into position 2: (X) = {E, G, R, S}.

Two additional light-directed arrays were synthesized to explore the generality of the approach. First, an array of ten 20-mer peptides (PEG2 at position 20) was synthesized on glass to determine if this approach could be used to detect longer peptides. The 20-mer peptide array was based on template sequence AHKVVPQRQMRHAY(X)RYGS, with the amino acid set {K, R, D, N, H, Y, S, W, F, V} substituted into position 15 (X) and all other residues fixed (Supporting Information, Figure 2b). The peptides were cleaved using acid vapor cleavage and analyzed with MALDI-MS. The resulting spectra from each array feature (Supporting Information, Figure 4) show the desired peak at most of the features, however, with a lower signal-to-noise ratio than was observed above with shorter peptides. Dilute calibration mixtures containing peptides in excess of 3000 Da produce very high signal-to-noise ratios when spotted on the array substrate and analyzed in situ by MALDI-MS. Therefore, the lower signalto-noise ratio is likely due to the low synthesis yield of 20-mer peptides. In addition, as was seen in the array in Figure 1, the major contaminant is the peptide from the first patterned feature (2542 Da). Second, to determine if this approach can be used to characterize microarrays on 3D surfaces, a light-directed porous polymer peptide microarray was synthesized.13 In this case, an array of three peptides of different lengths was tested: R(G)P, R(TG)P, and R(KTG)P (Supporting Information, Figure 2c). Each additional residue shown in parentheses adds another lightdirected synthesis step for a total of three light-directed steps in the longest peptide R(KTG)P. This array provides information on the degree of side-product accumulation that occurs when 1458 DOI: 10.1021/la903510y

multiple light-directed steps are used to synthesize an array. The resulting MALDI-MS spectra (Supporting Information, Figure 5) show the desired peptide product at each of the three features; however, it is apparent from the ion intensities relative to the background peaks (feature interstitial space peaks) that the amount of desired peptide product decreases with each additional light-directed synthesis step. Again, this is information that would not have been available by characterizing the array synthesis with direct fluorescence (because all amines would be labeled) or even antibody binding (because of the high-affinity interaction). Direct MALDI-MS Characterization of Electrochemically Synthesized Peptide Microarrays. Like the light-directed array synthesis, electrochemical peptide microarray synthesis was performed using a mixed protecting group approach. In the case of electrochemical synthesis, the mixed protection group strategy utilized base-labile/acid-labile (FMOC/trityl) protecting groups. This strategy allows for the use of commercially available FMOC amino acids in constant positions and highly acid-labile trityl groups in variable positions. Using a very acid-labile protecting group such as trityl allows for deprotection with electrochemically generated acids at low applied potentials at the electrode. Because electrochemical peptide array synthesis generates acid at the electrode surface, the acid-cleavable Rink linker cannot be used. Instead, an orthogonal photocleavable linker was used to attach peptides to the electrode surface. An electrochemical array of four peptides of different lengths; KAFGAFGAFG, K(G)AFGAFGAFG, K(FG)AFGAFGAFG, and K(AFG)AFGAFGAFG;was synthesized (Supporting Langmuir 2010, 26(3), 1456–1459

Greving et al.

Letter

Information, Figure 2d). Each additional residue in parentheses adds another electrochemical step; the first peptide without parentheses was synthesized at the feature using only FMOC chemistry. To simplify the task of visualizing array features during MALDI-MS analysis, the array was synthesized as a collection of several individual neighboring electrodes. After array synthesis, the dry chip was irradiated with UV light to cleave the peptides from the surface without diffusion. The MALDI matrix was applied to the surface as small drops with a pipet, and the chip (75 mm  25 mm dimensions) was mounted in the same MALDI target plate as used for the lightdirected glass slides that were then analyzed with MALDI-MS. In the case of electrochemical arrays, surface charging during MALDI spectrum acquisition was not a problem because of the conductive nature of the electrochemical chip. The resulting spectra (Supporting Information, Figure 6) from each array feature show the desired peptide as a dominant peak along with several side-product peaks. A second electrochemical array using the same array layout but with modified peptide sequences produced a similar result (Supporting Information, Figures 2e and 7).

Conclusions We have developed a simple, accurate, general MALDI-MSbased approach to characterize the chemical composition of an in situ-synthesized microarray directly on the array features. As demonstrated, this approach is not specific to a particular type of microarray substrate and has sufficient sensitivity to characterize array features synthesized on 2D surfaces such as glass. In

Langmuir 2010, 26(3), 1456–1459

addition, peptides with masses greater than 2500 Da can be detected. This approach provides significantly more information about microarray spot composition than can be gained from traditional approaches such as fluorescence and antibody binding. In addition to the in situ determination of amino acid composition, with MALDI-TOF-TOF (MS-MS), it is likely possible to obtain peptide sequence information directly from the microarray feature. We recognize that this approach, as described, has spatial limitations that make it difficult to analyze array features that are less than 100 μm in diameter. The two most significant factors limiting the spatial resolution of this work are the limit to which the MALDI laser can be focused and the limited precision in positioning the array features inline with the laser. Improvements in either of these limiting factors would significantly reduce the size of array features that can be analyzed with MALDI-TOF. Array spatial resolutions as small as 25 μm may be possible, as has already been demonstrated in tissue imaging.37 However, the characterization of synthesis efficiency is most important during optimization rather than production. Therefore, spatial resolution is likely less important than the ability to characterize the spot composition comprehensively because microarrays with larger feature sizes can be used during optimization. Supporting Information Available: Additional spectra and detailed experimental methods. This material is available free of charge via the Internet at http://pubs.acs.org. (37) Stoeckli, M.; Chaurand, P.; Hallahan, D. E.; Caprioli, R. M. Nat. Med. 2001, 7, 493–496.

DOI: 10.1021/la903510y

1459