- Email: [email protected]
Protein micro- and macroarrays: digitizing the proteome
Journal of Chromatography B, 787 (2003) 19–27 www.elsevier.com / locate / chromb
Review
Protein micro- and macroarrays: digitizing the proteome Mary F. Lopez*, Malcolm G. Pluskal Proteome Systems, 14 Gill St., Woburn, MA 01801, USA
Abstract The early applications of microarrays and detection technologies have been centered on DNA-based applications. The application of array technologies to proteomics is now occurring at a rapid rate. Numerous researchers have begun to develop technologies for the creation of microarrays of protein-based screening tools. The stability of antibody molecules when bound to surfaces has made antibody arrays a starting point for proteomic microarray technology. To minimize disadvantages due to size and availability, some researchers have instead opted for antibody fragments, antibody mimics or phage display technology to create libraries for protein chips. Even further removed from antibodies are libraries of aptamers, which are single-stranded oligonucleotides that express high affinity for protein molecules. A variation on the theme of protein chips arrayed with antibody mimics or other protein capture ligand is that of affinity MS where the protein chips are directly placed in a mass spectrometer for detection. Other approaches include the creation of intact protein microarrays directly on glass slides or chips. Although many of the proteins may likely be denatured, successful screening has been demonstrated. The investigation of protein–protein interactions has formed the basis of a technique called yeast two-hybrid. In this method, yeast ‘‘bait’’ proteins can be probed with other yeast ‘‘prey’’ proteins fused to DNA binding domains. Although the current interpretation of protein arrays emphasizes microarray grids of proteins or ligands on glass slides or chips, 2-D gels are technically macroarrays of authentic proteins. In an innovative departure from the traditional concept of protein chips, some researchers are implementing microfluidic printing of arrayed chemistries on individual protein spots blotted onto membranes. Other researchers are using in-jet printing technology to create protein microarrays on chips. The rapid growth of proteomics and the active climate for new technology is driving a new generation of companies and academic efforts that are developing novel protein microarray techniques for the future. 2002 Elsevier Science B.V. All rights reserved. Keywords: Reviews; Microassays; Macroassays; Proteins; Proteomes
Contents 1. Introduction ............................................................................................................................................................................ 1.1. What is a protein array and what have we learned from DNA microarray applications?........................................................ 2. Protein chips ........................................................................................................................................................................... 2.1. Antibody arrays / phage display / antibody mimics .............................................................................................................. 2.2. Affinity capture / MALDI TOF ......................................................................................................................................... 2.3. Protein / peptide arrays ..................................................................................................................................................... 2.4. Protein / protein interactions ............................................................................................................................................. *Corresponding author. Tel.: 11-781-932-9477x3309; fax: 11-781-932-9294. E-mail address: [email protected] (M.F. Lopez). 1570-0232 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1570-0232(02)00336-7
20 20 21 21 24 24 25
20
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
3. Complementary approaches ..................................................................................................................................................... 3.1. 2-D gels are bona fide protein macroarrays ....................................................................................................................... 3.2. Piezoelectric and ink-jet chemical printing ........................................................................................................................ 4. Future directions and challenges ............................................................................................................................................... 5. Notation ................................................................................................................................................................................. References ..................................................................................................................................................................................
1. Introduction The challenge ahead will no longer be the elucidation of the DNA coding sequence of the human genome, but understanding how its gene sequences are expressed. Studying this process at the protein level, termed Proteomics, is expected to be a more complex task than genome sequencing. To meet this challenge, many new technology tools are in development. Some of these new developments have roots in the successful application of microarray and macroarray fabrication. These were heavily influenced by technology, such as photolithography and ink-jet dispensing that was developed for semiconductor and computer hardware applications. Arrays of cDNA sequences and oligonucleotides are in routine use in large-scale analysis of gene expression and variations in gene sequence.
1.1. What is a protein array and what have we learned from DNA microarray applications? Currently, the study of the changes in the level of proteins in the cell, or Proteomics, is limited to serial processing of samples using high-resolution 2D gel technologies [1–5]. This approach, coupled with mass spectrometry, can rapidly yield structural information and hence identity on proteins resolved from complex mixtures [6–9]. Automation of this process will accelerate throughput in the future. Synergistic to this process would be a high-throughput, analytical screening tool to rapidly identify differences between large numbers of samples. This approach would facilitate the location of potential targets for further analysis with existing ‘‘classical proteomics’’ technology platforms. Such a screening tool would process many samples in parallel and be capable of detecting differentially expressed protein targets. Recently, numerous researchers have begun to develop technologies for the creation of microar-
25 25 25 25 26 26
rays of protein-based capture ligands to create such screening tools. The early applications of microarrays and detection technologies were clearly centered on DNAbased applications. This was a direct result of the effort being placed on genome sequencing and the necessity to analyze large numbers of samples in parallel. However, the nature of DNA lends itself very well to some of the technologies employed, such as PCR and labeling with fluorescent reporter molecules. The success of these approaches is clearly a function of the chemistry and enzymology of nucleic acids. In protein-based studies, it may not be possible to exploit similar tools. However, in examining the evolution of new analytical tools for protein-based applications, there are some valuable lessons to be learned from the successes with nucleic acids. Some examples of these applications are detailed below:
(i) Amplification of DNA with PCR is probably the single largest advance that enabled many applications to date [10]. Being able to amplify the target is a great asset in any analytical approach. The closest protein equivalent would be the specific affinity selection or isolation of a defined charge class of proteins by isoelectric focusing from a complex mixture, leading to its enrichment in the final sample [11–14]. This approach, termed complexity reduction or enrichment will be an essential component of future applications in Proteomics. It is not quite the same as PCR but it offers the ability to increase the level of the protein or target of interest by processing a larger volume sample and enriching a specific component (ii) Labeling with reporter molecules, such as fluorescent dyes allows for the increased sensitivity of detection, the ability to multiplex detection of several targets and easy interfacing
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
with detection instrumentation and automation [15–17]. Protein labeling with fluorescent probes has seen many applications in structural studies employing site-specific modifications of reactive amino acid side chains [18–20]. For less protein sequence dependent detection, a new generation of fluorescent dyes or modified surfactant molecules has been developed and offers some of the above attributes [21]. Being able to multiplex samples in future detection strategies will be key to opening up comparative studies of complex protein mixtures. The power of some of the gene expression chip-based approaches lies in their ability to allow for the comparison of control and experimental signals with these multiplexed detection schemes on the same analytical surface. (iii) More recently, MALDI TOF with the appropriate sample energy transfer matrix has opened up rapid analysis of DNA fragments (,150 bp) with extremely high precision sufficient to achieve the equivalent single base resolution of electrophoresis [22–24]. In the protein application, MALDI TOF and electrospray MS–MS have seen spectacular success in resolving peptides and proteins (,12 000 MW) with high mass resolution [25–27]. Larger proteins and nucleic acids are not as well resolved, with peaks being broader due to isotope effects and the influence of post-translational modifications in proteins. Hence, this approach has high intrinsic value as an analytical tool for fragments derived from intact proteins. For whole molecule analysis its application will be different, being more useful as a comparative-screening tool complementing other, more high-resolution approaches, such as 2-D gels or peptide fragmentation / MS.
2. Protein chips
2.1. Antibody arrays /phage display /antibody mimics The relative robustness and stability of antibody molecules as well as the extensive libraries already available have made antibody arrays a starting point for proteomic microarray technology. However, it
21
should be noted that the requirements of antibodies for protein chips might be significantly more rigorous than for other applications in terms of specificity. Indeed, a caveat of antibody arrays is the difficulty of recognizing all the isoforms of a particular protein and its post-translational modifications. Nevertheless, several researchers and companies have demonstrated the initial feasibility of this approach [28–32]. In one such study [29], the authors printed 115 pairs of antibody / antigen pairs on derivatized glass slides using a robotic arrayer. The researchers were able to detect 50% of the arrayed antigens and 20% of the arrayed antibodies in complex mixtures containing the cognate ligands at concentrations ranging from 1 ng / ml to 0.34 mg / ml. These results suggested that protein microarrays could be used for characterization of pattern variation in clinical or research applications. Other researchers have achieved similar results [32] and a number of new companies such as Zyomyx (Hayward, CA, USA), BioSite (San Diego, CA, USA), Hypromatrix and Diversys (Cambridge, UK) are developing protein chip strategies based on antibody arrays. Intact antibody molecules have several drawbacks for protein chip applications including their large size and potential lack of specificity. To overcome these problems, some researchers have turned to antibody fragments [30], antibody mimics, or phage display technology [33] to create libraries for protein chips. Phage display is a method for creating antibody libraries using genetically modified phages that express variable fragments of immunoglobulin on their surfaces. Several new companies such as Dyax (Boston, MA, USA), Cambridge Antibody Technology (UK) and Affitech (Oslo, Norway) are pursuing this approach or a modification of this approach (Phylos, Lexington, MA, USA). Unfortunately, phage displayed antibodies are subject to many of the same difficulties with specificity that are evidenced by conventional antibodies [34]. Display libraries are not necessarily limited to antibodies and can consist of non-biological antibody mimics and diverse molecules [33]. Even further removed from antibodies are libraries of aptamers, which are single-stranded oligonucleotides that express high affinity for protein molecules [34–36]. These aptamer libraries can easily be synthesized and are less labile than antibodies but are
22
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
Fig. 1. Detection of specific carbohydrate modifications of proteins in a microarray format (figure courtesy of K. Martin and W.F. Patton, Molecular Probes Inc.). Alpha-1 acid glycoprotein, horseradish peroxidase, immunoglobulin G, and soybean trypsin inhibitor were arrayed on to PVDF membrane, adhered to a glass slide, from a source plate (384-well) with a concentration of 0.0625–4 mg / ml protein in phosphate buffered saline. Arrays were spotted using a manual glass slide arrayer (V&P Scientific, San Diego, CA, USA) fixed with four rows of eight pins (32 total), |500 mm diameter spot size, 1125 mm horizontal pitch and 750 mm vertical pitch (pitch5center to center spacing of spots). Proteins were arrayed in replicates of six, resulting in an array of 168 spots. (A) The total protein content (left) of the array was labeled with 10 mM BODIPY TR-X, SE (Molecular Probes) in sodium borate buffer, pH 9.5 followed by specific glycoprotein detection with the alkaline phosphatase-conjugated concanavalin A (right). Concanavalin A binds to glycoproteins containing a-mannosyl and a-glucopyranosyl residues as found in horseradish peroxidase and immunoglobulin G. Arrays were blocked with mTBS (50 mM Tris pH 7.5, 150 mM NaCl) containing 0.25% MOWIOL 4-88 and 0.2% Tween-20 before being probed with 1 mg / ml of lectin conjugate in mTBS with 0.5 mM MgCl 2 and 1 mM CaCl 2 . Detection of the lectin was performed by incubating the array in the alkaline phosphatase substrate ELF 39 phosphate (Molecular Probes) at a concentration of 10 mg / ml in 10 mM Tris pH 9.5, 1 mM MgCl 2 . (B) The total protein content (left) of the array was labeled with 10 mM BODIPY FL-X, SE (Molecular Probes) in sodium borate buffer, pH 9.5 followed by specific glycoprotein detection with the alkaline phosphatase-conjugated wheat germ agglutinin (right). Wheat germ agglutinin binds glycoproteins containing N-acetylglucosamine and N-acetylneuraminic acid residues as found in a-1 acid glycoprotein and immunoglobulin G. Arrays were blocked with mTBS (50 mM Tris pH 7.5, 150 mM NaCl) containing 0.25% MOWIOL 4-88 and 0.2% Tween-20 before being probed with 1 mg / ml of wheat germ agglutinin in mTBS with 0.5 mM MgCl 2 and 1 mM CaCl 2 . Detection of the lectin was performed by incubating the array in the alkaline phosphatase substrate DDAO phosphate (Molecular Probes) at a concentration of 1.25 mg / ml in 10 mM Tris pH 9.5, 1 mM MgCl 2 .
Fig. 2. Identification of proteins resolved in a 2DE gel blotted to a PVDF membrane: applications of the Chemical Printer to Trypsin digestion and subsequent MALDI–MS analysis of peptides. Human plasma proteins (2.5 mg) prefractionated using the multi compartment electrolyzer (Proteome Systems, Sydney, Australia) were resolved on a pH range 5–6 IPG strip and then ran in a second dimension 6–15% linear gradient SDS–PAGE gel (GelChip, Proteome Systems, Woburn, MA, USA). The resulting separation was then transferred to a high surface area PVDF Western blotting membrane (Immobilon Psq, Millipore, Danvers, MA, USA) and stained with Direct blue (Sigma, St Louis, MO, USA). A replicate gel was stained directly in a colloidal Coomassie blue stain (Ez-Blue姠, Sigma) for direct in-gel digestion. The protein spots labeled on the stained PVDF blot (Panel A) were analyzed using the prototype Chemical Printer instrument. The entire membrane was positioned on filter paper saturated with 2.5 mM Tris–HCl, pH 8.5. Solutions required for digestion of proteins were applied to the membrane with a piezoelectric device; (A) Trypsin / detergent solution (0.5 mg / ml of Porcine Trypsin [Promega modified trypsin], 2.5 mM Tris–HCl, pH 8.5, 1% [v / v] octyl-glucopyranoside) was jetted onto the blotted proteins in 23100 drops then 23 999 drops to a final dispensed volume of 250 nl. Samples were then incubated in a humidified environment at 30 8C overnight. (B) Following digestion matrix solution (20 mg / ml a-cyano-4-hydroxy cinnamic acid in 70% [v / v] acetonitrile and 0.5% [v / v] formic acid) was jetted (23999 drops) using the same robotic coordinates used for dispensing the trypsin. The PVDF membrane was then transferred to a metal MALDI target and fixed in place using conductive adhesive and analyzed on a KRATOS Axima CFR instrument (Panel B). A parallel in gel digest of spots on the replicate gel was carried out using standard protocols, data shown in Panel C.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
23
24
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
nevertheless subject to nucleases that may be present in complex mixtures of proteins from body fluids used to screen chips [34].
2.2. Affinity capture /MALDI TOF An interesting variation on the theme of protein chips arrayed with antibody mimics or other protein capture ligands is that of affinity MS [37–40]. This technique, termed SELDI and pioneered by Ciphergen Biosystems (Palo Alto, CA, USA), consists of a MALDI target surface that has been modified with affinity ligands to capture proteins of interest. The target can then be inserted directly into the mass spectrometer for detection [41]. A different version of this approach uses chips that employ surface plasmon resonance to quantify the target capture. The chips may then also be analyzed by MALDI for characterization of the captured proteins [42,43]
brain cDNA library. The fusion proteins were arrayed onto glass and membrane surfaces and were probed with antibodies. Although many of the proteins were likely denatured, successful screening was demonstrated. Some companies are now taking a similar approach to generate collections of recombinant proteins. In one case, full-length proteins were expressed and immobilized via an affinity tag preserving full protein function [32]. In a further extension of these approaches, researchers have developed a technique for detecting post-translational carbohydrate modifications of proteins in microarrays (Fig. 1, courtesy of K. Martin and W.F. Patton, Molecular Probes, Eugene, OR, USA). In this experiment, dichromatic fluorescent dyes were used to detect and quantify the total microarrayed proteins and then, using different excitation / emission filters, reporter enzyme–lectin conjugates monitored the glycosylation status of the proteins.
2.3. Protein /peptide arrays Several groups have taken the approach of creating microarrays of proteins directly on chips [44– 46]. In a recent report [47], proteins were attached to a glass slide that was activated with a cross linking agent that reacts with primary amines. They tested protein–protein interactions, kinase substrate reactions and protein–ligand reactions using fluorescent probes and radiolabeled ATP. In one experiment, a single protein was identified within an array of 10 000 other proteins. This demonstrates the feasibility of using protein microarrays for largescale, high-throughput screening studies. In order to maintain the three-dimensional integrity of proteins spotted onto surfaces, some researchers have developed a technique for immobilizing the proteins in polyacrylamide gel pads fixed onto glass slides [48]. Other workers have developed a nanowell technology and assayed 117 different yeast kinases with 17 different substrates [49]. The results of this experiment suggested the existence of novel yeast tyrosine kinases. On a much larger protein expression scale, researchers at the Max Planck Institute [31,32,50] have created a library of approximately 15 000 nonredundant His-tagged proteins from a human fetal
Fig. 3. Chemical printing of a microarray of reagent spots on a macroarray protein target. High density printing of reagents in a microarray format on a single gel spot resolved in a 2DE macroarray separation, demonstrates the potential of the Chemical Printer. The entire membrane was positioned on filter paper saturated with 2.5 mM Tris–HCl, pH 8.5. Solutions required for digestion of proteins were applied to the stained post located on the membrane with a piezoelectric device; Trypsin / detergent solution (0.5 mg / ml of Porcine Trypsin [Promega modified trypsin], 2.5 mM Tris–HCl, pH 8.5, 1% [v / v] octyl-glucopyranoside) was jetted onto a single blotted protein spot in 23100 drops then 23 999 drops to a final dispensed volume of 250 nl. A 434 spot microarray format was over layered on a single spot.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
2.4. Protein /protein interactions In a ‘‘targeted proteomics’’ approach, the investigation of protein–protein interactions has formed the basis of a technique called yeast two-hybrid [51–53]. This method typically involves the generation of yeast ‘‘bait’’ proteins fused to transcriptionactivation DNA domains. The bait proteins can then be probed with other yeast proteins (‘‘prey’’) fused to DNA binding domains. Using this approach, researchers recently created an array of 5300 yeast bait proteins and identified 957 potential protein– protein interactions by probing with yeast prey proteins [52]. A company (Myriad Genetics, Salt Lake City, UT, USA) plans to use this approach to map all the protein–protein interactions in yeast. A recent report described the cloning and expression of 5800 yeast proteins with their subsequent deposition onto a microarray [54]. The high-density microarray was screened for interaction with other proteins and phospholipids. Using this method, the authors identified many novel calmodulin and phospholipid binding proteins.
3. Complementary approaches
3.1. 2 -D gels are bona fide protein macroarrays Although the current interpretation of protein arrays emphasizes microarray grids of proteins or ligands on glass slides or chips, it is interesting to note that 2-D gels (currently the workhorse of proteomics) are technically macroarrays of authentic proteins, i.e. charge, molecular mass and posttranslational modifications are preserved during electrophoresis. This allows direct analysis of protein isoforms that may be involved in particular metabolic or disease processes. Due to the true parallel nature of the technique, hundreds to thousands of proteins can be visualized simultaneously and when coupled with mass spectrometry, complex mixtures of proteins can be resolved and individual components identified [5,55–61]. In addition, the quantitative differences between proteins in mixtures can be determined from 2-D gel images allowing the direct detection of differentially expressed gene products [55–61]. Unfortunately, the perceived limitations of 2-D
25
gels in the detection of low abundance or very hydrophobic proteins [62] have fueled the search for more high-throughput, non-biased technologies. However, evidence presented in a recent study demonstrates that sample fractionation in combination with the correct solubilization reagents can overcome the bias of 2-D toward high-abundance, cytosolic proteins [60]. These new approaches to 2-D gel technology may expand the application of 2-D arrays since at present, 2-D gels are the only expedient way to analyze post-translational modifications.
3.2. Piezoelectric and ink-jet chemical printing In an innovative departure from the traditional concept of protein chips, some researchers (Proteome Systems, Sydney, Australia) are implementing microfluidic printing of arrayed chemistries on individual protein spots blotted onto membranes [60]. This ‘‘Chemical Printing’’ technique will effectively allow each spot on a 2-D gel blot to act as a protein chip (Fig. 2). The advantages of working with these 2-D ‘‘macroarrays’’ are several. First, the proteins blotted from 2-D gels are authentic, i.e. they represent the proteins as they occur in the cell, including post-translational modifications. Protein macroarrays on membranes are stable and can be archived for indefinite periods of time. Using the piezo-electric Chemical Printing technique, several chemistries can be deposited on a single spot allowing, for example, trypsin digestion and peptide mass fingerprinting, glycosidase digestion and glycopeptide analysis and other chemistries (Fig. 3). Alternatively, antibodies or antisera can be microarrayed onto a single polypeptide spot for the detection of antigenicity in diagnostic applications. Detection can be carried out by placing the macroarrayed membrane directly into a MALDI TOF mass spectrometer, or using fluorescent tags. Other researchers are using in-jet printing technology to create protein microarrays on chips [63].
4. Future directions and challenges Some of the challenges that will be faced by researchers attempting to implement high-throughput microarray technologies for proteomics include
26
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
maintaining diverse protein activities on chips, especially membrane proteins, creating sufficiently diverse antibody or ligand libraries and finding attachment strategies that allow three-dimensional access for binding. Although these hurdles may be difficult to surmount, it is likely that protein micro- and macroarrays will bear fruit, especially when complemented by other ‘‘classic’’ proteomics technologies. In summary, protein analytical applications in the future will benefit from the explosion of new DNAbased technologies in array and detection. The rapid growth of the proteomics market and the active climate for new technology is driving the new generation of companies and academic efforts that are developing novel protein microarray techniques for the future.
5. Notation 2-D, two-dimensional electrophoresis MS, mass spectrometry mass spec, mass spectrometry MALDI TOF, matrix assisted laser desorption / ionization time of flight LC ms / ms, tandem mass spectrometry IEF, isoelectric focusing IPG, immobilized pH gradient
References [1] P.H. O’Farrel, J. Biol. Chem. 250 (1975) 4007. [2] J. Klose, Humangenetik 26 (1975) 231. [3] A. Gorg, W. Postel, S. Gunther, Electrophoresis 9 (1988) 531. [4] S.M. Hanash, J.R. Strahler, J.V. Neel, N. Hailat, R. Melhem, D. Keim, X.X. Zhu, D. Wagner, D.A. Gage, J.T. Watson, Proc. Natl. Acad. Sci. USA 88 (1991) 5709. [5] M.R. Wilkins, C. Pasquali, R.D. Appel, K. Ou, O. Golaz, J.C. Sanchez, J.C. Yan, A. Gooley, G. Hughes, I. HumphriesSmith, K.L. Williams, D. Hochstrasser, Biotechnology 14 (1996) 61. [6] R. Aebersold, L.N. Amankwa, H. Nika, D.T. Chow, E.J. Bures, H.D. Morrison, D. Hess, M. Affolter, J.D. Watts, in: M.Z. Atassi, E. Appella (Eds.), Methods in Protein Structure Analysis, Plenum Press, New York, 1995, p. 3. [7] S. Patterson, R. Aebersold, Electrophoresis 16 (1995) 1791. [8] M. Shevchenko, M. Wilm, O. Vorm, M. Mann, Anal. Chem. 68 (1996) 850.
[9] H.W. Lahm, H. Langen, Electrophoresis 11 (2000) 2105. [10] K.B. Mullis, F.A. Faloona, Methods Enzymol. 155 (1987) 335. [11] G.L. Corthals, M.P. Molloy, B.R. Herbert, K.L. Williams, A. Gooley, Electrophoresis 18 (1997) 317. [12] M. Fountoulakis, M.F. Takacs, P. Berndt, H. Langen, B. Takacs, Electrophoresis 11 (1999) 2181. [13] K. Karlsson, N. Cairns, G. Lubec, K. Fountoulakis, Electrophoresis 14 (1999) 2970. [14] M. Fountoulakis, M.F. Takacs, B. Takacs, J. Chromatogr. A 833 (1999) 157. [15] G.L. Igloi, Methods Mol. Biol. 65 (1996) 23. [16] R.E. Cunningham, Methods Mol. Biol. 115 (1999) 271. [17] M. Schuelke, Nat. Biotechnol. 2 (2000) 233. [18] G. Allen, G. Lowe, Biochem. J. 133 (1973) 679. [19] E.A. First, S.S. Taylor, Biochemistry 28 (1989) 3598. [20] A.N. Kapanidis, Y.W. Ebright, R.H. Ebright, J. Am. Chem. Soc. 123 (2001) 12123. [21] K.N. Berggren, E. Chernoskalskaya, M.F. Lopez, J.M. Beechem, W.F. Patton, Proteomics 1 (2001) 54. [22] F. Kirpekar, E. Nordhoff, K. Kristiansen, P. Roepstorff, A. Lezius, S. Hahner, M. Karas, F. Hillenkamp, Nucleic Acids Res. 22 (1994) 3866. [23] Y. Zhu, L. He, J.R. Srinivasan, D.M. Lubman, Rapid Commun. Mass Spectrom. 11 (1997) 987. [24] E. Nordhoff, C. Luebbert, G. Thiele, V. Heiser, H. Lehrach, Nucleic Acids Res. 28 (2000) E86. [25] V. Egelhofer, K. Bussow, C. Luebbert, H. Lehrach, E. Nordhoff, Anal. Chem. 72 (2000) 2741. [26] J. Raymackers, A. Daniels, V. De Brabandere, C. Missiaen, M. Dauwe, P. Verhaert, E. Vanmechelen, L. Meheus, Electrophoresis 11 (2000) 2266. [27] K. Gevaert, J. Vanderkerckove, Electrophoresis 6 (2000) 1145. [28] B.B. Haab, Curr. Opin. Drug Discov. Dev. 4 (2001) 116. [29] B.B. Haab, M.J. Maitreya, P.O. Brown, Genome Biol. 2 (2001) 4. [30] C.A. Borrebaeck, S. Ekstrom, A.C. Hager, J. Nilsson, T. Laurell, G. Marko-Varga, Biotechniques 5 (2001) 1126. [31] D.J. Cahill, J. Immunol. Methods 250 (2001) 81. [32] R.E. Hayward, G. Cameron, R.Z. Kowzlowski, Drug Discov. Technol. 6 (2001) 1263. [33] M. Li, Nat. Biotechnol. 18 (2000) 1251. [34] R.E. Jenkins, S.R. Pennington, Proteomics 1 (2001) 13. [35] T. Hermann, D.J. Patel, Science 287 (2000) 820. [36] A.D. Ellington, J.W. Szostak, Nature 346 (1990) 818. [37] F. von Eggeling, K. Junker, W. Fiedle, V. Wollscheid, M. Durst, U. Claussen, G. Ernst, Electrophoresis 14 (2001) 2898. [38] M. Verma, G.L. Wright Jr., S.M. Hanash, R. Gopal-Srivastava, S. Srivastava, Ann. NY Acad. Sci. 945 (2001) 103. [39] J.D. Wulfkuhle, K.C. McLean, C.P. Paweletz, D.C. Sgroi, B.J. Trock, P.S. Steeg, E.F. Petricoin 3rd, Proteomics 10 (2001) 1205. [40] S. Wang, D.L. Diamond, G.M. Hass, R. Sokoloff, R.L. Vessella, Int. J. Cancer 6 (2001) 871. [41] M. Merchant, S.R. Weinberger, Electrophoresis 21 (2000) 1164.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27 [42] R.W. Nelson, D. Nedelkov, K.A. Tubbs, Electrophoresis 21 (2000) 1155. [43] C.P. Sonksen, E. Nordhoff, O. Jansson, M. Malmqvist, P. Roepstorff, Anal. Chem. 70 (1998) 2731. [44] H. Zhu, M. Snyder, Curr. Opin. Chem. Biol. 5 (2001) 40. [45] C.P. Paweletz, L. Charboneau, V.E. Bichsel, N.L. Simone, T. Chen, J.W. Gillespie, M.R. Emmert-Buck, M.J. Roth, E.F. Petricoin 3rd, L.L. Liotta, Oncogene 20 (2001) 1981. [46] D.M. Williams, P.A. Cole, Trends Biochem. Sci. 5 (2001) 271. [47] G. Macbeath, S.L. Schreiber, Science 289 (2000) 1760. [48] P. Arenkov, A. Kukhtin, A. Gemmell, S. Voloshcuk, V. Chupeeva, A. Mirzbekov, Anal. Biochem. 278 (2000) 123. [49] H. Zhu, J.F. Klemec, S. Chang, P. Bertone, A. Casamayor, K.G. Klemic, D. Smith, M. Gerstein, M.A. Reed, Nat. Genet. 26 (2000) 283. [50] D.J. Cahill, E. Nordhoff, J.O. O’Brien, J. Klose, H. Eickhoff, H. Lehrach, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 1. [51] T. Ito, K. Tashiro, S. Muta, R. Ozawa, T. Chiba, M. Nishizawa, K. Yamamoto, S. Kuhara, Y. Sakaki, Proc. Natl. Acad. Sci. USA 97 (2000) 1143. [52] P. Uetz, L. Giot, G. Cagney, T.A. Mansfield, R.S. Judson, J.R. Knight, D. Lockshon, V. Narayan, M. Srinivasan, P. Pochart, Nature 403 (2000) 623.
27
[53] A.J. Walhout, R. Sordella, X. Lu, J.L. Hartley, G.F. Temple, M.A. Brasch, N. Thierry-Mieng, M. Vidal, Science 287 (2000) 116. [54] H. Zhu, M. Bilgin, R. Bangham, D. Hall, A. Casamayor, P. Bertone, N. Ian, R. Jansen, S. Bidlingmaier, T. Houfek, T. Mitchell, P. Miller, R.A. Dean, M. Gerstein, M. Snyder, Science 293 (2001) 2101. [55] M.R. Wilkins, J.C. Sanchez, A.A. Gooley, R.D. Appel, I. Humphery-Smith, D.F. Hochstrasser, K.L. Williams, Biotechnology 13 (1995) 19. [56] M.J. Dunn, A. Gorg, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 43. [57] S.J. Fey, P.M. Larsen, Curr. Opin. Chem. Biol. 5 (2001) 26. [58] J. Klose, U. Kobalz, Electrophoresis 16 (1995) 1034. [59] M.G. Pluskal, A. Bogdanova, M.F. Lopez, S. Gutierrez, Proteomics 2 (2) (2002) 145. [60] B.R. Herbert, J.L. Harry, N.H. Packer, A. A Gooley, S.K. Pederson, K.L. Williams, Trends Biotechnol. 19 (2001) S3. [61] M.F. Lopez, S. Melov, Circ. Res. 90 (4) (2002) 380. [62] S.D. Patterson, R. Aebersold, D.R. Goodlett, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 87. [63] A. Roda, M. Guardigli, C. Russo, P. Parsini, M. Baraldini, Biotechniques 28 (2000) 492.
Review
Protein micro- and macroarrays: digitizing the proteome Mary F. Lopez*, Malcolm G. Pluskal Proteome Systems, 14 Gill St., Woburn, MA 01801, USA
Abstract The early applications of microarrays and detection technologies have been centered on DNA-based applications. The application of array technologies to proteomics is now occurring at a rapid rate. Numerous researchers have begun to develop technologies for the creation of microarrays of protein-based screening tools. The stability of antibody molecules when bound to surfaces has made antibody arrays a starting point for proteomic microarray technology. To minimize disadvantages due to size and availability, some researchers have instead opted for antibody fragments, antibody mimics or phage display technology to create libraries for protein chips. Even further removed from antibodies are libraries of aptamers, which are single-stranded oligonucleotides that express high affinity for protein molecules. A variation on the theme of protein chips arrayed with antibody mimics or other protein capture ligand is that of affinity MS where the protein chips are directly placed in a mass spectrometer for detection. Other approaches include the creation of intact protein microarrays directly on glass slides or chips. Although many of the proteins may likely be denatured, successful screening has been demonstrated. The investigation of protein–protein interactions has formed the basis of a technique called yeast two-hybrid. In this method, yeast ‘‘bait’’ proteins can be probed with other yeast ‘‘prey’’ proteins fused to DNA binding domains. Although the current interpretation of protein arrays emphasizes microarray grids of proteins or ligands on glass slides or chips, 2-D gels are technically macroarrays of authentic proteins. In an innovative departure from the traditional concept of protein chips, some researchers are implementing microfluidic printing of arrayed chemistries on individual protein spots blotted onto membranes. Other researchers are using in-jet printing technology to create protein microarrays on chips. The rapid growth of proteomics and the active climate for new technology is driving a new generation of companies and academic efforts that are developing novel protein microarray techniques for the future. 2002 Elsevier Science B.V. All rights reserved. Keywords: Reviews; Microassays; Macroassays; Proteins; Proteomes
Contents 1. Introduction ............................................................................................................................................................................ 1.1. What is a protein array and what have we learned from DNA microarray applications?........................................................ 2. Protein chips ........................................................................................................................................................................... 2.1. Antibody arrays / phage display / antibody mimics .............................................................................................................. 2.2. Affinity capture / MALDI TOF ......................................................................................................................................... 2.3. Protein / peptide arrays ..................................................................................................................................................... 2.4. Protein / protein interactions ............................................................................................................................................. *Corresponding author. Tel.: 11-781-932-9477x3309; fax: 11-781-932-9294. E-mail address: [email protected] (M.F. Lopez). 1570-0232 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1570-0232(02)00336-7
20 20 21 21 24 24 25
20
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
3. Complementary approaches ..................................................................................................................................................... 3.1. 2-D gels are bona fide protein macroarrays ....................................................................................................................... 3.2. Piezoelectric and ink-jet chemical printing ........................................................................................................................ 4. Future directions and challenges ............................................................................................................................................... 5. Notation ................................................................................................................................................................................. References ..................................................................................................................................................................................
1. Introduction The challenge ahead will no longer be the elucidation of the DNA coding sequence of the human genome, but understanding how its gene sequences are expressed. Studying this process at the protein level, termed Proteomics, is expected to be a more complex task than genome sequencing. To meet this challenge, many new technology tools are in development. Some of these new developments have roots in the successful application of microarray and macroarray fabrication. These were heavily influenced by technology, such as photolithography and ink-jet dispensing that was developed for semiconductor and computer hardware applications. Arrays of cDNA sequences and oligonucleotides are in routine use in large-scale analysis of gene expression and variations in gene sequence.
1.1. What is a protein array and what have we learned from DNA microarray applications? Currently, the study of the changes in the level of proteins in the cell, or Proteomics, is limited to serial processing of samples using high-resolution 2D gel technologies [1–5]. This approach, coupled with mass spectrometry, can rapidly yield structural information and hence identity on proteins resolved from complex mixtures [6–9]. Automation of this process will accelerate throughput in the future. Synergistic to this process would be a high-throughput, analytical screening tool to rapidly identify differences between large numbers of samples. This approach would facilitate the location of potential targets for further analysis with existing ‘‘classical proteomics’’ technology platforms. Such a screening tool would process many samples in parallel and be capable of detecting differentially expressed protein targets. Recently, numerous researchers have begun to develop technologies for the creation of microar-
25 25 25 25 26 26
rays of protein-based capture ligands to create such screening tools. The early applications of microarrays and detection technologies were clearly centered on DNAbased applications. This was a direct result of the effort being placed on genome sequencing and the necessity to analyze large numbers of samples in parallel. However, the nature of DNA lends itself very well to some of the technologies employed, such as PCR and labeling with fluorescent reporter molecules. The success of these approaches is clearly a function of the chemistry and enzymology of nucleic acids. In protein-based studies, it may not be possible to exploit similar tools. However, in examining the evolution of new analytical tools for protein-based applications, there are some valuable lessons to be learned from the successes with nucleic acids. Some examples of these applications are detailed below:
(i) Amplification of DNA with PCR is probably the single largest advance that enabled many applications to date [10]. Being able to amplify the target is a great asset in any analytical approach. The closest protein equivalent would be the specific affinity selection or isolation of a defined charge class of proteins by isoelectric focusing from a complex mixture, leading to its enrichment in the final sample [11–14]. This approach, termed complexity reduction or enrichment will be an essential component of future applications in Proteomics. It is not quite the same as PCR but it offers the ability to increase the level of the protein or target of interest by processing a larger volume sample and enriching a specific component (ii) Labeling with reporter molecules, such as fluorescent dyes allows for the increased sensitivity of detection, the ability to multiplex detection of several targets and easy interfacing
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
with detection instrumentation and automation [15–17]. Protein labeling with fluorescent probes has seen many applications in structural studies employing site-specific modifications of reactive amino acid side chains [18–20]. For less protein sequence dependent detection, a new generation of fluorescent dyes or modified surfactant molecules has been developed and offers some of the above attributes [21]. Being able to multiplex samples in future detection strategies will be key to opening up comparative studies of complex protein mixtures. The power of some of the gene expression chip-based approaches lies in their ability to allow for the comparison of control and experimental signals with these multiplexed detection schemes on the same analytical surface. (iii) More recently, MALDI TOF with the appropriate sample energy transfer matrix has opened up rapid analysis of DNA fragments (,150 bp) with extremely high precision sufficient to achieve the equivalent single base resolution of electrophoresis [22–24]. In the protein application, MALDI TOF and electrospray MS–MS have seen spectacular success in resolving peptides and proteins (,12 000 MW) with high mass resolution [25–27]. Larger proteins and nucleic acids are not as well resolved, with peaks being broader due to isotope effects and the influence of post-translational modifications in proteins. Hence, this approach has high intrinsic value as an analytical tool for fragments derived from intact proteins. For whole molecule analysis its application will be different, being more useful as a comparative-screening tool complementing other, more high-resolution approaches, such as 2-D gels or peptide fragmentation / MS.
2. Protein chips
2.1. Antibody arrays /phage display /antibody mimics The relative robustness and stability of antibody molecules as well as the extensive libraries already available have made antibody arrays a starting point for proteomic microarray technology. However, it
21
should be noted that the requirements of antibodies for protein chips might be significantly more rigorous than for other applications in terms of specificity. Indeed, a caveat of antibody arrays is the difficulty of recognizing all the isoforms of a particular protein and its post-translational modifications. Nevertheless, several researchers and companies have demonstrated the initial feasibility of this approach [28–32]. In one such study [29], the authors printed 115 pairs of antibody / antigen pairs on derivatized glass slides using a robotic arrayer. The researchers were able to detect 50% of the arrayed antigens and 20% of the arrayed antibodies in complex mixtures containing the cognate ligands at concentrations ranging from 1 ng / ml to 0.34 mg / ml. These results suggested that protein microarrays could be used for characterization of pattern variation in clinical or research applications. Other researchers have achieved similar results [32] and a number of new companies such as Zyomyx (Hayward, CA, USA), BioSite (San Diego, CA, USA), Hypromatrix and Diversys (Cambridge, UK) are developing protein chip strategies based on antibody arrays. Intact antibody molecules have several drawbacks for protein chip applications including their large size and potential lack of specificity. To overcome these problems, some researchers have turned to antibody fragments [30], antibody mimics, or phage display technology [33] to create libraries for protein chips. Phage display is a method for creating antibody libraries using genetically modified phages that express variable fragments of immunoglobulin on their surfaces. Several new companies such as Dyax (Boston, MA, USA), Cambridge Antibody Technology (UK) and Affitech (Oslo, Norway) are pursuing this approach or a modification of this approach (Phylos, Lexington, MA, USA). Unfortunately, phage displayed antibodies are subject to many of the same difficulties with specificity that are evidenced by conventional antibodies [34]. Display libraries are not necessarily limited to antibodies and can consist of non-biological antibody mimics and diverse molecules [33]. Even further removed from antibodies are libraries of aptamers, which are single-stranded oligonucleotides that express high affinity for protein molecules [34–36]. These aptamer libraries can easily be synthesized and are less labile than antibodies but are
22
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
Fig. 1. Detection of specific carbohydrate modifications of proteins in a microarray format (figure courtesy of K. Martin and W.F. Patton, Molecular Probes Inc.). Alpha-1 acid glycoprotein, horseradish peroxidase, immunoglobulin G, and soybean trypsin inhibitor were arrayed on to PVDF membrane, adhered to a glass slide, from a source plate (384-well) with a concentration of 0.0625–4 mg / ml protein in phosphate buffered saline. Arrays were spotted using a manual glass slide arrayer (V&P Scientific, San Diego, CA, USA) fixed with four rows of eight pins (32 total), |500 mm diameter spot size, 1125 mm horizontal pitch and 750 mm vertical pitch (pitch5center to center spacing of spots). Proteins were arrayed in replicates of six, resulting in an array of 168 spots. (A) The total protein content (left) of the array was labeled with 10 mM BODIPY TR-X, SE (Molecular Probes) in sodium borate buffer, pH 9.5 followed by specific glycoprotein detection with the alkaline phosphatase-conjugated concanavalin A (right). Concanavalin A binds to glycoproteins containing a-mannosyl and a-glucopyranosyl residues as found in horseradish peroxidase and immunoglobulin G. Arrays were blocked with mTBS (50 mM Tris pH 7.5, 150 mM NaCl) containing 0.25% MOWIOL 4-88 and 0.2% Tween-20 before being probed with 1 mg / ml of lectin conjugate in mTBS with 0.5 mM MgCl 2 and 1 mM CaCl 2 . Detection of the lectin was performed by incubating the array in the alkaline phosphatase substrate ELF 39 phosphate (Molecular Probes) at a concentration of 10 mg / ml in 10 mM Tris pH 9.5, 1 mM MgCl 2 . (B) The total protein content (left) of the array was labeled with 10 mM BODIPY FL-X, SE (Molecular Probes) in sodium borate buffer, pH 9.5 followed by specific glycoprotein detection with the alkaline phosphatase-conjugated wheat germ agglutinin (right). Wheat germ agglutinin binds glycoproteins containing N-acetylglucosamine and N-acetylneuraminic acid residues as found in a-1 acid glycoprotein and immunoglobulin G. Arrays were blocked with mTBS (50 mM Tris pH 7.5, 150 mM NaCl) containing 0.25% MOWIOL 4-88 and 0.2% Tween-20 before being probed with 1 mg / ml of wheat germ agglutinin in mTBS with 0.5 mM MgCl 2 and 1 mM CaCl 2 . Detection of the lectin was performed by incubating the array in the alkaline phosphatase substrate DDAO phosphate (Molecular Probes) at a concentration of 1.25 mg / ml in 10 mM Tris pH 9.5, 1 mM MgCl 2 .
Fig. 2. Identification of proteins resolved in a 2DE gel blotted to a PVDF membrane: applications of the Chemical Printer to Trypsin digestion and subsequent MALDI–MS analysis of peptides. Human plasma proteins (2.5 mg) prefractionated using the multi compartment electrolyzer (Proteome Systems, Sydney, Australia) were resolved on a pH range 5–6 IPG strip and then ran in a second dimension 6–15% linear gradient SDS–PAGE gel (GelChip, Proteome Systems, Woburn, MA, USA). The resulting separation was then transferred to a high surface area PVDF Western blotting membrane (Immobilon Psq, Millipore, Danvers, MA, USA) and stained with Direct blue (Sigma, St Louis, MO, USA). A replicate gel was stained directly in a colloidal Coomassie blue stain (Ez-Blue姠, Sigma) for direct in-gel digestion. The protein spots labeled on the stained PVDF blot (Panel A) were analyzed using the prototype Chemical Printer instrument. The entire membrane was positioned on filter paper saturated with 2.5 mM Tris–HCl, pH 8.5. Solutions required for digestion of proteins were applied to the membrane with a piezoelectric device; (A) Trypsin / detergent solution (0.5 mg / ml of Porcine Trypsin [Promega modified trypsin], 2.5 mM Tris–HCl, pH 8.5, 1% [v / v] octyl-glucopyranoside) was jetted onto the blotted proteins in 23100 drops then 23 999 drops to a final dispensed volume of 250 nl. Samples were then incubated in a humidified environment at 30 8C overnight. (B) Following digestion matrix solution (20 mg / ml a-cyano-4-hydroxy cinnamic acid in 70% [v / v] acetonitrile and 0.5% [v / v] formic acid) was jetted (23999 drops) using the same robotic coordinates used for dispensing the trypsin. The PVDF membrane was then transferred to a metal MALDI target and fixed in place using conductive adhesive and analyzed on a KRATOS Axima CFR instrument (Panel B). A parallel in gel digest of spots on the replicate gel was carried out using standard protocols, data shown in Panel C.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
23
24
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
nevertheless subject to nucleases that may be present in complex mixtures of proteins from body fluids used to screen chips [34].
2.2. Affinity capture /MALDI TOF An interesting variation on the theme of protein chips arrayed with antibody mimics or other protein capture ligands is that of affinity MS [37–40]. This technique, termed SELDI and pioneered by Ciphergen Biosystems (Palo Alto, CA, USA), consists of a MALDI target surface that has been modified with affinity ligands to capture proteins of interest. The target can then be inserted directly into the mass spectrometer for detection [41]. A different version of this approach uses chips that employ surface plasmon resonance to quantify the target capture. The chips may then also be analyzed by MALDI for characterization of the captured proteins [42,43]
brain cDNA library. The fusion proteins were arrayed onto glass and membrane surfaces and were probed with antibodies. Although many of the proteins were likely denatured, successful screening was demonstrated. Some companies are now taking a similar approach to generate collections of recombinant proteins. In one case, full-length proteins were expressed and immobilized via an affinity tag preserving full protein function [32]. In a further extension of these approaches, researchers have developed a technique for detecting post-translational carbohydrate modifications of proteins in microarrays (Fig. 1, courtesy of K. Martin and W.F. Patton, Molecular Probes, Eugene, OR, USA). In this experiment, dichromatic fluorescent dyes were used to detect and quantify the total microarrayed proteins and then, using different excitation / emission filters, reporter enzyme–lectin conjugates monitored the glycosylation status of the proteins.
2.3. Protein /peptide arrays Several groups have taken the approach of creating microarrays of proteins directly on chips [44– 46]. In a recent report [47], proteins were attached to a glass slide that was activated with a cross linking agent that reacts with primary amines. They tested protein–protein interactions, kinase substrate reactions and protein–ligand reactions using fluorescent probes and radiolabeled ATP. In one experiment, a single protein was identified within an array of 10 000 other proteins. This demonstrates the feasibility of using protein microarrays for largescale, high-throughput screening studies. In order to maintain the three-dimensional integrity of proteins spotted onto surfaces, some researchers have developed a technique for immobilizing the proteins in polyacrylamide gel pads fixed onto glass slides [48]. Other workers have developed a nanowell technology and assayed 117 different yeast kinases with 17 different substrates [49]. The results of this experiment suggested the existence of novel yeast tyrosine kinases. On a much larger protein expression scale, researchers at the Max Planck Institute [31,32,50] have created a library of approximately 15 000 nonredundant His-tagged proteins from a human fetal
Fig. 3. Chemical printing of a microarray of reagent spots on a macroarray protein target. High density printing of reagents in a microarray format on a single gel spot resolved in a 2DE macroarray separation, demonstrates the potential of the Chemical Printer. The entire membrane was positioned on filter paper saturated with 2.5 mM Tris–HCl, pH 8.5. Solutions required for digestion of proteins were applied to the stained post located on the membrane with a piezoelectric device; Trypsin / detergent solution (0.5 mg / ml of Porcine Trypsin [Promega modified trypsin], 2.5 mM Tris–HCl, pH 8.5, 1% [v / v] octyl-glucopyranoside) was jetted onto a single blotted protein spot in 23100 drops then 23 999 drops to a final dispensed volume of 250 nl. A 434 spot microarray format was over layered on a single spot.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
2.4. Protein /protein interactions In a ‘‘targeted proteomics’’ approach, the investigation of protein–protein interactions has formed the basis of a technique called yeast two-hybrid [51–53]. This method typically involves the generation of yeast ‘‘bait’’ proteins fused to transcriptionactivation DNA domains. The bait proteins can then be probed with other yeast proteins (‘‘prey’’) fused to DNA binding domains. Using this approach, researchers recently created an array of 5300 yeast bait proteins and identified 957 potential protein– protein interactions by probing with yeast prey proteins [52]. A company (Myriad Genetics, Salt Lake City, UT, USA) plans to use this approach to map all the protein–protein interactions in yeast. A recent report described the cloning and expression of 5800 yeast proteins with their subsequent deposition onto a microarray [54]. The high-density microarray was screened for interaction with other proteins and phospholipids. Using this method, the authors identified many novel calmodulin and phospholipid binding proteins.
3. Complementary approaches
3.1. 2 -D gels are bona fide protein macroarrays Although the current interpretation of protein arrays emphasizes microarray grids of proteins or ligands on glass slides or chips, it is interesting to note that 2-D gels (currently the workhorse of proteomics) are technically macroarrays of authentic proteins, i.e. charge, molecular mass and posttranslational modifications are preserved during electrophoresis. This allows direct analysis of protein isoforms that may be involved in particular metabolic or disease processes. Due to the true parallel nature of the technique, hundreds to thousands of proteins can be visualized simultaneously and when coupled with mass spectrometry, complex mixtures of proteins can be resolved and individual components identified [5,55–61]. In addition, the quantitative differences between proteins in mixtures can be determined from 2-D gel images allowing the direct detection of differentially expressed gene products [55–61]. Unfortunately, the perceived limitations of 2-D
25
gels in the detection of low abundance or very hydrophobic proteins [62] have fueled the search for more high-throughput, non-biased technologies. However, evidence presented in a recent study demonstrates that sample fractionation in combination with the correct solubilization reagents can overcome the bias of 2-D toward high-abundance, cytosolic proteins [60]. These new approaches to 2-D gel technology may expand the application of 2-D arrays since at present, 2-D gels are the only expedient way to analyze post-translational modifications.
3.2. Piezoelectric and ink-jet chemical printing In an innovative departure from the traditional concept of protein chips, some researchers (Proteome Systems, Sydney, Australia) are implementing microfluidic printing of arrayed chemistries on individual protein spots blotted onto membranes [60]. This ‘‘Chemical Printing’’ technique will effectively allow each spot on a 2-D gel blot to act as a protein chip (Fig. 2). The advantages of working with these 2-D ‘‘macroarrays’’ are several. First, the proteins blotted from 2-D gels are authentic, i.e. they represent the proteins as they occur in the cell, including post-translational modifications. Protein macroarrays on membranes are stable and can be archived for indefinite periods of time. Using the piezo-electric Chemical Printing technique, several chemistries can be deposited on a single spot allowing, for example, trypsin digestion and peptide mass fingerprinting, glycosidase digestion and glycopeptide analysis and other chemistries (Fig. 3). Alternatively, antibodies or antisera can be microarrayed onto a single polypeptide spot for the detection of antigenicity in diagnostic applications. Detection can be carried out by placing the macroarrayed membrane directly into a MALDI TOF mass spectrometer, or using fluorescent tags. Other researchers are using in-jet printing technology to create protein microarrays on chips [63].
4. Future directions and challenges Some of the challenges that will be faced by researchers attempting to implement high-throughput microarray technologies for proteomics include
26
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27
maintaining diverse protein activities on chips, especially membrane proteins, creating sufficiently diverse antibody or ligand libraries and finding attachment strategies that allow three-dimensional access for binding. Although these hurdles may be difficult to surmount, it is likely that protein micro- and macroarrays will bear fruit, especially when complemented by other ‘‘classic’’ proteomics technologies. In summary, protein analytical applications in the future will benefit from the explosion of new DNAbased technologies in array and detection. The rapid growth of the proteomics market and the active climate for new technology is driving the new generation of companies and academic efforts that are developing novel protein microarray techniques for the future.
5. Notation 2-D, two-dimensional electrophoresis MS, mass spectrometry mass spec, mass spectrometry MALDI TOF, matrix assisted laser desorption / ionization time of flight LC ms / ms, tandem mass spectrometry IEF, isoelectric focusing IPG, immobilized pH gradient
References [1] P.H. O’Farrel, J. Biol. Chem. 250 (1975) 4007. [2] J. Klose, Humangenetik 26 (1975) 231. [3] A. Gorg, W. Postel, S. Gunther, Electrophoresis 9 (1988) 531. [4] S.M. Hanash, J.R. Strahler, J.V. Neel, N. Hailat, R. Melhem, D. Keim, X.X. Zhu, D. Wagner, D.A. Gage, J.T. Watson, Proc. Natl. Acad. Sci. USA 88 (1991) 5709. [5] M.R. Wilkins, C. Pasquali, R.D. Appel, K. Ou, O. Golaz, J.C. Sanchez, J.C. Yan, A. Gooley, G. Hughes, I. HumphriesSmith, K.L. Williams, D. Hochstrasser, Biotechnology 14 (1996) 61. [6] R. Aebersold, L.N. Amankwa, H. Nika, D.T. Chow, E.J. Bures, H.D. Morrison, D. Hess, M. Affolter, J.D. Watts, in: M.Z. Atassi, E. Appella (Eds.), Methods in Protein Structure Analysis, Plenum Press, New York, 1995, p. 3. [7] S. Patterson, R. Aebersold, Electrophoresis 16 (1995) 1791. [8] M. Shevchenko, M. Wilm, O. Vorm, M. Mann, Anal. Chem. 68 (1996) 850.
[9] H.W. Lahm, H. Langen, Electrophoresis 11 (2000) 2105. [10] K.B. Mullis, F.A. Faloona, Methods Enzymol. 155 (1987) 335. [11] G.L. Corthals, M.P. Molloy, B.R. Herbert, K.L. Williams, A. Gooley, Electrophoresis 18 (1997) 317. [12] M. Fountoulakis, M.F. Takacs, P. Berndt, H. Langen, B. Takacs, Electrophoresis 11 (1999) 2181. [13] K. Karlsson, N. Cairns, G. Lubec, K. Fountoulakis, Electrophoresis 14 (1999) 2970. [14] M. Fountoulakis, M.F. Takacs, B. Takacs, J. Chromatogr. A 833 (1999) 157. [15] G.L. Igloi, Methods Mol. Biol. 65 (1996) 23. [16] R.E. Cunningham, Methods Mol. Biol. 115 (1999) 271. [17] M. Schuelke, Nat. Biotechnol. 2 (2000) 233. [18] G. Allen, G. Lowe, Biochem. J. 133 (1973) 679. [19] E.A. First, S.S. Taylor, Biochemistry 28 (1989) 3598. [20] A.N. Kapanidis, Y.W. Ebright, R.H. Ebright, J. Am. Chem. Soc. 123 (2001) 12123. [21] K.N. Berggren, E. Chernoskalskaya, M.F. Lopez, J.M. Beechem, W.F. Patton, Proteomics 1 (2001) 54. [22] F. Kirpekar, E. Nordhoff, K. Kristiansen, P. Roepstorff, A. Lezius, S. Hahner, M. Karas, F. Hillenkamp, Nucleic Acids Res. 22 (1994) 3866. [23] Y. Zhu, L. He, J.R. Srinivasan, D.M. Lubman, Rapid Commun. Mass Spectrom. 11 (1997) 987. [24] E. Nordhoff, C. Luebbert, G. Thiele, V. Heiser, H. Lehrach, Nucleic Acids Res. 28 (2000) E86. [25] V. Egelhofer, K. Bussow, C. Luebbert, H. Lehrach, E. Nordhoff, Anal. Chem. 72 (2000) 2741. [26] J. Raymackers, A. Daniels, V. De Brabandere, C. Missiaen, M. Dauwe, P. Verhaert, E. Vanmechelen, L. Meheus, Electrophoresis 11 (2000) 2266. [27] K. Gevaert, J. Vanderkerckove, Electrophoresis 6 (2000) 1145. [28] B.B. Haab, Curr. Opin. Drug Discov. Dev. 4 (2001) 116. [29] B.B. Haab, M.J. Maitreya, P.O. Brown, Genome Biol. 2 (2001) 4. [30] C.A. Borrebaeck, S. Ekstrom, A.C. Hager, J. Nilsson, T. Laurell, G. Marko-Varga, Biotechniques 5 (2001) 1126. [31] D.J. Cahill, J. Immunol. Methods 250 (2001) 81. [32] R.E. Hayward, G. Cameron, R.Z. Kowzlowski, Drug Discov. Technol. 6 (2001) 1263. [33] M. Li, Nat. Biotechnol. 18 (2000) 1251. [34] R.E. Jenkins, S.R. Pennington, Proteomics 1 (2001) 13. [35] T. Hermann, D.J. Patel, Science 287 (2000) 820. [36] A.D. Ellington, J.W. Szostak, Nature 346 (1990) 818. [37] F. von Eggeling, K. Junker, W. Fiedle, V. Wollscheid, M. Durst, U. Claussen, G. Ernst, Electrophoresis 14 (2001) 2898. [38] M. Verma, G.L. Wright Jr., S.M. Hanash, R. Gopal-Srivastava, S. Srivastava, Ann. NY Acad. Sci. 945 (2001) 103. [39] J.D. Wulfkuhle, K.C. McLean, C.P. Paweletz, D.C. Sgroi, B.J. Trock, P.S. Steeg, E.F. Petricoin 3rd, Proteomics 10 (2001) 1205. [40] S. Wang, D.L. Diamond, G.M. Hass, R. Sokoloff, R.L. Vessella, Int. J. Cancer 6 (2001) 871. [41] M. Merchant, S.R. Weinberger, Electrophoresis 21 (2000) 1164.
M.F. Lopez, M.G. Pluskal / J. Chromatogr. B 787 (2003) 19–27 [42] R.W. Nelson, D. Nedelkov, K.A. Tubbs, Electrophoresis 21 (2000) 1155. [43] C.P. Sonksen, E. Nordhoff, O. Jansson, M. Malmqvist, P. Roepstorff, Anal. Chem. 70 (1998) 2731. [44] H. Zhu, M. Snyder, Curr. Opin. Chem. Biol. 5 (2001) 40. [45] C.P. Paweletz, L. Charboneau, V.E. Bichsel, N.L. Simone, T. Chen, J.W. Gillespie, M.R. Emmert-Buck, M.J. Roth, E.F. Petricoin 3rd, L.L. Liotta, Oncogene 20 (2001) 1981. [46] D.M. Williams, P.A. Cole, Trends Biochem. Sci. 5 (2001) 271. [47] G. Macbeath, S.L. Schreiber, Science 289 (2000) 1760. [48] P. Arenkov, A. Kukhtin, A. Gemmell, S. Voloshcuk, V. Chupeeva, A. Mirzbekov, Anal. Biochem. 278 (2000) 123. [49] H. Zhu, J.F. Klemec, S. Chang, P. Bertone, A. Casamayor, K.G. Klemic, D. Smith, M. Gerstein, M.A. Reed, Nat. Genet. 26 (2000) 283. [50] D.J. Cahill, E. Nordhoff, J.O. O’Brien, J. Klose, H. Eickhoff, H. Lehrach, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 1. [51] T. Ito, K. Tashiro, S. Muta, R. Ozawa, T. Chiba, M. Nishizawa, K. Yamamoto, S. Kuhara, Y. Sakaki, Proc. Natl. Acad. Sci. USA 97 (2000) 1143. [52] P. Uetz, L. Giot, G. Cagney, T.A. Mansfield, R.S. Judson, J.R. Knight, D. Lockshon, V. Narayan, M. Srinivasan, P. Pochart, Nature 403 (2000) 623.
27
[53] A.J. Walhout, R. Sordella, X. Lu, J.L. Hartley, G.F. Temple, M.A. Brasch, N. Thierry-Mieng, M. Vidal, Science 287 (2000) 116. [54] H. Zhu, M. Bilgin, R. Bangham, D. Hall, A. Casamayor, P. Bertone, N. Ian, R. Jansen, S. Bidlingmaier, T. Houfek, T. Mitchell, P. Miller, R.A. Dean, M. Gerstein, M. Snyder, Science 293 (2001) 2101. [55] M.R. Wilkins, J.C. Sanchez, A.A. Gooley, R.D. Appel, I. Humphery-Smith, D.F. Hochstrasser, K.L. Williams, Biotechnology 13 (1995) 19. [56] M.J. Dunn, A. Gorg, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 43. [57] S.J. Fey, P.M. Larsen, Curr. Opin. Chem. Biol. 5 (2001) 26. [58] J. Klose, U. Kobalz, Electrophoresis 16 (1995) 1034. [59] M.G. Pluskal, A. Bogdanova, M.F. Lopez, S. Gutierrez, Proteomics 2 (2) (2002) 145. [60] B.R. Herbert, J.L. Harry, N.H. Packer, A. A Gooley, S.K. Pederson, K.L. Williams, Trends Biotechnol. 19 (2001) S3. [61] M.F. Lopez, S. Melov, Circ. Res. 90 (4) (2002) 380. [62] S.D. Patterson, R. Aebersold, D.R. Goodlett, in: S.R. Pennington, M.J. Dunn (Eds.), Proteomics from Protein Sequence to Function, Bios Scientific Publishers, Oxford, UK, 2001, p. 87. [63] A. Roda, M. Guardigli, C. Russo, P. Parsini, M. Baraldini, Biotechniques 28 (2000) 492.