A comprehensive understanding of protein pathways in cellular processes and in disease states, as well as the identification of novel disease biomarkers, will require the ability to monitor tens or even hundreds of thousands of protein species in parallel. This represents a significant technological challenge that will involve the fabrication of functional high-density micrometer- or submicrometer-scale arrays, and much attention has been paid to the development of suitable techniques to enable such investigations [1
]. While the problem of highly multiplexed monitoring is tractable at the level of DNA [4
], the problem becomes significantly more complex when seeking to interrogate proteins, for which several alternatively spliced variants may be expressed from a single gene, with each variant then being subject to posttranslational modifications that regulate its activity and conformation. Thus, the development of the protein equivalent of the DNA microarray – the protein array – faces several challenges. The first of these is the identification of specific, high-affinity robust probe molecules that can bind each of the range of conformations that native proteins can adopt. A second is preserving these conformations for analysis, which will require the development of label-free sensing strategies. A third is the fact that strategies must be suitable for the detection of low-abundance proteins in complex biological solutions. Finally, the fabrication of such arrays will require a technology capable of immobilizing specific probe molecules with micrometer or submicrometer accuracy, as the total surface area required to present very large numbers of probes will dictate the minimum sample volume that can be investigated.
Currently, most protein microarrays are produced by dot printing and consist of surface-immobilized antibodies; interactions with sample ligands are detected optically [6
]. However, the antibodies used are generally made against denatured, prokaryotically expressed proteins either by immunization of animals or selection in vitro
from phage display libraries. Consequently, the antibodies predominantly recognize linear epitopes, which severely hinders their usefulness in applications to detect native proteins in complex biological mixtures. Antibodies are also relatively fragile molecules. Furthermore, when immobilized on a surface they often exhibit affinities for a number of different ligands, and hence lack specificity [7
]. Alternative probe molecules, for example DNA and RNA aptamers [8
], have been used in protein arrays, and although likely to be more robust than antibodies, they are also traditionally selected for binding to prokaryotically expressed proteins that may not be correctly folded, and will not be correctly posttranslationally modified. Another problem with current methodologies is that, although several label-free detection strategies have been discussed, including surface-plasmon resonance [12
], mass spectroscopy [13
], and atomic force microscopy-based techniques [14
], the predominant method for the detection of probe-target interactions is based on fluorescent labeling of the target proteins. The fluorescent dyes are typically hydrophobic and may lead to conformational changes when covalently bound to a protein that will mask biologically relevant conformations [6
]. Finally, the dot-printing techniques applied to conventional protein arrays offer resolutions of around 0.1 mm. Although the resolution of these printing and sensing techniques has improved in recent years, the feature sizes required for high-density protein arrays capable of handling extremely small sample volumes are beyond the scope of such systems. Alternative techniques, which enable the immobilization of probe molecules on surfaces such as metal electrodes with resolutions in the micro- or nanometer range, are required.
Here we offer solutions to each of these problems. First, we use robust, in vivo
selected peptide aptamers as probe molecules to replace antibodies. Second, we use a novel molecular resist technology to functionalize different closely spaced gold electrodes with different peptide aptamers, resulting in feature sizes more than one order of magnitude smaller than current protein-array technologies. This functionalization technique is inherently scalable to submicrometer and nanometer feature sizes [16
]. Third, we use an electronic, label-free method to detect binding events occurring between target proteins in solution and peptide aptamers immobilized on high-density electrode arrays. The array devices are fabricated using conventional silicon microfabrication technology, where miniaturization is well established. This, coupled with the scalability of the electrode functionalization technique, yields the potential for submicrometer, high-density functional devices with integrated readout technology capable of performing the many simultaneous measurements required for proteome-wide studies.