| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Abstracts of Oak Ridge Posters |
1 SomaLogic, Inc., 1745 38th St., Boulder, CO 80301
aauthor for correspondence: fax 303-545-2525, e-mail hpetach{at}somalogic.com
To study and identify the complex protein expression patterns associated with a disease, efficient methods are necessary to detect and quantify hundreds of proteins simultaneously, many of which are present in exceedingly low concentrations. Photoaptamers are intriguing capture agents for multiplexed proteomics assays because they demonstrate extraordinary specificity and sensitivity toward protein analytes and can be used in the multiplexed array format. Photoaptamers are single-stranded DNA molecules that have the ability to form covalent bonds with their cognate proteins when they are electronically excited.
Photoaptamers have been discovered for proteins with a wide range of characteristics, including acidic, basic, large, small, glycosylated, chemically modified, and hydrophobic, and these photoaptamers may be used in a wide variety of formats, including most formats available to antibodies (1)(2)(3)(4)(5). The photoSELEX process has been successfully automated as a high-throughput process so that a wide range of proteins have yielded active photoaptamers that exhibit nanomolar or better affinities (6)(7)(8)(9). The photoaptamers described below were selected for the proteins thrombin and basic fibroblast growth factor (bFGF) (10)(11)(12)(13). The sensitivities, specificities, and cross-linking efficiencies of the photoaptamers suggest that they will be suitable capture agents for microarrays. For example, the bFGF photoaptamer has a Kd of 
16 pmol/L, a photo cross-linking yield of 50%, and specific binding to bFGF that is more than three orders of magnitude higher than binding to other heparin-binding proteins, such as vascular endothelial growth factor and platelet-derived growth factor (13). The bFGF photoaptamer characteristics and sequence have been described previously (13).
We identified photoaptamers in vitro by use of the photoSELEX process (12)(14) to select high-affinity sequences from a random pool of oligonucleotides. Photoaptamers were synthesized at SomaLogic, Inc. by use of standard phosphoramidite chemistry (15)(16), with a brominated deoxyuridine substituted for the thymidine (T) usually found in DNA; these photoactive residues participate in covalent bond formation. The photoaptamers were deprotected by use of a mild base to avoid the degradation of the brominated deoxyuridine, a chemically sensitive moiety.
Photoaptamers have previously been shown to cross-link to specific sites on their target proteins, as assessed by Edman and mass spectral sequencing methodologies (17). The array methods described below allowed the parallel assessment of several photoaptamers for both affinity and specificity. They also served as a model for how photoaptamer arrays can be used to measure the concentrations of numerous distinct proteins in a single 100-µL sample. For the multiplexed assays described below, photoaptamers were synthesized with an amine on the 5' terminus to provide a covalent anchor to an array surface.
Both the thrombin and bFGF photoaptamers were characterized using the same experimental approach as described previously (13). Because photoaptamers are chemically stable and renature readily, harsh conditions, including base, chaotropic agents, and detergents, can be used during immobilization and washing to create high-density aptamer loading with homogeneous orientation and activity.
Amine-substituted photoaptamers were spotted onto Motorola CodeLink activated slide surfaces (25 x 75 mm) from a phosphate spotting solution at pH 8.5 (containing 150 mmol/L sodium phosphate and 10 µmol/L oligonucleotide) by use of the Genemachine Accent contact spotter with a Telechem Stealth SMP4B pin. Spots were 100 µm in diameter and were separated by a spot-to-spot distance of 250 µm. After 60 min, the slides were finished by soaking in 20 mmol/L NaOH followed by a 0.2 g/L sulfo-NHS-acetate treatment to ensure that residual N-hydroxysuccinimide (NHS) and amine reactive groups were capped to prevent further reaction. Microarrays typically contained 200 photoaptamer spots, but could be configured to array from one to hundreds of photoaptamers on a single chip. The microarray was incubated with 100 µL of the relevant analyte solution (see discussion of multiplexed analytes below), and the analyte solution was flowed across the slide surface in a continuous loop. Photoaptamers have been specifically attached to a variety of surfaces, including multiplexed array surfaces, by incorporating appropriate reactive groups during their synthesis.
After incubation of the binding solution, the microarray was washed in HEPES buffer at pH 7.4 for several minutes to remove nonspecifically bound protein, then the array was irradiated at 308 nm with a XeCl excimer laser to photo cross-link the photoaptamer to its cognate protein. Exposing the protein/aptamer complex to ultraviolet light induces covalent bond formation between the photoaptamer and cognate protein. Harsh denaturing washes were applied to remove any noncovalently bound protein from the array. The microarray was then treated with 6.3 mg/L NHS-Alexa555 for 30 min or other amine-specific reagents coupled to enzymatic or fluorescent probes to label the lysines on the covalently linked proteins.
The processed microarray was read in an Applied Precision ArrayWoRxe Biochip Reader to quantify the fluorescence from covalently bound protein on each photoaptamer feature.
Microarrays were incubated with a protein analyte solution containing mixtures of proteins to test the photoaptamers for both specificity and sensitivity. We attempted to ensure that photoaptamers exhibiting nonspecific binding and cross-linking or other interferences were detected and eliminated from the photoaptamer collection by requiring each protein pair to be sampled in two ways: (a) in one solution, one protein was present in excess over the second protein; and (b) in the second solution, the second protein was present in excess of the first protein. For eight proteins distributed in eight samples, there are 28 such pairs in each sample (7 comparisons to the photoaptamer with no protein, 6 comparisons for the lowest concentration protein, and so forth). In the current example, each protein pair will differ by at least one log of concentration in at least one solution (on average, two), ensuring that cross-reactivity is detected in the assay.
In this experiment, each cognate protein concentration was varied from 0.01 to 10 nmol/L, whereas seven other proteins were varied over the same concentration range. The total concentration in each microarray experiment was 11.1 nmol/L protein, containing various concentrations of endostatin, bFGF, thrombin, angiogenin, tumor growth factor-ß1, interleukin-4, p-selectin, and serum amyloid p component.
To assess the quality of photoaptamers for binding to a specific protein target, the experiments must include two key qualities that are required for use on a proteomics chip: the dynamic range of the linear response and assessment of specificity. A multiplex assay for such an evaluation would allow for greater efficiency than testing photoaptamers in individual experiments.
To assess surface activity of the photoaptamers, eight-point binding curves for each photoaptamer were generated. The multiplexed analyte solutions were used to simultaneously evaluate dose–response curves for photoaptamers to eight protein targets using eight complex solutions. Specific binding curves are shown for both the thrombin photoaptamer and bFGF photoaptamer. The images of the microarray (Fig. 1A
) illustrate the increasing signals on photoaptamer features in response to increasing cognate protein concentrations. The quantification of the microarray images yielded the protein binding curves below (Fig. 1B
). The presence of multiple analytes during binding incubation on the array does not alter the specific binding curve for the cognate photoaptamer. The thrombin binding curve to its cognate aptamer demonstrates a limit of detection of 100 pmol/L; this limit was determined by slide background (noise), not the capabilities of the photoaptamers as capture agents.
|
Because each of the analyte solutions contained a different high-concentration protein, the binding curves demonstrate specificity of the photoaptamer reaction to its cognate protein. Nonspecific binding would be evident if some points along the binding curve were dramatically displaced, indicating that a noncognate protein interacted with the photoaptamer. The binding curves in this experiment (Fig. 1B
) demonstrate specific binding between photoaptamer and cognate protein.
Because many photoaptamers to the same cognate protein can be arrayed together, the performance of those individual photoaptamers can be compared directly for selection of the most sensitive and specific photoaptamer to a protein target. Specificity can be determined over a wide range of concentrations of noncognate proteins without evidence of interference.
The alluring feature of microarrays is the multiplexing capacity, which yields massively parallel analysis of complex mixtures of proteins from serum, plasma, cell lysates, and even tissue extracts separated by laser capture dissection (18). The advantages of using photoaptamers in microarrays instead of multiplexed antibodies are many, most notably their ease of manipulation and the elimination of cross-reactivity associated with the second dimension of an ELISA.
High specificity in a photoaptamer array is achieved by combining the intrinsic affinity of a photoaptamer for its target protein with the demand for photoactivated cross-linking between the brominated deoxyuridine in the photoaptamer and a specific amino acid residue of the target protein. The additional specificity of the second dimension (photo cross-linking) derives from the requirement that the target amino acid must be positioned appropriately both in distance and orientation to yield the covalent bond (13)(19). In this sense, the photo cross-linking event provides a second level of specificity analogous to the secondary antibody of an ELISA.
In addition to adding specificity to protein/aptamer pairs, the photoactivated covalent cross-linking effectively attaches each analyte molecule to its appropriate address on a microarray surface. The cross-link is stable to stringent wash conditions, allowing the removal of nonspecifically bound proteins before labeling and detection. The result should improve limits of detection and quantification, which will make protein microarrays truly useful tools by enabling the parallel measurement of even low-abundance proteins. Experiments are underway to test these arrays in complex biological solutions, but those data are not presented here.
Because photoaptamers are nucleic acids, a simple universal protein stain can be used to generate signal. A universal protein stain will chemically couple to protein moieties, e.g., lysine residues, with virtually no reactivity to photoaptamer features. This unique approach allows a single binding event, rather than a "sandwich" to generate sensitive, specific results in a multiplex array format. In the absence of the photo cross-linking event, we expect the background of the assay to be increased and adversely affect the lower limit of detection.
Arrays containing from one to hundreds of photoaptamers can be readily created for a wide variety of protein targets by use of the PhotoSELEX process for photoaptamer selection and the array protocols developed for those photoaptamers. These arrays should scale to any feature density without complications caused by secondary antibodies.
Acknowledgments
We thank the following individuals for their contributions to the assay experimental work: Steve Buhl, Steve Tyrrell, Jim Heil, Greg Husar, Drew Smith, and Glenn Foulds.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
