One of the central goals of this project is to develop protein-detecting arrays (PDAs). As will be discussed below, we will construct specific arrays containing hundreds of features in house as tools in understanding sleep biology. But the important point is that we generate general solutions to the major technical hurdles in the protein array area. With the methods we develop, the size of the array that one could construct would be limited only by the manpower and budget available. We hope that this work will pave the way for the construction of chips that rival the current level of sophistication of DNA microarrays, though this would almost certainly have to be done in a well-funded commercial enterprise. But with these methods, larger academic laboratories could employ these methods to construct modest-sized custom PDAs to address problems of particular interest to them.

Figure 1, sandwich-type assay.

Fig. 1. shows a schematic diagram of a sandwich-type assay. Here, the capture molecule is a single chain antibody, but in our work it will be a synthetic molecule (Figure provided courtesy of Prof. Brent Iverson, UT-Austin).

All of our efforts will employ the simple sandwich assay. Captured ligands with high affinity and specificity for a particular protein will be arrayed on an appropriate surface. The sample of interest will then be applied to the array, allowed to equilibrate, and unbound material will be washed off, in some cases after fixing binding by covalent trapping. To quantitate the amount of target protein bound to each feature of the array, a sandwich assay will be used. Labeled antibodies raised against the target proteins of interest will be added and after washing, the amount of label trapped at each feature will be quantitated using the appropriate scanner.

Many people in the proteomics area are focused on the development of much more sophisticated methods to detect protein binding to a capture ligand such as multi-site surface plasmon resonance (SPR) chips and other “smart surfaces” that intrinsically register a binding event . However, we believe that the “low tech” sandwich assay approach offers some important advantages over physical detection methods. Most importantly, two independent binding events to the same protein must occur to register a signal, providing high specificity. This will be very important when making measurements in cell extracts, where most proteins are associated with other factors. This will be a major complicating factor in obtaining quantitative data from SPR or other schemes that essentially register changes in mass at a particular array feature. Finally, the major drawback of the sandwich approach is that twice as many protein-binding compounds are required, but we believe that developments in our lab have reduced the significance of this issue greatly.

Production of sandwich antibodies
We believe that antibodies produced by genetic immunization (see toolbox ) will make excellent sandwich reagents. Unlike capture agents, sandwich compounds need not be rigorously purified. Indeed, hundreds or thousands of different sandwich reagents will have to be added simultaneously in a PDA experiment so purification is a moot point. We plan to simply purify an IgG fraction from the serum of immunized mice, biotinylate this antibody mixture, and use it as a sandwich reagent (along with streptavidin-phycoerythryin). All antibodies that do not bind immobilized proteins will simply be washed away. Since genetic immunization does not require purified proteins, the number of antibodies that can be made per unit time per co-worker is much higher than using standard methods. We will maintain a core facility throughout the lifetime of the project for the constant production of antibodies by genetic immunization, both to support the PDA efforts as well as other aspects of the project.

High-throughput production of synthetic capture agents
No matter what method is employed to detect protein binding to the microarray (sandwich assay, SPR, mass spectrometry, etc.) generation of thousands of useful binding agents is the major technical roadblock to the development of useful PDAs. Early efforts in this area have understandably focused on the use of antibodies, with some cutting edge work involving non-antibody protein aptamers and nucleic acid aptamers . Our work will be distinguished by focusing solely on the development of relatively low molecular weight synthetic capture agents, which have numerous advantages over biomolecules. They can be made in large quantities and purified rigorously much more easily and cheaply than biologically produced macromolecules. They are more robust than any protein aptamer or antibody since no stable tertiary structure must be maintained. Finally, a number of straightforward methods are available to screen libraries of peptides, peptidomimetics or other small molecules for protein ligands, many of which are amenable to high throughput. Unfortunately, synthetic ligands that come out of primary screens have affinities that are simply too low to be of practical utility. In general, peptides and small molecules bind proteins with at least 1000-fold less affinity than a good antibody. No one has yet developed a general strategy to bridge this formidable gap without resorting to tedious structure/activity relationship (SAR)-type methods. We intend to solve this problem and develop a facile route to synthetic protein-binding compounds that exhibit binding characteristics similar to that of a good antibody.

Citation Guidelines

Website use statistics

Web page maintained by Webmaster

Last Update: Friday, August 8, 2003.