Human cells have only about 20,000 genes. But thanks to alternative splicing and post-translational modification, their instructions can manifest in millions of ways. The trick, says Christian Loch, director of research and development at AVMBioMed, a service provider located in Malvern, Pa., is tapping that information.
Mass spectrometry is perhaps the most popular proteomics tool for doing that, but it’s not the only one. Despite rumors to the contrary, protein microarrays also are alive and well, says Guene Lynne Thio, research and development manager for protein biology at Thermo Fisher Scientific.
Protein microarrays—addressable arrays of proteins (or antibodies) on a flat surface, which typically are probed using antibodies or interacting proteins—can be used to interrogate biological samples for changes in relative protein abundance or post-translational modification, to examine protein-protein interactions and enzyme substrates and to screen for biomarkers and drug candidates.
“The microarrays are a unique tool in that they can be used to screen thousands of interactions in a single reaction in a day,” Thio says. “They are very easy to use, require very little material, and the arrays are very flexible and sensitive.”
“The microarrays are a unique tool in that they can be used to screen thousands of interactions in a single reaction in a day,”
The flip side is they are relatively difficult and expensive to manufacture. Unlike nucleic acids, which behave more or less identically regardless of sequence, proteins can vary dramatically in stability, activity and expression level. Obtaining reliable and consistent protein yields, and getting those proteins to play nicely together on a single surface, represent significant challenges for array designers. Nevertheless, a variety of arrays and related tools now are commercially available.
Array of options
Protein microarrays are available in two basic flavors. Some, like the HuProt™ v2.0 19K Human Proteome Microarray from ArrayIt Corp. and CDI Laboratories, and the ProtoArray® Human Protein Microarray from Thermo Fisher Scientific, comprise full-length cellular proteins—19,275 human proteins in the case of the HuProt array, and “over 9,000” human proteins in the case of the ProtoArray.
Researchers can use these to screen for protein-protein interactions, identify enzyme substrates and screen for autoimmunity biomarkers, among other applications. Service provider AVMBioMed offers both these arrays to its clients, Loch says, using them to scan for differences in post-translational modification. AVMBioMed clients use these data to predict pharmaceutical side effects, among other applications, he adds.
An alternative format is the antibody microarray, which researchers can use to track relative protein abundance.
RayBiotech, for instance, offers antibody arrays in multiple formats, including both membrane-based arrays that are treated like Western blots and higher-density, glass slide-based arrays that require dedicated microarray scanners. According to Bill Edens, associate vice president of operations, these arrays typically are developed using an antibody “sandwich” approach, with both capture and detection antibodies for each protein. But users also may biotinylate their protein probes directly, to avoid potential antibody cross talk. “This opens you up to larger and larger screens without worrying about antibody interactions,” he says.
Meso Scale Discovery’s MULTI-ARRAY technology provides essentially a cross between an antibody microarray and an ELISA. Here, each well of a 96-well microplate contains multiple capture antibodies, enabling researchers to quantify up to 10 proteins simultaneously via an electrochemiluminescent signal.
A hybrid approach
Joshua LaBaer, director of the Virginia G. Piper Center for Personalized Diagnostics at Arizona State University, has developed an alternative approach to building proteome arrays.
The nucleic acid-programmable protein array (NAPPA) is an array of up to 2,400 cDNAs on a planar surface, each coding for a different protein. Just prior to use, the researcher incubates the slide with an in vitro transcription/translation extract (available from New England Biolabs, Promega and Thermo Fisher Scientific, among others) to create localized pools of fresh protein, which are then immediately screened, as with any other proteome microarray.
According to LaBaer, this strategy offers several advantages: The proteins are fresh, can be created with mammalian post-translational modifications and are produced at relatively high and consistent levels. His team has successfully converted some 97% of tested cDNAs into protein, he says, including membrane proteins and proteins as large as 350 kDa, with 93% of proteins within 2% of the mean. “We get very consistent and reproducible yields.”
Another advantage of the NAPPA strategy, LaBaer says, is flexibility: As long as a cDNA is available, it can be arrayed—no protein purification required. “If someone comes to us with a new organism, if the genome is sequenced, we can make an array very quickly.”
Researchers can order NAPPA arrays through the Arizona State University Protein Array Core; they cost $175 apiece for external customers.
Suspension arrays
After completing an initial screen with protein microarrays, researchers commonly work to validate their findings. Enter suspension (or multiplexed bead) microarrays.
Based on Luminex xMAP technology, in which each capture antibody is bound to microspheres labeled with a discrete fluorescent signature, these arrays enable researchers to focus on a relatively small number of targets in a large pool of samples (for example, a 96- or 384-well plate).
In theory, says Daniel Braunschweig, global product manager for Bio-Plex® Assays at Bio-Rad Laboratories, xMAP allows researchers to multiplex up to 500 analytes per well. But optimal performance is achieved at close to 40 or 50, he says—a result of antibody cross-reactivity and potential background issues. Bio-Rad’s Human Chemokine Bio-Plex panel, capable of quantifying 40 analytes per reaction, is the company’s most multiplexed offering, though its new Human Inflammation Panel comes close at 37-plex. A new 48-plex human cytokine panel is expected to launch later this year.
MilliporeSigma offers a 41-plex human cytokine panel. Its newest offering is a 14-plex human myokine panel for use in muscle research.
Luminex arrays can be more sensitive than planar arrays, says Cindy Fry, product manager for MILLIPLEX products at MilliporeSigma, because they offer better mixing of sample and capture reagents than is typically possibly on a flat surface. They also are more quantitative, offering both high “n” values and standard curves, and they are highly customizable.
According to Braunschweig, researchers can easily select subsets of assay panels to suit their specific needs and reduce costs. Users can also mix and match beads from across panels, though additional optimization may be necessary. In that case, Bio-Rad recommends contacting technical support for guidance on appropriate testing protocols. Braunschweig advises researchers to run the different beads from different panels separately for best performance.
Luminex assays require dedicated readers, and both flow cytometry- and CCD-based instruments are available. (These instruments are all manufactured by Luminex, but they are sold and supported by Luminex partners.)
Bio-Rad’s Bio-Plex 200 and Bio-Plex 3D are flow-based instruments compatible with both polystyrene and magnetic Luminex beads. (The 3D offers up to 500-plex multiplexing, whereas the 200 is limited to 100-plex assays.) MilliporeSigma also offers these two Luminex models, as well as the Luminex MAGPIX, an inexpensive CCD-based system that is limited to 50-plex assays. According to Fry, the price of the MAGPIX is about $30,000, compared with $60,000 for the 200.
Peptoid microarrays
Thomas Kodadek, professor of chemistry and chemical biology at The Scripps Research Institute in Florida, uses microarrays to search for biomarkers of autoimmune disease.
“Our hypothesis was that in most any disease state, your immune system will produce antibodies that bind disease-specific molecules,” Kodadek explains. Those antibodies would represent “terrific biomarkers of disease, but we don’t know what the biomarkers might be, nor the antigens they recognize.”
Figuring that autoantibodies likely target “weirdly modified proteins which simply are not seen on proteome arrays,” Kodadek created what he calls “peptoid” microarrays to look for them.
“A peptoid is just an oligomer of glycine that has side chains coming off the nitrogen,” he explains. “And we purposely put all sorts of unnatural functionality off that side chain because we want to look at areas of chemical space outside normal biology.”
In 2011, Kodadek’s team demonstrated they could use an array of about 9,000 peptoids to capture autoantibodies associated with Alzheimer’s disease. But that strategy has several limitations, Kodadek says, including the amount of chemical space it is possible to probe at one time. “You have a very serious speed limit with arrays in terms of the number of compounds you can use.”
Recently his team has migrated to a suspension array approach, increasing the scans to cover a million compounds at a time. They also have altered the peptoid chemistry to make it sturdier and to reduce off-target binding.
Using this platform, Kodadek has identified autoantibodies associated with type 1 diabetes, and significantly, used the isolated antibodies to home in on the autoantigens themselves. Now, with funding from the Bill and Melinda Gates Foundation, he is applying the technology to search for signatures of latent tuberculosis.
“We finally know what we’re doing,” Kodadek says. “We’re turning this crank with increasing efficiency, and it looks like it’s going to be pretty cool.”
It seems the same could be said of protein microarrays in general.