AT A GLANCE
Research Scientist, Columbia Genome Center, head of the functional genomics division.
PhD in immunology and microbiology from Columbia University, 1992. Worked in the lab of Elvin Kabat, the scientist who identified gamma globulins as the proteins responsible for antibody activity and discovered the hypervariable regions that form antibody binding sites.
Postdoc at Columbia and Stanford, 1993 to 1997, studying HIV, genetics, and molecular biology.
Interests include the incorporation of immunological expertise in the development of tools for functional analysis of genes and the protein and post-translational levels, as well as the development of high-throughput nucleic acid, protein, and carbohydrate biochips; and to focus on specific models.
How did you get into making microarrays?
My background is in immunology. And after my postdoc at Columbia and Stanford in molecular biology and HIV research, I was recruited to work at the Columbia Genome Center to establish the functional genomics division of the center. Very soon after Pat Brown established the cDNA microarray, we first experimented with cDNA arrays. But when I began working on DNA microarrays, I immediately got the idea to incorporate my previous training of the antibody-antigen assay. And I was one of the earlier people working on antigen microarrays.
You published a paper in the March 2002 issue of Nature Biotechnology on carbohydrate microarrays for the recognition of pathogens, in which you spotted down the microbial polysaccharides on modified glass slides, then used these slides to probe for the presence of antibodies for these microbial antigens in the blood. Can you tell me how you came to develop these arrays?
In my hands, there is a collection of purified carbohydrate antigens, including polysaccharide antigens from some pathogens and blood group substances. This collection was begun a long time ago by Karl Landsteiner, the Austrian scientist who received the Nobel prize [in 1930] for discovering blood groups and by Michael Heidelberger [at Columbia], who discovered in 1917 for the first time that an antigenic substance of Pneumococcus was polysaccharide and not protein as thought previously. This work was continued by Elvin Kabat, who was Michael’s first PhD student. Heidelberger and Kabat developed a quantitative method called quantitative immunoprecipitation. It was the first study method that quantitatively studied immunology. I was Kabat’s last graduate student. When Kabat retired in 1994, he wrote me a very nice letter to authorize me to take over this collection. He wants the collection to benefit the next generation. In the Nature Biotechnology paper, we used about 50 of these purified polysaccharides and blood group substances to demonstrate the carbohydrate microarray technology. We covered all different categories of carbohydrate antigens.
So how does this carbohydrate microarray system work technically? How can you spot sugars onto a glass slide and get them to stick?
There is a nitrocellulose layer over the glass. The section of the antigen can stick there very stably. We tested a bunch of commercially available slides, including one now made by Schleicher and Schuell. We found this one gave the best results with carbohydrates. Our results were so good — unexpectedly. Nobody is thinking in this direction. People use the membrane for protein, for DNA, for RNA. The theory underlying this is that you need a hydrophobic surface of a biological molecule to be exposed to interact with the nitrocellulose membrane. And nitrocellulose is highly hydrophobic. And for that reason nobody thinks because carbohydrates are so hydrophilic that you can use this mechanism to immobilize them on the slide. But many of the carbohydrate-containing macromolecules can be used on this platform. Many people are trying it a different way, but this way is simple and reproducible.
How did you know if a polysaccharide from the collection actually worked on the array?
We tested them using the mixed antibody test, where you have a collection of antibodies to the antigen in question. For example, to see whether polysaccharide (1,6)dextran is immobilized and its sugar antigenic determinant is exposed, we used antigen-antibody tests, where we printed dextran preparations on the array and then incubated them with monoclonal antibodies specific for (1,6)dextran.
For any new targets, we must test them. It is an advantage to test them using the high throughput microarray platform.
Now in the paper, you spotted down 48 different polysaccharides, and said you could get up to 20,000 spots per slide. Are you looking to expand the number of different macromolecules on the slide?
Now we have slides with about 10,000 spots and 1,000 different antigens on a slide. We also have found that [our system] is good for printing protein antigens.
Are you looking for people to commercialize this sugar arraying process?
Before I published [the Nature Biotechnology article], Columbia requested for me to send out a patent application. It’s a Columbia patent. For academics, everyone can use this. But right now, several companies are talking with Columbia about licensing this technology.