Heeding the call from proteomics researchers to breed new antibodies for capturing proteins, a new human antibody initiative led by Stockholm-based professor Mathias Uhlén will collaborate with two other European antibody initiatives, and most likely also join HUPO’s umbrella initiative. Uhlén’s project seeks to design antibodies for every protein in the human proteome and create a public database describing the results.
Uhlén, a professor at Albanova University Center at the Royal Institute of Technology in Stockholm, launched the Human Proteome Resource on January 1 with an initial grant of $28 million over four years from the Swedish nonprofit Wallenberg Foundation. Uhlén wants to stretch this funding to analyze the proteins expressed by 10,000 genes, at a rate of 100 genes per week, although he admits that meeting this lofty goal will be a challenge. He hopes that Wallenberg or other collaborators will eventually dish out a total of $80 million, which he estimates is the minimum amount required to look at the whole proteome.
Uhlén describes his methods as “affinity proteomics,” defining this as “the systematic generation and use of protein-specific affinity reagents to functionally explore the proteome.” In other words, he’ll have to methodically express all the proteins encoded in the human genome, and then generate antibodies to profile all the proteins.
The approach is somewhat different from those taken by other well-known, antibody-focused proteomics scientists such as Marius Ueffing, leader of the German Society for Proteome Research’s monoclonal antibody initiative, and Dolores Cahill, who directs the Center for Human Proteomics at the Royal College of Surgeons in Dublin, Ireland, and works on antibody arrays. Instead of expressing the entire encoded protein, Uhlén uses homology search programs to find stretches of sequence most unique to a particular protein, and with the fewest transmembrane portions. He expresses in E. coli only these selected portions of the protein-producing gene fragments, dubbing them Protein Epitope Signature Tags. Uhlén then immunizes rabbits or chickens with the PrESTs and collects polyclonal antibodies — which are easier and cheaper to generate and better at binding non-native proteins such as PrESTs than monoclonal antibodies, he says. The antibodies will then be used for expression localization studies with tissue arrays.
The initial database will contain this localization information, but Uhlén hopes that — with the help and input of collaborators — the database will eventually also contain a wide range of information about every protein in healthy individuals, as well as in those with disease. So far, Uhlén has done “one pass of the genome,” with a 50 percent overall success rate for generating workable antibodies to a given PrEST. A second pass is on the way, though Uhlén admits that it is unlikely he would be able to generate antibodies to every encoded protein this way. “Some proteins are very hard to get — maybe five percent of the proteome we won’t be able to study this way,” he says. “But these are maybe not the most interesting proteins.”
Uhlén has repeatedly expressed his desire to collaborate, which he views as key to accomplishing such a broad task, and Ueffing and Cahill have answered his call: The three researchers have tentatively agreed to work together. Matthias Mann, professor of biochemistry and molecular biology at the University of Southern Denmark and head of the umbrella HUPO antibody initiative, sees Ueffing and Uhlén’s approaches as very complementary, and he would like to use HUPO as a vehicle for coordinating them. “This is not really a HUPO antibody initiative as such — it’s really to encourage the different initiatives to work together and to avoid duplication so it’s not everyone doing the same thing. In return, HUPO can assure funding bodies that this makes sense internationally — that it’s not just something that this particular researcher wants to do,” Mann says.
An extended version of this column appeared in ProteoMonitor’s September 19, 2003, edition.