12 May 2002

In the post-genome era, the next big challenge lies in mapping the proteome of every cell type, a daunting task indeed. One obstacle will be to identify trace amounts of proteins in cell or tissue samples. Time-of-flight mass spectroscopy, probably the biggest gun in the proteomics arsenal, is up to the task but can identify weakly expressed proteins only after they have been separated from the bulk of the sample. Because even partial purifications are time-consuming, and need to be tailored for every protein, this approach is not suitable for routine analysis. In this month's Nature Biotechnology, Ulf Landegren and colleagues at the Rudbeck Laboratory, Uppsala, Sweden, describe a method for the detection of zeptomole (yes, that's 10^-21 mole) amounts of protein in cell culture medium without the need for purification. Ironically, this protein-sleuthing method depends on the DNA-amplifying polymerase chain reaction.

Fredriksson et al. use special oligonucleotides, called aptamers, to identify target proteins by a process that they label "proximity-dependent DNA ligation." Aptamers are DNA sequences that bind strongly and specifically to unique protein structures, essentially behaving like antibodies. The researchers engineered aptamers to have tails complementary to a second oligonucleotide that acts as matchmaker and brings two aptamers together, whereupon they can be permanently ligated. In theory, two aptamers will only bind in the presence of a protein dimer, the matchmaker, and a ligase. Oligonucleotide ligation can then be detected using the polymerase chain reaction.

The authors used the platelet-derived growth factor B chain dimer (PDGF-BB) to test their assay. Proximity-dependent DNA ligation was three orders of magnitude more sensitive than ELISA, detecting as little as 10-20 moles of protein. The assay was specific, (for example PDGF-AA did not cross-react), and detected the target in complex biological fluids (for example fetal calf serum or cerebrospinal fluid) without compromising sensitivity. Probes designed for human a-thrombin gave similar results, but this time the aptamers were designed to bind to different sites of the same molecule, showing that the technique can be made to work for protein monomers, as well.

The only downside, as discussed by Michael Famulok, University of Bonn, Germany, in an accompanying News and Views, is the difficulty of making this generally applicable given that specific aptamers must be found for every protein one wants to assay.—Tom Fagan