Marcus Peter, of the German Cancer Research Center, Heidelberg, and Vishva Dixit, University of Michigan, Ann Arbor, found and sequenced a protein they call FLICE, which physically links the so-called death receptors on the cell surface to molecules such as ICE that have already been shown to be very early steps in apoptosis. The cell surface receptors-CD95 and TNFR-1-share a sequence called the death domain, which binds to a molecule called FADD. This molecule has a region, the death effector domain (DED), that is essential to set apoptosis in train, and it was believed to be an adaptor molecule that linked the receptor to effector enzymes, such as ICE, lower down the chain. However, the nature of the link between receptors and effectors was unknown.

   In their paper, Peter and Dixit and their colleagues show that FLICE is the missing link. It shares homologies with both FADD and ICE, and is itself a member of the ICE family of proteases. FLICE, thus, is the most upstream enzyme in the apoptosis pathway. Ordinarily, by this model, two molecules of FLICE bind together via their DED sites, which protects them from activation. An activated CD95 receptor binds to FADD via the death domain, changing the shape of FADD's DED and allowing it to disrupt the FLICE DEDs. CD95, FADD, and FLICE form a trimeric signaling complex that allows FLICE to autocatalyze and release the activated ICE-like section to set off the proteolytic cascade of ICE that eventually results in cell death.

   Neat. And interesting. And important. But what truly boggles is that this fine edifice of sequences, homologies, and family relationships was constructed on less than 0.5 pmol of protein. That's so little as to be almost unimaginable. The technique that made it possible is called nano-electrospray tandem mass spectrometry (nano-ES MS/MS for short) and was developed at the European Molecular Biology Laboratory (EMBL), in Heidelberg.

   Nano-ES MS/MS starts with a protein spot on a gel. That is cut out and digested into fragments with trypsin. EMBL has developed considerable expertise in concentrating the protein fragments into a volume less than 1 microliter. The liquid goes into a needle rather like those used for injecting stuff into cells, but coated with metal charged to 600 volts. Droplets sprayed from the needle become equally charged. The crucial element is that they're small-only 200 nanometers across-so the liquid evaporates very quickly, leaving behind a charged peptide fragment. "In fact, you can't even see them," says Matthias Mann, leader of the Protein and Peptide Group at EMBL, and a key developer of nano-ES MS/MS.

   The peptide ions are sucked into the opening of a mass spectrometer, which creates a conventional spectrum, and then each fragment is allowed to pass through a rest gas into the second spectrometer. "On the way through the gas they get bashed into and go to pieces," Mann explains. That creates a second spectrum of fragments that are one amino acid smaller than the originals.

   "One of the difficulties," Mann concedes, "is that you need to interpret the spectra to get a reliable sequence," but the team has developed software to assist and is confident that it works. From the protein sequence it's a hop, skip, and jump to search databases, find DNA clones, and get on with confirming details and interpreting function. The nano-ES MS/MS equipment developed at EMBL is now in a number of other labs, "but we have the highest sensitivity," according to Mann. That is borne out by the nine papers in Cell, Science, and Nature this year that they have collaborated on. "They wouldn't come to us if they could do this," Mann says.