Lost in translation : the signal hypothesis

It was cell biology's version of the ship in the bottle. How do proteins a cell intends to secrete end up in the endoplasmic reticulum? Winkling out the details of the translocation mechanism that spirits these proteins into the ER required more than 20 years and earned Günter Blobel of Rockefeller University the 1999 Nobel Prize in Physiology or Medicine.

Another Rockefeller laureate, George Palade, had demonstrated that ribosomes free in the cytoplasm manufactured nonsecreted proteins, whereas ribosomes stuck to the ER made proteins for export. Cell biologists searched in vain for distinctions between free and attached ribosomes that might explain their contrasting behavior. A new assistant professor at Rockefeller and Palade's protege, Blobel suspected that the difference must lie in the proteins themselves. He and colleague David Sabatini conjectured that secretory proteins might carry a short segment near the NH₂ terminus (Blobel and Sabatini, 1971). Once this sequence protruded from the ribosome during translation, a “binding factor” would hook onto the protein and guide it and the ribosome to the ER membrane. Continued translation would then thread the elongating protein into the ER's interior. “It was a beautiful idea,” says Blobel. It was also, he admits, “pure speculation.”

But it didn't take long for evidence of a “signal sequence” to start accruing. The cell-free translation

system concocted by Philip Leder and colleagues (Swan et al., 1972) churned out an antibody light chain that was 6 to 8 amino acids longer than the normal secreted version in the body. Tonegawa and Baldi (1973) and Schechter (1973) obtained similar results.

Unaware of Blobel and Sabatini's hypothesis, Cesar Milstein of Cambridge University proposed a similar idea based on his team's cell-free system. It also pumped out an overweight light chain, but when the researchers checked the output of microsomes (ER fragments), they found only the normal-sized protein (Milstein et al., 1972). Milstein speculated that the extra amino acids help direct the growing protein to the ER.

Despite this suggestive data, detractors argued that the protein's extra heft was an artifact of in vitro translation or isolation errors, Blobel recalls. To answer their complaints, he crafted a protein-synthesizing system with help from post-doc Bernhard Dobberstein (now at the University of Heidelberg). Using detergent, they dislodged ribosomes from rough microsomes, and then slipped the particles—which carried unfinished light chains—into a solution that allowed protein making to resume. Because the researchers also added a compound that blocks new translation, the ribosomes could only complete chains they had started.

At first, only the smaller, processed chain appeared (Blobel and Dobberstein, 1975a). These proteins came from ribosomes that were well into translation when they parted from microsomes, the researchers concluded, and the chains they held had already undergone pruning to remove the signal sequence. After a few minutes, however, the synthesis mixture started producing longer chains as well. The bulkier proteins emerged from ribosomes that had just started translating when isolated from microsomes. At the time, they bore stubby chains that hadn't yet shed their signal sequence. When translation restarted, these short chains didn't lose the sequence—evidence that the processing enzyme that removes the signal is part of the ER membrane.

In another key experiment, Blobel and Dobberstein let rough microsomes—which carry ribosomes and some associated mRNA—produce proteins. The scientists detected only the shorter version. Adding the protein-dissolving enzymes trypsin and chymotrypsin (which rarely enter the microsomes) did not digest most of the chains, confirming that the trimmed protein ends up tucked away within the microsomes, as the signal hypothesis predicted.

The next goal, Blobel recalls, was to build the “translation-translocation” mechanism from scratch, using isolated mRNA, small and large ribosome units, and microsomes. But the work stalled. No matter what animal the microsomes came from, they always stifled translation in the cell-free system. After numerous setbacks, Blobel was “prepared for failure” when he tried microsomes from dog pancreas. Instead, in December of 1974, the procedure finally worked.

The pair quickly showed (Blobel and Dobberstein, 1975b) that this combination produced mostly the short form of the light chain. If primed with the right mRNA, the system would also make globin, a nonsecreted protein. Unlike the processed light chain, globin fell victim to the protein-dissolving enzymes, indicating that it didn't slip into the microsomes. Moreover, if complete, oversized light chains were added after the microsomes, they didn't lose the signal sequence, verifying that the removal of the segment occurs during translation, not afterwards. That their Rube Goldberg concoction of mouse RNA, rabbit ribosomes, and dog ER actually synthesized proteins demonstrated something else, Blobel says. “[It] had the virtue of showing that this is a universal system.”