Naturally, this new research is controversial-paradigm shifts always are. But, says Brandeis University's Michael Rosbash, whose postdoctoral fellow is working on a similar problem, "Cook has gone a long way to making believers out of a lot of skeptics, including myself." Says Leslie Leinwand, chair, department of molecular, cellular, and developmental biology, University of Colorado, Boulder: "This goes against everybody's dogma so much that everybody, including myself, is saying, 'Whoa, this is just not the way you think about a cell.'" Cook was shocked himself. "[We] didn't believe the results [at first], and repeated them over and over and over again."
What Cook's work has done is invite new interpretations of long-standing dilemmas, most notably the puzzle of nonsense-mediated RNA decay. Moreover, if it is correct-and most researchers believe additional work is needed to definitively prove that nuclear translation occurs-it will likely change the way new experiments are interpreted. These results might even force scientists to rethink the role of the eukaryotic nucleus.
The work also has implications beyond pure science. Indeed, says Leinwand, Cook's data could directly affect the way some gene therapies are developed. "You would design vectors quite differently if you were targeting the nucleus versus the cytoplasm."
Time To Revise Textbooks?
|Courtesy of University of Colorado, Boulder|
Cook and his colleagues permeabilized cervical carcinoma cells, introduced labeled precursors, and allowed translation to proceed under suboptimal conditions. When they examined the cells, as expected, they found the majority of label in the cytoplasm. However, a significant amount of label was also found in the nucleus. But, says Cook, "we all know that translation only occurs in the cytoplasm." Even after doing all of the appropriate controls, they still thought they were dealing with some sort of contamination. The clincher, says Cook, was that they could inhibit the nuclear signal by inhibiting transcription. That suggested that the two processes are coupled in eukaryotes, just as they are in bacteria. The authors' propose that as much as 15 percent of cellular protein synthesis may occur in the nucleus.
Since then, Cook and his team have demonstrated the nuclear synthesis of a specific protein. In a series of unpublished experiments that Cook described as "torturous," the researchers detected nuclear CD2-a cell- surface receptor and the "most non-nuclear protein" they could imagine. Again, the protein's presence was coupled to transcription. The nucleus evidently has no endoplasmic reticulum (ER) surrogate, because the protein misfolded and was degraded. Based on this "sample of one," Cook speculates that many proteins are made in the nucleus. Furthermore, he notes that these results are not limited to cervical carcinoma cells; they've been confirmed in murine and simian cells.
This observation upends nearly 30 years of biology textbook wisdom, which states that eukaryotes segregate transcription and translation because their transcripts require post-transcriptional processing. Eukaryotic transcripts contain noncoding introns that are interspersed between the gene's coding exons. The introns must be removed prior to translation to ensure that the correct protein is synthesized. Teleologically speaking, this is why eukaryotic cells contain a nucleus in the first place. To counter this argument, Cook suggests that the nucleus is much like an assembly line, in which transcripts are translated only after they have already been spliced.
Saverio Brogna, a postdoctoral fellow in Rosbash's lab, has collected data that are consistent with Cook's hypothesis. According to Rosbash, Brogna has observed ribosomes associated with chromosomes in Drosophila salivary glands. However, no independent evidence exists that suggests these ribosomes are functional. "In fact," adds Rosbash, "the data we have are more compatible with their not being active." He tempers that statement by noting that these are negative data and that he cannot draw any conclusions yet.
As a practical matter, these results are not likely to change the way most people conduct their experiments, says Matthias Hentze, a senior scientist at the European Molecular Biology Laboratory in Heidelberg, Germany, and author of a perspective piece accompanying Cook's article.2 However, they might affect how researchers interpret their results, he adds. The data also provide a framework on which previously confusing results can be reevaluated. Foremost among these is the nonsense-mediated decay (NMD) hypothesis.
The cytoplasmic camp describes NMD as a ribosome-mediated process that somehow senses a premature-stop codon and appropriately destroys the offending transcript. But those in the nuclear camp counter that, at least in some cases, mRNAs degrade in the nucleus-or, more accurately, in purified nuclear preparations-and not in the cytoplasm. However, this position presents a conundrum, because translation, the most obvious mRNA-scanning method, occurs in the cytoplasm. "There's a logical impossibility here," says Cook, "so people in the field go through incredible contortions to ... come up with possible explanations of how this could possibly happen." Nuclear translation would go a long way toward explaining nuclear NMD, says Wilkinson.
Where NMD occurs in the cell has importance beyond pure scientific knowledge. It also has clinical implications, says Colorado's Leinwand. Her lab uses molecular genetic approaches to study cardiac and skeletal muscle diseases. For example, her group uses RNA-mediated approaches as "proofs of principle" for gene therapy. Specifically, Leinwand has been developing suppressor tRNA-based approaches to combat diseases caused by nonsense mutations.
A suppressor tRNA places an amino acid in the peptide chain being synthesized when a stop codon is encountered, instead of terminating peptide synthesis. In theory, if such a vector were introduced into a patient suffering from a nonsense codon-mediated disease, the disease could be functionally corrected at the protein level. Prior to Cook's research, Leinwand and her colleagues developed gene therapy vectors to produce the suppressor tRNAs in the cytoplasm. However, she says, "The assumption, I think, before Cook's paper, was that the cytoplasm was the way to go, and I'm no longer sure that that's true."
Cook's publication is not the first to float the idea of nuclear translation. He notes that a number of earlier publications pointed to this possibility, but scientists dismissed them based on the belief that cytoplasmic components were contaminating the nuclear preparations. The problem is akin to separating an egg yolk from an egg white: It's easy to collect pure egg whites, but it's nearly impossible to obtain the yolk sans any whites. Although the authors of these papers performed controls to account for this possibility, the scientific community largely ignored them. According to Cook, this might be because early interpretations of these data suggested that nuclear translation synthesized nuclear proteins, while cytoplasmic translation synthesized cytoplasmic proteins, a hypothesis that was clearly wrong. So nuclear translation fell by the wayside, and the notion that eukaryotic translation occurs only in the cytoplasm stuck.
Now Cook's publication has resurrected the argument, offering the most compelling results to date, says Hentze. "The significance of the Cook work is that it takes the rigor of analysis for nuclear translation to a new and much higher level," he says. However, he adds, "I don't think this paper provides rigorous proof beyond any level of doubt." Wilkinson and Leinwand concur, noting that Cook's experiments were conducted using permeabilized cells so that protein precursors could be introduced into the system, but permeabilization essentially kills the cell. And, "because you've broken the cell membrane, you never know if inappropriate, 'non-normal' things will happen," Wilkinson says.
Cook's research leaves many unanswered questions. For instance, what happens to those proteins synthesized in the nucleus? Are they functional or degraded? If they are functional, do they remain in the nucleus or are they exported to the cytoplasm? How many proteins are made off any single transcript, and how many are needed to determine whether the transcript is properly spliced? Are there biochemical differences between nuclear and cytoplasmic translation? Finally, Cook's work yields one very interesting question: Why do eukaryotes have a nucleus at all?
It was a well-accepted fact that only eukaryotes have nuclear membranes. In fact, the first sentence of the Cook manuscript calls the nuclear membrane "the defining feature of eukaryotes." Now, however, Cook says that even this distinction may be incorrect. In fact, he notes that he received a message following the publication of his article, noting that bacteria appear to contain membranes similar to nuclear membranes.
Maquat says the nucleus was always thought to compartmentalize different cellular processes. But, she adds, "the more we're learning about mammalian cell metabolism, the more we realize that there isn't as much compartmentalization." For example, many processes previously assumed to occur only in the cytoplasm, such as tRNA charging, are now known to also occur in the nucleus. Thus, "the differences between eukaryotes and prokaryotes are blurred," Cook says.
As he contemplated the years spent on this project, and the time it took to publish this work, Cook couldn't really explain why 30 years of evidence pointing at nuclear translation was essentially ignored. "Paradigms are hard to shift," he says. "It's often hard to see why they arise in the first place."
1.F.J. Iborra et al., "Coupled transcription and translation within nuclei of mammalian cells," Science, 293:1139-42, Aug. 10, 2001.
2. M.W. Hentze, "Believe it or not-translation in the nucleus," Science, 293:1058-9, Aug.10, 2001.