Finding the tail end: The discovery of RNA splicing

Major findings are sometimes hidden in small details. At least that was the case when, in a PNAS article, molecular biologist Phillip Sharp and his research team described a little strand of RNA that led to an understanding of how proteins are synthesized in cells (1).

Fig. 1.

Members of the MIT Center for Cancer Research (Robert Weinberg, Second Row from Bottom, Far Left; Susan Berget, Third Row from Bottom, Third from Left; Claire Moore, Back Row, Fourth from Left; Philip Sharp, Back Row, Far Right). Image courtesy of Robert Weinberg.

Fig. 2.

(A–C) Electron micrograph of a hexon mRNA hybridized to an adenovirus genome fragment. The arrows in the micrographs mark the single-stranded RNA tails at the ends of the RNA/DNA hybrid. The 5’ to 3’ orientation of the mRNA is indicated in C, and the displaced single-stranded viral DNA is shown as a thin line. Reproduced with permission from ref. 1.

Fig. 3.

(A) Phillip Sharp receives the 1993 Nobel Prize in Physiology or Medicine from King Carl XVI Gustaf of Sweden. Image courtesy of Tobbe Gustavsson/Reportagebild/TT/Sipa USA. (B) Richard Roberts receives the 1993 Nobel Prize in Physiology or Medicine from King Carl XVI Gustaf of Sweden. Image courtesy of AGIP/Rue des Archives/Granger, NYC.

During the 1970s, Sharp headed a laboratory in the Center for Cancer Research at the Massachusetts Institute of Technology (MIT). Having been a postdoctoral researcher for geneticist James Watson and then a staff member at Cold Spring Harbor Laboratory (CSHL) prior to his appointment at MIT, Sharp was drawn to studying genes and measuring chromosome sizes. The field of genetics was relatively primitive; Watson and Francis Crick had discovered DNA’s structure only 2 decades earlier.

Molecular biologists at the time worked almost exclusively with bacterial systems, which are easy to grow in a laboratory. Although Sharp primarily studied bacteria and published work about the Escherichia coli genome during his postdoctoral position at the California Institute of Technology with biochemist Norman Davidson (2, 3), he started exploring tumor biology and virology. When Sharp arrived at CSHL, he turned his attention toward DNA viruses known to infect animal cells. He was particularly curious about gene expression—the conversion of DNA into instructions for creating proteins—in human cells and began studying the transcriptional profile of a simian DNA virus called SV40. Found in both humans and monkeys, SV40 can generate tumors. Sharp’s work with the virus stemmed from a collaboration with virologist Joseph Sambrook. Meanwhile, an officemate, Ulf Pettersson, worked on DNA replication of adenovirus, a common virus with a double-stranded DNA genome that is known to cause tumors in rodents and a range of illnesses in humans.

“Ulf and I became friends,” Sharp says. “He was interested in adenovirus DNA replication; I was interested in adenovirus transcriptional activities and the location of genes.”

A few years earlier, microbiologists Daniel Nathans and Hamilton Smith discovered enzymes called restriction endonucleases (4–6), which earned them a shared Nobel Prize with Werner Arber. Following their insights, Sharp developed a method to purify restriction enzymes using gel electrophoresis and ethidium bromide staining. Eventually, Sharp, Pettersson, and others made restriction maps of the adenovirus 2 and 5 genomes as well as several other serotypes. These maps were used to identify the viral regions that contained cancer-causing genes.

“The transcriptional pattern was important for understanding how the virus created tumors,” Sharp says. “It also set the table for a long-term curiosity: The significance of heterogeneous nuclear RNA.”

When Sharp was recruited by MIT’s cancer center, he and postdoctoral fellow Jane Flint extended the transcriptional maps of adenovirus. One of Sharp’s colleagues, molecular biologist Robert Weinberg, ran an adjacent laboratory on the fifth floor and studied SV40. Both adenovirus and SV40 can produce multiple copies of their genomes in infected cells, enabling analysis.

“The reason that Phil Sharp, and I, and others worked on DNA tumor viruses, like adenovirus and SV40, was actually not because we were interested in cancer per se, but because these viruses clone their own genomes for us,” says Weinberg, who is currently a member of the Whitehead Institute for Biomedical Research and professor of biology at MIT.

In fact, Weinberg adds, not all groups at the center focused directly on cancer as a disease. “The MIT cancer center was built on the notion that curiosity-driven research is likely to yield great benefits.”

And it did. During the 1970s and 1980s, MIT’s Center for Cancer Research was a research powerhouse with notable scientists, including David Baltimore and Susumu Tonegawa, who won Nobel Prizes for discovering reverse transcriptase and the rearrangement of antibody genes, respectively.

Around 1975, Weinberg organized a weekly meeting for the center’s fifth floor laboratories to discuss their research. Little did Sharp know that those meetings would ultimately play an integral role in a landmark discovery.

In 1976, postdoctoral fellow Susan Berget began using electron microscopy to examine the relationship between cytoplasmic RNA and the DNA structure of adenovirus. Both Sharp and Berget worked closely with Claire Moore, a technician who ran MIT’s electron microscopy facility for the cancer center. Through her work with Sharp, Moore became familiar with a technique called R-loop analysis, which at the time was a new method to map RNA sequences on a genome. By creating optimal conditions using a combination of salt, formamide, and heat, Moore could make an RNA strand hybridize with its complementary DNA.

“You would see a bit of string, and there would be a bubble where the RNA had hybridized with the DNA and displaced the other strand,” says Moore, now a professor of developmental, molecular, and chemical biology at Tufts University. “Phil thought that would be a great way to map the adenovirus because you could precisely map the location of each gene.”

Adenovirus replicates efficiently, so when Berget infected a human cell line, the majority of the messenger RNA (mRNA) molecules would be of viral origin; as its name suggests, mRNA serves as an intermediary template, transmitting information from DNA to proteins. Berget purified the most abundant viral mRNAs, which encode a capsid protein, and in collaboration with Moore, subjected the mRNAs to hybridization conditions in order to form R-loops with specific restriction endonuclease fragments of the adenovirus genome. The resulting R-loops were visualized using electron microscopy, and the length of the R-loops and double-stranded DNA were measured to determine the gene’s location.

“My task was to prepare the samples,” Moore says. “There were a lot of tricks to getting a good spread.”

With a viscous solution of formamide and a ramp constructed from a slide that ran into another solution, Moore had to drop a sample on the slide at just the right place so it would spread onto a film. Once Berget and Moore collected the film on a tiny grid coated with plastic, they processed it for electron microscopy. This film was subsequently inserted into the electron microscope and examined for interpretable R-loop structures that were then photographed. To get accurate measurements, the micrographs were photographically printed, and a “little map measurer,” as Moore puts it, was traced along each strand. Throughout this process for dozens of micrographs, something seemed off.

“I was getting nice, uniform R-loops, but there were these little strands of single-stranded RNA sticking out at the ends of the R-loops, and I didn't know what to do with them,” Moore says.

Other laboratories had previously found that adenovirus RNA in the nucleus was far longer than cytoplasmic mRNA. The team wondered if these long, viral RNAs were related to observations of cellular heterogeneous nuclear RNAs. Similar to adenovirus RNAs, these nuclear RNAs containing gene sequences were considerably longer than more stable mRNAs in the cell cytoplasm.

Berget, Moore, and Sharp concluded that one of the extensions must be a polyadenosine tail that is added posttranscriptionally to the end of mRNA and has nothing with which to hybridize. Assuming that the other tail was an artifact, possibly caused by DNA rehybridizing and displacing RNA, the team conducted a myriad of experiments. They eliminated the opposite DNA strand so nothing could compete with the end of the RNA. However, when the RNA hybridized to a single-stranded bit of DNA, the tail was still present. Even trying different conditions of formamide and salt proved unsuccessful. Nevertheless, Sharp and his colleagues were determined to rule out other explanations for the tail before proceeding.

“Our first indication that there was something we didn't understand was when we were doing the microscopy of mapping the adenovirus mRNA on the genome, and we found this tail at the 5′ end of the RNA,” Sharp says. “It didn’t behave as if it were part of the sequences immediately adjacent to the genome, and we spent 3 months mostly trying to convince ourselves that it wasn’t an artifact.”

But 3 months and several experiments later, the extra tail was still a mystery.

“It was a puzzle until Sue presented the data at one of the floor meetings, and we got the idea that maybe it’s coming from a different region of the adenovirus,” Moore says.

So the team tried hybridizing the RNA to a longer piece of DNA.

“That’s when the little tail found its partner strands and pulled the DNA into the loops,” Moore says.

Although Sharp was unsure what to anticipate, he recalls: “When I saw those 3 loops, I knew what it was.”

It was the answer to the heterogeneous nuclear RNA mystery. Berget, Moore, and Sharp had discovered RNA splicing.

This stage in gene transcription, or the process by which DNA information is transcribed into mRNA, is similar to a message that needs to be decoded because of a sentence with extra letters. Once those unnecessary letters that make the sentence seem like gibberish are removed, a clear message becomes apparent. In the case of mRNA, the message conveys directions to guide synthesis of proteins.

The human genome has an estimated tens of thousands of genes that provide instructions for producing an even larger number of proteins, which are created via RNA splicing. When a gene is copied into RNA, it contains a long, jumbled message of nucleotide sequences called exons and introns. While exons compose a message with specific instructions from a particular gene, they are separated by introns that are interspersed in the RNA, rendering the message incomprehensible. Introns must be spliced out so exons can form a coherent message called a mature mRNA. This is the genetic information used to synthesize proteins in the cytoplasm. Thus, the concept of RNA splicing explains why nuclear mRNA is longer than cytoplasmic RNA.

The evidence for RNA splicing, which was published in PNAS in 1977, was groundbreaking (1).

“There were people who were tantalizingly close to discovering splicing, but being close is not good enough,” Weinberg says. “You either discover it, or you don't. And Phil did.”

Before knowledge of RNA splicing, the consensus was that all organisms have the same gene structure as bacteria, which lack introns. French biochemist Jacques Monod, who received the 1965 Nobel Prize in Physiology or Medicine, famously asserted that “anything found to be true of E. coli must also be true of elephants” (7). However, RNA splicing demonstrated that eukaryotic cells, with their discontinuous genes, are far more complex than bacterial cells. It also debunked the dogma that one gene produces one mRNA, and all mRNAs from a gene produce one protein. Only a few genes are needed to create various proteins via a process called alternative splicing. Similar to how rearranging a group of letters can form different words, the exons of a single gene can reorder to generate a multitude of proteins.

Sharp compares discovering RNA splicing to finding the Rosetta Stone. His revelation earned him the 1993 Nobel Prize in Physiology or Medicine, which he shared with Richard Roberts (8), a molecular biologist who independently made the same discovery and published the results in a 1977 issue of Cell (9).

While at CSHL, Roberts and his colleague Richard Gelinas focused their research on promoters, the areas of DNA where gene transcription into mRNA begins. They specifically studied adenovirus to determine whether bacterial and eukaryotic promoters share the same sequence characteristics.

“Our experiments gave us results that didn’t make sense if eukaryotic genes were the same as bacterial genes,” says Roberts, now the chief scientific officer at New England Biolabs. “I came up with the idea for an electron microscopy experiment that would shed light on all of the biochemistry we’d been doing and devised the experiment on a Saturday morning; the experiment was carried out by Louise Chow on a Tuesday morning, and by Tuesday afternoon, splicing was universally accepted as being discovered.”

But Sharp believes it was just a matter of time before RNA splicing would become known to the world.

“Somebody would have made the discovery,” he says. “We were fortunate to have made it first.”

Sharp, who is still a professor at MIT, remains in awe of what he accomplished with Berget and Moore more than 40 years ago. “To have the privilege of being part of discoveries that advance our understanding and our ability to help people is such an extraordinary, extraordinary experience.”