Hype and Reality in an Important Transcription Paper

An important paper has just been published in Nature Structural & Molecular Biology. The work examines the in vivo kinetics and dynamics of transcription in mammalian cells and it confirms some important parameters derived from in vitro studies (Darzacq et al., 2007).

Before looking at the results, let's see what the press release has to say. The experiments were mostly done in Robert Singer's lab at the Albert Einstein College of Medicine in New York. The college issued a press release that was relayed on the ScienceDaily site [Scientists Discover The Dynamics Of Transcription In Living Mammalian Cells]. Here's the hype ...
THEME:
Transcription

Transcription — the transfer of DNA’s genetic information through the synthesis of complementary molecules of messenger RNA — forms the basis of all cellular activities. Yet little is known about the dynamics of the process — how efficient it is or how long it takes. Now, researchers at the Albert Einstein College of Medicine of Yeshiva University have measured the stages of transcription in real time. Their unexpected and surprising findings have fundamentally changed the way transcription is understood.

Actually, a great deal is known about the dynamics of transcription including how efficient it is and how long it takes. There's nothing in the paper that's particularly unexpected or surprising. The results do not fundamentally change the way transcription is understood.

In my opinion, this sort of exaggeration does a great deal of harm. Not only does it misrepresent the way science is actually done, it degrades the credibility of the scientists who carry out the work—especially, as in this case, when the lead scientists are actually quoted in the press release. This is unfortunate because this is a nice piece of work whose significance has been tainted by an overly-enthusiastic attempt to frame it in a way that appeals to a wider audience.

Let's try and ignore the silly press release and look at the paper itself. Darzacq et al. (2007) made a construct of artificial genes that were transformed into human tissue culture cells. The genes could be located inside the cell nucleus because they contained 256 binding sites for a protein with a red fluorescent tag (lac repressor). An example is shown in the above figure on the left (k). The binding of RNA polymerase to the genes could be monitored in living cells because the RNA polymerase molecules (Pol II) were tagged with yellow fluorescent protein (l). By combining the two images, the authors could be sure they were monitoring transcriptional activity at this particular site.

In addition to measuring recruitment of RNA polymerase to the promoters in real time, the paper also measured the synthesis of mRNA by detecting it with oligonucleotide probes or a protein (MS2) that bound specifically to the nascent mRNA. Various drugs and chemicals were used to block transcription initiation or elongation in order to sort out the various stages of transcription as described by the authors in the introduction.

The process of transcriptional initiation involves several structural changes in the polymerase as the nascent transcript elongates⁶. Early in initiation, the polymerase can produce abortive transcripts^7,8. These abortive cycles have been observed with a single prokaryote polymerase (RNAP) releasing several transcripts without escaping the promoter^9,10. The elongation step can be regulated by pausing for various times, as demonstrated using prokaryotic polymerases in vitro^11,12. For eukaryotic cells, attempts have been made to calculate the endogenous elongation speed using run-on assays¹³, reverse-transcription (RT)-PCR14 or fluorescence in situ hybridization (FISH)15 on specific mRNAs, and these have yielded apparent elongation estimates ranging from 1.1 to 2.5 kilobases (kb) min-1. To date, no assay has been developed to measure the various steps of Pol II transcription in a living cell.

Note that the authors are perfectly aware of the fundamental properties of transcription as deduced from in vitro studies. They hope to confirm these results in vivo.

In a typical experiment, the RNA polymerase molecules bound to the genes are made to disappear by photobleaching. Then the reappearance of fluorescent labeled RNA polymerase is monitored by measuring the increase in brightness at the spot where the genes are known to be located. The rate of increase over 10 minutes or so is plotted and the data fitted to a curve with several parameters. The various constants correspond to the amount of time that RNA polymerase sits at the promoter, the number of times it falls off before initiating transcription, the rate of transcription, the length of pauses etc. Some of the parameters can be estimated in other experiments with various inhibitors.

The results are interesting even though they do not alter our fundamental understanding of transcription.

1. Transcription initiation is inefficient. The calculations suggest that only 1 in 90 bound RNA polymerase molecules actually proceeds to the elongation stage. It has been known for some time that initiation is the rate-limiting step in transcription. In vitro studies showed that some RNA polymerase molecules could not remain bound to the promoter long enough to from an open complex for initiation. Other studies showed a high level of abortive initiation where small bits of RNA were repeatedly synthesized but the transcription complex could not get started. It is well known that the initiation complex contains different proteins than the elongation complex and the exchange of proteins is required in order to progress to the elongation stage. This is some of the best data illustrating that these limitations apply in vivo as well.
2. About 10% of the transcription complexes on a gene are paused at any one time. Pauses can last up to 4 minutes. During a pause, the complex does not move along the gene and no RNA is synthesized. The nature of transcriptional pausing in bacteria is understood. It depends on certain sequences on the gene and, to some extent, secondary structure in the newly transcribed RNA. There are special proteins bound to the transcription complex that inhibit or reduce pausing. Some pause sites are required for proper gene regulation.
3. The rate of transcription in E. coli is known to vary between 30 and 85 nucleotides per second. In vivo measurements in eukaryotic cells suggested that the rate of transcription of chromatin was close to the lower limit of the rate in bacteria (i.e., 30 nucleotides per second). This study suggests that the rate of transcription of actively transcribing complexes is about 70 nucleotides per second, or much faster than previous crude estimates.
4. It takes a long time to transcribe a gene. It is well known that transcription of eukaryotic genes can be a time consuming process. By combining the rate of elongation and the pause times, the results indicate that a typical mammalian gene of 14,000 base pairs will be transcribed in 20 minutes.
5. Only a few transcription complexes can be active on a gene at any one time. This conclusion follows from the rate of initiation, the rate of elongation, and the measured production of mRNA. It makes sense since the long pauses during the elongation reaction would cause all following complexes to back up. Consequently, there's no point in loading up the gene with inactive transcription complexes. This result with class II genes (protein encoding) is remarkably different than the result with class I and class III genes that encode ribosomal RNA and tRNA. In those cases you can visualize the the transcription complexes on the genes inside the cell and they are lined up head to tail as they move down the gene. Similar EM pictures of other genes support the result found in this study.