by PhilipJ on 17 March 2006
And by the same group, no less!
β-gal is a commonly used molecule to detect gene expression, as a single β-gal molecule can produce a large number of fluorescent particles called fluorescein by hydrolysing a synthetic fluorogenic substrate called FDG, as so:
Cells, however, are clever. As these aren’t molecules naturally found in vivo, efflux pumps on the cell membrane actively work to reduce their concentration. In traditional experimental environments this leads to a rapid loss in concentrated fluorescence, making it quite difficult to accurately determine how many β-gal molecules are present.
This paper presents a microfluidics-based method to get around this limitation, by confining individual (or small numbers of) cells in a small volume. With a small confined volume (roughly 100×100×10 μm cubed), the high efflux rate, and efficient mixing present in the microfluidic chamber is sufficient to accurately determine enzymatic activity inside the cell.
An example of their data looks like this, where discrete jumps in β-gal are observed in living E. coli cells. The black curve is signal from dividing cells confined in a chamber, while the red curve is background signal from an empty chamber.
The bursts in expression are attributed to transient association/dissociation of the Lac repressor with its promotor, allowing the machinery of the central dogma to create mRNA and then β-gal molecules while the repressor has dissociated. The paper then goes on to characterise the bursts in terms of burst frequency and burst size, and ultimately proposing two regimes for stochastic gene expression.
Not to rest on their laurels, in this week’s issue of Science, the Xie lab has another paper (subscription required here as well) with a different method for measuring gene expression at the single molecule level in living E. coli cells using Yellow Fluorescent Protein (YFP), a derivative of GFP.
Trying to look at YFP or GFP itself is usually difficult — fast diffusion times lead to a spreading of the signal during an integration time of the camera used for detection. One way to get around this issue is to fuse YFP to a protein with much slower diffusive motion. The protein chosen for fusion with YFP variant valled Venus was Tsr, a membrane protein which will diffuse to the cell surface and then get anchored in the cell wall. This has the effect of drastically reducing the diffusion of the fluorescent signal, an example of which is shown below:
With a working gene expression reporter, a series of time-lapse experiments were carried out. Living E. coli cells were monitored for YFP fluorescence, with a typical experiment looking like this (also, for the physicists reading this, note the regularity with which the cells divide — even though chemistry is stochastic at its fundamental level, cells go about their business with remarkable timing):
The authors remark,
Several qualitative features are evident from these time traces. First, protein molecules are generated in bursts. Second, the number of protein molecules in each burst varies. Third, the bursts exhibit particular temporal spreads. Analysis of of the data allows us to address the following four questions: Do these gene expression bursts occur randomly in time? How many mRNA molecules are responsible for each expression burst under repressed conditions? What is the distribution of the number of protein molecules in each burst? And what is the origin of the temporal spread of the individual bursts?
The answers are perhaps not surprising: analysis of burst frequency with cell cycle shows that burst frequency is poisson distributed, a random process.
One mRNA molecule able to give rise to bursts in expression, and the probability for expression n proteins from a single mRNA molecule follows a geometric series:
where there is competition between ribosomes (which translate proteins) and RNAse E (which degrade RNA) for the mRNA binding site with probabilities ρ and ρ – 1, respectively. A histogram of the number of molecules per burst fits well to the above equation, giving a value of n ~ 4.2 Tsr-Venus constructs per burst.
The temporal spread of the bursts, characterized by the autocorrelation of the fluctuation in protein expression, gives a much less convincing answer, but given the speed of RNA transcription (~50bp/s) and protein translation (15 residues/s), the limiting rate constant of ~7.0 minutes is assigned to post-translational assembly of the Trs-Venus protein.
Given our incomplete understanding of gene expression, these two techniques will help open the door to further studies on low-copy number gene translation which, up to now, have been impossible to study. And doing science in vivo is just cool, too.