Stochastic protein expression in individual cells at the single molecule level

(This was a project report for a microfluidic class written by Ravi Peesera.)

Introduction

Microfluidics is a prominent modern technology which deals with the manipulation of fluids in very small quantities, and has a very large number of potential applications in many fields of science.^2,3 In the field of analysis, microfluidics offers distinct advantages in conferring the ability to handle minute quantities of samples, when coupled with powerful analytical techniques opens up a vast number of applications. Single molecule detection methods are the emerging technologies which carry out measurements at single molecule level, which remove the discrepancies in bulk, ensemble average sample measurements. Microfluidics, with small sample handling capability, makes single molecule measurements possible and further makes the experiments highly designable, in allowing varied designs of miniaturized microfluidic chip systems.

The publication “Stochastic protein expression in individual cells at the single molecule level”¹ is a relevant example of measurements carried out with the combination of microfluidics and single molecule technology, in understanding the events of protein production. The transcription of DNA to messenger RNA and the translation to protein synthesis that follows occurs in a stochastic way in living cells. This randomness makes the development of assays to identify processes during the protein production difficult on a large ensemble of cells, due to the lack of synchronization of these events between cells, thus necessitating the use of single cell level measurements to understand these processes correctly.

β-galactosidase (β-gal) is a hydrolase enzyme that catalyzes the hydrolysis of biologically bound galactosides into galactose monosaccharides in living systems.⁴ Further, β-gal has been shown in various studies to be the standard reporter for gene expression in prokaryotic and eukaryotic cells. In 1961, Rotman and coworkers demonstrated that β-gal is active only as a tetramer and that a single molecule of β-gal can produce a large number of fluorescent product molecules when treated with the synthetic FDG molecule shown in the figure below.⁵ As shown in the figure below, β-gal reacts with the synthetic FDG, which contains two galactose units, cleaves both of them and liberates two molecules of galactose and one molecule of the fluorescent fluoroscein molecule. By virtue of the enzymatic activity β-gal reacts with a large number of FDG molecules, breaks them down by hydrolysis, thus liberating a large number of the fluoroscein molecules. This leads to an accumulation of these fluorescent molecules in the cell which would serve as a marker for the presence and thus the expression of β-gal, which in turn pinpoints to events in protein synthesis. Thus the accurate measurement of this amplified fluorescence in the cell at the single molecule level is crucial to development of an assay to understand this process in real time.

This exiting design of the experiment however has a limitation biologically. After the hydrolysis of FDG’s by the β-gal expressed during protein production, the resulting foreign organic fluoroscein molecules are acted upon by the cellular mechanism and are pumped out of the cell cytoplasm by the efflux pumps on the cell membrane, actively and efficiently. The pumped out fluoroscien molecules diffuse away rapidly into the extracellular region, and so the amplification of the fluorescent signal gained in the cell is lost. So in order to suppress the loss of fluorescence signal, arresting the diffusion process and the localizing it to an area close to the cell is crucial for the measurement of this signal without loss of the enzymatic β-gal amplification.

Theory and Principles

Long Cai and coworkers at Harvard chose to address this problems by using microfluidics to develop an assay to measure the fluorescence without loss of the amplified signal.¹ The cells were trapped in closed microfluidic chambers which ensures that the fluorescent product expelled from the cell is accumulated in small volumes around the cell in these chambers thus stopping the diffusion loss of enzyme amplified fluorescence. The short mixing time of the FDG reagent and the rapid efflux rate at the cellular membrane ensure that the fluorescence signal outside the cells accurately reflects the enzymatic activity happening in the cell. The requisite microfluidic device is made using a soft polymer, polydimethylsiloxane (PDMS), and consist of a flow layer that contains the cells and a top control layer as shown in figure 1. The actuation of the two adjacent valves in the control layer forms an enclosure of the dimensions 100*100*10 µm³ in which the cells are trapped and cultured. The experiment was carried out by mounting the microfluidic chip onto an inverted fluorescence microscope and was translated by a motorized stage, thus allowing multiplexing of the data acquisition by repeated scans on the chamber, carrying out a single scan of 100 chambers in 2 min typically. As shown in figure 1 fluorescence is exited by a tightly focused laser beam that does not directly illuminate the cell, which ensures no cellular auto fluorescence and photo-damage to the cell do not happen.

Figure 1: Schematic diagram of the microfluidic chamber 

used for the enzymatic assay.¹

Initially experiments were carried out to test the system for single molecule measurements. This was done by injecting β-gal enzyme solutions which are very dilute into the chambers and then the fluoregenic fouorescien digalactoside( FDG) substrate was introduced into the chambers. It was observed that the fluorescent signals increase with time, and the slopes of curves shown in the figure 2 give the rates of the hydrolysis.  It was found from these plots that the distribution of hydrolysis rates measured in different chambers showed quantized and evenly spaced peaks which in turn was attributed to the presence of integer numbers of β-gal molecules. Also the spacing between the peaks is 60pMmin^-1., which tells the calibration for the increase in rate of fluoroscein concentration corresponding to one enzyme molecule in the chamber. 

Figure 2: FDG hydrolysis by purified β-gal.¹

Fluoroscein concentration increases with time in chambers. The discrete slopes are due to, in decreasing order, 3,2,1,0 (autohydrolysis) β-gal molecules. A histogram of hydrolysis rates (48 chambers). The red curve is a fit to the data with three gaussians of equal widths. The peaks are attributed to 0, 1, and 2 enzyme molecules per chamber. The distribution is well-fitted by a Poisson distribution (black line), with an average of 0.7 β-gal molecules per chamber.

In live cells, the cell wall acts as a barrier for FDG influx. So, to quantify this effect, they measured the hydrolysis rates for live cells compared to cells treated with chloroform, which completely permeabilizes cell membrane. The ratio of hydrolysis rates between these two case (permeability ratio) was measured to be R=13 at 300 µm FDG. They increased FDG influx by transforming E. coli cells with a plasmid conferring amplicillin resistance and grew the cells in media with β-lactam antibiotics, there by inhibiting the cell wall synthesis. Hence the permeability ratio of 2 + 0.3 was observed. So in determining the number of enzyme molecules in live cells, R=2 is used as a correction factor to the in vitro calibration.

Having established the ability of the system to analyze at single molecule level, gene expression in live E.Coli cells in real time was attempted. The lacZ gene on the chromosomal DNA, which is under the control of the lac  promoter, is responsible for the expression of β-gal.  In order to monitor these events the cells were cultured in glucose containing medium, and then an abrupt change in the hydrolysis rates was observed in chambers with dividing cells as shown in figure 3. The stepwise increase leads to the conclusion that the expression of the β-gal molecules is stochastic in nature. Thus this microfluidic based single molecule measurement shows that the expression of β-gal molecules which are the most common reporters for the protein expression cascade are expressed in a stochastic manner.

Figure 3: Quantitative real-time measurement of individual 

protein expression events in live E. coli cells.¹

The average frequency of expression bursts per cycle, and the average number of protein molecules per burst, indicated by a and b respectively, are the two key parameters which help characterize the expression of a protein from a given gene. These valued were measured for the live E.Coli system to be a = 0.11 + 0.03 bursts per cycle. Also the average burst size was measured to be 5 + 2 enzymes, or 20 + 8monomers per  burst. Knowing the number of β-gal molecules produced per burst, the number of enzymes produce per burst was found using figure 4, which is an exponential distribution fitted using the equation, P(n) = Cexp(-n/b), where n is the number of β-gal molecules per burst and C is the normalization constant. These two measurements of the number of β-gal molecules per burst and the number of enzymes produced per burst provide quantitative real time information in monitoring the gene expression in a live cell.

Figure 4: Histogram of copy number of β-gal molecules 

per burst.¹

In comparison, this assay developed has the advantage of real time monitoring of the protein expression phenomenon and does not necessitate fluorescent tagging experiments.⁶ In addition these studies performed involve low copy of protein transcription and thus very low expression of protein being studied which under single molecule experimental conditions will be a very minute concentration. In such cases fluorescent tagging would not give accurate measurements owing to a weak signal, and the current methodology circumvents this problem by not measuring the protein concentration and measuring the enzymatic activity of the β-gal, which being enzymatic provides the amplified signal of the fluoroscien molecule as shown above. Further microfuidics is pivotal in the experimental design for localizing the signal to prevent diffusion and successfully measuring the low copy number phenomena. This experiment thus successfully opens way for measuring cellular events by combining single molecule experiments with microfluidics.

Applications

This microfluidic assay can be applied to other cells types expressing β-gal as a reporter. They demonstrated this applicability by measuring low-level expression from repressed promoters in yeast and mammalian cells. This assay exhibits single molecule sensitivity for single prokaryotic and eukaryotic cells. This method, combined with other newly developed single cell and single molecule techniques at the mRNA and protein levels, has opened up possibilities for studying low-level gene expression in single cells of diverse cell types.

Conclusion:

In conclusion these series of assays developed to trace the gene expression pathway in real time rely upon microfluidics techniques. Key to the success of these experiments is the ability to stop the diffusion of the effluxed fluoroscein molecules form the cell cytoplasm by localizing them in microfluidic cell chambers thus preserving the enzymatic amplification signal. These experiments showcase not only the ability of microfluidics chips to be highly designable for various experimental requirement but also their use in single cell based measurements which are highly relevant tools in studying  complex biological processes.

References:

1.      Cai, L., Friedman, N & Xie, S. Stochastic protein expression in individual cells at the single molecule level. Nature, 440, 358-361 (2006).

2.      Whitesides, G. M., The origins and the future of microfluidics., Nature, 447, 368-373 (2006).

3.      Moerener, W.E. & Fromm, D.P. Methods for single molecule fluorescence spectroscopy and microscopy. Rev. Sci. Instrum., 74, 3517-3619 (2003).

4.      Fowler, A. V., & Zabin, I. The amino acid sequence of beta galactosidase, Isolation and composition of tryptic peptides. J. Biol. Chem. 245, 5032–41 (1970).

5.      Rotman, B. Measurement of activity of single molecules of b-D-galactosidase.

      Proc. Natl Acad. Sci. USA 47, 1981–-1991 (1961).

         6.   Ghaemmaghami, S. et al. Global analysis of protein expression in yeast. Nature

      425, 737–-741 (2003).