blog

We recently indexed an article by Shin-ichiro Kojima and Gary Borisy, about a method they developed that allows cell biologists to evaluate within single cells whether their RNA interference method is working. One of the current challenges in cell and molecular biology is to determine the biological function of genes, now that we know the genetic sequence of all of them. A popular method for this kind of work is RNA interference, which allows researchers to see what happens when a particular gene is no longer expressed. In practice, however, knockdown isn’t always efficient and there is a lot of variation between cells. I discovered this first hand during my own research, so I was excited to see that Kojima and Borisy developed a method to determine knockdown efficiency at the cellular level.

Below, they answered some of my questions about their article. (You may also remember Gary Borisy from the evening of talks F1000Research organised last year in Boston, where he talked about the role of government funding agencies in the move towards open access.)

 

Could you explain in a few sentences what your paper is about?

The paper is about an improved way of determining the effectiveness of shRNA-mediated knockdown of mRNA. The basic design concept is to quantitatively measure the fluorescence of two reporter constructs in individual cells–one construct expresses the candidate shRNA together with a reporter fluorescent protein, say green, while the other expresses a 19-nt target sequence, which serves as a proxy for the mRNA, and a second reporter, say red. If the shRNA is effective, the target RNA is degraded which is read out as a reduced red/green fluorescent ratio.

 

What inspired you to develop this assay? Did your own work bring to light the necessity for such a system?

We, like many laboratories, have used siRNA-mediated knockdown of mRNA to analyze the role of specific proteins in cell function—for us, the focus was on the actin cytoskeleton. A general and serious problem was heterogeneity in the cell population in terms of both knockdown level and cell phenotype. The combined heterogeneity introduced ambiguity in determining whether the knockdown was effective and therefore, what molecule could be connected with what phenotype. Therefore, it became important for us to have a method for evaluating the extent of knockdown in individual cells where we could also determine the phenotype. The ratiometric approach seemed desirable because it essentially provides an internal control.

Another issue was the effort involved in developing and evaluating the antibody reagents typically used to evaluate knockdown. A great deal of work is involved in making and evaluating a good reagent for a specific protein, but the work does not carry over to making a reagent for a different protein.  Since we wanted to perform RNAi systematically, we wanted to develop a method of evaluation of RNAi that is basically applicable to all genes. The dual fluorescence reporter assay provides a general approach for knockdown of any gene provided sequence information on the target mRNA is available.

 

I also have some practical questions. First, considering that the pSHIN-G/pREFLECT-R system itself uses two fluorophores, would it be possible to combine this with additional fluorophores? For example, I can imagine situations where a researcher might want to use immunofluorescence to look at downstream effects of knockdown in these same cells.

Actually, our assay uses three colors; besides the two colors of fluorescent proteins, we use a third color channel for DNA-staining in order to facilitate computationally-based image analysis of the cells.  Nevertheless, your point is well-taken. It would certainly be possible to combine our technique with additional labels—either by immunofluorescence or other affinity methods. Current spectral imaging technology allows the discrimination of many colors.

 

The three-plasmid system looks really useful, considering many researchers will already have shRNA constructs. However, as the referees of your paper also pointed out, wouldn’t it rely on a direct dosage correlation? Do you think this would be affected in any way by plasmid size or promoter type of the third plasmid?

We introduced the three-plasmid system because of the reason you said—it could be convenient for the many researchers who will already have shRNA constructs, but it definitely represents a trade-off. In the two-plasmid system, we have two key advantages: first, and most important, the molar ratio of the GFP reporter and the shRNA is fixed so that the reporter truly gives an accurate readout of how much shRNA is expressed in the same cell; secondly, we designed the shRNA and target constructs to be as similar as possible so that transfection and expression levels would also be similar and we in fact saw tight correlation in the results.

In the three-plasmid system, we do lose the tight coupling of reporter and shRNA expression. Also, because the third plasmid which contains the shRNA may differ in construction and promoter type, it is likely that transfection and expression levels may vary depending on the plasmid.  Nevertheless, our results did show that fluorescence intensities were linearly correlated though the expression level varied cell by cell.  So, although we agree with the reviewers’ concern that the three plasmid system will not always work, we thought that investigators who already have significant shRNA libraries might find the approach worth considering.

 

Finally, you mentioned in your paper the opportunity for using your assay in high-throughput screening or high-content analysis systems. To facilitate such work, could your two-plasmid system be used as a vector for shRNA libraries, or would the need of target sequences for pREFLECT-R complicate that?

High throughput screening was one of the applications we had in mind in developing the assay. We do see this as being made possible by the combination of the two technical advances in our method—the computational, image-based, dual fluorescence assay and the simplified method of construction of shRNA expression plasmids. However, the method does require a match between the shRNA sequence and the target sequence in the pREFLECT-R construct. So, for screening shRNA libraries, one would also have to construct cognate target sequence libraries. Although this is a complication, modern DNA synthesis technology is reducing the barrier to adoption.