Creating a population specific gene expression profile from heterogenous tissue

[Anand]

When attending the symposium on developmental disturbances in the nervous system, I was lucky enough to hear a talk by Ines Ibanez-Tallon. During her talk she mentioned a technique they were using to look at the specific gene expression of inner hair cells and outer hair cells within the cochlear. This tissue is rather hard to get to and thus they employ what I thought was a rather brilliant technique.

They created mice with a conditional mutation that caused them to express ribosomal proteins which were tagged with E-GFP. After processing the heterogenous tissue they were able to specifically select the ribosomes that were tagged with E-GFP using antibodies to E-GFP. After further processing they were able to get the mRNA that these ribosomes were translating and apply them to a gene chip to get the specific gene expression of the population of neurons expressing the E-GFP tagged ribosomal protein.

I do wonder how tagging the ribosomal protein would affect the ability of the ribosome to bind and translate the mRNA. Would there be specific mRNA that would not be able to bind this modified ribosomes? I wonder if this has been checked. That said this does prove to be an intriguing way of specifically targetting a specific population of cells and finding out its gene expression.

[Wing]

It is a common technique in molecular field. RNA tagging is more conveniently done by using a FLAG or other small tag linked PABP. It is simply like doing a RT-ChIP with anti-FLAG. I do not know whether ribosome-EGFP will influence gene expression but I think PABP is better, since it only leads to mRNA.

Most of tagging techniques recovers only partial set of RNA. But tagging is less difficult than single cell analysis.

Axel lab pioneered single cell RTPCR by labelling specific cells, and pick up with either pipette or laser afterwards. Some years latter with the aid of T7 linear amplification people were enabled to perform single cell microarray. Currently these single cell techniques could be adopted to be (and less difficult) multi-cell, single population, hence solved the problem you mentioned.