Peter Mombaerts, M.D., Ph.D.
Director, Department of Molecular Neurogenetics | Max Planck Institute of Biophysics - Frankfurt, Germany
No fewer than 1,200 genes comprise the mouse odorant receptor (OR) gene repertoire – making it by far the largest gene family in the mouse genome, and perhaps in any genome.
Dr. Peter Mombaerts works on the neurobiology of smell in mice. Among his research interests is “the difficult question of how each olfactory sensory neuron in the mouse manages to express just one gene from the repertoire of these 1,200 genes.”
In particular, the mechanisms that regulate the probability of OR gene choice are not well understood. Understanding these mechanisms could have broader implications for understanding gene regulation in the mammalian genome.
“Because we have so many genes to look at, and OR genes are often very similar to each other in sequence, specificity is a major concern for an expression assay,” said Dr. Mombaerts.
In 2008, he was delighted to learn about the NanoString technology while giving a seminar at Caltech, an early adopter of the NanoString platform. Later he acquired the first NanoString system to be installed in Germany.
In November 2009 Dr. Mona Khan and Dr. Mombaerts had a proof of principle project with the technology done at NanoString headquarters in Seattle, WA. “The results that came back were very encouraging,” said Dr. Mombaerts. “The assay had both the sensitivity and specificity that we need to look at the diverse tissue that the olfactory epithelium represents - 1,200 types of cells mixed together, each of them expressing a different OR gene. Sensitivity and specificity are usually opposing demands, and this was just right.”
In a paper published in the 11/11/11, issue of Cell,1 Dr. Mombaerts and his colleagues Dr. Mona Khan and Dr. Evelien Vaes detail how they applied the NanoString platform of fluorescent barcodes and digital readout to measure RNA levels of 577 OR genes in a single reaction.
“We made a strategic decision early on to only use coding sequence for probe design because the non-coding sequence of OR genes is unreliable, often only computationally determined, and so we didn’t want to rely on it,” he said. “The fact that NanoString can measure up to 800 genes in parallel is a sweet number for us. With probes designed against coding sequence, we can measure expression of half of the OR gene repertoire in one go.”
In the Cell paper, the Max Planck team describes a new regulatory element they call the P element. Using a knockout mouse that is lacking it, they sought to determine the effect of the P element. Using a NanoString CodeSet codenamed Gorilla they were able to show that the effect was limited to the nearby OR genes, and also only on the chromosome on which it resides (in cis).
“We show that only 10 OR genes are differentially expressed in the absence of this tiny DNA sequence,” he explained. “NanoString was particularly helpful here to give us a genome-wide view, and we can convincingly say that the effect of the P element is local.”
“The P element regulates the probability of OR gene choice,” he continued. “Without this element there are fewer cells expressing OR genes of the cluster, but each cell expresses RNA for its OR gene on average at normal levels. Therefore, RNA expression is normal at the single cell level but at the tissue level expression is decreased because there are fewer cells.”
Powerful control experiments further convinced Dr. Mombaerts and colleagues that they were on the right track with NanoString for their research. The researchers used a cell sorter to collect pools of ~1000 olfactory sensory neurons that express an OR gene tagged with a GFP construct. As the nCounter assay does not require purified RNA, they were able to evaluate small populations of cells by the direct analysis of cell lysates. The only OR gene in the CodeSet to give a high number of barcode counts in the sorted sample was the tagged OR gene. In a reciprocal experiment, they found that RNA samples from a mouse strain lacking an entire cluster of OR genes as a result of genetic engineering showed only background levels of barcode counts for the deleted genes.
One of the best things about the NanoString technology, according to Dr. Mombaerts, is the time savings. “The hands-on time is really very, very minimal – you get a lot of data by spending almost no time after the RNA samples are collected. And we don’t have to do technical replicates either.”
He stressed that this technology was truly enabling. “This is not a validation of another technique; we’ve done something we couldn’t do before with any other technology.”
Dr. Mombaerts concluded: “The question of how one neuron only expresses one gene out of 1,200 is a simple question to ask, but a very difficult one to answer. To do that we need a very good assay - an expression assay that is reproducible, sensitive, and specific. And we have that now, and the field has it now too, as our probe sequences are of course publicly available.”
1 Khan et al., Regulation of the Probability of Mouse Odorant Receptor Gene Choice, Cell (2011), p. 907-921, doi:10.1016/j.cell.2011.09.049