Fast isogenic mapping-by-sequencing of EMS-induced mutant bulks: an interview with Dr. Turck, MPI, Germany
Deep candidate resequencing (dCARE) using the Ion PGM™ Sequencer increased coverage at the candidate region
Franziska Turck, PhD, is a research group leader at the Max Planck Institute for Plant Breeding Research in Cologne, Germany. Dr. Turck’s group studies the interplay of DNA-binding transcription factors and chromatin structuring complexes in the orchestration of gene expression. In particular, the group studies the regulation of key developmental genes, which are mostly under the control of repressive chromatin complexes. Dr. Turck and her team recently published the paper “Fast isogenic mapping-by-sequencing of EMS-induced mutant bulks” in Plant Physiology and introduced deep candidate resequencing (dCARE) using the Ion PGM™ Sequencer to their mutant identification pipeline. In this interview, Dr. Turck answers questions about her work.
1. Can you tell us the broad goal of the work described in the recent Plant Physiology article?
In plants, induced mutagenesis provides a magnificent tool to elucidate new components in any regulatory pathway. For our particular case, there is a mystery still to be solved around a chromatin factor called LHP1 (LIKE-HETEROCHROMATIN PROTEIN 1), which should have much stronger effects on plant development if mutated than it actually has. This implies that there are some undiscovered redundancies or compensatory pathways. As Arabidopsis geneticists, an obvious approach for us was to initiate a suppressor/enhancer screen. In such a screen we use the original mutant, already defective in LHP1, and hit it again with a chemical mutagen. Among the offspring of these plants, we identify individuals that are either worse off than the parental mutant or look more like a nonmutant plant. The goal is to identify the causative genes and study their connection to LHP1.
2. Why is this goal important? Who will ultimately benefit from this work, and how? For example, how will it matter to breeders or farmers?
The results are interesting from a purely mechanistic point of view because, as mentioned above, there is still a mystery around this. As the pathways we are working on are conserved in insects, humans, and plants, we may uncover something new in plants with corresponding implications for other species. There are many examples of this cross-fertilization between fields in the past. In animals, “our” pathway is highly relevant for stem cell research.
In plants, this pathway controls genes that are important for the correct seasonal timing of flowering. Mutant plants that are defective in the pathway are not so interesting for breeding purposes, because they have lost the ability to delay the expression of flower-promoting genes, which are usually only turned on after the perception of inductive environmental signals. However, we are interested in the detailed “how” of LHP1 and its effect on gene control. As we learn more about this, we become able to turn a few screws in the process. For example, we can make flower-promoting genes more or less sensitive to external signals. Being able to subtly change such environmental responses can have tremendous impact on breeding in the context of a changing environment. Even better, we can predict effects on naturally occurring mutations, which are immediately useful for breeding.
3. What has enabled you to do this experiment now? Has our knowledge or the technology changed? What is ‘first” here?
It is not a matter of first, but of much-improved in this case. The mapping of mutations in enhancer/suppressor screens has several bottlenecks that next-generation technology helps us to overcome.
4. Can you explain isogenic mapping? How does it work? What was the previous approach? Why is it better?
In the past, in order to map a suppressor/enhancer mutation, it was necessary to have the original mutant available in two independent genetic backgrounds that could be crossed. Quite often, the crosses between the different backgrounds generated so much variation in the offspring that the actual phenotype of interest was lost. With NGS technology, we and our bioinformatic collaborators from Korbinian Schneeberger’s group realized that we do not need two independent backgrounds since we can use the mutations introduced during the mutagenesis to identify our gene of interest.
5. Can you briefly walk us through the steps of the experiment?
- In a pool of seeds that are already mutants, more random mutations are induced with the help of a chemical mutagen. The seeds are grown up and their offspring are scored for additional phenotypes.
- Once interesting candidates are identified, these are crossed back to the original mutant. Subsequently, the offspring of the cross are again scored and selected for the phenotype. This can be repeated one more time, but that is optional.
- Then many plants (~250) are grown—soil is inexpensive—and leaves from many individuals with the phenotype are harvested. The leaves are pooled and DNA is harvested from the pool. As a reference, DNA is extracted from the original mutant.
- As a first sequencing step, an Illumina platform was used to obtain 20- to 40-fold coverage of the genome.
- Mutations that are not selected by phenotype segregate in the pool of twice-backcrossed plants, therefore it’s necessary to identify a region of interest with nearly homozygous mutations.
- The pool of 250 plants contains more precise mapping information than can be assessed by a 40-fold coverage. Candidate mutations from the focus region are then selected, primers that amplify over these regions are designed, and DNA is amplified by PCR using the original pool.
- The Ion PGM™ is then used for amplicon sequencing, and the SNP rates are counted. In our case, we nailed the mutated gene with deep candidate resequencing (dCARE).
6. What role do you think isogenic mapping-by-sequencing will play in plant research in the future?
Isogenic mapping will replace more conventional approaches that require distinct genetic backgrounds for gene identification. I predict a new wave of interesting genes being cloned that could previously not be mapped because the phenotypes disappeared during mapping. Isogenic approaches also facilitate the use of transgenic marker lines as phenotypic trait. In addition, nothing stops us now from doing tertiary mutagenesis or mutagenesis of, let’s say, quadruple mutants.
7. Could dCARE replace other methods, e.g., TILLING?
Yes, dCARE can replace CEL1-based TILLING immediately. One would have to think of a cost-effective barcoding system for the Ion PGM™ System.
8. Why did you choose the Ion PGM™ System for this experiment?
We can go from PCR to sequences in two days, whereas we would have had to wait up to a couple of months for results with other sequencers. And some of these would have been overkill with respect to sequence depth, but that is more an argument decided through the final cost of the experiment.
9. What are the advantages of Ion Torrent™ semiconductor sequencing for this application?
We didn’t need to think about any complicated way of making libraries as we included all required sequences in the amplification primers.
10. Do you see this workflow being adopted by the research community, and if so, for which applications?
The Ion PGM™ System could be useful in marker-assisted breeding—once the markers of interest are known and can be amplified.
11. Do you see any obstacles for this?
Mostly people do not have an Ion PGM™ System available, and therefore don’t even consider using it.
12. What's the next step for your research in this area?
Well, we have now mapped half a dozen genes with our approach and will probably be busy for a couple of years studying their precise functions. Technology-wise, I would like to pick up the challenge and map some of the mutants that seem to be affected in two genes at once, because the phenotype segregates accordingly. This should be possible in a very similar way.
13. Where would you like to take the Ion PGM™ technology for your laboratory?
The Ion PGM™ could become a great tool for measuring allele-specific expression—not so much as global transcriptome analysis, but for more targeted approaches. Again, a nicely functioning barcoding system would be very helpful here. Such a high-power barcoding system would also promote other applications (such as TILLING, marker-assisted breeding, etc.).