Last author

Genomic imprinting is a controversial mechanism because it contradicts the idea of mendelian inheritance. In this process, only one of the two alleles is expressed, so only one parent influences expression of the gene. To investigate the necessity of imprinting in seed development, Arp Schnittger at the University of Cologne in Germany and his colleagues crossed paternal mutants that could not contribute genes to endosperm development with maternal mutants that decode genetic imprinting patterns (see page 312). The seeds that resulted were viable, lending credence to botanist Eduard Strasburger's hypothesis, first raised in 1900, that the endosperm is of maternal origin.

What was it like studying a hypothesis first put forward in 1900?

I am awestruck by the insights of Eduard Strasburger — especially as his understanding was based on comparative morphology. We used elaborate genetics and advanced tools and techniques, and came to the same conclusion.

How did your research in this area begin?

The major interest of my group is the crosstalk between cell-cycle control and development. We developed an Arabidopsis mutant in which only one male gamete is generated, rather than the two typically produced in flowering plants. This single gamete can still fertilize female gametes, but only the egg cell, and not the other female gamete, the central cell. With this single fertilization we can now compare the input from the mother with that of the father.

Do your findings challenge the biological necessity of genomic imprinting?

We think that our data will help to refine the issue of how necessary biological imprinting is. We have shown that imprinting is not required for seed development, but as it is an essential mechanism in wild-type plants it is tempting to speculate that a 'naïve' genome, which does not have imprinting marks, might not be stable in evolution.

What are the possible applications for plant breeding?

The most exciting potential application is for the generation of apomictic plants — those that produce seeds without fertilization.

What's your next step?

We now want to analyse which genes from the mother and father are imprinted in the developing seed. The combination of the mutant with genome-wide transcriptional profiling will hopefully provide us with a new way to unravel differentially expressed genes.