Why self-incompatibility in the Brassicaceae is totally cool
Self-incompatibility (SI)
is a reproductive system by which hermaphrodite flowering plants
recognize and specifically reject self-pollen. Here are three reasons making SI an outstanding model system for evolutionary biology.
1. A clearly defined phenotype
Incompatibility phenotypes are highly reproducible and easy to measure experimentally, either from direct counting of pollen tubes germination along the style (Figure 1) or from fruit elongation after controlled pollination (Figure 2). While there can be some overlap in some circumstances, the distribution of these two measures typically allows the definition of a clear cut-off value distinguishing compatible from non-compatible crosses (Figure 2).
Figure 1. Count of germinated pollen tubes on pistils provides a powerful way to distinguish compatible (right) from non-compatible (left) crosses.
|
Figure 2. Measurement of pistil elongation after controlled pollen deposition provides an alternative way to tell apart compatible and non-compatible crosses.
|
2. Balancing selection and inbreeding depression : a clear phenotype-to-fitness map
A major asset of SI is the whealth of theoretical work that has been done to explore the fitness consequences of changes in the genetic composition of the population. Beside loss-of-function mutations, whose fate is mostly related to the level of inbreeding depression in the population, the major determinant of fitness is relative frequency of all other individual recognition phenotypes in the population. As a result, SI has been THE textbook exemple of balancing selection, in the form of negative frequency-dependent selection (Figure 3). A number of high profile population geneticists have contributed to the development of the field, starting by S. Wright as early as in 1939, more recently followed by our colleagues M Uyenoyama, D Charlesworth, X Vekemans, M Schierup and S Billiard. Based on these models, it is possible to predict the evolutionary trajectories for a range of different mutations either disturbing the recognition process, restoring it after it has been disturbed (i.e. compensatory mutations), altering the dominance hierarchy or modifying the level of inbreeding depression (Figure 3).
3. A molecular lock-and-key and a regulatory network based on small RNAs: a clear genotype-to-phenotype map
In parallel to the long-standing interest of evolutionary biologists, another scientific community has been studying SI for years, providing an exquisitely detailed mechanistic understanding of the molecular players controlling the phenotype (Figure 3). In the Brassicaceae, SI is based on a molecular lock-and-key mechanism involving two tightly linked genes: one encoding a transmembrane protein displayed at the surface of stigmatic papilla cells (SRK, the female receptor) and one encoding a short peptide expressed in the anther tapetum and deposited on the surface of pollen grains (SCR, the male ligand). Both genes display large allelic series (at least 50 S-alleles within species) and are tightly linked in a small non-recombining region (Guo et al. 2011; Goubet et al. 2012), which ensures strict haplotypic association between co-adapted SCR and SRK alleles. Pollen germination is inhibited when pollen coated by a given SCR allele lands on a stigma expressing its cognate SRK allele, ultimately preventing fertilization. In contrast, molecular docking is impossible between SCR and SRK proteins that have been produced from distinct haplotypes, allowing for successful cross-fertilization provided that the two mating partners carry and express distinct S-alleles (Figure 4).
Figure 3. A major asset of SI is the detailed understanding we now have of both the phenotype-to-fitness and the genotype-to-phenotype maps. This enables us to link molecular variation down to the nucleotide level, up to their fitness consequences in natural populations.
|
Figure 4. Self-incompatibility in outcrossing Arabidopsis species is controlled by a molecular lock-and-key mechanism, and entails a specific co-evolutionary process between the pistil receptor (SRK) and its pollen ligand (SCR). From Castric et al. (2014)
|
Unsolved puzzles on the evolution of SI
The specific interaction between the SCR and SRK genes is a spectacular example of co-evolution at the molecular level, whereby one SCR allele recognizes one (and only one) SRK allele
in order to maintain self-rejection while in the same time optimizing
cross-fertilization opportunities. The driving force of the diversifying
process in this system is the fitness cost associated with pistil
rejection of non-self pollen, as this cost is
inversely related to the number of co-occurring allelic specificities.
The very fact that so many S-alleles segregate in natural populations
implies that this force is strong and that functional novelty must have
emerged repeatedly in this system. Surprisingly however, the process by which new SCR-SRK combinations arise remains a major evolutionary puzzle,
precisely because a mutation in just one of the two genes would break
down the interaction and thus lead to non-functional low fitness
intermediates (Uyenoyama 2001, Gervais 2011). Solution to this puzzle
will require to unscrew the co-evolutionary process in order to
understand how the fitness valleys in the fitness landscape may be
crossed.
A
second fascinating feature of this system is the complex network of
pairwise dominance/recessivity interactions among alleles of the SCR gene (expressed in pollen). While the SCR gene is expressed in the anther tapetum of the paternal plant (i.e. a
diploid tissue) rather than by the haploid genome of the pollen grain
itself, in most cases heterozygote plants only express one of their two SCR alleles
through a phenomenon of monoallelic gene silencing. We have recently
discovered that these dominance/recessivity interactions are controlled
by a collection of small non-coding RNAs produced from hairpin
precursors in the non-recombining S-locus region (Figure 2) that
collectively explain 93% of all observed dominance interactions with a
fairly low rate of false positives (7%). In this system, alleles
that are high in the dominance hierarchy tend to have sRNA precursors
that have a broader targeting spectrum, while alleles that are more
recessive tend to have more sRNA targets. Interestingly, only a
minor fraction of pairs of S-alleles show co-dominance, suggesting that
this kind of regulatory mechanism has evolved repeatedly in the course
of the diversification process of the S-alleles. The driving force of
this regulatory system is again the fitness cost associated with pistil
rejection of non-self pollen, as expressing a single allele in
heterozygous pollen-producing plants will minimize pistil recognition
and rejection and will therefore be advantageous. Evolution of this
regulatory mechanism also entails a clear co-evolutionary mechanism
between the sRNAs and their target sites in a way that we are just
beginning to understand.
|
Figure 5. The dominance hierarchy observed between SCR alleles (shown left) is governed by tightly linked small non-coding RNAs produced from hairpin precursors (color-coded by family, on the left) that target specific sequences of the SCR alleles (on the right). Lines join sRNA precursors to their predicted target sequence. Black lines indicate target predictions that are consistent with the phenotypic observation (from a dominant to a more recessive allele), while red lines indicate target predictions that are not consistent with the phenotype (from a recessive to a more dominant allele). This regulatory mechanism also entails a co-evolutionary process between the sRNAs and their target sites.
|
Emergence of functional novelty at the biological systems level.
Overall, self-incompatibility is an outstanding model system in which both the primary phenotype (specificity of the recognition proteins) controlled by the molecular lock-and-key system AND the inheritance system ultimately determining the phenotype (dominance/recessivity interactions) are evolving rapidly. The high diversity of the molecular lock-and-key and its complex regulatory machinery involve repeated emergence of functional novelty for both of these aspects, hence providing the opportunity to study recently evolved molecular interactions in a context where the co-evolutionary constraint is an integral part of the mechanism by which the system functions. A major asset is that both the genomic and ecological contexts in which these genetic systems evolve have been explicitly clarified, providing the unique opportunity to investigate the co-evolutionary process down from nucleotide variation all the way up to their fitness consequences. By building upon these two case studies, our objectives are to 1) decrypt the molecular alphabet of the interaction between the co-evolving nucleotide sequences governing these genetic systems; 2) predict and evaluate the fitness landscapes upon which the two co-evolutionary processes are taking place and 3) exploit natural variation of the genomic region containing these two molecular systems to unveil the kind of co-evolutionary process taking place in natural populations. Our projects therefore aim at shedding light on a fundamental biological process: the emergence of functional novelty at the biological systems level.