Many flowering plants are hermaphrodite, posing the problem of self-fertilization and the subsequent loss of the genetic fitness of the offspring. To prevent this, many plants have developed a genetically controlled mechanism called self-incompatibility (SI). When the male and female S-determinants match, self (incompatible) pollen is recognized and rejected before fertilization can occur. In poppy (Papaver rhoeas), the pistil S-determinant (PrsS) is a small secreted protein that interacts with incompatible pollen, initiating a Ca2+-dependent signalling network. SI triggers several downstream events, including depolymerization of the cytoskeleton, phosphorylation of two soluble inorganic pyrophosphatases and an MAPK (mitogen-activated protein kinase). This culminates in PCD (programmed cell death) involving several caspase-like activities. The recent discovery of the Papaver pollen S-determinant PrpS marks a significant step forward in the understanding of the Papaver SI system. PrpS encodes a ~20 kDa predicted transmembrane protein which has no homology with known proteins. It is specifically expressed in pollen, linked to the pistil S-determinant, and displays the high polymorphism expected of an S-locus determinant. The present review focuses on the discovery and characterization of PrpS which strongly support the hypothesis that Papaver SI is triggered by the interaction of PrsS and PrpS.
Successful sexual reproduction in many flowering plants involves a genetically controlled mechanism called self-incompatibility (SI). SI is controlled by a multi-allelic S-locus that encodes linked pollen and pistil S-determinants that allow self-recognition and rejection. When pollen of a certain haplotype lands on a stigma of the same haplotype, the pollen is recognized as ‘self’ or ‘incompatible’, and pollen tube growth, and thus fertilization, is inhibited. Pollen carrying S-alleles that are different from those of the stigma is able to germinate and fertilization is achieved. SI therefore acts to prevent inbreeding and maintain genetic diversity within a species. The three main SI systems that are well-characterized at the molecular level to date are: the Brassicaceae, the Papaveraceae and the S-RNase-type SI, which includes the Solanaceae, Plantaginaceae and Rosaceae. The reader is referred to other reviews for detailed information on the different mechanisms of SI [1–3]. Both the female and male S-determinants of the Brassicaceae, Solanaceae, Rosaceae and Plantaginaceae have been identified . The female S-determinant of poppy (Papaver rhoeas) (PrsS; P. rhoeas stigma S) was identified in 1994 . However, the male S-determinant PrpS (P. rhoeas pollen S) has only recently been identified and characterized . Nevertheless, owing to a robust in vitro bioassay that has been developed in our laboratory, whereby SI is induced in growing pollen tubes by addition of recombinant PrsS proteins, the Papaver SI system has been intensively studied at the molecular and cellular level. This makes it one of the most well-characterized SI systems with respect to mechanisms involved in inhibition of incompatible pollen tube growth. The present review highlights some of the major intracellular events that are triggered during Papaver SI and focuses on the recent identification of the long-sought-after pollen S-determinant, PrpS.
Downstream targets of Papaver SI
The earliest identified physiological event triggered by SI is the almost instantaneous and specific increase in the level of cytosolic free calcium ([Ca2+]i) in incompatible pollen tubes [6–8]. The elevated [Ca2+]i functions as a second messenger, initiating a complex set of events within incompatible pollen that result in the rapid inhibition of pollen tube growth and, ultimately, in the specific destruction of incompatible pollen by PCD (programmed cell death). In this section, we briefly review these SI-specific events, which are summarized in Figure 1.
A model for the targets of the Papaver SI signalling network
Within 90 s of SI induction, two sPPases (soluble inorganic pyrophosphatases), Pr-p26.1a and Pr-p26.1b, exhibit S-specific and Ca2+-dependent phosphorylation . sPPases are essential enzymes for cellular anabolism, catalysing the hydrolysis of PPi to Pi, providing a thermodynamic advantage for many biosynthetic reactions involved in regulating biosynthesis . It has been shown that both Ca2+ and phosphorylation reduce the pyrophosphatase activity of Pr-p26.1a/b and that inhibition of sPPase activity causes arrest of pollen tube growth , supporting the hypothesis that Pr-p26.1a/b are involved in polymer biosynthesis required for pollen tube growth.
Another rapidly modified target is the actin cytoskeleton. The actin cytoskeleton plays a critical role in regulating pollen tube growth  and in modulating signal–response coupling . SI induces rapid (within 1 min) alterations in F-actin (filamentous actin) organization, eventually leading to accumulation of F-actin in ‘punctate foci’ , as well as sustained depolymerization of actin filaments . An actin-binding protein, PrABP80 (P. rhoeas actin-binding protein 80), with properties of a gelsolin has been identified as a candidate to mediate SI-induced actin depolymerization because of its potent Ca2+-dependent severing activity . SI also targets the MT (microtubule) cytoskeleton, triggering very rapid depolymerization of cortical microtubules . Unlike actin, MTs do not reorganize into punctate foci. The use of specific drugs that alter the polymerization status of F-actin showed that actin depolymerization triggers depolymerization of cortical MTs. However, artificial depolymerization of MTs does not affect actin, suggesting that there is one-way ‘cross-talk’ from the actin to the tubulin cytoskeleton . An MAPK (mitogen-activated protein kinase), called p56, is activated in SI-induced pollen . MAPK cascades act as universal signal-transduction networks that connect many diverse signalling cascades , thus implicating MAPK signalling in Papaver SI. However, activation of the p56-MAPK peaked 10 min after SI-induced inhibition of pollen tube growth, suggesting a function in later events.
PCD is triggered by SI
Many of the earlier events triggered by SI contribute to the onset of PCD, the end-point of SI-induced events. Evidence for the involvement of PCD includes the detection of DNA fragmentation  and the activation of caspase-like activities  in incompatible pollen tubes. Caspases, a specific family of cysteine proteases, are essential for the execution of PCD in mammalian cells as they cleave specific target substrates, leading to cell death . Characterization of the caspase-like activities involved in Papaver SI identified a caspase-3like activity as the major caspase activity [21,23]; this is the predominant executioner caspase used in mammalian cell apoptosis. Localization studies showed that the DEVDase activity, characteristic of caspase 3, is localized first in the cytosol (1–2 h after SI) and later also in the nuclei. Measurement of the pollen tube cytosolic pH revealed that SI induces a dramatic acidification. The resulting decreased pH corresponds well with the acidic pH optima for the SI-induced caspase activities . Recently obtained data provide preliminary evidence that the vacuolar integrity is lost early in the SI response, which suggests that disintegration of the vacuole might lead to cytosolic acidification, allowing the activation of caspase-like activities (K.A. Wilkins, M. Bosch and V.E. Franklin-Tong, unpublished work).
Notably, major changes to the actin polymerization status can trigger caspase-like activities in pollen tubes, resulting in PCD , suggesting that the SI-mediated actin depolymerization participates in the activation of a PCD signalling network. Also, use of a highly selective inhibitor (U0126) of MAPK cascades, but not the inactive analogue U0124, to inhibit SI-induced p56 activation alleviated SI-induced caspase-3-like activity, providing evidence that the MAPK p56 participates in initiating the PCD signalling cascade .
In summary, SI induces an extensive and complex network of signalling events which can be divided into ‘phase 1’, contributing to the rapid inhibition of pollen tube growth, and ‘phase 2’, events associated with triggering and execution of PCD. Although we know much about the downstream events induced upon SI, it is essential to have information about the specific pollen–stigma interaction event, which is governed by the male and female S-determinants. Such information is crucial to build a full picture of the Papaver SI system and to resolve how it operates. In the next section, we review data relating to these two components.
The pistil S-determinant, PrsS
The female S-determinant of Papaver SI is a small (~15 kDa), soluble protein secreted by the papillae cells of the stigma [4,26,27]. These ‘S-proteins’ were recently renamed PrsS proteins (for P. rhoeas stigma S). The three PrsS alleles that have been fully cloned to date (PrsS1, PrsS3 and PrsS8) are highly polymorphic, with the predicted amino acid sequence only sharing between ~56 and 64% identity [26,27]. Despite the low level of identity between PrsS proteins, they are predicted to assume virtually identical secondary structures that consist of six β-strands with two α-helices at the C-terminal end; these are all linked by seven predicted hydrophilic surface loops [26,27]. There are four conserved cysteine residues that are thought to form disulfide bridges, which would maintain the PrsS protein's tertiary structure and, presumably, functional activity [26,28]. Site-directed mutagenesis revealed that both variable and conserved residues in surface loop 6 and also conserved residues in loop 2 of PrsS1 are essential for recognition of incompatible pollen [28,29]. At the time of their identification, the PrsS proteins from Papaver showed no significant homology with any previously reported sequence, suggesting that Papaver employed a different mechanism of SI from that of the Brassicaceae and Solanaceae . Subsequent studies, discussed above, have shown that Papaver SI is indeed distinct from the other SI systems studied to date. However, a whole family of plant-specific genes has been identified from the Arabidopsis genome which shares homology with the Papaver PrsS proteins . These proteins, called S-protein homologues (SPHs), do not play a role in SI, but it has been hypothesized that they act as small extracellular signalling molecules during other plant responses .
The pollen S-determinant, PrpS
The male S-determinant (PrpS) was recently identified in Papaver by sequence analysis of a cosmid clone containing the previously identified female PrsS gene . An open reading frame was discovered 457 bp downstream of PrsS1. Subsequently two further S-alleles, PrpS3 and PrpS8, were cloned using PCR. Their linkage to PrsS, the fact that their expression was limited to pollen and the considerable polymorphism between alleles (40–50%) suggested that the gene (PrpS) was a good candidate for the P. rhoeas pollen S-determinant.
Further analysis revealed that the PrpS genes encoded novel proteins of ~20 kDa. There are no domains within the protein that exhibit homology with well-characterized domains in other proteins; indeed, there is no convincing homology with any sequences in the databases in prokaryotic or eukaryotic organisms (including other plants). The PrpS proteins are, however, highly hydrophobic, and topology predictions suggest that PrpS proteins represent a class of Type I transmembrane proteins. Predictions using TMHMM  and Predict Protein  indicate between three and four transmembrane domains in each of the three proteins (Figure 2). Western blot analysis and immunolocalization showed that PrpS is present in the membrane fraction of pollen tube protein extracts and it was localized to the plasma membrane.
Possible structural topologies of PrpS1
All three PrpS proteins have a predicted extracellular domain of ~35 amino acids. This therefore represents a good candidate region of PrpS that could interact with PrsS proteins. To test this, competition experiments using peptides designed to the extracellular domain of PrpS were carried out using the in vitro bioassay described above. These experiments resulted in the effects of the PrsS proteins being titrated out and pollen tube growth being ‘rescued’, demonstrating that the extracellular domain is involved in PrpS–PrsS interaction and indicating PrpS functions in regulating pollen tube growth.
As P. rhoeas is not easily transformable, as-ODNs (antisense oligonucleotides) were used in combination with the in vitro bioassay to investigate whether PrpS was functionally involved in SI. The key results are summarized in Figure 3. Briefly, pollen from plants of S-haplotype S1S3 was successfully inhibited in the presence of recombinant PrsS1 and PrsS3, as expected. Addition of as-ODN designed for PrpS1 alleviated the SI response (~60% increase in length compared with SI-induced pollen), whereas addition of either the control s-ODN (sense oligonucleotide) designed to PrpS1 or as-ODN designed to PrpS8 failed to alleviate the response. Therefore knockdown of PrpS in the pollen resulted in alleviation of the SI response and, crucially, this was shown to be S-specific, demonstrating conclusively that PrpS is the male S-determinant of Papaver. Although PrpS has now been identified, as it is a novel protein, this leaves us with little idea as to its function. It is not a typical receptor or receptor-kinase like the Brassica S-determinant , so further work will need to address how PrpS functions in Papaver SI.
as-ODN experiments demonstrate that PrpS is functionally involved in determining S-specific pollen inhibition
A great deal of work has been carried out on the downstream events that are triggered in the pollen tube following the initiation of SI. This has allowed the Papaver SI system to be extremely well-characterized at the molecular level and has provided us with a good model system for studying signalling pathways involved in an important physiological response and for studying PCD in plant cells. The identification of PrpS as the long-sought-after male S-determinant creates the possibility of investigating the nature of the PrsS–PrpS interaction. We also still need to establish the nature of this unknown novel protein. Moreover, the basis for allelic specificity and crucial residues/domains involved in recognition still need to be established. This will probably require structural information on the proteins. We also have further work to establish other events triggered downstream of this interaction and how they are integrated.
Cell–Cell Communication in Plant Reproduction: A Biochemical Society Focused Meeting held at University of Bath, Bath, U.K., 14–16 September 2009. Organized and Edited by James Doughty (Bath, U.K.) and Rod Scott (Bath, U.K.).
mitogen-activated protein kinase
programmed cell death
Papaver rhoeas actin-binding protein 80
Papaver rhoeas pollen S
Papaver rhoeas stigma S
soluble inorganic pyrophosphatase
Work in the laboratory of V.E.F-T. is funded by the Biotechnology and Biological Sciences Research Council.
Present address: Warwick HRI, Wellesbourne, Warwick CV35 9EF, U.K.
Present address: Institute of Biological, Environmental and Rural Sciences (IBERS), Aberystwyth University, Plas Gogerddan, Aberystwyth SY23 3EB, U.K.