Self-incompatibility in Papaver

Self-fertilization, which results in reduced fitness of offspring, is a common problem in hermaphrodite angiosperms. To prevent this, many plants utilize self-incompatibility (SI), which is determined by the multi-allelic S locus, that allows discrimination between self (incompatible) and non-self (compatible) pollen by the pistil. In poppy (Papaver rhoeas), the pistil S-determinant (PrsS) is a small secreted protein which interacts with the pollen Sdeterminant PrpS, a ~20 kDa novel transmembrane protein. Interaction of matching pollen and pistil S-determinants results in self-recognition, initiating a Ca-dependent signalling network in incompatible pollen. This triggers several downstream events including alterations to the cytoskeleton, phosphorylation of soluble inorganic pyrophosphatases (sPPases) and a Mitogen Activated Protein Kinase (MAPK), increases in ROS and NO, and activation of several caspase-like activities. This results in the inhibition of pollen tube growth, prevention of self-fertilization and ultimately programmed cell death (PCD) in incompatible pollen. This review will focus on our current understanding of the integration of these signals with their targets in the SI-PCD network. We also discuss our recent functional expression of PrpS in Arabidopsis thaliana pollen. Introduction Successful sexual reproduction in many flowering plants involves a genetically controlled mechanism called self-incompatibility (SI). SI is genetically controlled by a multi-allelic Slocus that encodes linked pollen and pistil S-determinants that provide self-recognition. Interactions between pollen and pistil of the same haplotype initiate an SI response, which causes inhibition of “self” pollen tube growth, so fertilization is inhibited. Pollen carrying Sdeterminants that are different to those expressed on the surface of the recipient stigma are able to germinate, pollen tube growth proceeds and fertilization is achieved. The SI mechanism therefore acts to prevent inbreeding and increase genetic diversity. The SI systems that are well characterized at the molecular level to date are: the Brassicaceae, the Papaveraceae and the S-RNase-type which includes the Solanaceae, Plantaginaceae and Rosaceae. The reader is referred to other reviews for detailed information on the mechanisms of Brassicaceae [1] and S-RNase-type SI [2]. Here we focus on recent advances in our understanding of the Papaveraceae SI system in poppy (Papaver rhoeas). The two S-determinants of the poppy S-locus have been well characterized. PrpS encodes a novel transmembrane ~20 kDa protein that is specifically expressed in pollen [3], while PrsS is a ~14 kDa protein secreted to the pistil surface [4]; see Poulter et al, 2010 for a recent review [5]. A robust in vitro SI bioassay that uses recombinant PrsS proteins to induce the Papaver SI response in germinating pollen has been developed in our laboratory, and this has enabled investigations at the molecular and cellular level, making it one of the best characterized SI systems at the mechanistic level. In brief, upon the cognate interaction of PrsS and PrpS, Ca influx is triggered, which increases intracellular levels of Ca ([Ca]i), and induces a signalling cascade, resulting in the pollen tube growth inhibition and ultimately, PCD of incompatible pollen. Ca-dependent phosphorylation of sPPases has been observed in the initiation stage of this SI response. Both Ca and phosphorylation reduce their activities, causing arrest of pollen tube growth [6, 7]. Another rapidly modified target is the actin cytoskeleton. SI induces rapid (within 1 min) alterations in F-actin organization, and a sustained depolymerization of actin filaments [8], followed by an accumulation of F-actin into “punctate foci” [9]. Between 90 s and 10 min post-SI induction, a MAPK, p56, is activated [10]. MAPKs are known to play a key role in mediating signals to PCD in the plantpathogen hypersensitive response and also have been described as universal signal transduction networks that connect many diverse signalling cascades [11], suggesting a potential role of the p56 MAPK in Papaver SI signalling pathway. Studies based primarily on the use of a MAPK cascade inhibitor in combination with PCD markers, implicated an important role for MAPKs in signalling to PCD in incompatible pollen [12]. Moreover SI resulted in a cytosolic acidification, providing the optimal pH for the SI-induced caspase-like activities detected in the pollen tube during SI [13]. The ultimate mechanism of poppy selfpollen rejection is through PCD [14]; evidence in support of this includes cytochrome c leakage, caspase-like activities and DNA fragmentation in incompatible pollen [13-15]. In recent years, further components have been identified as being involved in the SI signalling network, contributing to the pollen tube growth inhibition and PCD. In this review, we will highlight some of the major intracellular events that are triggered during a Papaver SI response and focus on recent findings that contribute to our understanding of the integration of the signals and targets of this intricate SI-initiated PCD pathway. We will also describe the functional transferral of poppy SI system into selfcompatible Arabidopsis thaliana and its implications. A model, based on our current understanding of the SI signalling cascade, is provided. SI triggers membrane channel activation permeable for divalent and monovalent cations The earliest identified physiological event triggered by SI recognition is the almost instantaneous increase in [Ca]i in incompatible pollen tubes [16-18]. These studies formed the basis of a hypothesis, proposing that a receptor-ligand type of interaction, involving PrsS and PrpS of the same haplotype, triggers a Ca dependent signalling cascade within incompatible pollen. More recently, studies using an electrophysiological approach, have demonstrated evidence not only for Ca influx but also a large influx of K in incompatible pollen [19] (Figure 1). Currents were identified by means of whole-cell patch clamping of pollen protoplasts; plasma membrane ion channel activation was measured during an S-allelespecific SI response in pollen protoplasts exposed to either Ca or K. For the divalent cation Ca, after exposure to recombinant incompatible PrsS protein, the current-voltage (I-V) relationship revealed a large increase in the current (Figure 1). A similar result was obtained for the monovalent cation K, where I-V relationship showed that SI stimulates a high conductance for K, with greater amplitude than that for Ca (Figure 1). These studies suggest that the SI-induced conductance(s) is ligand-gated and may involve a Non-Specific Cation Channel (NSCC). Other studies have shown that the elevated [Ca]i functions as a second messenger [16-18], 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. Involvement of cytoskeleton in SI-mediated PCD Integrity of the actin cytoskeleton is crucial for pollen tube growth as it is important for both regulating growth and modulating signal-response coupling [20, 21]. SI induction triggers several striking alterations to the actin cytoskeleton. The typical longitudinal F-actin bundles in normally growing pollen tubes (Figure 2a) are rapidly depolymerized in SI-induced pollen (Figure 2b), and subsequently, the F-actin aggregates into “punctate foci” (Figure 2c), which increase in size over time [8, 9]. An actin-binding protein, PrABP80, with properties of a gelsolin was identified as a candidate to mediate SI-induced actin depolymerization due to its potent Ca-dependent severing activity, probably together with profilin [22]. Immunolocalization studies have shown that the punctate F-actin foci are associated with the actin-binding proteins Actin Depolymerizing Factor (ADF/cofilin; Figure 2d-e) and Cyclase Associated Protein (CAP) [23]. Although the name ADF suggests its filament severing activity under normal cellular conditions, it has been reported that ADFs exhibit pH-sensitive activity, with actin depolymerization at an alkaline pH and binding to and stabilizing F-actin at an acidic pH [24]. Thus, in the context of SI, with an SI-induced acidification [13], ADF is likely to be playing a role in the formation and stabilization of the SI-induced actin foci. SI also targets the microtubule cytoskeleton, triggering very rapid depolymerisation of cortical microtubules [25]. However, unlike actin, the microtubules do not reorganize into punctate foci. The use of specific drugs that alter the polymerization status of F-actin showed that actin depolymerization triggered the depolymerization of cortical microtubules. However, artificial depolymerization of microtubules did not affect actin, suggesting that there is one-way “cross-talk” from the actin to the tubulin cytoskeleton [25]. Both actin and microtubule alterations are implicated in mediating PCD. Several studies have shown that either actin depolymerization or stabilization can influence whether a eukaryotic cell goes through an apoptotic pathway [26, 27]. Investigations examining the effect of the actindepolymerizing drug, latrunculin B and the actin-stabilizing/polymerizing drug, jasplakinolide in Papaver pollen revealed that both treatments stimulated high levels of DNA fragmentation which was mediated by a caspase-3 like/DEVDase activity [28]. This demonstrated that disturbance of actin polymer dynamics can trigger PCD in pollen. These data implicated both actin depolymerization and stabilization being functionally involved in the initiation of SI-induced PCD, and established a causal link between actin polymerization status and initiation of PCD. Thus, the rapid and substantial actin depolymerization triggered by SI signalling not only results in the rapid inhibition of incompatible pollen tip growth, but it also activates a caspase-3-like/DEVDase activity, triggering PCD. Therefore the formation of the F-actin foci and their association with ADF, appears to be an active process that is also involved in signalling towards PCD; they are also a reliable marker for SI-induced PCD in poppy pollen, and their formation is alleviated concomitantly with alleviation of PCD [29]. A role for ROS and NO signalling in SI-mediated PCD In plants, ROS and NO have been shown to play a key role in a variety of cellular responses including PCD [30, 31]. Live cell imaging has recently been used to visualize ROS and NO in growing Papaver pollen tubes during SI. Around 5 min after SI-induction, rapid and transient increases in cytosolic ROS was observed in incompatible pollen tubes, and later, transient increases in NO [29]. Preventing ROS and NO increases in pollen tubes prior to SIinduction, using a pre-treatment with both the NADPH oxidase inhibitor, DPI (diphenyleneiodonium) and the NO scavenger cPTIO (2-(4-Carboxyphenyl)-4,4,5,5tetramethylimidazoline-1-oxyl-3-oxide) substantially decreased the caspase-3-like/DEVDase activity (Figure 3a), which is a key feature of the poppy SI-PCD response. Neither of these scavengers alone could significantly reduce the amount of caspase-3-like/DEVDase activity (Figure 3a). This suggests that ROS and NO act either in concert or tandem to signal towards SI-induced PCD through activation of DEVDase activity. Evidence also suggested that ROS and NO increases during SI also contribute to increased numbers of pollen tubes containing punctate actin foci [29]. Figure 3b shows that untreated pollen tubes had negligible actin foci, SI induced pollen tubes had high levels, and when pollen was pre-treated with the ROS scavenger TEMPOL (4-Hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl) and/or c-PTIO prior to SI-induction, the number of actin foci was significantly reduced in all three cases [29]. As ROS and NO scavengers can alleviate the formation of punctate actin foci and also prevent the activation of caspase-3-like/DEVDase activity, this suggests that those ROS and NO signals act upstream of these SI markers. Recruitment of signalling for SI events in other distantly related species Recent studies have begun to explore the possibility for functional transfer of the Papaver SI determinants into other species. The Papaver rhoeas pollen S-determinant, PrpS, was expressed as a GFP-fusion protein and demonstrated to be functional in self-compatible A. thaliana pollen [32]. When A. thaliana pollen expressing PrpS-GFP was treated with recombinant PrsS protein, pollen tube growth was inhibited and its viability decreased in an S-specific manner, demonstrating functional “SI” in A. thaliana (Figure 4a). This S-specific decrease in pollen viability could be rescued by pre-treatment with the caspase-3 inhibitor, Ac-DEVD-CHO (Figure 4a), demonstrating the involvement of PCD [32]. Recombinant PrsS protein treatment also stimulated S-specific actin foci formation in At-PrpS-GFP pollen in an S-specific manner (Figure 4b-c), demonstrating that a similar signalling cascade to that in Papaver SI response was triggered in the transgenic A. thaliana pollen. Successful SI trait transferral between species was first obtained nearly a decade ago by the Nasrallah laboratory [33, 34]. They demonstrated that S-haplotype pairs from A. lyrata and Capsella grandiflora (which are diverged from A. thaliana around ~5 [35], and ~6.2–9.8 million years ago (MYA) [36], respectively), confer self-incompatibility in A. thaliana. However, as these are closely related species that share a mechanistically common SI ancestor, this does not provide major insights into the evolution of SI signalling across angiosperm families. Our findings provide a breakthrough in this area, as functional transferral of the Papaver pollen S-determinant into A. thaliana is obtained between highly diverged species with an evolutionary distance ~144 MYA [37], and which has no orthologues of the Papaver S-determinants. This represents the first demonstration that an SI system can be functionally transferred into a very distantly related species with a different ancestral SI system. Our data indicate that the Papaver SI system works in A. thaliana due to the ability to execute and co-ordinate more than one endogenous component at the same time and can act in signalling networks that they do not normally operate in. The Papaver SI signalling networks (e.g. Ca) and targets (e.g. the actin cytoskeleton) appear to be universal and may be present in most of the angiosperms from the very early times. If these “common” cellular elements can be recruited to operate under the control of a newly introduced system (as we have shown with introduction of PrpS), it appears that a novel and functional signalling pathway can be initiated that results in a specific, predictable physiological outcome (PCD in this case). This could be similar to situations where gene redundancy and plasticity operate. MAPK cascade components are classic examples of signalling components that can participate in more than one signalling network in certain situations. For example, in Saccharomyces cerevisiae (yeast), many components can play a role in more than one pathway; (see [38-40] for further discussion of this phenomenon). Examples of dual functioning or ‘multi-tasking’ have been cited in the context of innate immune signalling pathways [41]. We think this is a likely explanation of why PrpS functions in A. thaliana pollen. Our data provides good evidence that A. thaliana possesses proteins that can be recruited to form a new signalling network for SI response which does not normally operate in A. thaliana. This also suggests that the Papaver SI system uses ubiquitous cellular targets (such as the cytoskeleton and sPPases), and that endogenous signalling components (such as Ca ROS, NO, MAPKs) in the host species can somehow be recruited to elicit an appropriate “SI” response in a species that has most likely never had this type of SI system. This has potentially important implications. Crop F1 hybrid seed production currently relies on artificial emasculations, which are timeand labour-consuming. Transferral of Sdeterminants from poppy into crop plants to make them self-incompatible provides a novel and efficient way to obtain hybrid seeds naturally. Our findings also are of interest from a scientific viewpoint. Studies of the evolution of self-/non-selfrecognition systems have generally focused on the receptor-ligand recognition, rather than the downstream signalling networks triggered by their interaction. Our findings suggest either conservation of an ancient signalling system or recruitment of signalling components to mediate the downstream responses for SI recognition. Our data suggest that the postulated parallels between SI and plant-pathogen resistance, with the idea that SI may utilize some of these signalling networks, may not be as unlikely as it initially seems. Summary and future perspectives Research on the Papaver SI system has shown that an S-specific interaction between the male and female S-determinants triggers an SI signalling network which integrates several cellular components in incompatible pollen (see Figure 5). These signalling events can be classified into two categories. First, signal initiation events which include Ca and K influx, inhibition of sPPase activity, depolymerization of F-actin and inhibition of incompatible pollen tube growth. Second, later suicide signalling events, are involved in commitment to PCD, through the activation of caspase-3-like/DEVDase activities. These include activation of the p56 MAPK, increases in ROS, and the formation of F-actin foci. Collectively, these signal to the “gateway” which incompatible pollen must pass to become irreversibly inhibited, by setting up self pollen “suicide”. Once cytosolic acidification and caspase activation occur, the incompatible pollen is programmed to die and disassemble. Together, these mechanisms coordinate the prevention of self-fertilization. Although key components have been identified in the signalling network of the Papaver SI system, many unanswered questions remain, and further efforts are required to fully elucidate the mechanisms involved in the poppy SI response. Moreover, the demonstration of wide transgenera functionality of the Papaver SI system in A. thaliana, suggests that the transfer of the Papaver SI system to unrelated crop species, is potentially feasible. Transfer of this SI system could be used as a tractable SI system and heralds the possibility, at least in principle, of using this as an application to increase efficiency in the production of F1 hybrids in crop plants. Acknowledgements Work in the lab of V.E.F-T. is funded by the Biotechnology and Biological Sciences Research Council (B.B.S.R.C.). TH is supported by a Commonwealth Scholarship; CF is supported by a CONICYT; ZL and NT are supported by China Scholarship Council.