The flowering plant Solanum chacoense uses an S-RNase-based self-incompatibility system in order to reject pollen that shares the same genes at the S-locus (S-haplotype) with the style (an incompatible reaction). Two different models have been advanced to explain how compatible pollen tubes are protected from the cytotoxic effects of the S-RNase, sequestration of the S-RNase in a vacuolar compartment or degradation of the S-RNase in the cytoplasm. Here, we examine the subcellular distribution of an S11-RNase 18 and 24 h post pollination (hpp) in compatible and incompatible crosses by immunogold labeling and transmission electron microscopy. We find that the S-RNase is present in the cytoplasm of both compatible and incompatible crosses by 18 hpp, but that almost all the cytoplasmic S-RNase is degraded by 24 hpp in compatible crosses. These results provide compelling evidence that S-RNases are degraded in compatible but not in incompatible pollen tubes.
Self-incompatibility is the most important and widespread genetic mechanism used by flowering plants to promote outcrossing and thus avoid the deleterious effects of inbreeding. Self-incompatibility operates to block the formation of zygotes after self-pollination or crosses between genetically related individuals (de Nettancourt, 1977). The phenomenon involves interactions between gene products expressed in the pollen, and those expressed in specialized cells of the pistil. Present in ∼60% of all angiosperm species, self-incompatibility is assumed to have played a key role in the successful evolution of the angiosperm, as assessed by current phylogeny studies (Allen and Hiscock, 2008). In the vast majority of species characterized by self-incompatibility, the interactions between pollen and pistil are under the control of elements of a single, highly polymorphic locus, the S-locus (de Nettancourt, 2001), that encodes for both the male and female determinants to self-incompatibility.
In Solanaceae, Rosaceae and Plantaginaceae, the breeding behavior of the pollen is determined by the S-allele constitution of each individual pollen grain. Pollen rejection occurs when the S-haplotype of the pollen grain matches any of the S-alleles expressed in the pistillar tissues. In these families the pistillar gene product that mediates self-incompatibility is an extremely polymorphic ribonuclease termed an S-RNase (McClure et al., 1989). S-RNases are synthesized by the specialized cells of the transmitting tissue, secreted into the surrounding extracellular matrix (ECM) where the pollen tubes grow, and taken up from the ECM by the pollen tubes. Rejection of incompatible pollen thus occurs in the style.
Electron microscopy studies have shown differences after compatible and incompatible crosses in the pistil ultrastructure (Herrero and Dickinson, 1979), as well as in pollen tubes during either intraspecific (de Nettancourt et al., 1973a; Herrero and Dickinson, 1980; Herrero and Dickinson, 1981) or interspecific crosses (de Nettancourt et al., 1974; de Nettancourt et al., 1973b), whereas immunolocalization studies have resolved the longstanding debate concerning whether S-RNases entered into the pollen tubes in a haplotype-specific manner or whether they were able to enter both compatible and incompatible pollen tubes. Using S-RNase-specific antibodies, Grey et al. (Gray et al., 1991) were the first to show that S-RNases enter Nicotiana alata pollen tubes in an S-haplotype-independent manner. However, these studies were performed on pollen tubes grown in vitro, and an in vivo confirmation was not provided until almost ten years later (Luu et al., 2000). These authors used immunogold electron microscopy to demonstrate accumulation of S-RNases in both compatible and incompatible pollen tubes of Solanum chacoense, although in that study no distinction was made between different subcellular compartments and only a single time post-pollination was used. The in vivo results were later extended to N. alata using confocal microscopy (Goldraij et al., 2006). Interestingly, observations made in this latter species indicated that S-RNases remain sequestered in a vacuole space after their entry inside compatible pollen tubes (see McClure et al., 2011), and were only released into the cytoplasm during incompatible crosses. These observations led to development of the sequestration model for S-RNase-based gametophytic self-incompatibility, in contrast to the degradation model derived from results using Petunia hybrida, which proposes degradation of presumably cytoplasmic S-RNases in compatible pollen tubes (see Meng et al., 2011).
Intriguingly, previous studies on the self-incompatibility reaction of S. chacoense have furnished support for both sequestration and degradation models. On the one hand, measurements of the amount of RNA and protein in pollen tubes as a function of time during compatible and incompatible crosses appeared to support the sequestration model (Liu et al., 2012). This comes from studies using transgenic pollen expressing GFP, where compatible and incompatible crosses only developed a marked difference in the amounts of both GFP mRNA and protein between 18 and 24 h post pollination (hpp), a result suggestive of a sudden release of S-RNase into the cytoplasm at this crucial time. Furthermore, when the amount of S-RNase inside pollen tubes was determined by immunogold electron microscopy at 18 hpp, both compatible and incompatible pollen tubes were found to have accumulated S-RNases to a similar extent (Luu et al., 2000), thus arguing against the degradation model. On the other hand, when the total S-RNase load in the styles was measured at 24 hpp, compatible crosses had ∼30% less S-RNase than did incompatible crosses (Liu et al., 2009), thus supporting the degradation model. In order to reconcile these apparently contradictory results, we have used immunogold electron microscopy to determine the amount and subcellular localization of the S11-RNase in both compatible and incompatible crosses at the crucial time interval between 18 and 24 hpp. The observations show the presence of S-RNase in the cytoplasm of both compatible and incompatible pollen tubes at 18 hpp and the disappearance of S-RNase from the cytoplasm of compatible pollen tubes by 24 hpp. These results provide convincing support for the degradation model.
RESULTS AND DISCUSSION
Similar to other Solanaceae such as Petunia (Herrero and Dickinson, 1979), S. chacoense has a compact style with a well-defined transmitting tissue occupying its central region that is formed by solid columns of cells communicating to each other by numerous plasmodesmata and surrounded by intercellular space through which the pollen tubes will grow towards the ovary. Transverse sections of pollinated styles observed by electron microscopy, revealed that the cells of the transmitting tissue were typically round, larger than 5 µm in diameter and are surrounded by electron-opaque intercellular spaces forming the ECM (Fig. 1A). At the electron microscopy level, these cells showed a complex pattern of darker and lighter areas corresponding to cytoplasm and membrane-bound compartments, respectively. In contrast, the pollen tubes were composed of irregularly shaped cells, which were smaller than the cells of the transmitting tissue and possessed a complex wall with numerous in-foldings (Fig. 1A). Most regions of the pollen tube cell wall are rich in callose, unlike the cell wall of any of the stylar cells, and this molecular correlate can thus be used to unambiguously identify pollen tubes immunocytochemically (Meikle et al., 1991). The use of a mouse anti-callose and a rabbit anti-S11-RNase antibody allows simultaneous labeling of sections and clearly identifies the S-RNase (Fig. 1B, 20-nm gold particle labeling) inside pollen tubes surrounded by callose (Fig. 1B, 10-nm gold particle labeling). S-RNase labeling was also seen over the ECM and the vacuolar space of the transmitting tissue cells, consistent with secretion of the S-RNase into the ECM by these cells. In contrast, the labeling observed over pollen tubes could encompass both cytoplasmic and vacuolar compartments.
Once they enter the transmitting tissue, ∼6 hpp, pollen tubes start to grow at different rates depending on whether they are compatible or incompatible with the style (Liu et al., 2012). Thus, the region of the style containing the apical tip of the pollen tubes was determined for all crosses by Aniline Blue staining to ensure that the stylar regions chosen would contain pollen tube tips for the two times selected (18 and 24 hpp) and the two types of crosses (compatible and incompatible). An examination of pollen tubes after both compatible (Fig. 2) and incompatible (Fig. 3) crosses revealed that they contained both cytoplasmic and vacuolar regions. This indicates that the stylar areas selected for all samples do indeed contain the growing apical tip, as proximal regions of the pollen tube (behind the tip) have collapsed, do not contain cytoplasm and are separated from the growing region by a callose plug (Steer and Steer, 1989).
We used measurements of S-RNase label density (gold particles per µm2) as a surrogate for S-RNase concentration in our sections. When regions representing the transmitting tissue cells, the ECM and the pollen tubes (Fig. 4A) are compared, the average label density in the stylar cells or in the ECM did not differ significantly between any of the four conditions. However, label density over the pollen tubes differed markedly among some of the different conditions. When the label density at 18 hpp for compatible crosses was compared to crosses observed at 24 hpp, S-RNase was found to be significantly lower in compatible pollen and significantly higher for incompatible pollen. Incompatible pollen tubes observed at 18 hpp were not significantly different from compatible pollen tubes at this same time, principally owing to the large variability in label density observed for the compatible pollen tubes 18 hpp. Curiously, incompatible pollen tubes 24 hpp often lacked callose staining, as illustrated (Fig. 3B), but are still recognizable as pollen tubes by morphology and substantial S-RNase accumulation.
In order to define more precisely the fate of S-RNase inside the compatible and incompatible pollen tubes, S-RNase label density was assessed for both cytoplasmic and vacuolar regions within the pollen tubes (Fig. 4B). We first note that there was a generally higher range of label densities observed for different pollen tubes in compatible crosses at 18 hpp (n = 12) for both the cytoplasm and vacuolar space. No significant differences in label density over either of these compartments were observed when compatible and incompatible pollen tubes were compared at 18 hpp. However, by 24 hpp the label density over compatible and incompatible pollen tubes was strikingly different. In compatible pollen tubes, the label density over the vacuolar space had decreased significantly by 24 hpp, and no labeling was observed over background for the cytoplasm. In contrast, there was a significant increase in label density in the cytoplasm of incompatible pollen tubes by 24 hpp. The data shown in Fig. 4 has not been corrected for the background S-RNase antibody labeling of 1.2±1 (mean±s.d.) gold particles per µm2 observed over regions of the sections lacking tissue or over regions of the epidermal cells (which do not synthesize S-RNase).
A previous study using immunogold labeling to assess S-RNase entry into compatible and incompatible pollen tubes showed that S-RNase label density inside both types of pollen tubes 18 hpp is roughly 5-fold greater than that found in the surrounding ECM (Luu et al., 2000). This level of accumulation is similar to that found for the total pollen label in the 18 hpp crosses reported here (∼3- to 4-fold for compatible and incompatible pollen tubes, respectively). Importantly, the present observations confirm the lack of a significant difference in pollen tube labeling between the compatible and incompatible crosses at 18 hpp. However, this similarity in label density at 18 hpp contrasts sharply with what was observed at 24 hpp, where compatible pollen tubes displayed markedly lower label densities than incompatible ones. This decrease in S-RNase immunogold labeling in compatible pollen tubes at 24 hpp is consistent with the marked S-RNase degradation previously observed by western blot measurements of whole styles, where total S-RNase extracted from pollinated styles has been shown to decrease by 30% in compatible compared to incompatible crosses (Liu et al., 2009). Thus, the experiments performed here resolve the disparity between the western blot and the previous immunolocalization experiments, and ascribe the differences between the two to a remarkable change in pollen tube behavior occurring between 18 and 24 hpp.
A striking difference in the behavior of compatible and incompatible pollen tubes has also been observed to occur between 18 and 24 hpp with respect to levels of both pollen tube RNA and protein. It is during this time interval that the RNA and protein levels in compatible pollen tubes were reported to increase markedly compared to incompatible pollen tubes (Liu et al., 2012). The results described here thus suggest that this increase in pollen tube RNA might be due to the disappearance of the S-RNase from the cytoplasm. If so, this might indicate that RNA is undergoing constant degradation in both compatible and incompatible pollen tubes up until 18 hpp. Although this would certainly explain the similarity in levels of pollen tube RNA in both types of crosses prior to this time, to date there is no information about levels of RNA turnover in growing pollen tubes. There is also no indication of why S-RNase degradation should not occur until 18 hpp. Another possibility is that in S. chacoense the S-RNase remains sequestered until about 18 hpp, and that degradation occurs only when the S-RNase is released to the cytoplasm of compatible pollen tubes. This would be consistent with the view that compartmentalization and degradation are not mutually exclusive processes (McClure et al., 2011) but might in fact work to provide a ‘layered’ defense against the S-RNases.
One peculiar aspect of pollen tube behavior, namely the prominent increase in the growth rate of compatible pollen tubes after 6 hpp (Liu et al., 2012), cannot yet be understood in terms of S-RNase levels. With the S-RNase and S-locus F box (SLF) proteins being the only known components defining the haplotype-specific self-incompatibility reaction, it seems likely that in order to engender such a haplotype-specific response, these two components must come into contact shortly after 6 hpp. Logically then, the presence of S-RNase in the cytoplasm by this time, should serve as the signal that distinguishes compatible from incompatible pollen tubes. However, this reintroduces the conceptual difficulty of understanding why S-RNase–SLF interaction would not immediately result in S-RNase degradation. Clearly this is not the case, as cytoplasmic S-RNases are indeed observed at 18 hpp. It would be interesting in future work to follow pollen tube S-RNase from times as early as 6 hpp.
Despite the proposal that degradation and sequestration are not mutually exclusive processes (McClure et al., 2011), our findings provide more support for degradation that they do for sequestration. First, we directly infer degradation in compatible crosses from the decrease in S-RNase labeling that occurs between 18 and 24 hpp. Second, the sequestration of S-RNase predicted for compatible crosses does not result in any significant difference in the amount of S-RNase in the cytoplasm of compatible and incompatible pollen tubes at 18 hpp. Finally, sequestration would be expected to retain S-RNase within the vacuolar compartment of compatible crosses, yet we observe a significant decrease in S-RNase levels in the compartment between 18 and 24 hpp. Taken together, these observations suggest that the mechanism of the self-incompatibility system in S. chacoense is different from that in Nicotiana alata, where sequestration of the S-RNase in the vacuolar compartment of compatible pollen tubes appears sufficient to account for their self-compatibility phenotype (Goldraij et al., 2006).
MATERIALS AND METHODS
Material and sample preparation
Styles from Solanum chacoense line V22 [with S haplotype S11S13 (Qin et al., 2001)] were crossed with freshly collected pollen from transgenic lines tGFP2548 (S12S12) or tGFP1022 (S11S13) (Liu et al., 2009) expressing GFP under the control of the pollen-specific Lat52 promoter (kindly provided by Sheila McCormick, University of California Berkley, CA).
Pollinated styles were collected at either 18 or 24 hpp and placed in 0.1 M PBS (pH 7.4). For compatible crosses (V22×tGFP2548), styles were cut at either one-half (18 hpp) or five-eighths (24 hpp) along the total style length and placed in PBS containing 4% freshly made paraformaldehyde for 2 h under a vacuum. For incompatible crosses (V22×tGFP1022), styles were cut at either one-quarter (18 hpp) or one-third (24 hpp) along the total style length and fixed as for compatible crosses. These distances correspond to the position of the tips of most of the pollen tubes in the styles as determined using Aniline Blue staining of several styles pollinated in parallel. All samples were dehydrated through an ethanol series [1 h each of 30%, 50%, 70%, 95% and 100% (v/v) ethanol in water] then transferred through an LR White dilution series (1 h 25%, 2 h 50%, 1 h 75%, 1 h 100% and 24 h 100% LR White in ethanol). Samples were polymerized for 24 h at 60°C in fresh 100% LR White. Sections (70 nm) were cut using a Leica Ultracut and placed on microscope grids treated with a solution of 0.5 g Formavar in 100 ml chloroform.
Immunolabeling and microscopy
Sections were rehydrated through an ethanol dilution series [1 min each of 100%, 95%, 70%, 50%, 30% and 0% (v/v) ethanol in water] then blocked for 15 min in PBS containing 3% BSA and 0.1% Tween 20. Blocked sections were incubated with either a rabbit anti-S11-RNase prepared as described previously (Luu et al., 2000) or a commercial mouse anti-callose antibody (Meikle et al., 1991; catalog no. 400-2, Biosupplies Australia Pty, Parkville, Australia) for 24 h at 4°C, washed three times for 1 min with blocking buffer, and then incubated with either a goat anti-rabbit-IgG labeled with 20-nm gold particles, or a goat anti-mouse-IgG labeled with 10-nm gold particles (Ted Pella) for 2 h at room temperature. After incubation with the secondary antibody, the sections were washed three times 1 min in blocking buffer then incubated for 5 min in distilled water to remove salts. Grids were allowed to dry in air. All antibodies were used at a dilution of 1:100 in blocking buffer. All samples were observed using a JEOL JEM-1010 operating at 80 kV. Images were analyzed using ImageJ software, to determine surface areas and number of gold particles.
We thank Louise Pelletier (Biology department, University of Montreal, Canada) for excellent assistance with the transmission electron microscopy, and Jeffrey Labovitz (Bolinas, CA) and Firas Bou Daher (Sainsbury Laboratory, University of Cambridge, UK) for constructive comments on the manuscript.
The study was conceived and designed by M.C. and D.M. Experiments were performed by N.B. and analyzed by M.C. and D.M. who also wrote the paper with the help of N.B.
The work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) to M.C.
The authors declare no competing interests.