Pear (Pyrus pyrifolia L.) has a S-RNase-based gametophytic self-incompatibility (SI) mechanism, and S-RNase has also been implicated in the rejection of self-pollen and genetically identical pollen. No studies, however, have examined the extent of organelle alterations during the SI response in Pyrus pyrifolia. Consequently, this study focused on the alterations to mitochondria and nuclear DNA in incompatible pollen tubes of the pear. Methylthiazolyldiphenyl-tetrazolium bromide was used to evaluate the viability of pollen tubes under S-RNase challenge. The results showed that the viability of the control and compatible pollen tubes decreased slightly, but that of the incompatible pollen and pollen tubes began to decline at 30 min. The mitochondrial membrane potential (Δψmit) was also tested with rhodamine 123 30 min after SI challenge, and was shown to have collapsed in the incompatible pollen tubes after exposure to S-RNase. Western blotting 2 h after SI challenge confirmed that the Δψmit collapse induced leakage of cytochrome c into the cytosol. Swollen mitochondria were detected by transmission electron microscopy as early as 1 h after SI challenge and the degradation of nuclear DNA was observed by both 4,6-diamidino-2-phenylindole and transferase-mediated dUTP nick-end labeling. These diagnostic features of programmed cell death (PCD) suggested that PCD may specifically occur in incompatible pollen tubes.
Pear (Pyrus pyrifolia L.), belonging to the family Rosaceae, has a S-RNase-based GSI system. Ten stigmatic S-RNase alleles (S1–S10) have been identified in Japanese pear cultivars (Kim et al., 2006; Takasaki et al., 2004). The SI system of the pear has been investigated by our team for many years. During this process we have succeeded in isolating S-RNase from the style, and through an in vitro system have identified the characteristics of S-RNase that specifically inhibit self-pollen germination and tube elongation (Zhang and Hiratsuka, 1999, 2000a, b; Hiratsuka et al., 2001). We have developed a protocol for the successful isolation of spheroplasts from the pollen tube of the pear, identified a hyperpolarization-activated cation channel using the patch-clamp technique (Qu et al., 2007) and discovered the relationship between the Ca2+-permeable channels of the pollen tube and self-S-RNase activity (Xu et al., 2007). Recently, it has also been confirmed that S-RNase induces depolymerization of the actin cytoskeleton of self-generated pollen tubes (Liu et al., 2007). Several researchers have reported that either the stabilization or the depolymerization of the actin cytoskeleton is adequate to induce programmed cell death (PCD) in yeast and some animal cells (Janmey, 1998; Morley et al., 2003). Supporting this, Thomas et al. (2006) found that actin depolymerization was sufficient to induce PCD in the SI pollen of Papaver rhoeas. Therefore, we questioned whether PCD is also triggered by S-RNase in incompatible pollen tubes in Pyrus.
Programmed cell death is widely used by many organisms to eliminate unwanted cells in a precisely regulated manner. The process has key diagnostic features, including the leakage of cytochrome c from the mitochondria into the cytosol, mitochondrial tumefaction and nuclear DNA fragmentation (Nagata et al., 2003; Raff, 1998). In caspase-dependent death pathways, the collapse of the membrane potential (Δψmit) induces leakage of cytochrome c into the cytosol, which activates the caspase family of proteinases in an autoprocessing cascade, including caspase-3 which is instrumental in the cleavage of nuclear DNA (Yang et al., 1997; Crompton, 1999; Salvesen and Dixit, 1999; Wolf and Green, 1999; Wolf et al., 1999). It is generally accepted that apoptosis does not occur in plant cells, but PCD in plants does demonstrate some apoptotic features (Thomas and Franklin-Tong, 2004).
Effects of S-RNase on pollen grain and pollen tube viability
The pollen grains and pollen tubes were stained with methylthiazolyldiphenyl-tetrazolium bromide (MTT) to assess their viability in a test which detects the presence of dehydrogenase. Figure 1(a) shows typical images of pollen grains and pollen tubes stained with MTT. Positive staining, which indicated that the pollen grains or tubes were viable, showed up as a deeper color than negative staining. The viability of pollen grains and tubes under S-RNase challenge was examined at 10, 30, 60 and 120 min (Figure 1b). The time lines indicated that the viability of control and compatible pollen tubes decreased slightly during the period from 10 to 120 min, whereas that of incompatible pollen grains and tubes declined quite obviously at 60 min, from 68.92% (SE ± 0.036) at 30 min to 42.32% (SE ± 0.042) at 60 min. Furthermore, the decline continued and reached 31.23% (SE ± 0.046) at 120 min.
S-RNase caused the Δψmit of incompatible pollen tube to collapse
Figure 2(a) shows representative images of pollen tubes that underwent different treatments and were then stained with rhodamine 123 (RH-123). The intensity of fluorescence of tubes which were incubated with either self-S-RNase or valinomycin was weaker than that of tubes in the control and compatible treatments.
Figure 2(b) provides data on the mean fluorescence intensity of pollen tubes in each treatment. The mean fluorescence of tubes incubated with valinomycin (mean = 24.39, SE ± 11.16, n = 15, P < 0.01) was lower than that of the control tubes (mean = 66.58, SE ± 12.49, n = 15). The fluorescence of the incompatible tubes was similar to that of the valinomycin-treated tubes at only half that of the controls (mean = 25.07, SE ± 9.24, n = 15, P < 0.01), whereas the compatible tubes had a fluorescence similar to that of the controls (mean = 60.34, SE ± 7.66, n = 15, P > 0.05).
Self-incompatibility triggered the release of cytochrome c into the cytosol of incompatible pollen tubes
Figure 3 indicates that leakage of cytochrome c into the cytosol was stimulated by the addition of S-RNase in incompatible pollen tubes. A high cytochrome c concentration was detected in cytosolic extracts from incompatible pollen 120 min after addition of S-RNase but was not detected in either the compatible pollen tubes or the controls.
Self-incompatibility caused striking alterations to the structure of mitochondria
In order to characterize the mitochondrial alterations in pollen tubes following SI challenge, transmission electron microscopy (TEM) was used to assess the change in ultrastructure. The control pollen tubes showed a fairly electron-translucent cytoplasm filled with mitochondria which appeared to be electron-dense and showed dense packing of well-developed cristae 1 h after pre-culture (Figure 4a). By 3 h and 6 h, the electron density of the cytoplasm had increased, and the mitochondrial ultrastructure was not as clear as at 1 h (Figure 4b, c). At these time points there were also fewer differences in ultrastructure between the control and the compatible tubes (Figure 4d–f).
One hour after treatment, the cytoplasm of incompatible pollen tubes showed an increase in electron density compared with the controls (Figure 4g). At this point, two types of mitochondria were seen to exist simultaneously within the same incompatible tubes. One type appeared to be normal whereas the other was accompanied by a reduction in the number of cristae. By 3 h, the mitochondria had swollen dramatically and a large blebbing appeared (Figure 4h). It was common to see such mitochondria fused. At 6 h post-SI induction, the blebs had become more prominent (Figure 4i).
The mitochondria of valinomycin-treated tubes began to swell as early as 1 h after treatment (Figure 4j) and commonly fused into aggregations (Figure 4k, l).
Self-incompatibility-induced nuclear DNA degradation in incompatible pollen tubes
The DNA fluorescent stain 4,6-diamidino-2-phenylindole (DAPI) was used to evaluate the extent of degradation of nuclear DNA. Figure 5(a) shows representative images of three types of pollen tubes stained with DAPI. Binucleate tubes were regarded as normal, whereas the chromatin of generative cell nuclei was more highly condensed and more deeply stained with DAPI than that of vegetative nuclei. The presence of a single nucleus suggested that the vegetative nuclear DNA had degraded entirely, whilst the complete absence of nuclei indicated that all nuclear DNA had degraded.
The samples were fixed at 30 min after S-RNase challenge and then stained with DAPI to count the three types of tube (Figure 5b). The percentage of binucleate tubes, single-nucleate tubes and tubes with no nucleus relative to the total number of tubes in controls were 71.42% (SE ± 0.051), 17.63% (SE ± 0.024) and 10.95% (SE ± 0.032), respectively. In the compatible treatment, the percentages were 70.38% (SE ± 0.062), 16.76% (SE ± 0.031) and 12.86% (SE ± 0.059), respectively. There was no obvious difference between the control and compatible treatment. In the incompatible treatment, however, the percentage of binucleate tubes relative to the other types decreased markedly to only 11.05% (SE ± 0.045), while the percentage of single-nucleate tubes and those lacking a nucleus reached 25.22% (SE ± 0.041) and 63.73% (SE ± 0.046), respectively. These percentages indicated that SI caused the complete degradation of both the vegetative nucleus and the generative cell in incompatible pollen tubes. Moreover, cytochalasin B (CB), an actin-depolymerizing agent, can also induce degradation of nuclear DNA. Relative to the total number of tubes, the percentages of binucleate tubes and those with a single nucleus or no nucleus were 25.39% (SE ± 0.057), 23.44% (SE ± 0.032) and 51.17% (SE ± 0.038), respectively.
The terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method was adopted to further assess and confirm the degradation of nuclear DNA in incompatible pollen tubes. During the early period after SI challenge, only the vegetative nucleus appeared to be TUNEL-positive (Figure 5c, 10 min). When the DNA of the vegetative nucleus had been degraded, the generative cell nucleus also appeared to be TUNEL-positive (Figure 5c, 30 min). These results indicate that the vegetative nucleus was degraded in advance of the generative cell nucleus.
With the Japanese pear it is difficult to produce plants homozygous for the S locus. Therefore, in this series of experiments, pollen and S-RNase from heterozygous, rather than homozygous, plants were used to assay the effect of S-RNase on pollen. Two types of S gene pollen were present in the basal medium, S4 and S5. In the incompatible treatment, pollen grains/tubes of S4 and S5 were challenged with S4- and S5-RNase. The S4 pollen grains/tubes could theoretically encounter both S4-RNase (self-S-RNase) and S5-RNase (non-self-S-RNase) in the basal medium, and, similarly, the S5 pollen grains/tubes could encounter both types of S-RNase. In the S-RNase degradation model (Ushijima et al., 2003, 2004; Qiao et al., 2004a, b; Sijacic et al., 2004) or the S-RNase compartmentalization model (Goldraij et al., 2006), the non-self-S-RNase was degraded or compartmentalized in the compatible pollen tubes in which it could have no effect. Furthermore, compared with the control treatment, slight changes were apparent in the pollen grains/tubes of the compatible treatment in this series of experiments. These results indicate that the non-self-S-RNase performed no function in agreement with the assumptions of the two S-RNase models mentioned above. Based on these findings we concluded that the alterations noted in the incompatible pollen grains/tubes were only caused by self-S-RNase.
Fluorescence staining of the mitochondrial membrane potential (Δψmit)
Rhodamine 123 is a form of cationic dye that is electrophoretically distributed in the mitochondrial matrix in response to an electric potential across the inner mitochondrial membrane. The accumulation takes place as a consequence of its charge and of its solubility in both the inner membrane lipids and the matrix aqueous space. Because of this, RH-123 has been widely used to assess the mitochondrial Δψmit of isolated mitochondria from both plants and animals (Braidot et al., 1998; Petit, 1992; Zamzami et al., 2001). In animals, RH-123 had also been used in cortical neurons (Sensi et al., 2002) and Caco-2 cells (Bolduc et al., 2004) to detect the dissipation of Δψmit in intact cells. Here, we also made successful use of RH-123 in intact plant cells to evaluate Δψmit. This method was suitable for use with pollen tubes because they are single-celled structures, allowing RH-123 to penetrate into the cytoplasm. In addition, pollen tubes have a higher mitochondrial density compared with somatic cells (Derksen et al., 1995), which means that treatment differences in RH-123 fluorescence intensity are easily observable. As a lipophilic cation, RH-123 has a 10-fold accumulation in the cytoplasm and a 1000–10 000-fold accumulation in mitochondria at equilibrium, compared with its concentration in the incubation buffer (Chen, 1989). Therefore, differences in fluorescence intensity in the pollen tubes in our research mainly resulted from the accumulation of RH-123 in the mitochondria. The rate of RH-123 fluorescence decay has been found to be proportional to the decrease in Δψmit (Ronot et al., 1986). To confirm these results, pollen tubes were incubated with valinomycin, which removes the membrane potential by equilibrating the potassium ion concentrations on both sides of the mitochondria. The results showed that the mean fluorescence intensity of tubes incubated with valinomycin was lower than that of the control tubes. These findings further validate the methods chosen in this study.
Mitochondrial alterations and DNA degradation in incompatible pollen tubes
The MTT test, detecting the presence of the respiratory chain enzyme dehydrogenase (Rodriguez-Riano and Dafni, 2000), indicated that the mitochondria of incompatible pollen tubes were specifically affected by S-RNase. Furthermore, the collapse in Δψmit was observed in incompatible pollen tubes at 30 min after SI challenge. In animal cells, Δψmit collapse induces the release of mitochondrial pro-apoptotic factors into the cytosol (Zamzami and Kroemer, 2003). These are followed by the release of cytochrome c which drives the assembly of the apoptosome that culminates in the activation of the executioner caspase-3 (Liu et al., 1996). Cytochrome c was detected by western blotting in the cytosol extraction of incompatible pollen tubes at 120 min in the present study. The high concentration found suggests that cytochrome c began to leak into the cytosol earlier than the point at which we detected it. Translocation of cytochrome c from mitochondria into the cytosol is a classic marker for PCD in many organisms, including plants (Balk et al., 1999; Adrain and Martin, 2001). Furthermore, swollen mitochondria accompanied by normal mitochondria were observed in incompatible pollen tubes 1 h after SI challenge, and the change in mitochondria had become more prominent and universal when observed at 3 h and 6 h. These dramatic changes included swelling, loss of cristae, loss of electron density in the matrix and blebbing. In addition, the mitochondria of pollen tubes treated with valinomycin showed some similar alterations. These results indicate that collapse of Δψmit can induce alterations in mitochondrial structure which are similar in nature in incompatible pollen tubes to the swelling of mitochondria described in several animal systems during apoptosis (Korostoff et al., 2000; Kwong et al., 1999; Petit et al., 1995, 1998). Mitochondrial swelling has also been observed in SI-challenged poppy pollen (Geitmann et al., 2004), microspore mother cells in male-sterile sunflowers (Laveau et al., 1989) and in the tapetal cells in sunflower and maize lines with cytoplasmic sterility (Hoener, 1977;Warmke and Lee, 1977). It has been generally accepted that PCD is triggered in these cells (Balk and Leaver, 2001; Thomas and Franklin-Tong, 2004). The above observations suggest that mitochondrial swelling might be associated with PCD. Mitochondria play an essential role in the regulation of apoptosis and are considered to be strategic centers in the control of cell death (Ferri and Kroemer, 2001; Green and Reed, 1998). The biochemical processes taking place in the mitochondria in apoptotic cells have been suggested to represent the ‘point of no return’ in cell death signaling cascades in animal cells (Green and Reed, 1998), indicating their critical role in the orchestration of PCD (Dubin and Stoppani, 2000; Gottlieb, 2000). The observation that the cytoplasmic electron density of the control and compatible pollen tubes was increased at 3 h and 6 h was possibly caused by the decreased growth rate of the pollen tubes at this time.
The nuclear DNA fragment, a hallmark feature of apoptosis, was detected by DAPI at 30 min. The pollen of the pear is binucleate, possessing a vegetative nucleus and a generative cell. Our results showed that the percentage of binucleate pollen tubes compared with that of all tubes observed decreased markedly after SI challenge to nearly one-seventh of the number in the control treatment. This strongly suggests that nuclear DNA fragmentation and/or the formation of apoptotic bodies occurred in incompatible pollen tubes. To confirm our conclusions, the TUNEL method was adopted to further assess the nuclear DNA fragmentation of incompatible pollen tubes. Some TUNEL-positive pollen tubes were first observed at 10 min after SI challenge and the vegetative nucleus was degraded in advance of the generative cell nucleus. This result indicates that the vegetative nucleus was more susceptible to degradation than the generative nucleus. A similar observation has been made in Papave rhoeas under SI challenge (Jordan et al., 2000). There are two possible explanations for this difference in susceptibility. Firstly, the chromatin of the generative nucleus was more highly condensed, making it less accessible to endonucleases, and secondly, the generative cell nucleus might have additional ‘insulation’ due to its own cell wall (Jordan et al., 2000).
Did PCD occur in the S-RNase-based GSI response?
It is widely accepted that the growth of pollen tubes is inhibited by S-RNase degrading the RNA of incompatible pollen tubes. However, RNA degradation may have only been the beginning of the SI response, not the end. Previous work has shown that alterations in actin occur prior to the arrest of pollen tube growth after induced incompatibility (Liu et al., 2007). Here, the collapse of Δψmit and DNA degradation were detected at 30 min, and the viability of the incompatible pollen tubes was compromised by 60 min. These data indicate that alterations in actin, the collapse of Δψmit and DNA degradation were the cause, not the result, of growth arrest of the pollen tubes. Actin cytoskeletal degradation has always been observed to be accompanied by apoptosis or PCD (Janmey, 1998; Morley et al., 2003; Thomas et al., 2006), and we found that CB, an actin-depolymerizing agent, can directly induce degradation of nuclear DNA. It seems that a relationship exists between self-S-RNase, actin depolymerization and DNA degradation. Cytotoxins universally induce apoptosis in animal cells (Hickman and Boyle, 1997) and S-RNase has been widely thought to function as a specific cytotoxin (Goldraij et al., 2006). Theoretical support existed to suggest that S-RNase could trigger PCD in incompatible pollen tubes and, in this study, we have provided some preliminary evidence that PCD may occur. However, further studies are required to fully understand the GSI system.
Adult pear (Pyrus pyrifolia L.) trees planted in the orchards of Nanjing Agricultural University, Jiangsu, China were used. The cultivars and their S-genotypes were Kosui (S4S5) and Imamuraaki (S1S6). Flowers from each cultivar were collected a few days before anthesis, and the styles were detached, weighed and stored in liquid nitrogen. Kosui anthers were also collected, dehisced, dried in bottles containing desiccants and stored at −20°C.
Preparation, concentration and activity of S-RNase
Kosui pollen grains were pre-cultured for 2 h at 25°C in a basal medium and in darkness based on the method of Hiratsuka et al. (2001). The basal medium consisted of a 2-(N-morpholine)-ethanesulfonic acid (MES)-NaOH buffer supplemented with 10% sucrose, 15% polyethylene glycol 4000, 0.01% H3BO3, 0.07% Ca(NO3)2 4H2O, 0.02% MgSO4 7H2O and 0.01% KNO3, pH 6.0–6.5.
After pre-culture, Kosui stylar S-RNase was added to the medium as an SI challenge, Imamuraaki stylar S-RNase was added to the medium as a compatible treatment and medium without S-RNase was used as a control. The final activity of the S-RNases in the basal medium was 0.15 U.
Pollen grain and pollen tube viability
The pollen was cultured and SI was induced as described above. To determine pollen grain and pollen tube viability at 30 min after SI challenge MTT tests were used as described by Rodriguez-Riano and Dafni (2000). The test solution contained a 1% concentration of the MTT substrate (Fluka, Sigma-Aldrich; http://www.sigmaaldrich.com/) in 5% sucrose. After 30-min incubation at 37°C the pollen grains/tubes stained with MTT were examined with a Zeiss Axioakop40 microscope (Carl Zeiss, http://www.zeiss.com/). At least 100 pollen grains/tubes were observed per treatment, and each treatment was repeated three times. Images were captured with a Nikon COOLPIX4500 digital camera (http://www.nikon.com/) mounted on the microscope and were processed by Adobe Photoshop CS (http://www.adobe.com/).
Mitochondrial membrane potential (Δψmit)
Pre-cultured pollen tubes were treated with S-RNases or valinomycin. The S-RNases were added as described above. Valinomycin was added in solution with 10% dimethyl sulfoxide (DMSO) to the basal medium at a concentration of 1 μm. Thirty minutes after treatment, all samples were stained in the dark for 30 min with RH-123 (Calbiochem, EMD Chemicals Inc., http://www.emdbiosciences.com/home.asp) at a final concentration of 1 μm in the basal medium at 25°C. Rhodamine 123 had previously been dissolved in 10% DMSO. The stained samples were centrifuged at 100 g for 5 min, washed three times in MES-NaOH buffer (pH 6.2), then placed onto a cover slide and investigated with a laser scanning confocal microscope (Leica TCS SP2, Leica Microsystems Group, http://www.leica-microsystems.com/). At least 10 pollen tubes were photographed per slide using the same laser scanning confocal microscope parameter. Each treatment included three independent experiments. The mean fluorescence intensity of the pollen tubes was quantified using Leica Confocal Software Lite.
Detection of cytochrome c in pollen tube extracts
For the SI treatments, pollen was grown in the basal medium for 1 h at 25°C and then SI was induced as described above. Mitochondrial and cytosolic fractions were prepared as described by Balk et al. (1999) with some modifications. All the subsequent steps were performed at 0–4°C. The pollen tubes were collected by centrifugation at 100 g, and mitochondrial extraction buffer was added (50 mm HEPES, 600 mm sorbitol, 5 mm EDTA, 0.1% BSA, 1% PVP 40, 1 mm DTT, 5 mm KH2PO4, pH 7.4). The pollen tubes were then ground and filtered. The filtrate was homogenized and centrifuged for 15 min at 2000 g to remove whole cells and other heavy debris. The supernatant was then centrifuged again for 40 min at 18 000 g and the resulting supernatant represented the cytosolic fractions. The deposit was resuspended in 2 ml of mitochondrial extraction buffer and then layered on top of a discontinuous Percoll gradient consisting of the following steps: 3 ml each of 13.5, 21 and 45% (v/v). Each step also contained 50 mm HEPES-KOH (pH 7.4), 0.5 m sucrose, 5 mm EDTA, 0.1% BSA and 5 mm KH2PO4. The gradients were then spun in a Beckman Optimal L80-XP ultracentrifuge (Beckman Coulter Inc, http://www.beckmancoulter.com/) with a SW 41 TI rotor at 80 000 g for 45 min. The pure mitochondrial fractions were located in the middle of the 21 and 45% Percoll gradients. Equal concentrations of protein from both fractions were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting. Blots were probed with a 1:1000 dilution of a monoclonal antibody (clone 7H8.2C12; BD Biosciences, http://www.bdbiosciences.com/) and then with an anti-(mouse horseradish peroxidase) secondary antibody (Boster Biotechnology, http://www.boster.com.cn) before finally being detected by a DAB kit (yellow; Boster Biotechnology).
For electron microscope observation, samples of different treatments were fixed for 3 h in 2.5% glutaraldehyde in a 0.1 m phosphate buffer at pH 7.0. They were then embedded in agar, fixed in 2.5% glutaraldehyde for a further 12 h, fixed in OsO4 in cacodylate buffer at pH 7.4 and then dehydrated and embedded in epoxy resin. Sections stained with uranyl acetate and lead citrate were examined with a Hitachi H-7650 transmission electron microscope (http://www.hitachi.com/).
Nuclear DNA degradation
To assess the extent of nuclear DNA degradation, pollen tubes were fixed in 95% ethanol:glacial acetic acid (3:1) for 1 h at 4°C, transferred into 70% ethanol at −20°C for at least 4 h and then stained with 0.05 μg ml−1 DAPI (Sigma) in citrate buffer at pH 4.1 for 2 h. The concentration of CB was 5.2 μm.
The DeadEnd Fluorometric TUNEL System (Promega, http://www.promega.com/)was used for independent assessment of nuclear DNA degradation in incompatible pollen tubes. The protocol used followed the manufacturer’s instructions.
The stained samples were investigated with a Zeiss Axioskop40 fluorescence microscope. At least 100 pollen tubes were counted in each treatment and the experiment was repeated three times. Images were captured with a laser scanning confocal microscope (Leica TCS SP2) and processed by Adobe Photoshop CS.
This work was supported by the Science Foundation of the Doctoral Subject Point of the Chinese Ministry of Education (no. B200523) and National Science and Technology Support Project of the Chinese Government (no. 2006BAD01A1704-11). Also, we express our gratitude to Professor Betsy Drager of the College of Foreign studies, Nanjing Agricultural University for improving the English language in the manuscript.