Pre-zygotic interspecific incompatibility (II) involves an active inhibition mechanism between the pollen of one species and the pistil of another. As a barrier to fertilization, II effectively prevents hybridization and maintains species identity. Transgenic ablation of the mature transmitting tract (TT) in Nicotiana tabacum resulted in the loss of inhibition of pollen tube growth in Nicotiana obtusifolia (synonym Nicotiana trigonophylla) and Nicotiana repanda. The role of the TT in the II interaction between N. tabacum and N. obtusifolia was characterized by evaluating N. obtusifolia pollen tube growth in normal and TT-ablated N. tabacum styles at various post-pollination times and developmental stages. The II activity of the TT slowed and then arrested N. obtusifolia pollen tube growth, and was developmentally synchronized. We hypothesize that proteins produced by the mature TT and secreted into the extracellular matrix inhibit interspecific pollen tubes. When extracts from the mature TT of N. tabacum were injected into the TT-ablated style prior to pollination, the growth of incompatible pollen tubes of N. obtusifolia and N. repanda was inhibited. The class III pistil-specific extensin-like protein (PELPIII) was consistently associated with specific inhibition of pollen tubes, and its requirement for II was confirmed through use of plants with antisense suppression of PELPIII. Inhibition of N. obtusifolia and N. repanda pollen tube growth required accumulation of PELPIII in the TT of N. tabacum, supporting PELPIII function in pre-zygotic II.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Inter-specific reproductive barriers are important for genetic isolation of populations and genesis of new species (Moyle and Graham, 2006; Howard, 1999; Rieseberg and Blackman, 2010). Geographical, ecological, mechanical and physiological barriers all contribute to reproductive isolation (Tiffin et al., 2001). In order to achieve double fertilization, pollen must be received by a compatible pistil, hydrate, germinate and grow through the stigma and style to the ovary (Lord and Russell, 2002). Pollen–pistil interactions play a critical role in controlling fertilization from compatible pollen sources and preventing fertilization from incompatible pollen sources (Kikuchi et al., 2007; Lee et al., 2008; Figueroa-Castro and Holtsford, 2009). Inter-specific pollen–pistil barriers may drive speciation events or form as a consequence of speciation and genetic divergence (Rieseberg and Willis, 2007; Liedl and Anderson, 1993). This study focuses on elucidating pollen–pistil barriers in the transmitting tract (TT) of the Nicotiana tabacum pistil. Understanding interspecific reproductive barriers will provide insight into species evolution, maintenance and gene flow (Howard, 1999).
The pistil functions in both intraspecific and interspecific pollen incompatibilities. Self incompatibility (SI) is a highly studied mechanism of intraspecific incompatibility that prevents self pollination and promotes outcrossing, thereby reducing inbreeding (Kao and McCubbin, 1996; Takayama and Isogai, 2005; Goldberg et al., 2010). Nicotiana shows gametophytic SI, controlled by the multi allelic S locus that encodes the pollen-specific S–locus F–box and the style-specific S–RNase proteins (Chen et al., 2010). Inter-specific pollen–pistil barriers are the result of either incongruity or incompatibility. Incongruity is defined as a passively evolved mismatch of genetic information between two species (Hogenboom, 1973; Liedl and Anderson, 1993; Kuboyama et al., 1994), and interspecific incompatibility (II) is an active rejection of interspecific pollen (Liedl and Anderson, 1993; McClure et al., 2000). When II is associated with SI, where SI females reject pollen from self-compatible males and the reciprocal cross is compatible, it is termed unilateral incompatibility (Pandey, 1968; Lewis and Crowe, 1958; de Nattancourt, 1997; Bedinger et al., 2011). There is evidence that SI and unilateral incompatibility are controlled by both common and unique mechanisms depending on the species involved in the interaction (McClure et al., 1999; Li and Chetelat, 2010; Covey et al., 2010). Two proteins involved in SI, S–RNase and HT (small asparagine-rich protein), have been shown to be required for some unilateral incompatibility interactions in Nicotiana (Murfett et al., 1996; McClure et al., 1999, 2000; Bedinger et al., 2011). However, self-compatible species have also been shown to reject interspecific pollen, which demonstrates SI-independent mechanisms for II (Kuboyama et al., 1994; Bedinger et al., 2011). There is disagreement in the literature about whether inhibition of interspecific pollen is due to incompatibility or incongruity, as the two concepts are not mutually exclusive; here we use II to refer to inhibition of interspecific pollen tube growth by the N. tabacum style.
Rapidly growing pollen tubes use TT resources to support their growth, and the TT is the location of gametophytic SI interactions in Nicotiana (Johnson and Lord, 2006). The Nicotiana TT is a solid, specialized, metabolically active file of cells that is continuous from the stigma to the ovary (Jansen et al., 1992). Pollen tubes grow through the TT extracellular matrix, which comprises carbohydrates, amino acids, glycolipids, glycoproteins, lipids and proteins (Swanson et al., 2004; Lind et al., 1994). Proline-rich proteins are predominant in the TT extracellular matrix, and are known to control pollen tube growth (Wu et al., 2001). In Nicotiana, three hydroxyproline-rich proteins, transmitting tissue-specific proteins (TTS), 120 kDa protein (120 K) and class III pistil extensin-like proteins (PELPIII), have been identified as major components of the TT extracellular matrix. The proteins are soluble, developmentally accumulated and highly glycosylated, and share conserved cysteine-rich C–terminal domains (Bosch et al., 2001; Wu et al., 2001). The TTS proteins have been shown to promote pollen tube growth in vivo and in vitro in N. tabacum (Cheung et al., 1995). The 120 K protein, isolated from Nicotiana alata, is taken up into the cytoplasm of pollen tubes and has been shown to be required for S–specific pollen rejection, although its function in self-compatible N. tabacum has not been reported (Hancock et al., 2005; Lind et al., 1996). The PELPIII proteins were shown to be incorporated into pollen tube walls of both compatible and incompatible pollen tubes, but their function has not been identified (Goldman et al., 1992; de Graaf et al., 2003; Bosch et al., 2003; de Graaf et al., 2004; Cruz-Garcia et al., 2005). In N. alata, these proteins were shown to interact with S–RNase, which suggests that they may function in SI or unilateral incompatibility (Cruz-Garcia et al., 2005).
The TT and its components are well studied, but the mechanisms that control pollen tube growth are still largely unknown (Edlund et al., 2004; Swanson et al., 2004; Hogenboom, 1973, 1984). In Arabidopsis and Nicotiana, although the TT does support optimal pollen tube growth, it is not essential for pollen tube growth of compatible pollen (Crawford et al., 2007; Eberle et al., 2012). Kuboyama et al. (1994) previously showed that Nicotiana obtusifolia and Nicotiana repanda pollen was incompatible with N. tabacum pistils, and we further investigated the role that the TT plays in II using a transgenic N. tabacum line with ablated mature TT cells and a hollow style (TT-ablated) (Gardner et al., 2009). We show here that the N. tabacum TT has an essential function in inhibiting interspecific pollen tube growth of N. obtusifolia and N. repanda. As N. tabacum is self-compatible and contains no active S–RNase (Golz et al., 1998), the II mechanism is independent of SI. We hypothesized that the inhibitory molecule of the TT that was responsible for II is a soluble protein present in the extracellular matrix. We used the TT-ablated style in a pollen tube growth assay (Figure S1) (Eberle et al., 2012) to measure the inhibitory activity of proteins produced by the normal style toward II pollen, by injection into the hollow space of the TT-ablated style followed by pollination.
Nicotiana obtusifolia pollen tube growth in normal and TT-ablated N. tabacum styles
The N. obtusifolia pollen tube length was significantly shorter in normal N. tabacum styles than in TT-ablated styles from 4 to 88 h post-pollination, with the difference in pollen tube length between the two style types increasing with increased incubation time (Figure 1). At 40 h post-pollination, the pollen tubes of N. obtusifolia looked morphologically similar to N. tabacum pollen tubes in both normal and TT-ablated styles (Figure S1). Nicotiana tabacum styles were 26.4 ± 0.13 mm long. At 64 h post-pollination, the N. obtusifolia pollen tubes had reached the end of the TT-ablated N. tabacum styles (Figure 1). In comparison, at 64 h post-pollination, N. obtusifolia pollen tube growth was arrested in normal N. tabacum styles with a mean length of 19.9 ± 0.7 mm, approximately 6 mm short the end of the style (Figure 1). TT ablation resulted in a loss of interspecific inhibitory activity toward N. obtusifolia pollen tubes. Increased temperature was also found to reduce the II interaction between the N. tabacum style and N. obtusifolia (Figure S2).
Developmental accumulation of II
The N. tabacum pollen tube length was not significantly different between normal and TT-ablated styles in 15 mm, 35 mm and post-anthesis developmentally staged styles (Figure 2a). In 25 mm styles, the pollen tube length of N. tabacum was significantly (P =0.05) longer in TT-ablated styles (8.9 ± 1.1) than in normal styles (5.4 ± 1.2). In pre-anthesis styles, the N. tabacum pollen tube length was significantly longer in normal styles than in TT-ablated styles. For N. tabacum pollen tubes, there was a relationship between style length and pollen tube length, with an increase in pollen tube lengths as the style length increased, but at no stage did the N. tabacum pollen tubes consistently grow the entire style length by 40 h post-pollination.
The N. obtusifolia pollen tube length in the normal style was significantly different to that in TT-ablated styles at all developmental stages, except 25 mm (Figure 2b). At the 15 mm bud stage, N. obtusifolia pollen tubes were significantly longer in normal than in TT-ablated styles. At the 25 mm stage, the pollen tube length in the normal style was not significantly different from that in the TT-ablated style. At the three most mature stages of pistil development, N. obtusifolia pollen tube lengths were significantly longer in the TT-ablated style relative to the normal styles (Figure 2b). This was the opposite to the early developmental stages where the pollen tube length was longer in normal styles than in TT-ablated styles. At the 35 mm developmental stage, a clear transition from compatible to incompatible pollen tube growth occurs in the normal style, with an N. obtusifolia pollen tube length that was 2.6 mm longer in the TT-ablated style than in normal style. This continues through maturation, with the greatest difference between the two style types (2.9 mm) occurring post-anthesis. These data show that N. obtusifolia pollen tube II increased with pistil maturation of the normal N. tabacum styles.
Activity of TT proteins in the pollen tube growth assay
The N. tabacum pollen tube grew further in the normal style than in the TT-ablated style and the Nicotiana sylvestris pollen tube grew equally in the normal and TT-ablated styles (Figure 3). N. tabacum and N. sylvestris were used as compatible pollen controls. The interspecific pollen of N. sylvestris was not inhibited by the N. tabacum TT and provided an interspecific compatible pollen source. Transmitting tract eluate (TTE) made from normal N. tabacum styles and injected into the TT-ablated style specifically inhibited pollen tube growth of N. obtusifolia and N. repanda but not N. tabacum or N. sylvestris (Figure 3). The N. tabacum pollen tube length was 18.3 ± 1.8 mm in the TT-ablated style compared to 25.6 ± 0.3 mm in the normal style (Figure 3). The pollen tube lengths of N. tabacum in styles injected with modified complete medium (MCM; see 'Developmental accumulation of II') or TTE were 20.5 ± 0.8 and 18.6 ± 1.0 mm, respectively, and were not significantly different from the pollen tube length in the TT-ablated style (Figure 3). Pollen tubes of N. sylvestris in normal and TT-ablated styles grew to 23.9 ± 1.6 and 23.3 ± 0.9 mm, respectively. Ablation of the TT had no effect on N. sylvestris pollen tube growth. The N. sylvestris pollen tube lengths in MCM- and TTE-injected styles were not significantly different from those in the TT-ablated style. N. obtusifolia pollen tubes were 5.4 mm shorter in the TTE-injected style than the MCM-injected style (14.2 ± 1.1 and 19.6 ± 0.5 mm, respectively; Figure 3). The N. obtusifolia pollen tube lengths in the normal and TTE-injected styles were not significantly different from each other (12.2 ± 0.3 and 14.2 ± 1.1 mm, respectively), but were significantly shorter than the pollen tube lengths in TT-ablated and MCM-injected styles (21.5 ± 0.5 and 19.6 ± 0.5 mm, respectively; Figure 3). Inhibitory activity of the TTE toward N. obtusifolia pollen tube growth was dependent on the protein concentration of the TTE (Figure S3). Injection of TTE also significantly reduced the pollen tube length of N. repanda by 7.9 mm compared to MCM-injected styles (13.5 ± 0.9 and 21.4 ± 1.1 mm, respectively; Figure 3), although length in the normal style was significantly shorter than in the TTE-injected style. For both N. repanda and N. obtusifolia pollen tubes, injection of TTE into the TT-ablated style specifically and significantly inhibited pollen tube growth compared to the length in the MCM-injected styles (Figure 3).
Specific inhibition of N. obtusifolia pollen tube growth by purified TTE proteins
Use of cation fast protein liquid chromatography (FPLC) effectively separated the protein species and II activity of the TTE into three fractions (F1, F2 and F3; Figure 4). The pollen tube length of N. tabacum in the normal style was significantly longer than the pollen tube length for all the injected style treatments, which were not significantly different from each other (Figure 4a). The N. obtusifolia pollen tube length was significantly shorter in the normal style than in the MES-injected style (13.7 ± 1.1 and 18.7 ± 0.4 mm, respectively; Figure 4a). The mean pollen tube length in the TTE-injected style was 1.5 mm longer than in the normal style and was 3.5 mm shorter than in the MES-injected style. F1 injected at 13 μg μl−1 resulted in an N. obtusifolia pollen tube length of 16.7 ± 0.7 mm, which was 3 mm longer than the normal style and 2 mm shorter than the length of the MES-injected style. Both the TTE and F1 injections caused partial inhibition of pollen tube growth, but the inhibition was not significant compared to the length in either the normal or MES-injected style. Injection of F2 at 13 μg μl−1 or F3 at 20 μg μl−1 had a significant inhibitory effect towards N. obtusifolia relative to MES-injected styles, with mean lengths of 13.1 ± 1.6 and 13.1 ± 1.1 mm, respectively. When F3 was diluted to 13 μg μl−1 (equal to the concentrations of F1 and F2) and injected into the TT-ablated style, N. obtusifolia pollen tubes grew to 18 ± 0.9 mm, which was not significantly different from the MES-injected styles but was significantly longer than in the normal style. Thus, at 13 μg μl−1, F2 had inhibitory activity towards N. obtusifolia, while F3 had inhibitory activity at 20 μg μl−1 but not 13 μg μl−1.
Each FPLC fraction had a unique and less complex protein profile compared to the TTE as shown by SDS–PAGE (Figure 4b). The TTE had a complex profile from 6 to 200 kDa, with a high molecular weight smear from 50 to 200 kDa. F1 had three major bands at 35, 40 and 56 kDa that were easily visualized, as well as other lower-abundance bands that showed weaker staining. F2 had two major bands at 50 kDa (low molecular weight) and 115–200 kDa (high molecular weight). The majority of the F3 protein was contained in a high-molecular-weight smear at 55–200 kDa, with weakly stained bands at 50 and 40 kDa.
Identification of F2 and F3 proteins
Protein immunoblots were used to test for the presence of the PELPIII, 120 K and TTS proteins in TTE, F1, F2 and F3 (Figure 4c). These proteins are major components of the TT extracellular matrix, and have molecular weights from 55 to 200 kDa. PELPIII was detected at high abundance from 120 to 200 kDa in the TTE, F2 and F3, but not in F1 (Figure 4c). Detection of PELPIII in F2 showed an additional high-molecular-weight signal above 200 kDa, which was not detectable in the TTE and weakly detected in F3. Additionally, in F3, a low-molecular-weight band at approximately 90 kDa was also detected with the PELPIII antibody. The variable detection across fractions was probably caused by the heterogeneous nature of PELPIII glycosylation. The 120 K and TTS proteins were only detected in the TTE and F3. Two distinct bands were detected using the 120 K antibody, consistent with the results for N. tabacum stylar proteins obtained by Cruz-Garcia et al. (2005). Presence of PELPIII in F2 and F3 was confirmed by tandem MS/MS analysis (Table S1). The presence of PELPIII in F2 and F3, combined with the specific inhibitory activity of F2 and F3, implied involvement of PELPIII in II.
Inter-specific inhibitory activity of PELPIII on N. obtusifolia and N. repanda pollen tube growth
The role of PELPIII in inhibition of N. obtusifolia and N. repanda pollen tube growth was tested by pollinating PELPIII antisense N. tabacum ‘Petit Havana’ SR1 lines F81, C81 and C64 (Bosch et al., 2003) with N. tabacum, N. obtusifolia and N. repanda pollen (Figure 5). Protein immunoblots of stylar extracts from wild-type, C64, C81 and F81 showed no detectable PELPIII in C64 and C81 extracts, whereas 120 K and TTS were detected in all four lines (Figure 5b) (Bosch et al., 2003). The F81 line was PCR-positive for amplification of the neomycin phosphotransferase gene sequence (a selectable marker included in the antisense plasmid construct; Bosch et al., 2003). However, there was still detectable PELPIII protein present in F81 style extracts (Figure 5b).
The pollen tube length of N. tabacum was 24.7 ±0.4 mm in the wild-type styles, which was significantly shorter than in the F81 and C64 styles (29.8 ± 0.4 and 27.4 ± 0.8 mm, respectively) but not the C81 style (26.8 ± 0.9 mm) (Figure 5a). N. obtusifolia pollen tubes were significantly shorter in both wild-type and F81 styles (14.0 ± 0.3 and 14.9 ± 0.2 mm) compared to C64 styles (19.3 ± 1.5 mm). In C81 styles, the N. obtusifolia pollen tube length was 17.8 ± 0.6 mm, which was not significantly different from F81 or C64, but was significantly longer than wild-type styles (Figure 5a). The pollen tube length of N. repanda was also significantly shorter in wild-type and F81 styles (2.7 ± 1.0 and 1.9 ± 0.4 mm) relative to C64 styles (16.4 ± 3.3 mm), but these reduced lengths were not significantly different from those in C81 styles (10.1 ± 3.5 mm) (Figure 5a). The pollen tube length of N. obtusifolia and N. repanda was significantly longer in C64 styles than wild-type and F81 styles, but the pollen tube length of N. tabacum did not follow the same trend, indicating a specific loss of interspecific incompatible pollen tube growth in the C64 style. Overall, detection of PELPIII correlated with a reduced mean pollen tube length of N. obtusifolia and N. repanda but not N. tabacum (Figure 5).
The mature TT of N. tabacum is required for II with N. obtusifolia and N. repanda pollen. The TT-specific PELPIII was identified as a major component of the II interaction. We show that semi-purified PELPIII injected into the TT-ablated style was able to re-establish the II interaction. When PELPIII accumulation was reduced by antisense suppression, the II interaction was also suppressed. Thus, PELPIII plays an important role in II.
The mature N. tabacum TT is required for II
Ablation of the N. tabacum TT increased N. obtusifolia pollen tube length relative to normal styles at all times from 4 to 64 h post-pollination (Figure 1). Increased pollen tube growth by TT ablation was specific to II pollen and was not observed for N. tabacum pollen, which grows at a slower rate in the TT-ablated style (Eberle et al., 2012).
The II activity of the N. tabacum style was developmentally dependent, with N. obtusifolia pollen tube growth inhibition measureable at the 35 mm developmental stage (Figure 2). Kuboyama et al. (1994) also reported a developmental dependency of II by N. tabacum styles by testing whether pollen tubes were able to reach the ovaries 2 days after pollination. Their results also show developmental synchronization of II by the N. tabacum style (Kuboyama et al., 1994). Goldman et al. (1992) showed that PELPIII mRNA began to accumulate at stage 3 (approximately 10 mm) of N. tabacum ‘Petit Havana’ SR1 flower development, and de Graaf et al. (2003) showed that PELPIII protein began to accumulate at stage 3, with greater accumulation at stage 5 (approximately 20 mm). We estimate that our 35 mm N. tabacum ‘Samsun’ buds were at approximately stage 8 (Koltunow et al., 1990). There may be some developmental differences between the ‘Samsun’ and ‘Petit Havana’ SR1 cultivars that explain the lag between accumulation of PELPIII (20 mm, stage 5; de Graaf et al., 2003) and the development of II (35 mm, stage 8; Figure 2b). Furthermore, when activity of the TTE was tested in the TT-ablated style, there was a relationship between inhibitory activity and total TTE protein concentration (Figure S3), indicating that II activity was concentration-dependent. The TT-specific developmental accumulation of PELPIII, combined with our results showing that the TT is essential for II and that II is developmentally synchronized, demonstrate an association between PELPIII and II.
PELPIII is correlated with II activity in the pollen tube growth assay
Inter-specific inhibition of N. obtusifolia and N. repanda pollen tube growth by the N. tabacum style in vivo was reproduced by injection of TTE in the TT-ablated style (Figure 3). The TTE did not inhibit either N. tabacum or N. sylvestris pollen tube growth, showing that the inhibitory activity of the TTE was specific to II pollen and did not inhibit all pollen tube growth or all interspecific pollen tube growth. This result demonstrated the utility of the TT-ablated style as a functional pollen tube growth assay, for elucidation of the effect of exogenous compounds on pollen tube growth. The N. repanda pollen tube length in TTE-injected styles was not as reduced as in the normal style (Figure 3). This may be a result of the TTE having a lower concentration of proteins than required for in vivo inhibition, or a biological difference in the pollen tube growth response of N. repanda in the pollen tube growth assay compared with in vivo. Complete inhibition of N. repanda may also require additional factors that are not required for the more complete inhibition N. obtusifolia. The equal inhibition by TTE injection and the normal style on N. obtusifolia pollen tube growth indicates that all of the factors required for in vivo inhibition of N. obtusifolia are eluted and active in the TTE.
Separation of the TTE proteins showed that the majority of II activity occurs in F2 and F3, and PELPIII was present in both fractions (Figure 4). The higher protein concentration required for F3 activity in the pollen tube growth assay was probably the result of a lower level of PELPIII relative to other proteins in F3. Protein immunoblots showed detection of PELPIII in both F2 and F3, but TTS and 120 K were only detected in F3 (Figure 4c). Based on our hypothesis that the same II protein species occurs in both F2 and F3, we concluded that 120 K and TTS are not involved in II activity as they were not detected in F2. The 120 K protein is required for SI in N. alata, but was found to have no function in II (Hancock et al., 2005). Additionally, based on the protein immunoblot, F2 appeared to have higher abundance of PELPIII than F3 (Figure 4c). The putative higher relative abundance of PELPIII in F2 correlates with the higher II activity of F2, making PELPIII a probable candidate in II.
In N. tabacum, TTS was shown to have function in pollen tube growth promotion and guidance in vitro and in vivo (Cheung et al., 1995). The presence of TTS in both the TTE and F3 did not significantly promote growth of N. tabacum in the pollen tube growth assay (Figure 4). These data indicate that the pollen tube growth assay does not fully mimic the N. tabacum pistil. Wu et al. (1995) reported that N. tabacum styles accumulate TTS with an increasing glycosylation gradient from the stigma to the ovary. Pollen tubes were shown to deglycosylate TTS protein and did not respond to deglycosylated protein in vitro. They proposed that pollen tubes detect the glycosylation gradient in the style, and that the gradient is involved in supporting pollen tube growth (Wu et al., 1995). When TTE and F3 are injected into the TT-ablated style, the in vivo glycosylation gradient of TTS is probably not reproduced, and this may be the reason why we did not observe promotion of N. tabacum pollen by injections containing TTS.
Antisense suppression of PELPIII increased II pollen tube growth
The function of PELPIII in II was confirmed by the fact that both N. obtusifolia and N. repanda pollen tubes grew significantly longer in the C64 antisense styles, which did not accumulate detectable PELPIII, than in wild-type styles after 40 h pollination time (Figure 5). The pollen tube length of N. obtusifolia in C81 styles was significantly longer relative to wild-type but not F81 styles at α = 0.05 but not at α = 0.07. It is possible that there was low PELPIII accumulation in the C81 style that was below the immunoblot detection level but caused a slight decrease in pollen tube length. Additionally, the F81 style may have a reduced level of PELPIII relative to wild-type styles that may account for the slight increase in N. obtusifolia pollen tube length relative to the wild-type style. The pollen tube length of N. repanda in C81 styles was also intermediate; it was shorter than in C64 and longer than in F81 and wild-type (Figure 5). If there are variable levels of PELPIII in the various lines, it would appear that N. repanda pollen tubes are inhibited by lower levels of PEPLIII in vivo than N. obtusifolia pollen tubes. The pollen tube growth of N. repanda in C64 and C81 styles also had a larger variation (SE) than in wild-type and F81 styles. If N. repanda is more sensitive to low levels of PELPIII in vivo, and the C64 and C81 styles have variable low levels, this may cause individual styles to be more or less inhibitory depending on whether a critical PELPIII threshold is reached, as well as explaining the large variation (SE). The higher sensitivity of N. repanda pollen to PELPIII in vivo is in contrast to what was observed in the pollen tube growth assay, in which N. repanda pollen decreased inhibition by the TTE relative to normal styles. The response of N. repanda in vivo and in the pollen tube growth assay may be variable as a result of biological differences between the two environments. These data also indicate that the II interaction of N. repanda is different from the interaction of N. obtusifolia, although both required the presence of PELPIII in the style. The complete inhibition of N. repanda may require secondary factors or physical environmental conditions that are not required for inhibition of N. obtusifolia pollen tubes, and were not reproduced by the TTE in the pollen tube growth assay.
The PELPIII protein was identified in N. tabacum and shown to be soluble and localized specifically to the TT (Goldman et al., 1992; Bosch et al., 2001; de Graaf et al., 2003). After pollination, PELPIII was shown to be translocated from the TT extracellular matrix to the pollen tube wall and callose plugs of both N. tabacum and N. obtusifolia pollen tubes (de Graaf et al., 2003, 2004). When pollen tubes were cultured in vitro in the presence of purified PELPIII, PELPIII was still translocated into the pollen tube wall and maintained the same localization pattern, indicating that other TT factors may not be required for this movement (Bosch et al., 2003). The uptake of PELPIII by N. obtusifolia pollen, which does not accumulate PELPIII mRNA or protein in its own style (de Graaf et al., 2004), further implies involvement of PELPIII in II. The protein has both extension-like and arabinogalactan protein-like properties. Extensins function in cross-linking and strengthening cell walls, with a potential function in growth and defense (Wilson and Fry, 1986; Buchanan et al., 2000). Arabinogalactan proteins function in numerous biological processes (Seifert and Roberts, 2007), and have a reported function in pollen tube growth (TTS; Wu et al., 1995). The function of PELPIII in slowing and arresting the pollen tube growth of N. obtusifolia and N. repanda may be caused by an incompatible interaction with the pollen tube wall. de Graaf et al. (2004) determined the presence of PELPIII in 12 diverse Nicotiana spp., and detected its presence in species across four sub-genera. The protein was not detected in four of the species tested: N. paniculata, N. rustica, N. undulata and N. obtusifolia, which are in two sub-genera (de Graaf et al., 2004). The phylogenetic relationship of PELPIII proteins may provide insight into its role in speciation.
PELPIII had conserved function as an II protein for both N. obtusifolia and N. repanda, but there appears to be variation in the II response. Future studies will focus on elucidating the mechanisms by which PELPIII inhibits interspecific pollen. The role of PELPIII in speciation of Nicotiana is not clear. The presence of the PELPIII gene within the genus Nicotiana, and its correlation with II should be determined. By analyzing the phylogeny of the PELPIII gene throughout the genus, the role of PELPIII in speciation events may be determined with regard to its function as a reproductive barrier after species divergence.
Plant material and growth conditions
All plants were grown in a greenhouse (University of Minnesota, St Paul, MN; latitude 45N) under a 14 h photoperiod with supplemental lighting from 600 W high-pressure sodium lamps at photosynthetically active radiation 200 μmol m−2 sec−1 and an average temperature of 24.4°C. N. tabacum ‘Samsun’ wild-type, TT-ablated and male-sterile (MS) lines were as previously described (Eberle et al., 2012; Gardner et al., 2009). Male-sterile flowers were used for ‘normal’ style pollinations, referring to the presence of a normal TT in the style. Wild-type N. tabacum ‘Samsun’ plants were used as the N. tabacum pollen source in pollinations of normal and TT-ablated ‘Samsun’ styles. Seeds of N. obtusifolia (synonym Nicotiana trigonophylla; plant introduction [PI] 555543) and N. repanda (PI 555551) were obtained from the Germplasm Resources Information Network (http://www.ars–grin.gov/). The seed from N. sylvestris was a laboratory stock seed originally acquired from Thompson & Morgan (http://www.thompson-morgan.com/).
PELPIII antisense N. tabacum ‘Petit Havana’ SR1 lines C81, C64 and F81 were developed by Maurice Bosch (Bosch et al., 2003), and seed was provided by Bruce McClure (University of Missouri, Columbia, MO). Wild-type N. tabacum ‘Petit Havana’ SR1 (wild-type) seed was provided by Celestina Mariani (University of Nijmegen, The Netherlands). Wild-type and antisense ‘Petit Havana’ SR1 plants were emasculated prior to pollen shed. Wild-type ‘Petit Havana’ SR1 plants were used as the N. tabacum pollen source in pollinations of wild-type and antisense ‘Petit Havana’ SR1 styles.
Normal and TT-ablated N. tabacum ‘Samsun’ styles were pollinated, after stigma removal, as described by Eberle et al. (2012). Injections of TT-ablated styles were performed prior to pollination. Pollen was collected and either mixed with trilinolein (Wolters-Arts et al., 1998) to a concentration of 20 000 pollen grains/μl (concentration determined using a hemocytometer) and applied to the cut-style surface, or trilinolein was first applied to the cut style followed by addition of pollen. Pollinations of N. tabacum ‘Petit Havana’ SR1 lines were all performed on the stigma using 1 μl pollen suspended in trilinolein at 20 000 pollen grains μl−1. Pollinated flowers were incubated at room temperature (23°C) for 40 h unless otherwise noted. Developmentally staged pollinations of N. tabacum ‘Samsun’ normal and TT-ablated flowers were performed at five floral developmental stages (15 mm, 25 mm, 35 mm, pre-anthesis and post-anthesis). Developmental stage flower measurements were taken from the receptacle to the end of the corolla during flower collection. The style length at each developmental stage was shorter than the whole-flower measurement. Pollinations were performed after removal of the corolla, sepals and anthers for access to the pistil. The stigma was removed with a razor blade 1 mm below the stigma, and 0.5–1.0 μl trilinolein was added to the cut style surface. Pollen from N. tabacum or N. obtusifolia was added to the trilinolein. After pollination, styles were fixed and pollen tube lengths were measured from the cut end of the style to the pollen tube growth front for 10 pollinated styles per treatment (Figure S1). Styles with fewer than 10 visible pollen tubes were not included in mean pollen tube length calculations. Mean separations were performed using Tukey's honestly significant difference (HSD) test (α = 0.05).
Transmitting tract eluate from normal N. tabacum styles was prepared as described by Wang et al. (1993) from longitudinally bisected styles incubated in citric acid extraction buffer [84 mm citric acid, 2 mm Na2S2O5 pH 3.0, 1.5% w/v polyvinylpolypyrrolidone, 0.05% v/v protease inhibitor cocktail (Sigma-Aldrich; http://www.sigmaaldrich.com/united-states.html)] at 0°C for 2 h. The buffer with eluted material was centrifuged at 10 000 g for 10 min, and the supernatant was concentrated using a 3 kDa (Figure 3) or 50 kDa (Figure 4) Centricon concentrator (Millipore/Amicon; http://www.millipore.com/). The 50 kDa filter was used after it had been determined that the TTE activity was retained in the higher-molecular-weight filter (data not shown). A buffer exchange was accomplished by diluting the concentrated TTE 1:50 with control medium (modified complete medium, MCM) (Eberle et al., 2012), or 15 mm MES, pH 6.0, followed by reconcentration using the same filter. The TTE that was retained in the filter was used in the pollen tube growth assay.
Whole style extracts were produced from 10 post-anthesis pistils (with stigma and ovaries removed) of wild-type and antisense N. tabacum ‘Petit Havana’ SR1. Styles were placed in liquid nitrogen and ground in a pre-chilled mortar and pestle, followed by extraction with citric acid extraction buffer. Protein was precipitated in 80% v/v ice-cold acetone, centrifuged for 15 min at 10 000 g at 4°C, and the protein pellet was resuspended in 62.5 mm Tris/HCl pH 6.8, 2% SDS, 10% glycerol solution. Protein concentration was determined using the BCA protein assay kit (Thermo Scientific; http://www.thermofisher.com/global/en/home.asp) according to the manufacturer's instructions.
The 50 kDa TTE in 20 mm MES pH 6.0 was fractionated on a Mono S™ 5/50 GL (Tricorn™) cation exchange column (GE Healthcare; http://www.gelifesciences.com/webapp/wcs/stores/servlet/Home/en/GELifeSciences/US/). Proteins were eluted in 5 ml of 20 mm MES pH 6.0, followed by a linear gradient of 0.0–0.4 m NaCl over 30 ml and 0.4–1 m NaCl over 10 ml in 20 mm MES pH 6.0 buffer at 1 ml min−1. Fractions were collected in 10 ml volumes. Fractions 1, 2 and 3 contained proteins eluted using 0–0.15, 0.15–0.25 and 0.25–0.4 m NaCl, respectively. Fractions were concentrated and buffer-exchanged into 15 mm MES buffer using a 50 kDa Centricon concentrator.
SDS–PAGE gels and protein immunoblots
Protein samples were loaded at 40 μg (whole style extracts) or 10 μg (TTE), and separated by SDS–PAGE using PAGEr™ Gold 4–20% Tris-glycine gels according to the manufacturer's instructions (Lonza; http://www.lonza.com/). Proteins were visualized using a Coomassie-based Imperial protein stain according to the manufacturer's instructions (Thermo Scientific). Separated proteins were electrophoretically transferred to poly(vinylidene difluoride) membranes in buffer containing 192 mm glycine, 25 mm Tris and 0.05% SDS. Membranes were blocked using Tris-buffered saline (20 mm Tris, 500 mm NaCl, pH 7.5) containing 0.2% non-fat dry milk (Bio–Rad; http://www.bio-rad.com/) for 45 min at 23°C. Membranes were washed with gentle agitation at 23°C for 5 min in a solution of Tris-buffered saline + 0.1% Tween–20 (TBST). After washing, membranes were incubated with antibodies against PELPIII, 120 K or TTS (kindly provided by B. McClure) (Cruz-Garcia et al., 2005) in blocking solution for 90 min, rinsed twice in TBST, followed by incubation with alkaline phosphatase-conjugated goat anti-rabbit antibody (Bio–Rad Immun-StarTM Goat Anti-Rabbit alkaline phosphatase) for 60 min at 23°C. Membranes were rinsed three times in TBST, and detected using an Immun-StarTM chemiluminescent substrate (Bio–Rad) by exposure to X–ray film.
We thank Bruce McClure, Department of Biochemistry, University of Missouri and Celestina Mariani, Department of Molecular Plant Physiology, University of Nijmegen for providing N. tabacum ‘Petit Havana’ SR1 seed and 120 K, TTS and PELPIII antibodies. This project was supported by the US National Science Foundation (grant number IOS-0920114), the Minnesota Agricultural Experiment Station, the University of Minnesota Plant Biological Sciences Graduate Program, and the University of Minnesota Microbial Plant and Genomics Institute at the University of Minnesota.