Pollen Proteins Bind to the C-terminal Domain of Nicotiana alata Pistil Arabinogalactan Proteins*

Pollen tube growth is influenced by interaction between pollen proteins and the pistil extracellular matrix. The transmitting tract-specific glycoprotein (NaTTS) and 120-kDa glycoprotein (120K) are two pistil arabinogalactan proteins (AGPs) that share a conserved C-terminal domain (CTD) and directly influence pollen tubes in Nicotiana alata. 120K and other extracellular matrix proteins are taken up and transported to vacuoles of growing pollen tubes. We hypothesize that signaling and trafficking processes inside pollen tubes are important for controlling pollen tube growth. We performed a yeast two-hybrid screen of pollen cDNAs using sequences from 120K and NaTTS as baits. We found that an S-RNase-binding protein (SBP1), a C2 domain-containing protein (NaPCCP), and a putative cysteine protease bound to the AGP baits. SBP1 from Petunia hybrida and Solanum chacoense is a putative E3 ubiquitin ligase that binds to S-RNase and other proteins. C2 domain-containing proteins bind lipids and can regulate myriad cellular processes. Cysteine proteases are often associated with the degradation of vacuolar proteins. Expression analysis revealed that transcripts for these proteins are expressed in mature pollen. NaPCCP and NaSBP1 were characterized further because of their potential roles in signaling and trafficking. In vitro pull-down assays verified binding between maltose-binding protein (MBP) fusions, MBP::NaPCCP or MBP::NaSBP1 and glutathione S-transferase (GST), GST::AGP CTD fusions. NaSBP1 binds to the AGP CTDs through its helical and RING domains. NaPCCP binds through its C-terminal region. Binding between NaPCCP and NaSBP1 and the pistil AGPs may contribute to signaling and trafficking inside pollen tubes growing in planta.

Angiosperm sexual reproduction relies on the transfer of pollen to a stigma, followed by pollen hydration and germination and then growth of a pollen tube through the pistil. Pollen tubes grow through the pistil extracellular matrix (ECM) 2 toward the ovary, where fertilization occurs. Pollen tubes are in constant contact with the ECM as they grow. Therefore, pollen tube success requires interaction with as well as interpretation of chemical signals provided by the pistil in the ECM (1,2).
Pollen tubes elongate through regulated and coordinated growth pulses. These involve coordination of physiological processes including exocytosis of wall materials and enzymes required for cell wall synthesis (3), endocytosis and processing of ECM components through the pollen tube endomembrane system (4), cytoskeletal functions (5,6), intracellular calcium oscillations (7), and regulation of Rho family GTPases (8).
Pollen tubes internalize ECM materials as they grow through the pistil (9,10). Interaction between pollen tubes and pistil proteins can affect pollen tube growth either positively or negatively. For example, pistils promote tube growth, because growth occurs much faster in planta than in culture (11). On the other hand, pistils also can provide molecules inhibitory to pollen tube growth, such as self-incompatibility (SI) proteins (12). Indubitably, interactions between pollen proteins and pistil ECM factors must contribute to the control of pollen tube growth. However, little is known about how pollen tubes take up and process ECM materials or about how signaling pathways inside pollen tubes are affected by interactions with ECM components.
Several pollen-pistil interactions have been investigated using pollen tube attraction assays carried out in culture and using molecular approaches to detect protein-protein interactions. Chemocyanin, a small, basic protein present on lily stigmas, attracts pollen tubes (13). The transmitting tract-specific glycoprotein (TTS) in Nicotiana species serves as a guidance molecule for pollen tubes (14,15). The S-RNase-binding protein (SBP1) in Petunia pollen was found to interact with a pistil SI protein, S-RNase, in yeast two-hybrid and pull-down assays (16). S-RNase was also found to interact with the SLF (SI-specific pollen S-locus F-Box) protein (17). In tomato pollen tubes, yeast two-hybrid experiments were used to identify the interaction between the extracellular domain of a pollen-specific protein kinase (LePRK) and a small cysteine-rich protein, LeSTIG1, expressed in the pistil (18).
As noted, pistils can provide molecules that inhibit the growth of undesirable pollen tubes. Studies of SI systems in several plant families have provided direct evidence that pollenpistil interactions affect the success of pollination. SI mecha-nisms are diverse, but they all entail active rejection of closely related pollen (12). In Brassica, for example, interaction of the pollen-specific S-locus cysteine-rich ligand with the stigmatic S-receptor kinase determines compatibility. Signaling pathways in the pistil papillar cells allow acceptance or rejection of individual pollen grains; self-pollen is either prevented from hydrating or its growth halted shortly after germination (19). In Papaver, the stigma produces SI proteins that, upon contact with self-pollen tubes, induce a programmed cell death response (20). SI species in Solanaceae display S-RNase-based SI; pollen rejection is controlled by S-RNase proteins expressed in the pistil and SLF proteins expressed in the pollen (12,21).
Nicotiana pistil cells secrete abundant arabinogalactan proteins (AGPs) into the ECM. Some of these AGPs promote pollen tube growth, and others are associated with inhibitory pathways. The TTS protein, for example, is thought to contribute positively to pollen tube growth by providing a chemical guidance signal (14). Pollen tubes deglycosylate TTS and possibly use the arabinogalactan as a source of energy (15). TTS displays higher glycosylation levels at the base of the style near the ovary (15,22), and this may contribute to guidance. The 120-kDa arabinogalactan protein (120K) implicated in the SI response (23) is taken into growing pollen tubes (4,24), where it accumulates on a pollen tube vacuole membrane or in the cytoplasm (4).
Although it is clear that pollen tube growth through the pistil is required for fertilization and that growth is influenced by interplay between pollen proteins and pistil ECM factors, few of these interactions have been characterized at the molecular level. Here, the goal was to identify pollen proteins that interact with pistil ECM proteins, thus identifying proteins that could contribute to pistil-pollen signaling. We performed a yeast twohybrid screen of pollen cDNAs using the C-terminal domains (CTDs) of NaTTS and 120K as baits. Three proteins were identified: an S-RNase-binding protein (NaSBP1), a pollen-specific C2 domain-containing protein (NaPCCP), and a putative cysteine protease (CP).

Plant Materials and Growth Conditions
Nicotiana plumbaginifolia (43B), Nicotiana tabacum cv Praecox (Inventory number TI-1347), and Nicotiana alata S A2 S A2 have been described previously (25). Herbarium voucher specimens are located at the Dunn-Palmer Herbarium at the University of Missouri (Columbia, MO). Vouchers were prepared from putative species: "Rastroensis" (UMO 190742

Gene Constructs
GST Fusions-Sequences encoding the CTDs of 120K and NaTTS were amplified using primers designed to introduce an upstream BamHI site and a downstream SmaI site. Primers are reported in supplemental Table S1. The PCR products were cloned into pGEX 4T-3 and sequence-verified.
Yeast Two-hybrid Constructs-NaTTS and 120K CTD sequences were cut out of the pGEX-4T-3 constructs with SalI and BamHI and cloned into the pGBKT7 DNA-binding domain vector from the Matchmaker TM two-hybrid library construction and screening kit (Clontech, Mountain View, CA). Following the manufacturer's directions, a cDNA library, made from N. alata pollen, was cloned into the activation domain plasmid supplied with the kit, pGADT7-rec. After the NaSBP1 sequence was acquired, a full-length clone was inserted into the pGADT7 activation domain plasmid using EcoRI digests.
MBP Constructs-Full-length NaPCCP and NaSBP1 were amplified from an N. alata S A2 S A2 pollen cDNA library and cloned into pGEM-T-Easy (Promega, Madison, WI). NaPCCP and NaSBP1 sequences were cloned into pMAL-C2-x (New England Biolabs, Ipswich, MA) as EcoRI-HindIII fragments. The primers used to amplify and clone the MBP fusions are listed in supplemental Table S1. Construct sequences were confirmed by DNA sequencing.

Yeast Two-Hybrid Assay, Clone Identification, and Sequencing
A yeast two-hybrid screen was performed using NaTTS and 120K CTD baits. Yeast cotransformations were performed in AH109 cells, according to the manufacturer's instructions (Matchmaker TM ; Clontech). Transformed cells were plated on SD 4DO medium (ϪAde, ϪTrp, ϪLeu, ϪHis) plates with X-␣-gal (5-bromo-4-chloro-3-indolyl-␣-D-galactopyranoside). Each bait screened 2.2 ϫ 10 6 cotransformants. Plasmids were purified and sequenced from positive clones. The interactions were confirmed by cotransforming both bait and prey plasmids into yeast strain Y187 and then growing them on SD ϪTrp ϪLeu to select for cotransformants. Cells from each cotransformation were streaked onto SD 4DO medium with X-␣-gal. NaSBP1, identified in the two-hybrid screen, was sequenced and cloned; poly(A) RNA was isolated from N. alata pollen, and then cDNA was prepared using the SMART RACE kit (BD Biosciences, Palo Alto, CA). NaSBP1 was amplified using the following oligodeoxynucleotides: SBP1 RACE antisense, 5Ј-CAC GCT CAA TAT CAT CAC CCA CAA GCC CAA G-3Ј; and SBP1 RACE sense, 5Ј-TAC TAG GAG GAA CAT GGA GCT TG-3Ј. A full-length clone was recovered, cloned into pGEM-T-Easy, and sequenced. The NaPCCP recovered from this screen was cloned and sequenced from N. alata cv Breakthrough (27) cDNA.

Fusion Protein Expression and Pull-down Assays
GST fusion protein expression vectors were transformed into Escherichia coli strain BL21 (B2685; Sigma), and MBP fusion protein expression vectors were transformed into E. coli strain TB1 (New England Biolabs). Transformed cells were grown in 2ϫ YT medium (28), induced with isopropyl-1-thio-␤-D-galactopyranoside at 16.5°C and shaken at 250 rpm for 4 h. The cells were washed (phosphate-buffered saline for GST fusions, column buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA) for MBP fusions), and lysed with a French press at 1,200 p.s.i. (29). Lysates were cleared by centrifugation (15,000 rpm, 25 min, F13S rotor (Piramoon Technologies, Santa Clara, CA) at 4°C) and then incubated with reduced glutathione-agarose (Sigma) or amylose resin (New England Biolabs) for isolation of GST or MBP fusions, respectively (4 h at 4°C). Pull-down assays were performed with GST or GST fusion proteins bound to reduced glutathione agarose after washing five times in phosphate-buffered saline. For consistent results, bound fusion proteins were used within 2 days of preparation. MBP fusions on amylose resin were washed five times with column buffer and then eluted with column buffer with 10 mM maltose. MBP fusions (2 g) were added to immobilized GST fusions in 1 ml (50 mM Tris-HCl, pH 7.2, 200 mM NaCl, 10 g/ml bovine serum albumin, 5 mM EDTA) and incubated for 2 h at room temperature in a tube rotator. After centrifugation (6,000 rpm, 1 min at room temp in an F45-24-11 rotor (Piramoon Technologies)), the beads were kept on ice and washed five times for 5 min each in 50 mM Tris-HCl, pH 7.2, 200 mM NaCl, 5 mM EDTA, 0.1% (w/v) Triton-X-100. The washed beads were boiled in 4ϫ loading buffer (0.25 M Tris, pH 6.8, 8% SDS (w/v), 40% glycerol (v/v), 200 mM dithiothreitol, 0.004% bromphenol blue) for 3 min and separated in 10% Tris-Tricine gels (30). Proteins were transferred to nitrocellulose and immunostained with anti-MBP (E8030S, New England Biolabs), anti-NaTTS, anti-GST, or anti-120K antibodies as described (31). All of the binding experiments were repeated in triplicate.

Identification of Pollen-specific Pistil AGP-binding Partners-
Yeast two-hybrid assay was utilized to identify pollen proteins that interact with pistil proteins known to impact pollen tube growth. The CTDs were chosen as baits because of their sequence conservation and apparent lack of glycosylation. Each bait was screened against 2.2 ϫ 10 6 N. alata pollen cDNAs. We recovered 74 clones that grew on selective media and were galactosidase-positive; these clones were then purified and sequenced.
Three classes of prey clones interacted with both NaTTS and 120K CTD baits: a putative cysteine protease, a C2 domain-containing protein, and an orthologue of the Petunia hybrida S-RNase-binding protein (PhSBP1) (16). Interactions were verified in the yeast strain Y187 grown on limiting media with X-␣-gal (Fig. 2). Positive interactions grew and turned blue, whereas growth was inhibited in cells with no interaction. Cells harboring the p53 and SV40-positive control plasmids showed robust growth, and negative control transformations did not exhibit galactosidase activity (NaSBP1) or grow at all.
NaSBP1 was recovered twice, once with each bait. Both clones contained the helical region in the middle of the protein; one of the clones included the C-terminal C3HC4 RING domain of SBP1 as well (supplemental Fig. S1). Full-length SBP1 was cloned into pGADT7 and then used to verify the interaction (Fig. 2). supplemental Fig. S1 shows that SBP1 proteins are highly conserved (88%) and possess a glutamine-rich region in their N-terminal domains. The Arabidopsis thaliana orthologue is 63% conserved and does not contain the glutamine-rich region (supplemental Fig. S1).
NaPCCP (a pollen-specific C2 domain-containing protein) was recovered once with 120K and three times with NaTTS. NaPCCP cDNAs were amplified and sequenced from Nicotiana section Alatae spp. and N. tabacum pollen using genespecific primers. Sequence analysis shows that PCCP is a 188amino acid, 21-kDa protein that is very highly conserved among Nicotiana species (97% sequence identity across all species). NaPCCP contains a 29-amino acid N-terminal domain, a C2 domain consisting of 82 residues, and a C-terminal domain 77 residues long (supplemental Fig. S2). All NaPCCP clones recovered in the yeast two-hybrid screen contained the C-terminal region of the protein. One of the four original NaPCCP clones contained the C2 domain as well. A putative cysteine protease was also recovered in the yeast two-hybrid screen; twice with the 120K CTD and once with the NaTTS CTD.
Pollen-specific Expression-We performed reverse transcription-PCR to determine expression patterns of the three sequences (CP, NaPCCP, and NaSBP1) recovered in the yeast two-hybrid screen. An actin primer set was used as a positive control. Single bands with the expected sizes were recovered for NaSBP1 and NaPCCP PCRs. Although the CP primer set yielded a gene-specific product of the expected size (Fig. 3), several nonspecific products were also recovered, which is consistent with a family of similar sequences (data not shown).
Expression profiles were investigated using a variety of N. alata organs (Fig. 3A), developing reproductive organs (Fig.  3B), and from pollen of nine Nicotiana species (Fig. 3C). Fig. 3A shows that CP, NaPCCP, and NaSBP1 transcripts are expressed in N. alata pollen. The CP is predominantly expressed in pollen but is detectable in every organ surveyed. NaPCCP cDNA is only detectable in pollen. NaSBP1 PCRs were amplified for more cycles because of low expression levels previously reported in Petunia (16). The NaSBP1 transcript is expressed at low levels in every organ examined.
Developmental expression patterns were assayed from both male and female reproductive organs from four stages of flowers, ranging from buds to mature flowers. Fig. 3B shows that both NaPCCP and the CP transcripts are present at the highest  levels in mature pollen but are not detectable in developing pistil tissues. NaSBP1 is expressed at a low level at all stages of development. Fig. 3C shows that NaPCCP and NaSBP1 transcripts are expressed in similar amounts in mature pollen of the Nicotiana species investigated. Both transcripts are present in closely related SI and self-compatible Nicotiana section Alatae species and the more distantly related species, N. tabacum. The CP transcript is detectable in pollen from all but two of the section Alatae species, N. bonariensis, and Rastroensis. The cysteine protease was not investigated further because its expression pattern is not consistent with a specific role in pollination and because of the ambiguity caused by expression of multiple similar genes. Taken together, the yeast two-hybrid assay and reverse transcription-PCR results suggest that NaPCCP and NaSBP1 are good candidates for pollen proteins with a meaningful interaction with the 120K and NaTTS CTDs.
NaPCCP and NaSBP1 Bind to 120K and TTS CTDs-Pulldown assays verified binding of NaPCCP and NaSBP1 to the 120K and NaTTS CTDs. For these assays, NaPCCP and NaSBP1 were expressed as MBP fusions and pulled down with immobilized GST fusions of the 120K or NaTTS CTD. These assays were conducted in the presence of excess bovine serum albumin to prevent nonspecific binding. Fig. 4A shows the fulllength, single domains (N-ter, C2, C-ter) and domain combinations of NaPCCP that were expressed as MBP fusions. Fig. 4B shows that full-length NaPCCP protein binds to the CTD of both NaTTS and 120K; NaPCCP was pulled down with GST::TTS and GST::120K but not with GST alone. Likewise, MBP alone does not interact with GST::TTS or GST::120K.
We performed pull-down assays using MBP fusions with single domains and domain combinations to determine which regions of NaPCCP mediate these interactions. Fig.  4C shows that the C-terminal domain of NaPCCP is sufficient for binding, whereas the C2-and N-terminal domains are not. However, all combinations of NaPCCP domains interact with both the 120K and NaTTS CTDs (Fig. 4D). The C2 ⌬ and N-ter ⌬ MBP fusions contain the C-terminal domain of NaPCCP and would be expected to bind. The C-ter ⌬ binding was unexpected.
NaSBP1 also interacts with the CTDs of both NaTTS and 120K (Fig. 5). Fig. 5A shows full-length NaSBP1, individual domains (N-ter, Helical, and RING), and combinations of domains that were expressed as MBP fusions. Like NaPCCP, full-length NaSBP1 binds to both the CTDs but not to GST alone (Fig. 5B).  OCTOBER 3, 2008 • VOLUME 283 • NUMBER 40

JOURNAL OF BIOLOGICAL CHEMISTRY 26969
To determine which domain(s) of NaSBP1 mediates the interaction, the individual domains of NaSBP1 (Fig. 5C) and combinations of domains (Fig. 5D) were expressed as MBP fusions and used in pull-down assays. The results show that the RING and helical domains are sufficient for binding to 120K and NaTTS CTD. Consistent with the individual domain interaction experiments, all combinations of NaSBP1 domains were pulled down by the GST-CTD fusions. This was expected because all NaSBP1 domain combination MBP fusions contained either the RING or helical domains that are sufficient for interaction alone.

DISCUSSION
We used yeast two-hybrid assay to identify pollen proteins potentially involved in deciphering chemical signals provided by the pistil. The screen utilized two pistil AGPs (NaTTS and 120K) known to interact with growing pollen tubes (4,15). These AGPs possess similar CTDs (Fig. 1) that, we speculate, may contribute to their ability to interact with pollen tubes. 120K localizes to the lumen and vacuolar membranes in N. alata pollen tubes in planta (4). NaTTS has been observed associating with the pollen tube wall (15). Although it has not been directly demonstrated, we infer that 120K enters the pollen through endocytosis and is then trafficked to a vacuole. Pollen proteins that interact with 120K are potentially involved in trafficking or SI. However, because pollen tubes need to metabolize pistil ECM materials for use in pollen tube growth, they also may be involved in metabolic pathways.
We identified three pollen proteins that bind to the 120K and NaTTS CTDs: a putative cysteine protease (CP), a C2 domaincontaining protein (NaPCCP), and an S-RNase-binding protein NaSBP1 (Fig. 2). All three proteins are expressed in mature pollen (Fig. 3) and therefore could interact with pistil ECM components. However, we chose not to investigate the cysteine protease further because of ambiguities introduced by the presence of a CP gene family and because CP expression was not detected in all species of interest (Fig. 3C).
In vitro pull-down assays were used to verify the interactions observed in yeast (Figs. 4 and 5). Each of the pollen proteins identified contains at least one cysteine residue, and the AGP CTD baits contain six cysteine residues. However, although the reducing conditions in the yeast two-hybrid and pull-down assays are very different, binding was observed in both assays. Moreover, pull-down assays contained excess bovine serum albumin, a cysteine-rich protein, as a competitor. Together, the yeast two-hybrid and pull-down assays provide good evidence for binding between the AGP CTD and pollen proteins.
Implications of NaSBP1 Binding to Pistil AGPs-Our results showing that NaSBP1 binds to the CTD of pistil AGPs (Fig. 5) suggest a potential link with ubiquitylation-regulated processes in pollen tubes. NaSBP1 is a RING domain protein (supplemental Fig. S1). This domain is often present in E3 ubiquitin ligase enzymes that catalyze polyubiquitylation of proteins targeted for degradation by the 26 S proteasome (32)(33)(34). Many biological processes (from cell cycle regulation to transcription factor processing) are regulated by degradation of polyubiquitylated proteins (35,36). E3 ligases can function alone or in a SCF complex (33). A canonical SCF complex, one of the best known E3 ligase complexes, is comprised of Skp1, a protein that recruits the F-box protein, Cul1, and an F-box protein, which provides specificity for the ubiquitylation substrate (32). Ubiquitylation can also provide an endocytic trafficking signal for transmembrane receptor proteins and proteins in the trans-Golgi destined for delivery to a vacuole (37,38). RING domain proteins have been associated with endocytosis of receptors. For example, the RING domain of Neuralized protein in Drosophila, for instance, is associated and required for endocytosis of Notch ligands in eye development (39 -41).
SBP1 is implicated in the Petunia S-RNase-based gametophytic SI pollen rejection system (17). This pollen rejection system is controlled by the S-locus; pollen is rejected when its S-haplotype matches either of the S-haplotypes of the diploid pistil (12). The S-locus encodes two specificity genes, SLF, which is expressed in pollen (21), and S-RNase, which is expressed in the pistil (42). Pollen recognition in SI is mediated by interaction between SLF and S-RNase inside growing pollen tubes, resulting in growth inhibition of self-pollen tubes. PhSBP1 was originally identified in a yeast two-hybrid screen for S-RNase-binding proteins (16). In Petunia inflata, PiSBP1 interacts with PiSLF, S-RNases, PiCul1-G, and an E2 conjugating enzyme (17). Hua and Kao (17) suggested that an SBP1-containing complex is directly involved in protecting pollen tubes from nonself S-RNase in the cytosol by targeting it for ubiquitylation and degradation. As of yet, there is no direct evidence for ubiquitylation of S-RNase by an SBP1-containing complex.
We showed that NaSBP1 binds to AGPs from the pistil ECM (Figs. 2 and 5). However, we also found that NaSBP1 interacts with the p53 control prey plasmid, although the interaction is not sufficient for expression of the colorimetric reporter (Fig.  2). Others have shown that SBP1-like proteins bind to unrelated proteins (43). This, together with the ubiquitous expression of NaSBP1 (16,44) and SBP1-like proteins, suggests that their function is not restricted to pollination. The expression and binding results are consistent with housekeeping functions such as degradation of misfolded protein or targeting proteins to various cellular locations. SBP1-containing complexes may be involved in the general removal of misfolded proteins in many types of plant cells. Clearly, more information is needed about how pistil proteins are taken up, processed, and transported in pollen tubes before the potential role of binding between NaSBP1 and pistil proteins can be understood. Nevertheless, the possibility that NaSBP1 could be involved with the processing or trafficking of pistil AGPs is intriguing.
Several ubiquitin-dependent trafficking pathways have been described involving different E3 ubiquitin ligases (37,38). Other SCF components, such as the yeast F-box protein Rcy1p, are involved in endocytic membrane trafficking and recycling (45). Rcy1p associates with Skp1, but it does not form a canonical SCF complex. Instead, it associates with a SNARE protein, Snc1p (46). Perhaps, the noncanonical SCF complex comprised of SBP1, SLF, and Cul1-G, reported by Hua and Kao (17), could participate in trafficking of endocytic cargo. Like Rcy1p, this pollen SBP1/SLF-containing complex could be directly involved in trafficking of endocytic vesicles. Alternatively, the complex could provide a ubiquitin tag to its substrates, targeting them to the multivescular body and, ultimately, a vacuole. Although there is no direct evidence of a SBP1 complex-linked trafficking pathway, a bulk of the S-RNase and 120K is trafficked to vacuoles in compatible pollen tubes (4).
Although the interaction between the AGP CTDs and NaSBP1 presents a topological problem, it also presents possibilities for a role in pollen-pistil interactions. Presumably, SBP1 (like other E3 ubiquitin ligases) resides in the cytosol. Thus, binding requires the AGPs to be taken into growing pollen tubes and the CTD to be presented to the cytosolic compartment. In this regard, it is significant that immunolocalization results using an antibody against the 120K CTD show localization to the tonoplast (4). If the highly basic CTD penetrates the membrane, NaSBP1 could recruit a pollen SBP1-complex (17) and potentially mediate an S-specific interaction between SLF and S-RNase.
Implications of NaPCCP Binding to Pistil AGPs-NaPCCP is implicated in intracellular transport of pistil AGPs because of its binding to the ECM AGP CTDs (Figs. 2 and 4). Our results show that the NaPCCP C-terminal domain is sufficient for CTD binding. The C2 domain comprises much of the NaPCCP sequence (supplemental Fig. S2). C2 domains are modular, lipid-binding domains (47) found in a variety of proteins with functions that include vesicular transport, GTPase regulation, lipid modification, protein phosphorylation, and ubiquitylation (47)(48)(49)(50)(51). The lipid binding properties of many C2 domains are regulated by cytosolic calcium (49). However, there are examples of the Ca 2ϩ -independent C2 domain binding to polyphosphoinositide lipids (52,53).
Many aspects of pollen tube growth could be affected by a C2 domain-containing protein such as NaPCCP that interacts with pistil AGPs. We speculate that NaPCCP plays a role in the trafficking of the pistil AGPs through the pollen endomembrane system. 120K is found on the membrane and in the lumen of vacuoles in pollen tubes (4). For 120K to arrive at a vacuole, it must travel from the ECM through the pollen tube endomembrane system. Recent work has highlighted two distinct endocytic pathways in pollen tubes (54,55). Both clathrin-coated and noncoated vesicle endocytic pathways exist; clathrincoated vesicles traffic to the trans-Golgi, whereas noncoated vesicles travel to the multivescular body and a vacuole. However, there is evidence that these two pathways are not mutually exclusive, suggesting that there is some degree of sorting of endocytic vesicles and their cargo (54). If endocytic vesicles contain many molecules with different destinations, these molecules need to be sorted consistently if they are to reliably arrive to their final destination. Thus, NaPCCP may serve as a sorting molecule to help deliver 120K to the vacuole. Some C2 domain proteins, including protein kinase C␣, reside in the cytosol but bind membranes only when a threshold Ca 2ϩ concentration is reached (56). If NaPCCP binds to endocytic membranes, it could potentially coordinate endocytic traffic with the oscillating tip-localized calcium fluctuations associated with pollen tube growth (57).
Sequence analyses of PCCP cDNAs from self-compatible and SI species in section Alatae and N. tabacum show that PCCP sequences are highly conserved; no consistent differences between self-compatible and SI species were observed (supplemental Fig. S2). Although PCCP may be involved with the general trafficking of the SI factor 120K through the endomembrane system, this function is most likely not unique to SI. Clearly, the yeast two-hybrid and pull-down experiments reported here require follow-up to establish biological function. Experiments testing the lipid binding potential of NaPCCP and its association with the pollen tube endomembrane system are planned.
Pollen uptake and processing of pistil ECM proteins suggests that pollen-pistil signaling can occur inside pollen tubes. For example, SI recognition interactions most likely occur in the pollen tube cytoplasm, because pollen SLF is a cytoplasmic protein. Less is known about other systems that affect pollen tube growth, but it appears that pistil proteins that effect tube growth both positively and negatively are taken up in bulk. Little is known about how pollen tubes internalize these pistil proteins or how they are targeted to endomembrane or cytosolic compartments. NaSBP1 and NaPCCP are good candidates for mediating these processes because they clearly bind to NaTTS and 120K and have potential roles in ubiquitylation and the pollen tube endomembrane system, respectively. Clearly, the in vitro studies reported here need to be extended to determine whether the interactions we describe are important in pollen tubes growing in planta.