Elaborate spatial patterning of cell-wall PME and PMEI at the pollen tube tip involves PMEI endocytosis, and reflects the distribution of esterified and de-esterified pectins

Authors


(fax +49 6221 545859; e-mail sgreiner@hip.uni-heidelberg.de).

Summary

In dicots, pectins are the major structural determinant of the cell wall at the pollen tube tip. Recently, immunological studies revealed that esterified pectins are prevalent at the apex of growing pollen tubes, where the cell wall needs to be expandable. In contrast, lateral regions of the cell wall contain mostly de-esterified pectins, which can be cross-linked to rigid gels by Ca2+ ions. In pollen tubes, several pectin methylesterases (PMEs), enzymes that de-esterify pectins, are co-expressed with different PME inhibitors (PMEIs). This raises the possibility that interactions between PMEs and PMEIs play a key role in the regulation of cell-wall stability at the pollen tube tip. Our data establish that the PME isoform AtPPME1 (At1g69940) and the PMEI isoform AtPMEI2 (At3g17220), which are both specifically expressed in Arabidopsis pollen, physically interact, and that AtPMEI2 inactivates AtPPME1 in vitro. Furthermore, transient expression in tobacco pollen tubes revealed a growth-promoting activity of AtPMEI2, and a growth-inhibiting effect of AtPPME1. Interestingly, AtPPME1:YFP accumulated to similar levels throughout the cell wall of tobacco pollen tubes, including the tip region, whereas AtPMEI2:YFP was exclusively detected at the apex. In contrast to AtPPME1, AtPMEI2 localized to Brefeldin A-induced compartments, and was found in FYVE-induced endosomal aggregates. Our data strongly suggest that the polarized accumulation of PMEI isoforms at the pollen tube apex, which depends at least in part on local PMEI endocytosis at the flanks of the tip, regulates cell-wall stability by locally inhibiting PME activity.

Introduction

Pollen tubes are a unique cell type, which exhibit extremely fast directional growth restricted to the tip. This tip growth requires massive localized incorporation of new cell-wall material. Targeted secretion at the apex, balanced by endocytic membrane retrieval at the flanks of the tip, is required for pollen tube growth, and is controlled by RAC/ROP signalling, a tip-focused cytoplasmic Ca2+ gradient and signalling lipids (Helling et al., 2006; Hepler et al., 2001; Kost et al., 1999; Parton et al., 2001; Pierson et al., 1996). The apical pollen tube cell wall, to which directional expansion is confined, is largely deprived of cellulose and hemicellulose (Ferguson et al., 1998), asserting pectins and callose as the primary load-bearing components (Taylor and Hepler, 1997). Therefore, pollen tubes are not only a well-established experimental system to study directional growth and growth-related signalling events, but are also an excellent tool to explore highly dynamic growth-related changes of pectin structure and function. Thus, although at the pollen tube tip the cell wall has to remain expandable to allow rapid growth, wall rigidity is required laterally. As enzyme-mediated changes in the esterification status of homogalacturonan (HGA), a major pectin component, are known to strongly impact on the biophysical properties of the cell wall, it is an intriguing possibility that spatial control of HGA esterification status is essential for pollen tube tip growth.

Homogalacturonan, a linear polymer composed of (1,4)-α-d-galaturonic acid residues, is synthesized in the Golgi apparatus by galacturonosyltransferases and becomes methylesterified in the same compartment (Micheli, 2001; Sterling et al., 2006). Subsequently, HGA is delivered to the extracellular space in secretory vesicles. Methyl groups can be removed by pectin methylesterase (PME, EC 3.1.1.11, CE8 of CAZy, http://www.cazy.org), leaving carboxyl groups on the pectin chain and releasing methanol and protons. It is assumed that, depending on the mode of PME action, de-esterification may affect the properties of HGA in different ways. Blockwise de-esterification of 10 or more adjacent galacturonic acid residues allows the formation of Ca2+ bridges between the free carboxyl groups of adjacent HGA chains, leading to pectin gelation and hence causing cell-wall rigidification (Willats et al., 2001). Alternatively, PMEs may also act in a non-blockwise fashion. The mode of PME activity is relevant for pectin turnover, as distinct patterns of de-esterification are necessary to render HGA a target for polygalacturonase-mediated degradation (Wakabayashi et al., 2003). Thus, PME activity has a profound impact on the mechanical properties of cell walls, and must be precisely regulated. At present, it is unclear whether individual PME isoforms are capable of both modes of action described above. Local PME activities in cells and tissues may be determined via the transcriptional control of the expression of distinct isoforms (e.g. 66 annotated Arabidopsis loci in TAIR, http://www.arabidopsis.org). In addition, the activity of different PME isoforms was shown to depend on the pH and on the availability of divalent cations (Catoire et al., 1998; Denes et al., 2000; Goldberg et al., 2001; Moustacas et al., 1991).

Recently, it became apparent that the expression of small proteinaceous PME inhibitors (PMEIs) may represent an additional control mechanism (Camardella et al., 2000; Giovane et al., 2004). PMEIs have been isolated from kiwi fruit and Arabidopsis, and the ability of two Arabidopsis PMEIs to reduce PME activity has been demonstrated in vitro and in planta (Lionetti et al., 2007; Raiola et al., 2004; Wolf et al., 2003). Meanwhile, analysis of PMEIs and PME–PMEI interactions has been extended to the structural level, and the structure of a PME–PMEI complex has been solved (Di Matteo et al., 2005; Hothorn et al., 2004). However, the precise physiological functions of PMEIs have remained unclear. Localized secretion of specific PMEIs could represent an efficient mechanism for the precise temporal and/or spatial control of PME activity. Remarkably, a large proportion (43 of 66) of the Arabidopsis PME genes are predicted to encode pre-pro-proteins, with the N-terminal pro-regions showing moderate yet significant homology to PMEIs (Pelloux et al., 2007). In contrast to these PMEs, which have been designated type I, type-II PMEs lack the pro-region. It has been proposed that the pro-region of type-I PMEs might act as an autoinhibitory domain and prevent untimely PME activity during transport in pectin-containing secretory vesicles (Micheli, 2001). Similarly, separate PMEIs may be required to block the activity of type-II PMEs within secretory vesicles. In addition, such PMEIs might also be transported independently of PMEs to certain cell-wall sites to locally modulate cell-wall biogenesis.

Different PME and PMEI isoforms are highly expressed in pollen and pollen tubes, consistent with the proposed key role of a dynamic pectin metabolism in the polar growth of this cell type (Pina et al., 2005). Recently, Arabidopsis mutants and post-transcriptional gene silencing in tobacco were employed to functionally characterize PME isoforms (Bosch and Hepler, 2006; Jiang et al., 2005; Tian et al., 2006). In general, reduced PME expression led to severe defects in pollen tube growth. Cultured Arabidopsis pollen tubes carrying mutant alleles of PME genes (vgd1, atppme1) were not only shorter than wild-type pollen tubes, but also displayed unusual shapes and reduced stability. This strongly supports the hypothesis that PME activity is necessary to rigidify the cell wall of pollen tubes (Jiang et al., 2005; Tian et al., 2006). Interestingly, whereas de-esterified HGA dominates in lateral regions of the cell wall, esterified HGA is prevalent at the pollen tube apex, where wall expansibility is required for cell growth (Bosch and Hepler, 2005; Bosch et al., 2005; Parre and Geitmann, 2005). These findings raised the following questions. Do PMEIs physically interact with PMEs in the pollen tube cell wall? If so, is this interaction essential for the control of cell-wall stability during pollen tube growth? And finally, does the spatial control of PME activity by PMEIs play a key role in establishing the pattern of the cell-wall rigidity and extensibility required for polar cell growth at the pollen tube tip?

To investigate the localization and function of pollen tube PMEIs, we transiently expressed YFP-tagged versions of AtPMEI1 and AtPMEI2 (Raiola et al., 2004; Wolf et al., 2003), and of the pollen tube type-II PME isoform AtPPME1 (Tian et al., 2006), in Nicotiana tabacum pollen tubes. A type-II PME isoform was opted for as the processing required for type-I PME maturation would have added an additional level of complexity. Tobacco pollen tubes were chosen as an experimental system because they grow well in culture, and can be transformed effectively. In marked contrast to the uniform distribution of AtPPME1:YFP throughout the cell wall, PMEI–YFP fusion proteins were detected exclusively at the pollen tube tip. Together with the demonstration that AtPMEI2 binds to and inhibits the activity of AtPPME1 in vitro, this strongly suggests a function of AtPMEI2 in locally inhibiting PME activity at the pollen tube apex. Interestingly, in contrast to AtPPME1:YFP, AtPMEI:YFP accumulated in intracellular compartments induced by treatment with Brefeldin A (BFA), or by the expression at high levels of a XFP:FYVE fusion protein. This indicates that endocytic internalization of AtPMEI2 is at least in part responsible for the specific accumulation of this protein at the pollen tube apex.

Results

AtPMEI1 and AtPMEI2 bind to and inhibit AtPPME1

AtPMEI1 (at1g48020) and AtPMEI2 (at3g17220) have been identified as in vitro and in planta inhibitors of PME activity, and are both strongly expressed in Arabidopsis pollen (Lionetti et al., 2007; Raiola et al., 2004; Wolf et al., 2003). As recombinant proteins, these inhibitors are able to inhibit PME activities from various plant sources, pointing to a broad spectrum of potential PME partners. A pollen-specific type-II PME from Arabidopsis, AtPPME1 (at1g69940), a potential in vivo target of AtPMEI1 and AtPMEI2, was isolated recently, and was shown to significantly contribute to pollen tube elongation, stability and, most importantly, PME activity (Tian et al., 2006). To demonstrate binding of AtPMEI1 and AtPMEI2 to AtPPME1, we performed pull-down experiments using recombinant His-tagged PMEIs fused to thioredoxin. Nickel-nitrilo-triacetic-acid (Ni-NTA) agarose beads loaded with PMEI fusion proteins were incubated with extracts from Nicotiana benthamiana leaves, in which AtPPME1 C-terminally tagged with a triple hemagglutinin tag (3HA) was transiently expressed. Beads loaded with a recombinant invertase inhibitor protein were used as a control. Invertase inhibitors and PME inhibitors belong to the same protein family and share a very similar structure, yet they inactivate their respective class of target enzymes specifically (Hothorn et al., 2004; Rausch and Greiner, 2004). After extensive washing of the beads, bound proteins were analysed by immunoblotting using HA antiserum. A distinct signal at approximately 38 kDa was detected when either TrxA:AtPMEI1 or TrxA:AtPMEI2 were used as a bait, but not in the control experiment with the TrxA:invertase inhibitor protein. This established that AtPMEI1 and AtPMEI2 specifically bind to AtPPME1:3HA in vitro (Figure 1, upper panel). For the following experiments, we concentrated on AtPMEI2 only.

Figure 1.

 AtPMEI1 and AtPMEI2 interact with AtPPME1:3HA.
Immunological detection of AtPPME1:3HA after a pull-down assay with immobilized recombinant invertase inhibitor (Inv), AtPMEI1 (1) and AtPMEI2 (2); In, input. Immunosignals obtained with anti-HA antiserum are displayed in the upper panel. The lower panel shows the amido black-stained recombinant proteins (TrxA:Inv, TrxA:AtPMEI1 and TrxA:AtPMEI2) on the western blot membrane.

To confirm that AtPMEI2 not only binds to AtPPME1 but also inhibits its enzymatic activity, we developed a functional assay. As all attempts to produce enzymatically active recombinant PME protein for biochemical characterization failed (expression systems tested: Escherichia coli, Saccharomyces cerevisiae, Pichia pastoris), we expressed AtPPME1:3HA in N. benthamiana leaves. For this purpose, we generated a binary construct bearing the full-length AtPPME1 cDNA under control of the cauliflower mosaic virus 35S promoter. The construct was mobilized in Agrobacterium tumefaciens and used to transiently transform cells in N. benthamiana leaves at one side of the middle vein. The other half of the leaf was transformed with a control construct conferring expression of a secreted version of GFP (secGFP), also under the control of the 35S promoter. After 48 h, leaves were harvested and PME activity extracted from each half leaf was measured in a coupled enzymatic assay (Grsic-Rausch and Rausch, 2004). Several independent leaves were analysed: expression of AtPPME1 consistently led to an increase in PME activity, ranging from 104% to 128% (mean, 114.15%; SD, 8.48%) presumably because of different levels of total PME activity and expression of the transgene. A representative result of one leaf measured in three replicates is shown in Figure 2. The transient expression of AtPPME1:3HA increased the total PME activity by about 15%. Addition of recombinant AtPMEI2 completely inactivated total PME activity in extracts from control and AtPPME1-expressing tissue, demonstrating that AtPMEI2 inhibited both the endogenous tobacco PME isoforms and the overexpressed AtPPME1. Using the same procedure, we also expressed AtPMEI2 in N. benthamiana leaves to confirm that this protein was as active when synthesized in leaves as it was when purified from E. coli. As expected, in extracts from transformed leaves expressing AtPMEI2 endogenous PME activity was completely abolished.

Figure 2.

 AtPMEI2 is able to silence AtPPME1-derived pectin methylesterase (PME) enzyme activity.
PME activity extracted from Nicotiana benthamiana leaves transformed with the control secGFP, AtPMEI2 or AtPPME1. Addition of recombinant AtPMEI2 protein (AtPMEI2r) completely silences endogenous and AtPPME1-derived activity.

Expression of AtPMEI2 and AtPPME1 affects tobacco pollen tube elongation

Having established that AtPMEI2 not only interacts with AtPPME1 but also with endogenous PMEs in N. benthamiana, we explored the possible effects of transient expression of AtPMEI2 and AtPPME1 in tobacco pollen tubes. AtPMEI2 was transiently co-expressed with the fluorescent marker YFP in N. tabacum pollen tubes using particle bombardment and expression vectors based on the Lat52 promoter (Twell et al., 1990). The length of transformed tubes was measured 6 h after bombardment and compared with that of control tubes expressing YFP only at comparable levels (based on fluorescence intensity). As depicted in Figure 3(a), pollen tubes transformed with AtPMEI2 were significantly longer than control pollen tubes. Together with the observation that AtPMEI2 also completely inhibited total PME activity in tobacco pollen tube extracts (data not shown), this finding is consistent with the assumption that transiently expressed AtPMEI2 promotes pollen tube growth by reducing the activity of PME isoform(s), and by thereby promoting cell-wall expansibility. In contrast, transient expression of AtPPME1 had a mild inhibiting effect on pollen tube growth (Figure 3b). Thus, AtPMEI2 and AtPPME1 were clearly active in tobacco pollen tubes, and appeared to have opposing effects, although the expression of these proteins affected the growth rate of these cells only moderately, and did not cause morphological changes.

Figure 3.

 Effects of transient AtPMEI2 and AtPPME1 expression on tobacco pollen tube length.
(a) Statistical analysis of pollen tube length 6 h after bombardment with cytosolic YFP and AtPMEI2. Error bars indicate 95% confidence intervals (n = 110, P < 0.003). (b) Effect of AtPPME1 on pollen tube length compared with cytosolic YFP. Error bars indicate 95% confidence intervals (n = 90, P < 0.003).

AtPMEI2:YFP and AtPPME1:YFP display distinct distribution patterns in the pollen tube cell wall

Nicotiana tabacum pollen tubes were transiently transformed by particle bombardment with a chimeric gene construct encoding a AtPMEI2:YFP fusion protein under the control of the pollen-specific Lat52 promoter. At 6–8 h after bombardment, transformed pollen tubes were analysed by confocal laser scanning microscopy. Although AtPMEI2:YFP weakly labelled intracellular structures, most of the protein clearly localized to the apical cell wall (Figure 4a). Plasmolyzed pollen tubes displaying a retracted protoplast confirmed that the cell wall was strongly labelled at the apex (Figure 4c). The AtPMEI2:YFP labelling always declined sharply at the flanks of pollen tube tips, and was never observed laterally in the shank of these cells. Identical distribution patterns were also observed when AtPMEI1:YFP or AtPMEI2:YFP were transiently expressed in tobacco pollen tubes under the control of the Lat52 promoter or the endogenous AtPMEI2 promoter, respectively (data not shown, Figure 4b), although the AtPMEI2 promoter was clearly less active than the Lat52 promoter. A secreted form of GFP (secGFP), used as a control, could be detected in cytoplasmic organelles, but was never observed to accumulate in the cell wall (Figure 4d). To confirm that the fusion of AtPMEI2 and YFP was still able to interact with PME, we performed inhibition assays after transient expression of AtPMEI2:YFP under the control of the 35S promoter in N. benthamiana leaves, as in Figure 2. Although less efficient than AtPMEI2 alone, AtPMEI2:YFP was able to silence the major part of PME activity from the leaves. In marked contrast to the tip-restricted localization of AtPMEI2:YFP, AtPPME1 with YFP attached to its C-terminus always labelled the cell wall of transiently transformed tobacco pollen tubes at the tip and in the shank to a similar extent (Figure 4e). Together, the observed distribution patterns of the pectin methylesterase AtPPME1, and of its inhibitor AtPMEI2, establish the presence of both proteins in the cell wall at the pollen tube apex, and of only the pectin methylesterase in the shank. This is in excellent agreement with the exclusive detection of esterified pectins at the tip, and of de-esterified pectins in the shank of pollen tubes (Bosch et al., 2005; Parre and Geitmann, 2005).

Figure 4.

 Localization of transiently expressed AtPMEI2:YFP and AtPPME1:YFP in tobacco pollen tubes.
Single confocal sections of pollen tubes 6–8 h after microprojectile bombardment are presented.
(a) Pollen tube expressing AtPMEI2:YFP under the control of the Lat52 promoter.
(b) AtPMEI2:YFP expressed under the control of the AtPMEI2 promoter. Retraction of the protoplast in (c) confirms wall localization of Lat52:AtPMEI:YFP-derived fluorescence.
(d) SecGFP.
(e) AtPPME1:YFP. Scalebars = 10 μm.

In contrast to AtPPME1, AtPMEI2 is recycled by endocytosis

Several mechanisms may contribute to the intriguing and selective decline of AtPMEI2 levels at the flanks of the pollen tube tip, including rapid degradation, dissipation into the medium (possibly after dissociation from PME) and endocytic uptake. To explore the possibility that AtPMEI2 is endocytically internalized to prevent it from accumulating laterally in the pollen tube cell wall, we determined whether AtPMEI2:YFP localizes to BFA-induced aggregates of endosomal compartments. The fungal toxin BFA not only blocks transport from the endoplasmic reticulum (ER) to the Golgi apparatus, but also induces the formation of large aggregates containing endosomal membranes near the pollen tube apex and in other cell types. Although these aggregates are generally referred to as BFA compartments, they have been named BFA-induced aggregates (BIAs) in pollen tubes because of their distinct features in these cells (Parton et al., 2003). The styril dye FM4-64, an established tracer of endocytic membrane traffic in plant cells, has been shown to accumulate in BFA compartments and in BIAs (Dettmer et al., 2006; Lam et al., 2007; Meckel et al., 2004). When applied to pollen tubes in the absence of BFA, this dye first accumulates in the plasma membrane, is then internalized by endocytosis at the flanks of the tip, and within 20 min labels the cone-shaped apical ‘clear zone’, which is filled with secretory vesicles (Parton et al., 2001). When BFA was applied to FM4-64-labelled pollen tubes transiently expressing AtPMEI2:YFP, BIAs were formed within 60 min about 5–20 μm behind the apex, in which AtPMEI2:YFP and FM4-64 both accumulated to high levels (Figure 5b). In contrast, control experiments showed that free cytosolic YFP was clearly excluded from these compartments (Figure 5a, middle panel, and merged in lower panel). In marked contrast to AtPMEI2:YFP, AtPPME1:YFP (Figure 5c) and secGFP (not shown) accumulated only weakly in BIAs under the same experimental conditions. Intracellular AtPPME1:YFP fluorescence, presumably derived from protein present in the ER, retracted behind the BIA as previously observed with ER-resident GFP after BFA treatment (Parton et al., 2003).

Figure 5.

 Transiently expressed AtPMEI2:YFP is trapped in Brefeldin A (BFA)-induced aggregates.
(a) BFA treatment on pollen tubes results in the formation of an FM4-64-positive membranous subapical aggregation (BIA, upper panel), from which cytosolic YFP is excluded (middle panel, and merged in the lower panel). (b) In AtPMEI2:YFP-expressing pollen tubes, the YFP fluorescence (middle panel) co-localizes with the FM4-64-stained BIA (upper panel, and merged in the lower panel). (c) AtPPME1:YFP fluorescence is not enriched in BFA-induced aggregates (FM4-64 staining in upper panel, YFP fluorescence in middle panel, and merged in the lower panel). Scalebars = 10 μm.

At least in some cell types, the trans-Golgi-network (TGN) appears to be incorporated into BFA-induced intracellular aggregates (Nebenfuhr et al., 2002; Ritzenthaler et al., 2002). It can therefore not be excluded that transport of AtPMEI2:YFP along the secretory system contributed to some extent to the observed accumulation of AtPMEI2:YFP in BIAs, although AtPPME1 and secGFP, which are also passing through the TGN on their way to the cell surface, were barely detected in these structures.

We therefore adopted a complementary approach to confirm the results obtained with BFA. The FYVE domain specifically binds to phosphatidylinositol-3-phosphate, a lipid that accumulates to high levels in animal and plant endosomal membranes (Burd and Emr, 1998; Corvera et al., 1999; Gaullier et al., 1998; Kutateladze et al., 1999; Stenmark et al., 1996; Vermeer et al., 2006). At low expression levels, fluorescent FYVE domain fusion proteins co-localize with Rab5 homologues to endosomal membranes without disrupting membrane traffic (Vermeer et al., 2006). However, expression of such fusion proteins at high levels induces the formation of large endosomal aggregates in various cell types. In pollen tubes, FM4-64 and signalling lipids, which are normally recycled back to the plasma membrane at the apex, were shown to be trapped within these aggregates (Gillooly et al., 2000; Helling et al., 2006; Voigt et al., 2005). As reported in the literature, strong expression of a mRFP:FYVE fusion protein (Helling et al., 2006) caused the formation of large endosomal aggregates with which this fusion protein remained associated (Figure 6a and b). Interestingly, in these aggregates co-expressed AtPMEI2:YFP accumulated to high levels (Figure 6a), whereas AtPPME1:YFP was never detected (Figure 6b).

Figure 6.

 AtPMEI2:YFP, but not AtPPME1:YFP, is trapped in mRFP:FYVE-induced endosomal aggregates.
Single confcocal sections of pollen tubes expressing mRFP:FYVE at high levels (a and b). In (a), overexpression of mRFP:FYVE (upper panel) and AtPMEI2:YFP (middle panel) leads to co-localization to intracellular aggregates (lower panel). No co-localization was observed after co-expression of mRFP:FYVE (b, upper panel) and AtPPME1:YFP (middle panel, and merged in the lower panel). Bars = 10 μm.

Together, these observations demonstrate that AtPMEI2 is endocytically internalized, and strongly suggest that endocytic uptake at the flanks of the tip plays an important role in sustaining the accumulation of this protein, specifically at the apex of growing pollen tubes.

Discussion

Pollen-expressed PMEs and PME inhibitors interact in vitro and in vivo

Plant PMEs have been implicated in diverse processes like pollen tube growth, root border cell separation, fruit maturation, pollen tetrade separation, systemic movement of tobacco mosaic virus, leaf growth polarity, ion balance and seed maturation (Chen et al., 2000; Dorokhov et al., 1999; Francis et al., 2006; Jiang et al., 2005; Phan et al., 2007; Pilling et al., 2004; Ren and Kermode, 2000; Tian et al., 2006; Wen et al., 1999). Similar to PMEs, PMEIs appear to be ubiquitously expressed in higher plants. Based on their strong and specific effects on PME activity, they are assumed to provide an effective post-translational control mechanism for PME activity; however, the precise role of their interplay with PMEs is far from clear, and a loss-of-function effect of any PMEI protein remains to be shown. At present, the pollen tube represents the best available system to gain insight into the role of the PME–PMEI interaction: (i) it is a simple experimental system amenable to advanced in vivo imaging techniques, (ii) both proteins (including multiple isoforms) are highly expressed in pollen (Pina et al., 2005), and (iii) regulated pectin modification is essential for pollen tube growth (Bosch and Hepler, 2005, 2006; Bosch et al., 2005; Jiang et al., 2005; Parre and Geitmann, 2005; Tian et al., 2006).

Previous biochemical analysis of recombinant or purified PMEIs has revealed a rather broad specificity towards PME activities. Thus, recombinant AtPMEI1 and AtPMEI2 inactivated PME enzymes from orange peel, Arabidopsis flowers, siliques and leaves and from tobacco BY-2 cells (Wolf et al., 2003; unpublished results), and a complex of a PMEI from kiwi fruit and a PME purified from tomato was recently crystalized (Di Matteo et al., 2005). Here, we demonstrate that the pollen-specific isoform AtPMEI2 completely inhibited PME activity in the leaves of N. benthamiana, either when ectopically expressed in this tissue or when added to leaf extracts as E. coli-expressed recombinant protein. Ectopic expression of AtPPME1 (a recently identified pollen-specific PME isoform; Tian et al., 2006) in N. benthamiana leaves resulted in a significant increase in total PME activity, which was completely inhibited by recombinant AtPMEI2, demonstrating that the two proteins functionally interact. Furthermore, the ability of AtPMEI2 to physically bind to AtPPME1 was independently confirmed with pull-down experiments using recombinant TrxA:AtPMEI2 fusion protein and HA-tagged AtPPME1 expressed in N. benthamiana leaves. In summary, these results support the notion that the pollen-expressed inhibitors AtPMEI1 and AtPMEI2 are regulators of the pollen-expressed PME isoform AtPPME1, and also inhibit pollen-expressed tobacco PME isoforms.

AtPMEI2:YFP is localized exclusively at the cell wall of the pollen tube tip, whereas AtPPME1:YFP is present at the tip and along the flanks

As Arabidopsis pollen tubes are not amenable to cell biological analysis, we have used in our study established protocols for the analysis of gene function in tobacco pollen tubes. Our choice of functionally characterized, pollen-expressed Arabidopsis PMEI and PME proteins was dictated by the fact that only a type-I PME has been shown to be expressed in tobacco pollen so far, which contains a PMEI-related putative autoinhibitory domain. Furthermore, no tobacco proteins with confirmed PMEI activity have been identified to date.

Confocal laser scanning microscopy of growing pollen tubes showed that AtPMEI2 fused to YFP is present in the apical cell wall, but the fluorescence is greatly reduced at the lateral walls. On the contrary, AtPPME1:YFP is found throughout the pollen tube cell wall, with no pronounced decrease of fluorescence intensity behind the tip. Localization to the lateral pollen tube wall has also been reported for the type-I PME VGD1 (Jiang et al., 2005). Mislocalization as a consequence of transgene-induced changes in pollen tube growth are unlikely, regarding the modest effects observed on pollen tube length upon transient expression of AtPMEI2 or AtPPME1. Our results support the hypothesis that complexing of PME(s) with PMEI(s) at the pollen tube tip prevents premature pectin de-esterification, thereby keeping the tip wall extensible. Premature de-esterification is likely to block pollen tube elongation, as it was shown that exogenously added PME causes a thickening of the apical cell wall and cessation of growth (Bosch et al., 2005; Parre and Geitmann, 2005). Conversely, pectin de-esterification is certainly required behind the apex to convey mechanical stability required for the maintenance of pollen tube shape during penetration of the reproductive tract. Thus, the presence of large quantities of PMEI protein at the tip, but not at the flanks, is perfectly in line with these physiological requirements. In addition, several independent studies have demonstrated the presence of esterified HGA almost exclusively at the pollen tube tip. Whereas earlier studies reported a uniform distribution of unesterified pectin along the entire length of the pollen tube, it was shown recently that in Solanum chacoense and N. tabacum unesterified pectins are absent from the apical wall (Bosch et al., 2005; Parre and Geitmann, 2005). Interestingly, our data indicate that the previously described border region between both pectic epitopes is located in the region where AtPMEI2 levels sharply decline.

The interpretation described above is based on the assumption that in growing pollen tubes PMEI(s) control(s) the activity of PME(s) performing blockwise de-esterification, because it is this pectin modification that conveys wall rigidity. However, this does not exclude that pollen-expressed PME isoforms may also perform random de-esterification. Numerous studies have shown that the mode of PME action is affected by the availability of cations, local pH, the presence of hydrolases, and interaction with cell-wall pectin (Bordenave, 1996; Bordenave and Goldberg, 1994; Catoire et al., 1998; Charnay et al., 1992; Denes et al., 2000; Goldberg et al., 2001; Kim et al., 2005; Limberg et al., 2000; Moustacas et al., 1991), suggesting that microenvironment rather than differences between isoforms affect the mode of PME action (Bosch and Hepler, 2005; Pelloux et al., 2007).

AtPMEI2, but not AtPPME1, undergoes endocytic trafficking

Differential localization of AtPMEI2 and AtPPME1, as observed in transformed tobacco pollen tubes, may explain how the requirement for spatially defined changes in pectin methylation status is met in growing pollen tubes. However, to prevent lateral accumulation of PMEI, which is expected to occur as a direct consequence of the dynamic growth process, this protein either has to be internalized by endocytosis or specifically degraded behind the apex. Alternatively, in the absence of intrinsic affinity to the cell wall itself, PMEI not bound to PME may be diluted via diffusion. Providing evidence for endocytic internalization of AtPMEI2, we have localized AtPMEI2:YFP in BIAs and mRFP:FYVE-induced endosomal aggregates. As AtPPME1 does not seem to undergo endocytic trafficking, it is tempting to speculate that AtPMEI2:YFP is constantly, and selectively, internalized at a region 5–15 μm behind the tip (Parton et al., 2001), and, consequently, accumulates exclusively at the apex, where it prevents premature activity of newly secreted PME(s).

Localized PME regulation via PMEI: impact on pollen tube growth

Although it remains to be demonstrated that PME–PMEI complexes are indeed formed and maintained exclusively at the growing pollen tube tip, circumstantial evidence supports this scenario. Thus, exogeneous application of PME to pollen tubes has an inhibitory effect on growth (Bosch et al., 2005; Parre and Geitmann, 2005), whereas a lack of PME activity also had a detrimental effect on pollen tubes and fertility (Bosch and Hepler, 2006; Jiang et al., 2005; Tian et al., 2006), indicating that PME activity has to be balanced and regulated. Upon which signal(s) do PME–PMEI complexes dissociate? And how is the selectivity of PMEI (versus PME) internalization ensured? Here, we can only speculate. First, little is known about the factors impacting on in vivo formation (and dissociation) of PME–PMEI complexes. A number of ion gradients and fluxes are well established in pollen tube growth, including an intracellular Ca2+ gradient, an oscillating apical efflux and non-oscillating lateral influx of Cl, an oscillating apical influx of K+ and a lateral localized efflux of H+ behind the clear zone (Feijo et al., 1999; Holdaway-Clarke and Hepler, 2003; Lovy-Wheeler et al., 2006). However, for the stability of PME–PMEI complexes controversial results have been reported. Whereas biochemical analysis of PMEI from Kiwi fruit indicated a pronounced pH dependence of its inhibitory activity (D’Avino et al., 2003; Raiola et al., 2004), the complex of AtPMEI2 with a tomato PME was extremely stable over a wide pH range (Raiola et al., 2004). Second, apart from PME–PMEI complex dissociation, the mechanism of AtPMEI2 endocytosis remains at present unresolved. Possibly, AtPMEI2 endocytosis is functionally linked to the retrieval of other wall components. Thus, previous studies have shown that in meristematic root cells partially esterified (up to 40%) HGA pectins may undergo endocytosis (Baluska et al., 2002). Clearly, the next challenges ahead are to unravel the factors that regulate PME–PMEI complex formation and dissociation, and to elucidate possible interactions of these proteins alone and in combination with different pectin types or other wall components.

Experimental procedures

Plant material

Pollen material was harvested from tobacco plants (N. tabacum Petite Havana SR1) grown in a greenhouse under standard conditions. N. benthamiana plants were maintained in a growth chamber under a light regime of 16 h and 300 μE.

Plasmid construction

35S:AtPMEI2 in pB2GW7 was generated by amplification of the AtPMEI2 coding sequence from genomic DNA (AtPMEI2 is intronless) with sense primer 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTAATGGCAGCATACCTGACGAAC-3′ and antisense primer 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTTTCACATCATGTTTGAGATGAC-3′. Gateway cloning into pDONR201 and subsequently into pB2GW7 was performed according to the manufacturer’s instructions (Invitrogen, http://www.invitrogen.com) except that BP and LR reactions were downscaled to a quarter of the reaction volume. AtPPME1:3HA was generated by amplification of the AtPPME1 coding sequence without stop codon from a pollen cDNA preparation with sense primer 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTATGGGATACACAAATGTGTC-3′ and antisense primer 5′-ATCTCAGGAGCTCCAAAGCGGGTGGTGGGAG-3′, and amplification of 3HA from vector pREP3-adh using sense primer 5′-ATCTCAGGAGCTCATCTTTTACCCATACGATGTTC-3′ and antisense primer 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCCCGGGTCGACTAAGCAGCG-3′. After restriction digest with SacI (sites underlined), the two PCR fragments were ligated and subsequently incorporated into pDONR 201 and pB2GW7 through Gateway cloning, as described above.

AtPMEI2:YFP was generated by amplification of the AtPMEI2 coding sequence from genomic DNA with sense primer 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTAATGGCAGCATACCTGACGAAC-3′ and antisense primer 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTTCATCATGTTTGAGATGACAAG-3′. The PCR product was cloned into pDONR221. The Lat52 promoter was amplified from plasmid Lat52:YFP using sense primer 5′-GGGGACAACTTTGTATAGAAAAGTTGCTAAGCTTGTCGACATACTCG-3′ and antisense primer 5′-GGGGACTGCTTTTTTGTACAAACTTGGGTTTAAATTGGAATTTTTTTTTTTGG-3′ and incorporated into pDONRTM P4-P1R. YFP was amplified from vector pK7YWG2 using sense primer 5′-GGGGACAGCTTTCTTGTACAAAGTGGCTATGGTGAGCAAGGGCGAGG-3′ and antisense primer 5′-GGGGACAACTTTGTATAATAAAGTTGGTTACTTGTACAGCTCGTCC-3′, and incorporated into pDONRTM P2R-P3. All three entry clones were recombined into pB7m34GW by Multisite Gateway technology (Invitrogen). AtPMEI2:YFP under control of its own promoter and AtPMEI2 under control of the Lat52 promoter without translational fusion to YFP were created accordingly, except that the AtPMEI2 promoter was amplified with sense primer 5′-GGGGACAACTTTGTATAGAAAAGTTGCTTTCTTTATGTATCTTTCACCTG-3′ and antisense primer 5′-GGGGACTGCTTTTTTGTACAAACTTGGCTTTGCTTCTTTCTTTCTTATG-3′ and integrated in pDONRTM P4-P1R, or AtPMEI2 was incorporated with stop codon in pDONR221 by using antisense primer 5′-GGGGACTGCTTTTTTGTACAAACTTGGGTTTAAATTGGAATTTTTTTTTTTGG-3′. AtPPME1:YFP was created likewise by amplification of the AtPPME1 coding sequence without stop codon using sense primer 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTAATGGGATACACAAATGTG-3′ and antisense primer 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTTCAAAGCGGGTGGTGGG-3′ and incorporation into pDONR221. Profilin promoter-driven secGFP, Lat52:YFP and mRFP:FYVE have been described elsewhere (Helling et al., 2006). Gateway destination vectors are described in (Karimi et al., 2002, 2005) and at http://www.psb.ugent.be/gateway/index.php.

Bacterial expression and pull-down experiment

Expression of TrxA- and His-tagged AtPMEI1 and AtPMEI2 protein was described earlier (Wolf et al., 2003), as was expression of invertase inhibitor protein (Hothorn et al., 2003), except that E. coli Rosetta-gami cells were used in all cases. AtPPME1:3HA containing extract was obtained by grinding frozen leaf tissue (for transformation, see below) 48 h after infiltration with a homogenizer, and thawing in 2.5 ml g−1 fresh weight of extraction buffer [100 mm Tris/100 mm maleic acid, pH 7, 1 m NaCl, 1 mm PMSF and protease inhibitor cocktail (P9599, diluted 1 : 100; Sigma-Aldrich, http://www.sigmaaldrich.com)]. After 30 min of incubation in an end-over-end shaker at 4°C, the extract was cleared by centrifugation at 10 000 g in a cooled table-top centrifuge for 20 min. For the pull-down experiment purified recombinant fusion protein was coupled to 50 μl of Ni-NTA agarose (Invitrogen) and incubated for 45 min with 300 μl of AtPPME1:3HA containing extract in an end-over-end shaker at 4°C. The suspension was centrifuged at 800 g for 10 min and the supernatant was discarded. The beads were washed six times with 1 ml of extraction buffer and boiled in 100 μl of SDS-PAGE sample buffer. A 20-μl portion was loaded on a 13% polyacrylamide gel, and after electrophoresis the proteins were transferred to a polyvinylidene difluoride (PVDF)-membrane. AtPPME:3HA was detected with monoclonal anti-HA antiserum (Covance, http://www.covance.com) diluted 1 : 20 000 in Tris-buffered saline Tween-20 (TBST) and 1% skim milk powder using standard procedures.

Agrobacterium-mediated transformation of N. benthamiana leaves

Agrobacterium tumefaciens cells (strain C58 C1) were grown overnight in 50 ml of YEB-medium supplemented with carbenicillin (50 μg ml−1), rifampicin (100 μg ml−1) and spectinomycin (50 μg ml−1) until the stationary phase. After centrifugation at 3000 g for 30 min at room temperature, the cells were suspended in 10–15 ml of infiltration buffer [10 mm 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.9, 150 μm acetosyringone] and incubated with gentle agitation for at least 2 h. Subsequently, the cell suspensions were adjusted to an optical density of one with infiltration buffer and infiltrated into the lower epidermis of 8- to 12-week-old N. benthamiana leaves with a needleless 1-ml syringe.

Pollen transient expression and drug treatment

Pollen transformation and application of drugs was performed as described previously (Helling et al., 2006; Kost et al., 1998). Transformed pollen tubes were cultured in PT medium (1 mm CaCl2, 1 mm KCl, 0.8 mm MgSO4, 1.6 mm H3BO4, 30 μm CuSO4, 0.03% casein acid hydrolysate, 5% sucrose, 12.5% PEG-6000, 0.3% MES, pH 5.9, solidified with 0.25% phytagel) for 6–8 h at room temperature. BFA at a concentration of 50 μm was applied for 20–60 min, FM4-64 was applied at 7.6 μm and was added 10 min before BFA treament.

Microscopy and image analysis

Pollen tube lengths after micropojectile bombardment were analysed using a Leica DM IRB epifluorescence microscope equipped with a digital cooled camera (DFC350FX R2; Leica, http://www.leica.com) and ImageJ software (http://rsb.info.nih.gov/ij). Confocal laser scanning images were taken with an LSM 510 Meta inverted microscope (Zeiss, http://www.zeiss.com). YFP fluorescence was excited with a 514-nm laser line, and emitted fluorescence was collected using a 530- to 600-nm bandpass filter. RFP fluorescence was excited with the 543-nm laser line and was collected with a 560-nm long pass filter.

Acknowledgements

We would like to thank Katja Piiper for assistance with pollen bombardment and Sabine Strahl for the pREP3-adh plasmid. SW is the recipient of a fellowship from the state of Baden-Wuerttemberg (Landesgraduiertenfoerderung).

Ancillary