All chemicals (unless otherwise indicated) were obtained from Sigma–Aldrich (Milan, Italy). Mature pollen of pear (P. communis cv. Williams) was collected from plants grown in experimental plots at the University of Bologna (Department of Agricultural Sciences, University of Bologna). Handling, storage, pollen hydration, and germination were performed as previously reported (Bagni et al., 1981; Del Duca et al., 1997).

Pollen viability was tested by MTT (2,5-diphenyl tetrazoliumbromide). The test solution contained 1% concentration of the MTT substrate in 5% sucrose. After 15 min incubation at 30°C the pollen samples were visualized under a light microscope (Nikon Eclipse E600) equipped with a digital camera (Nikon DXM1200). Pollen was considered viable if it turned deep purple.

At 60 min from the start of germination, Spm (100 μM) was added to the growth medium.

The Ca²⁺-channel inhibitors LaCl₃ and GdCl₃ were tested at 1, 10, 50, or 100 μM; alternatively, the Ca²⁺-chelating agent EGTA was supplied at 0.4, 1, and 5 mM. These compounds were added to the growth medium 20 min prior to Spm supplementation (i.e., at 40 min of germination).

Pollen tube length was measured using ImageJ software.

For each localization and immunolocalization experiment, at least 30 pollen tubes of approximatively identical length for each developmental stage (control, balloon, snake, and shovel) were imaged and analyzed. For each stage of development, the data shown in figures is highly representative (more than 90% of pollen tubes showed an identical signal distribution).

Pollen tubes were observed using a Nikon inverted microscope Diaphot TMD. Extreme care was taken in observing only pollen tubes that grew in line with the focal plane in order to avoid focusing problems. Video clips were captured using a CCD camera C2400-75i Hamamatsu (Hamamatsu Photonics) connected to Argus-20 (Hamamatsu) and converted into MPEG-2 files (25 frames per second) by a video capture system (PCTV Center) working at a resolution of 720 × 576 pixels (Cai et al., 2000). MPEG-2 files were converted into AVI (MJPEG compression) by VirtualDub (http://virtualdub.org/). Video files were opened in ImageJ software (http://rsbweb.nih.gov/ij/index.html) and analyzed by a plug-in kymograph to measure the speed of moving objects in a series of images. The kymograph analyses and measures the gray values in a region of interest (ROI) selected manually for each video frame. A graphical representation of spatial position over time was generated; the X-axis was the time axis (the unit is the frame interval) and the Y-axis indicated the movement rate of each ROI (the unit of measurement is the distance covered by the object expressed in pixels). The speed of objects was measured directly by the plug-in. At least 30 pollen tubes for each sample were analyzed.

A BCECF-AM ester probe was used for visualizing proton levels (i.e., pH) in pear pollen tubes (Qu et al., 2012). A final concentration of 5 μM was obtained from a 1 mM dimethyl sulfoxide (DMSO) stock solution; the required volume was directly added to the germination medium containing the re-suspended pollen grains. The cytosolic pH was immediately visualized after addition of the probe to prevent uptake of the pH probe by organelles. In experiments with Spm, the pH probe was either added concomitant with Spm or after progressive 5-min incubation steps following its addition. In doing so, all significant growth stages of pollen tubes after Spm treatment were detected. Samples were observed with a Zeiss AxioImager fluorescence microscope equipped with structured illumination in the FITC filter.

Pollen (10 mg) was germinated in growth medium supplemented with 5 μM coelenterazine (Prolume) for 1 h in the dark, washed three times with 3 volumes of fresh medium, and then incubated with 30 μM TAT-aequorin for 10 min in the dark. After extensive washing as above, 100 μl of germinated pollen were transferred in the luminometer chamber. Luminescence measurements were performed with a custom-built luminometer (Electron Tubes Ltd.) containing a 9,893/350A photomultiplier (Thorn EMI), as previously described (Zonin et al., 2011). Ca²⁺ measurement assays were carried out in control conditions (injection, by a light-tight syringe, of an equal volume of germination medium into the sample) or after addition of 100 μM Spm. Luminescence emitted by whole pollen tubes was monitored continuously (every second) for 10 min, and then residual aequorin was discharged by addition of 0.33 M CaCl₂. The light signal was calibrated off-line into Ca²⁺ concentration values by using a computer algorithm based on the Ca²⁺ response curve of aequorin (Brini et al., 1995).

To prepare total protein extracts, 1-h germinating pollen samples (10 mg) were lysed with 200 μl of ice-cold Cell Disruption Buffer from the PARIS™ kit (Life Technologies) supplemented with protease inhibitors (0.5 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 1 μM leupeptin) in three cycles of 30 s each using a motorized pestle (Sigma-Aldrich), each one followed by 30 s incubation on ice. Samples were then subjected to three cycles of 30 s of sonication (3,510 Branson), each one followed by 30 s incubation on ice. Homogenates were centrifuged at 13,000 g for 5 min at 4°C and the protein concentration of the supernatant was determined by the Bradford assay (Bio-Rad). Proteins were separated by SDS-PAGE (12.5% polyacrylamide), transferred onto a polyvinylidene difluoride (PVDF) membrane and processed for immunoblotting, as described by Zonin et al. (2011). A monoclonal anti-polyhistidine antibody (Sigma-Aldrich) was used at a 1:1,500 dilution.

For labeling actin filaments, pollen tubes were fixed and permeabilized in 100 mM Pipes buffer, pH 6.9, containing 5 mM MgSO₄, 0.5 mM CaCl₂, 0.05% Triton X-100, 1.5% formaldehyde, and 0.05% glutaraldehyde for 30 min according to Lovy-Wheeler et al. (2005). Samples were washed twice with the same buffer, but at pH 7, containing 10 mM EGTA and 6.6 μM Alexa 543-phalloidin (Invitrogen). Samples were placed on slides and covered with a drop of Citifluor as anti-fading agent.

Growing pollen tubes were imaged for at least 60 s (or until the organelle was in focus), with frames acquired every 0.5 s. Video clips were captured by the Axiovision software (version 4.3; Zeiss) as ZVI files. Files were then converted into AVI files (MJPEG compression) and then opened in ImageJ software. Image stacks were calibrated with the scale command using the scale bar generated by the Axiovision software. Image (stack) sequences were analyzed by ImageJ using the Manual Tracking plug-in and the trajectories of selected organelles were superimposed onto the initial frame. For each organelle, the absolute velocities were calculated and averaged over the entire trajectory.

Labeling of cell wall-secreted material was performed using propidium iodide (PI). Evidence that PI labels pectins has been previously described (Rounds et al., 2014; Parrotta et al., 2015). As reported, PI fluorescence matches the fluorescent signal of GFP-labeled pectin methyl esterase; PI also competes with Ca²⁺ in binding to pectins, suggesting that PI binds pectin and can be used as a dye for pectins. Callose and cellulose were labeled as previously described (Cai et al., 2011) using Aniline Blue and Calcofluor White, respectively. Measurement of PI, Calcofluor White, and Aniline Blue fluorescence was performed along the cell edge using the Segmented Line tool of ImageJ. The selection width was about 2 μm (adequate to cover the fluorescence signal). Several pollen tubes of almost identical length were measured. The background was measured outside pollen tubes and subtracted from the average of measurements. As a control of pectin labeling with PI, we also monitored the distribution of high and low methyl-esterified pectins by standard antibodies such as JIM7 and JIM5. The method used was previously described (Del Duca et al., 2013). JIM5 and JIM7 antibodies (obtained from PlantProbes, http://www.plantprobes.net) were used at dilution of 1:50. As a secondary antibody, an Alexa 595 conjugated goat anti-rat IgG (Invitrogen) was used at dilution of 1:50.

Indirect immunofluorescence microscopy was performed according to standard procedures (Cai et al., 2011). Briefly, samples were fixed with 3% paraformaldehyde in PM buffer (50 mM PIPES, pH 6.9, 1 mM EGTA, 0.5 mM MgCl₂) for 30 min, washed with PM for 10 min and then incubated with 1.5% cellulysin (Sigma) for 7 min in the dark. After two washes in PM buffer, samples were incubated with the primary antibodies to callose synthase (Cai et al., 2011) at a dilution of 1:50. Antibodies were incubated at 4°C overnight. After two washes with PM buffer, samples were incubated in the dark with a goat anti-rabbit secondary antibody conjugated to Alexa Fluor 488 (Invitrogen) diluted 1:150 for 45 min. After two washes in PM buffer, samples were placed on slides and covered with a drop of Citifluor. Observations were made using a microscope Zeiss AxioImager equipped with structured illumination and a 63x objective; images were captured with an AxioCam MRm camera using the software AxioVision. In controls, primary antibodies were omitted.

Oryzalin at 1 μM was used in combination with Spm and treated samples were observed after 30–60 min; the concentration used is able to depolymerize most microtubules in pollen tubes (Åström et al., 1995; Gossot and Geitmann, 2007). Taxol at 10 μM, a concentration known to stabilize microtubules in pollen tubes (Åström, 1992), was supplied in combination with Spm for 30–60 min. Brefeldin A (BFA) was used at 5 μg/ml⁻¹ as previously described (Rutten and Knuiman, 1993; Parton et al., 2003) and observations were made after 30–60 min. All drugs were prepared as concentrated stock solutions in DMSO. In controls, pollen tubes were analyzed either in standard medium or in medium supplemented with equivalent concentrations of DMSO. No differences were observed.

An overview of the experimental procedure is illustrated in Figure 1.

Differences between data sets were determined by two-way analysis of variance (ANOVA), with a threshold P-value of 0.05, performed using GraphPad Prism.

Untreated pear pollen developed a regular cylindrical tube (Figure 2A). The diameter of control pollen tubes was 7.5 (±0.7) μm while it increased up to 10.6 (±0.6) μm in the enlarged portions of treated pollen tubes (n = 50; Figure 2B, arrow). The kymograph revealed a constant growth velocity of 3.7 (±1.1) μm min⁻¹ in control pollen tubes (Figure 2C). When treated with Spm, pollen tubes were characterized by a lower growth velocity, 0.41 (±0.15) μm min⁻¹; moreover, the tube apex isotropically enlarged into the so-called balloon stage within the first 15 min (Figure 2D). The apex then modified its shape and entered the so-called “snake stage” and after 20–30 min it developed into an enlarged tubular form, described as “shovel stage” (Figure 2D). Concomitantly, the growth speed increased to 1.95 (±0.7) μm min⁻¹, but never recovered the original growth rate. For each measurement of growth speed, at least 50 pollen tubes were analyzed. In all cases, the differences in diameter of pollen tubes before and after spermine treatment were statistically significant, as well as differences in growth rates before and after treatment (p < 0.01, student test).

To analyze and precisely quantify the alteration in Ca²⁺ fluxes induced by Spm in germinating pollen, Ca²⁺ dynamics were monitored in pear pollen tubes by using the TAT-aequorin method (Zonin et al., 2011). The translocating properties of the cell-penetrating peptide TAT were used to deliver into germinating pollen tubes the covalently-linked bioluminescent Ca²⁺ reporter aequorin. Western blot analyses of total protein extracts from 1 h germinated pollen, incubated for 10 min with the TAT-aequorin fusion protein (30 μM), confirmed the internalization of the recombinant protein, which was absent in control samples (Figure 3A). TAT-aequorin-based Ca²⁺ measurement assays demonstrated that the cytosolic free Ca²⁺ concentration ([Ca²⁺]_cyt) in germinating pear pollen was maintained at about 0.50 μM (Figure 3B). Upon treatment with 100 μM Spm, pear pollen tubes responded with a rapid [Ca²⁺]_cyt increase, that reached a peak value of about 0.90 μM after 20 s and then gradually decreased to almost basal values within 10 min (Figure 3B, black trace). On the other hand, the mechanical perturbation caused by the injection of an equal volume of germination medium (touch control) induced only a modest [Ca²⁺]_cyt change characterized by a very limited amplitude (about 0.55 μM, Figure 3B, gray trace). Notably, the recorded Ca²⁺ levels represent mean values between [Ca²⁺]_cyt at the tip and at the base of pollen tubes, because the aequorin method allows monitoring Ca²⁺ dynamics along the entire length of the pollen tube. After 2 h of pollen germination, a two-fold increase in the basal [Ca²⁺]_cyt was observed, with a concomitant loss of responsiveness to Spm, in agreement with a physiological and progressive decrease of pollen viability (Supplementary Figure 1).

To check the effect of Spm while avoiding the influx of external Ca²⁺, lanthanum (La³⁺) and gadolinium (Gd³⁺) were supplied 20 min before the PA. When these compounds were added to the germination medium (containing 1.27 mM Ca²⁺), no morphological changes were recorded over 2 h, including no tip swelling or increase in tube diameter (Supplementary Figure 2). By contrast, tube elongation was significantly inhibited by addition of 10, 50, or 100 μM La³⁺ or Gd³⁺ and 0.4, 1, 5 mM EGTA after the first hour of germination (Supplementary Table 1). Moreover, pretreatment with La³⁺, Gd³⁺, and EGTA counteracted, in a dose-dependent manner, the extension of the shovel-shaped apical region and thus the effects of Spm (Supplementary Figure 1). In addition to the Ca²⁺ gradient, the proton gradient is also a distinctive feature of growing pollen tubes and a low pH value at the tube tip is supposed to be necessary for optimal growth. Pollen tubes of some plant species (e.g., Lilium longiflorum) are also characterized by an increase in pH behind the tip domain (the so-called alkaline band; Feijo et al., 1999). However, tobacco does not display a distinct alkaline band (Michard et al., 2008, 2017). In growing pear pollen tubes, the clear zone was demonstrated to be acidic and the subapical zone alkaline (Qu et al., 2012). We therefore analyzed potential changes in pH values during Spm treatment at different times. In control pollen tubes, the lowest pH values were observed in the extreme tube region (the tip). The peak of fluorescence, representing the zone with the lowest pH, covered approximately the first 5 μm of the tip region; the signal then dropped rapidly and stabilized at background values (Figures 4A,B). Spm-treated pollen tubes at the balloon stage displayed the highest proton concentration in a wider area at the apex, in accord with the new cell shape (Figure 4C). During transition to the so-called “shovel shape,” i.e., when pollen tubes resumed growth, the proton gradient was dissipated. Indeed, areas with very low pH could hardly be observed at the apex. Instead, a relatively high concentration of protons was visible throughout the pollen tube whereas low pH areas were observed in distal regions (Figure 4D).

Perturbation of pollen tube growth might suggest that changes to the cytoskeleton also occurred. Moreover, because the Ca²⁺ gradient is likely involved in the formation of the swollen pollen tube tip and a key regulator of cytoskeleton dynamics, the first attempt was to analyze the actin cytoskeleton. Figures 5A1,A2 illustrates typical images of actin in pear pollen tubes visualized by fluorescence microscopy. A delicate pattern of longitudinal actin filaments is present in the shank of pollen tubes. Actin filaments occasionally showed a helical arrangement (arrow in Figure 5A1) and extended up to the tube tip; nevertheless, the very tip region was always devoid of prominent actin labeling. In addition, actin in the tip appeared to be less organized and neither fringe-like nor collar-like structures were discernible (Figure 5A3).

When pollen tubes were treated with Spm, changes to the actin cytoskeleton mirrored the changes in morphology and growth rate. Thus, the subapical actin network lost its level of organization and filaments appeared shorter and disorganized compared to controls; furthermore, the presence of several fluorescent spots suggested aggregation and depolymerization of actin (Figure 5B1). In the subapical region, actin filaments appeared as partially damaged while in the shanks they seemed relatively normal. Therefore, when Spm-treated pollen tubes assumed the balloon shape, the actin array in the apical dome appeared as disordered, which likely correlates with the dissipation of the Ca²⁺ gradient and the isotropic expansion of the tubes. An ordered pattern of actin filaments was partially restored as soon as pollen tubes resumed growth and the pollen tube apex changed from the balloon into the snake shape. In this new condition, a delicate fibrillar pattern might form in the expanded tip, with actin fibrils penetrating the new apex (Figure 5B2). Moreover, actin filaments in the new developing tube underwent conformational changes consisting of actin filaments coming from the distal tubular portion (not enlarged) that penetrated in the expanded portion of pollen tubes and dispersed thereby adapting to the new shape (Figure 5B3). This new pattern of actin filaments was more evident when pollen tubes started assuming a shovel shape (Figure 5B4) and became progressively clearer while the shovel-shaped pollen tube expanded (Figure 5B5). In the expanded portion of the tube, actin filaments formed bundles that were seemingly thicker in central and cortical regions (arrows in Figure 5B5). Actin filament bundles in the unchanged part of the pollen tube shank did not seem to be affected by Spm treatment.

The involvement of the microtubular cytoskeleton was also investigated to understand if it takes part in the swelling process. In control pollen tubes, no effects due to oryzalin or taxol treatment were observed and growth of pollen tubes was comparable to untreated ones in terms of growth rate and morphology (data not shown). Neither taxol (Figures 6A,B) nor oryzalin (Figures 6C,D) affected the morphological changes induced by Spm. Both timing of changes and shape were comparable in the presence of inhibitors to those observed in Spm-treated pollen tubes (Figures 6A,B). A statistical analysis showing the null effect of taxol and oryzalin on both morphology and growth rate of pollen tubes after Spm treatment is illustrated in the Supplementary Figure 3.

F-actin is the major structural factor supporting long-distance organelle transport in pollen tubes. Given that Spm stimulated F-actin to undergo major changes that involved actin depolymerization followed by the formation of thick actin filaments, alterations to organelle distribution and dynamics after Spm treatment were investigated. Observation of the shank region in normally-growing pollen tubes revealed rapid movement of organelles parallel to the longitudinal axis of the cell (Figure 7A). Larger organelles were rapidly transported toward the growing apex with speeds that could be clustered in a distinct range. Most of the organelles exhibited a speed of 0.3–0.5 μm s⁻¹ but some specific organelles showed speeds up to 0.7 μm s⁻¹ (histograms in Figure 7C). In the apical area, larger organelles were then transported back in a basipetal direction resuming the same speed as before. Small vesicles supposedly crossed this area and accumulated in the apex where they could not be distinguished clearly due to their small size (Figures 7A,B). After Spm supplementation, cytoplasmic streaming was not arrested (Figure 7B) and organelles moved significantly faster, with most exhibiting a speed of 0.8 μm s⁻¹ (histograms in Figure 7C). Organelles followed the normal streaming direction seen in control pollen tubes; however, in most cases, they were marginalized in the cortical region, thus avoiding crossing the central expanded area (see Supplementary Movie 1). This sometimes produced a central stationary area in which many organelles moved slowly or not at all. Organelles moving quickly along actin cables could enter the stationary area, stopping once inside and then getting back on track (see Supplementary Movie 1).

Given that Spm drastically altered pollen tube morphology, the effects of the co-treatment with Spm and Brefeldin A (BFA) was investigated. The latter is known to affect the secretory and endocytotic pathways. As shown in Figure 8A, this combined treatment did not give rise to the typical balloon-shaped apex even after 30 min of treatment. Results were essentially the same until 60 min, suggesting an essential role of vesicle turnover in the process of apical swelling. As shown in Figure 8B, co-treatment with Spm and BFA prevented the expansion of the apical dome but also caused a strong reduction or even arrest of pollen tube growth. In the timeframe in which control pollen tubes had grown to about 100–150 μm, those treated with Spm+BFA grew much more slowly (not more than 20–30 μm after 1 h of treatment). As also indicated by the data in Figure 8B, the inhibitory effect on growth is most likely due to BFA; in addition, the ability of this molecule to inhibit pollen tube growth is known and described in the literature (Wang et al., 2005). Moreover, the co-treatment prevented deformation of the apex. Pollen tubes grown for 60 + 30 min with Spm+BFA were significantly different from those grown for 60 min or for 60 + 30 min in control conditions. On the contrary, pollen grown for 60 + 60 min with Spm+BFA showed no significant differences when compared to the sample treated for 30 min with Spm+BFA. This indicates that treatment with Spm+BFA affected growth during the first 30 min of incubation (Figure 8B). Again, the inhibitory effect on growth was due more to BFA because the presence of BFA only determined a significant reduction of growth.

Co-treatment with Spm and BFA also affected the deposition of cell wall components, as demonstrated by propidium iodide (PI), which vitally stains plant cell walls by binding to carboxyl residues on pectins. This allows the analysis of the link between cell wall deposition and cell growth. PI-stained pectins appeared uniformly distributed in the pollen tube border and no specific accumulation could be observed in the tip (Figure 8C; see Figure 9A below for control). Because pectins are secreted by secretory vesicles, this finding indicates that the accumulation and fusion of secretory vesicles might be altered. On the other hand, the deposition of callose in the cell wall was not modified (Figure 8D) because callose appeared as a uniform sheet surrounding the pollen tube except for the tip region.

The altered distribution of pectins by co-treatment with Spm and BFA may simply reflect the absence of tube growth. Therefore, the secretion of pectins in Spm-treated pollen tubes was investigated to check if the PA affected the deposition pattern of newly synthetized cell wall components. We measured the fluorescence intensity of PI from the tip down along the pollen tube perimeter (as shown in Figure 9A). Therefore, the values given below refer to the half of the pollen tube surface. In untreated pollen tubes, PI fluorescence was localized along the pollen tube walls mostly in a 7.2 ± 0.9 μm wide apical zone (Figure 9A; n = 40). After Spm application, at the balloon stage, newly deposited methoxylated pectins were no longer depleted from the subapical region and could be observed along a wider circumference (18.5 ± 1.6 μm; n = 35; Figure 9B). In some cases, staining with PI extended to the periphery of the entire tube portion involved in the formation of the swollen apex (Figure 9C). This suggests that the apex surface involved in the exocytosis of methoxylated pectins increased considerably in comparison with control pollen tubes. As pollen growth continued, distribution of PI fluorescence changed in Spm-treated pollen tubes; as soon as the balloon shape changed to the snake apex, the PI fluorescence was progressively refocused into a 7.4 ± 0.7 μm wide apical region (n = 35; Figures 9D,E), which represented the new growth site. When the newly-growing pollen tube reached the shovel shape, PI accumulated again, generating a new tip-focused secretion area of about 10.5 ± 0.8 μm (n = 40; Figures 9F,G). By measuring the relative fluorescence intensity in Spm-treated pollen tubes, the differential distribution of PI-labeled pectins concomitant with the typical morphological stages is shown in Figure 9H.

When pollen tubes were analyzed with JIM7 and JIM5 antibodies directed against high and low methyl-esterified pectins, respectively, the resulting data was consistent with that obtained by PI. In control samples (Supplementary Figure 4A), high methyl-esterified pectins were localized essentially in the apical region, where they are secreted. JIM5-labeled low methyl-esterified pectins (hereafter referred to as “acidic pectins”) showed a uniform pattern along the tube edge, including the apex (Supplementary Figure 4B). After treatment with Spm, at balloon stage the JIM7 signal expanded over the entire swollen surface of the tube apex, again suggesting that secretion involved a larger area (Supplementary Figure 4C). At a corresponding stage of treatment, the JIM5 signal was also found in the swollen apex but with decreased content in the tip (reasonably due to the higher secretion rate of esterified pectins; Supplementary Figure 4D). At the onset of shovel stage (Supplementary Figure 4E1), the JIM7 signal was relatively uniform in the bulged apex but, as soon as tube growth restarted, high methyl-esterified pectins again accumulated consistently in the apex (Supplementary Figure 4E2) with the signal level that rapidly decreased in older regions. Correspondingly, acidic pectins redistributed uniformly along the tube edge (Supplementary Figure 4F), including the apex, and this pattern was maintained for the entire duration of the shovel stage.

In control pollen tubes, cellulose appeared to be distributed according to a standard pattern characterized by a relatively uniform distribution along the pollen tube cell wall (Figure 10A). As soon as the pollen tube tip enlarged after Spm treatment, cellulose appeared as more or less uniformly deposited in the swollen apex (Figure 10B) and no specific accumulation of cellulose was observed. The cellulose pattern changed distinctly when pollen tubes shifted from the balloon to the snake stage. In this case, a prominent accumulation of cellulose was observed in the so-called neck region of swollen tubes (Figure 10C). Relative quantization of the fluorescence signal clearly showed that cellulose is quite uniformly distributed in control pollen tubes while it apparently increases in the swollen tip at the balloon stage; during the transition toward the snake stage, accumulation of cellulose was mainly detected ca. 20 μm from the tip (Figure 10D). When enlarged pollen tubes resumed growth (at the transition between snake and shovel stages), cellulose deposition was observed along the border of the shovel-shaped tube; deposition of cellulose was not uniform because a marked accumulation was still found in both the neck region and the swollen portion (arrows in Figure 10E). Later, at the shovel stage (Figure 10F1), cellulose was detected in the enlarged pollen tubes with a more consistent accumulation in the subapex and neck regions, as shown by pseudo-colored imaging of cellulose fluorescence merged with a DIC-imaged pollen tube (arrows in Figure 10F2). This result was confirmed by relative measurement of fluorescence intensity along the cell border (graph in Figure 10G) showing that, at the end of the snake stage, cellulose still peaked at 15–20 μm from the tip, while the shovel stage was characterized by two distinct accumulation areas, one in the subapex, the other in the neck region.

While in control pollen tubes callose was uniformly distributed except in the very tip region (Figure 11A), it accumulated in the neck of Spm-treated pollen tubes at the onset of treatment (Figure 11B, arrows). Accumulation of callose in the neck region was more prominent at the snake stage (Figure 11C); callose levels remained constantly higher in the neck during the transition from the snake to the shovel stage (Figure 11D, square bracket). This observation was confirmed by measuring the relative fluorescence intensity in pollen tubes at both control and shovel stages (graph in Figure 11E). While control pollen tubes exhibited no callose deposition in the tip and a progressive accumulation starting from 20 to 30 μm, shovel-shaped pollen tubes were characterized by a consistent accumulation of callose in the neck region, which corresponds to ca. 30 μm from the tip (Figures 11D,E).

As the distribution of callose was altered by Spm, we checked, by immunofluorescence microscopy, if this might depend on the irregular localization of callose synthase. As shown in Figure 12A, in control pollen tubes the enzyme was detected along the entire periphery of pollen tubes, most likely in the plasma membrane. Distribution of callose synthase was not uniform, but can be described as organized in patches, more abundant in the apical zone (the hypothetical insertion site). After Spm treatment, at the balloon stage, callose synthase accumulated in the plasma membrane of the swollen apex (Figure 12B) again with a patchy pattern. Labeling was not limited to the swollen apex but also extended to distal segments (Figures 12C1,C2). When pollen tubes resumed growth (snake stage), callose synthase started to accumulate consistently in the neck region (Figures 12D1–D3) thus mirroring the neck-localized deposition of callose.

In some cases, deposition of callose synthase took the form of annulus-like structures surrounding the neck region (arrows in Figures 12E,F). Three reconstruction images from a Z-stack showing the pollen tube regions before the annulus, at the annulus level and after the annulus are shown in the insets of Figure 12E. A significant amount of callose synthase accumulation is evident in the central inset (i.e., annulus). The annulus was not uniform but was characterized by relatively more or less intense areas. The so-called “annulus” was never observed in control samples. In samples treated with spermine, presence of the annulus was not an absolute feature but it was still very frequent. By counting the analyzed pollen tubes, it turned out that ~60–65% of pollen tubes showed the structure defined as annulus. Currently, we do not know why this structure is not present in all treated pollen tubes but we are confident that it is neither an artifact nor an incidental structure. A video clip built with a series of distal-to-proximal Z stacks (with respect to the tube tip) illustrating the presence of the callose synthase annulus is available as Supplementary Movie 2.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.