The apical actin fringe contributes to localized cell wall deposition and polarized growth in the lily pollen tube.

In lily (Lilium formosanum) pollen tubes, pectin, a major component of the cell wall, is delivered through regulated exocytosis. The targeted transport and secretion of the pectin-containing vesicles may be controlled by the cortical actin fringe at the pollen tube apex. Here, we address the role of the actin fringe using three different inhibitors of growth: brefeldin A, latrunculin B, and potassium cyanide. Brefeldin A blocks membrane trafficking and inhibits exocytosis in pollen tubes; it also leads to the degradation of the actin fringe and the formation of an aggregate of filamentous actin at the base of the clear zone. Latrunculin B, which depolymerizes filamentous actin, markedly slows growth but allows focused pectin deposition to continue. Of note, the locus of deposition shifts frequently and correlates with changes in the direction of growth. Finally, potassium cyanide, an electron transport chain inhibitor, briefly stops growth while causing the actin fringe to completely disappear. Pectin deposition continues but lacks focus, instead being delivered in a wide arc across the pollen tube tip. These data support a model in which the actin fringe contributes to the focused secretion of pectin to the apical cell wall and, thus, to the polarized growth of the pollen tube.

Pollen tubes provide an excellent model for studying the molecular and physiological processes that lead to polarized cell growth. Because all plant cell growth results from the regulated yielding of the cell wall in response to uniform turgor pressure Rojas et al., 2011), the cell wall of the pollen tube must yield only at a particular spot: the cell apex, or tip. To accomplish the extraordinary growth rates seen in many species, and to balance the thinning of the apical wall due to rapid expansion, the pollen tube delivers prodigious amounts of wall material, largely methoxylated pectins, to the tip in a coordinated manner. Recent studies suggest that the targeted exocytosis increases the extensibility of the cell wall matrix at the tip, which then yields to the existing turgor pressure, permitting the tip to extend or grow (McKenna et al., 2009;Hepler et al., 2013). There are many factors that influence exocytosis in growing pollen tubes; in this study, we investigate the role of the apical actin fringe.
For many years, it has been known that an actin structure exists near the pollen tube tip, yet its exact form has been a matter of some contention (Kost et al., 1998;Lovy-Wheeler et al., 2005;Wilsen et al., 2006;Cheung et al., 2008;Vidali et al., 2009;Qu et al., 2013). The apical actin structure has been variously described as a fringe, a basket, a collar, or a mesh. Using rapid freeze fixation of lily (Lilium formosanum) pollen tubes followed by staining with anti-actin antibodies, the structure appears as a dense fringe of longitudinally oriented microfilaments, beginning 1 to 5 mm behind the apex and extending 5 to 10 mm basally. The actin filaments are positioned in the cortical cytoplasm close to the plasma membrane . More recently, we used Lifeact-mEGFP, a probe that consistently labels this palisade of longitudinally oriented microfilaments in living cells Fig. 1A, left column). For the purposes of this study, we will refer to this apical organization of actin as a fringe.
Our understanding of the process of exocytosis and pollen tube elongation has been influenced by ultrastructural images of pollen tube tips, which reveal an apical zone dense with vesicles Heslop-Harrison, 1987;Lancelle et al., 1987;Steer and Steer, 1989;Lancelle and Hepler, 1992;Derksen et al., 1995). It has long been assumed that these represent exocytotic vesicles destined to deliver new cell wall material. This model of polarized secretion has been challenged in recent years in studies using FM dyes. Two groups have suggested that exocytosis occurs in a circumpolar annular zone (Bove et al., 2008;Zonia and Munnik, 2008). However, other studies, using fluorescent beads attached to the cell surface, indicate that the maximal rate of expansion, and of necessity the greatest deposition of cell wall material, occurs at the apex along the polar axis of the tube (Dumais et al., 2006;Rojas et al., 2011). Similarly, our experiments with propidium iodide (PI; McKenna et al., 2009;Rounds et al., 2011a) and pectin methyl esterase fused to GFP (McKenna et al., 2009) show that the wall is thickest at the very tip and suggest that wall materials are deposited at the polar axis, consistent with the initial model of exocytosis (Lancelle and Hepler, 1992). Experiments using tobacco (Nicotiana tabacum) pollen and a receptor-like kinase fused to GFP also indicate that exocytosis occurs largely at the apical polar axis (Lee et al., 2008).
Many researchers argue that apical actin is critical for exocytosis (Lee et al., 2008;Cheung et al., 2010;Qin and Yang, 2011;Yan and Yang, 2012). More specifically, recent work suggests that the fringe participates in targeting vesicles and thereby contributes to changes in growth direction (Kroeger et al., 2009;Bou Daher and Geitmann, 2011;Dong et al., 2012). In this article, using three different inhibitors, namely brefeldin A (BFA), LatB, and potassium cyanide (KCN), we test the hypothesis that polarized pectin deposition in pollen tubes requires the actin fringe. Our data show that during normal growth, pectin deposition is focused to the apex along the polar axis of the tube. However, when growth is modulated, different end points arise, depending on the inhibitor. With BFA, exocytosis stops completely, and the fringe disappears, with the appearance of an actin aggregate at the base of the clear zone. LatB, as shown previously , incompletely degrades the actin fringe and leaves a rim of F-actin around the apical dome. Here, we show that, in the presence of LatB, pectin deposition continues, with the focus of this activity shifting in position frequently as the slowly elongating pollen tube changes direction. With KCN, the actin fringe degrades completely, but exocytosis continues and becomes depolarized, with pectin deposits now occurring across a wide arc of the apical dome. This dome often swells as deposition continues, only stopping once normal growth resumes. Taken together, these results support a role for the actin fringe in controlling the polarity of growth in the lily pollen tube.

The Fringe and the Thickened Tip Are Stable yet Dynamic
Although the cell wall at the pollen tube tip oscillates in thickness (McKenna et al., 2009), it is important to recognize that, even at its thinnest point, the apical cell Figure 1. The actin fringe and the thickened pollen tube tip wall are stable, although dynamic, structures during pollen tube growth. A, The left column shows a pollen tube transformed with Lifeact-mEGFP imaged with a spinning-disc confocal microscope. Maximal projections from every 15 s are shown. The right column shows epifluorescence images of a pollen tube stained with PI. Again, images captured every 15 s are shown. Bars = 10 mm. B, The data from the pollen tube in A expressing Lifeact-mEGFP were subjected to kymograph analysis using an 11-pixel strip along the image's midline. C, The first three frames from the pollen tube in A and B were assigned the colors red, blue, and green, respectively, and then overlaid. Areas with white show the overlap of all three. The fringe is stable, but most of its constituent actin is not shared between frames.
wall always remains thicker than the cell wall along the shank. We monitor the changes in wall thickness using the fluorescent signal from PI, with which the pollen tubes have been stained. The efficacy of this approach is based on the studies of McKenna et al. (2009) showing that the changes in PI fluorescence are virtually identical to the changes in cell wall thickness measured by light microscopy during oscillatory growth in the same lily pollen tubes. Subsequently, the same group determined in tobacco pollen tubes that the changes in PI fluorescence closely matched the fluorescent signal derived from the secretion of pectin methyl esterase fused with GFP (McKenna et al., 2009). Finally, in more recent work, Rounds et al. (2011a) established that PI competes with Ca 2+ in binding to demethoxylated pectin. We conclude from these studies that PI binds pectin and that it faithfully indicates the appearance of new cell wall material. PI fluorescence thus serves as a sensitive marker for pectin deposition in pollen tubes (Fig. 1A, right column; Supplemental Video S1; Rounds et al., 2011a).
As a first step in testing our hypothesis that the actin fringe regulates pectin deposition, we asked whether there was a corresponding change in the fringe that reflected cell wall dynamics during growth. To image actin in lily pollen tubes expressing Lifeact-mEGFP ), we used spinning-disc confocal microscopy. The improved time resolution of the spinningdisc confocal microscope relative to the conventional laser scanning confocal microscope permits us to obtain an image of both the longitudinal filaments along the shank (Fig. 1A, left column; Supplemental Video S2) and the apical actin fringe at close time intervals. A kymographic analysis of the pollen tube shown in Figure 1B documents the constancy of the fringe over time. In the kymograph, a subtle line runs immediately behind the tip (Fig. 1B, arrow). Its slope closely matches that of the tip itself, indicating that the fringe maintains a stable position relative to the tip despite the constant growth.
Actin filaments have been shown to be dynamic in plant cells . To examine whether the actin fringe is composed of stable filaments, we overlaid the first three images of the series shown in Figure 1A. Each image was given a different color, and then the images were superimposed on one another. Areas with coincident actin appear white, whereas areas with no coincident actin appear in one of the individual colors. Because very few white domains are seen, we conclude that, although the structure of the actin fringe as a whole is stable, the individual microfilaments are constantly changing.

BFA Blocks Exocytosis and Destroys the Actin Fringe
The fungal macrocyclic lactone BFA blocks the activity of an ADP ribosylation factor G nucleotide exchange factor, ultimately resulting in the inhibition of membrane trafficking (Nebenführ et al., 2002). Previous studies on pollen tubes have established that the drug halts pollen tube growth and dramatically alters vesicle trafficking. Recent data in particular indicate that both endocytosis and exocytosis are affected (Rutten and Knuiman, 1993;Parton et al., 2003;Hörmanseder et al., 2005;Wang et al., 2005;Zhang et al., 2010b). None of these studies actually examined deposition, instead focusing on internal membrane markers and membraneassociated proteins. BFA also alters the actin cytoskeleton, specifically disrupting the apical actin structure in Arabidopsis (Arabidopsis thaliana) pollen tubes (Zhang et al., 2010b). Earlier work on lily pollen tubes showed that the addition of BFA led to the appearance of a cytoplasmic aggregate behind the tip, called the BFAinduced aggregate, or BIA (Parton et al., 2001(Parton et al., , 2003. It was further shown that the formation of the BIA depended on an intact actin cytoskeleton, because antiactin drugs, including cytochalasin and jasplakinolide, either blocked or dispersed the formation of the BIA. Zhang et al. (2010b) described a dissipation of apical actin, whereas Parton et al. (2003) documented the appearance of an actindependent structure. These two descriptions are in some ways contradictory. We wished to further investigate the effects of BFA directly on the fringe in lily pollen tubes.
We asked what happens to the structure and distribution of apical actin and to changes in cell wall pectin deposition in pollen tubes treated with BFA. As the drug takes effect, growth slows and PI fluorescence becomes significantly less intense and more uniform over the apical dome. Although BFA takes several minutes (e.g. 20 min) to impact cell growth, once the pollen tube had stopped, marked alterations are clear, which are evident as a substantial decline in the PI intensity at the cell apex ( Fig. 2A, compare differential interference contrast [DIC] and fluorescence images before and after BFA-induced growth inhibition; Supplemental Video S3). Note especially that the intensity along the sides of the pollen tube does not change significantly when compared with the apex (compare the two traces in Fig. 2B). We interpret the concurrent changes in PI fluorescence and growth rate to mean that expansion of the apical cell wall continues for a short time, but the absence of vesicles due to the action of BFA makes exocytosis of new pectin impossible. The cell wall at the tip becomes more resistant to expansion, possibly due to the action of pectin methyl esterase on existing wall pectins, which leads to greater calcium cross-linking and ultimately the cessation of growth. These data clearly show that BFA blocks cell wall deposition, but they also further support the idea that exocytosis is confined largely to the pollen tube tip.
BFA also exhibits a profound effect on actin organization. When we treat cells expressing Lifeact-mEGFP with BFA, the apical fringe collapses and actin microfilaments appear as a star-like aggregation of GFP immediately behind the tip ( Fig. 2C; Supplemental Videos S4 and S5), in approximately the same position as the BIA, as described by Parton et al. (2003). While the apical actin fringe degrades, at least some of the actin cables in the shank of the tube remain, and these drive the streaming that is evident.

LatB Leads to Focused Points of Pectin Deposition That Frequently Change Position
LatB, which blocks actin polymerization by sequestering globular actin, has been shown to disrupt the localization of RabA4b, a Rab GTPase involved in exocytosis, in both root hairs and pollen tubes (Preuss et al., 2004;Zhang et al., 2010b). Previous work has also shown that the actin fringe dissipates in the presence of LatB (Cárdenas et al., 2005). In more recent studies on cells expressing Lifeact-mEGFP, Vidali et al. (2009) have shown that 2 nM LatB disrupts the actin fringe per se; however, a thin rim of cortical fluorescence arises and spreads across the apical dome (see figure 7, C and D, and Supplemental Movie S5 in Vidali et al., 2009). While some of the fluorescence in this rim may be due to Lifeact associated with globular actin, a fraction is due to randomly oriented cortical F-actin . It seems likely that these cortical actin filaments drive the streaming that is observed within the tube apex. Here, we see all organelles, including the amyloplasts, which heretofore had been excluded from the apical dome, moving through the apex, obliterating the clear zone.
When we treated pollen tubes with 2 nM LatB and monitored changes in pectin deposition as measured by PI fluorescence, the pollen tubes exhibited various defects in growth before the clear zone collapsed and cells stopped growing, in agreement with previous results (Vidali et al., 2001). Before completely stopping, cells generally began turning frequently, and the PI signal tended to move around the apical region rather than staying centered at the apex. A zone of intense PI fluorescence often presaged the formation of a new growing tip. Frequently, pollen tubes developed more than one deposition zone, as evidenced by the presence of two focused points of PI fluorescence in Figure  3, bottom right (also see figure 7C in Vidali et al., 2001).

KCN Reversibly Inhibits Growth
With a desire to achieve rapid inhibition together with relatively quick and full recovery of pollen tube growth, we built upon earlier studies showing that inhibitors of the mitochondrial electron transport chain, in particular KCN, were particularly effective Figure 2. BFA reduces PI signal at the pollen tube tip and alters the actin fringe. A, BFA (15 mM) was added to a growing pollen tube stained with 20 mM PI. The top pair of images represents the state of the pollen tube before the addition of BFA. The first image is DIC, and the second is PI fluorescence. The second pair represents the same pollen tube after BFA inhibition of exocytosis and growth. Note the dramatic reduction in PI fluorescence at the apex. B, Data drawn from the same experiment as shown in A. The red trace represents the PI fluorescence at the cell apex. The blue trace represents the PI signal 16 mm behind the apex along the cell's edge. A.U., Arbitrary units. C, A similar experiment carried out with a pollen tube expressing Lifeact-mEGFP. The image on the left represents the cell before BFA inhibition, and the right image shows the cell after BFA inhibition. The fringe is prominent on the left and is no longer present on the right; instead, a star-like formation is clear. Scale bar 5 10 mM. (Rounds et al., 2010;Obermeyer et al., 2013). Accordingly, lily pollen tubes were grown in a flow-through chamber, which allowed the quick addition and removal of KCN. Shortly after the addition of 200 mM KCN, pollen tubes stopped growing (approximately 1 min). The KCN was washed out with fresh growth medium immediately after growth stoppage, and within a few minutes the cells began growing again. In Figure 4A, a representative pollen tube is shown before, during, and after inhibition (Supplemental Videos S6 and S7). The left column shows DIC images, whereas the right column shows the fluorescent NAD(P)H signals, where the latter serves as a monitor of mitochondrial electron transport chain activity (Cárdenas et al., 2006). Notice that the intensity of the signal after washout is roughly equivalent to that before inhibition. Figure 4B provides a quantification of the data for the same pollen tube. The blue line represents growth, and the red line shows the NAD(P)H signal 20 mm behind the tip. Note that after recovery, the growth rate is similar to that before inhibition. Also, the NAD(P)H signal clearly oscillates after recovery. Taken together, these data support the use of KCN as a quick-acting, fully reversible inhibitor of pollen tube growth.

Pectin Deposition Continues during KCN-Dependent Growth Inhibition
To monitor apical pectin deposition, we used PI during cyanide-dependent growth inhibition. Figure 5A . LatB leads to focused points of pectin deposition. Pollen tubes treated with 2 nM LatB lose their polarity as growth slows markedly. Notably, the deposition of pectin, as indicated by PI fluorescence, no longer remains strictly at the polar axis but wanders over the apical region. Often, deposition can be observed simultaneously at two locations (bottom two image pairs). Eventually, the pollen tube stops elongating completely. The effects shown are typical, but overall there is considerable variability. Scale bar 5 10 mM. . KCN inhibition is fully reversible. Pollen tubes were grown on microscope slides, and then at 36 s, 200 mM KCN was added via a peristaltic pump. At 120 s, shortly after growth had stopped, washout of KCN was begun. A, DIC and NAD(P)H signals for a pollen tube going through KCN inhibition. The top image pair is from the beginning of the experiment, while the middle pair shows the pollen tube during inhibition. Note the high NAD(P)H signal. The bottom pair shows the recovered pollen tube with a lower NAD(P)H signal. Scale bar 5 10 mM. B, Data drawn from the experiment shown in A are displayed graphically. The blue trace represents growth rate, and the red trace represents the NAD(P)H signal from a 10-mm box behind the tip. Note that both signals oscillate both before and after, but not during, KCN inhibition. AU, Arbitrary units.
shows that the PI signal first drops sharply during growth inhibition but then increases dramatically at the cell tip (Supplemental Video S8). To quantify the PI signal, we measured the intensity along a midline during the entire movie. This procedure describes the most intense signal at the very tip of the pollen tube. These data are charted as the red line in Figure 5B against the growth rate (in blue). As expected, before growth inhibition, the PI signal oscillates, as does the growth rate (McKenna et al., 2009). Following the application of cyanide, both the growth rate and the apical PI signal decline dramatically. Remarkably, the PI signal almost immediately reverses and begins trending upward again, as indicated by the increase in fluorescence (Fig. 5A, middle image). Shortly before growth resumes, the PI signal declines somewhat and then, along with the growth rate, resumes oscillating.
As the PI signal has been shown to measure the amount of material in the cell wall by competing with Ca 2+ for binding to demethoxylated pectins (Rounds et al., 2011a), these data indicate that pectin deposition continues when growth is inhibited by KCN. This is somewhat surprising, as one would expect exocytosis, and perhaps the maintenance of the actin fringe, to be a significant energy sink and thus unable to occur in the absence of oxidative phosphorylation (Rounds et al., 2011b). However, a sufficient supply of ATP may derive from glycolysis and aerobic fermentation, which increases when the mitochondrial electron transport chain is blocked (Rounds et al., 2010;Obermeyer et al., 2013). Interestingly, despite these noticeable changes in cell wall deposition, the DIC image does not show a dramatic change in intracellular morphology; although the cell swells somewhat, the clear zone does not disappear, and the general cell structure of a normal uninhibited tube remains (Fig. 4A, left column).

The Actin Fringe Dissipates during KCN Inhibition
To determine the effect of metabolic inhibition on the actin fringe, we subjected pollen tubes expressing Lifeact-mEGFP to inhibitory levels of KCN as described above and then monitored the GFP signal using a spinning-disc confocal microscope. The cortical actin fringe is apparent at the beginning of the experiment (Fig. 6A, top; Supplemental Video S9), but during growth inhibition it disappears, although the longitudinal actin filaments remain (Fig. 6A, middle; Supplemental Video S9). As growth restarts, the actin fringe reforms (Fig. 6A, bottom; Supplemental Video S9) and is stable once again as growth recovers. To better visualize this phenomenon, data from the same experiment shown in Figure 6A are displayed as a kymograph (Fig. 6B). The slope of the line going from the left represents the growth. Under normal conditions, the fringe can be visualized as a stripe of fluorescence immediately behind the tip. However, when the pollen tube growth stops, and the line representing the apex becomes nearly vertical, the stripe indicating the fringe disappears. It is also noteworthy that actin organization overall at the tip is reduced markedly (Fig. 6B, asterisk). Taken together, these observations show that the fringe is completely absent during growth inhibition by KCN; however, exocytosis continues (Fig. 5).

Pectin Deposition Loses Polarity in the Presence of KCN
Although pectin deposition continues despite the absence of the actin fringe (Fig. 5), it seems possible that its directional specificity might be altered. That is, vesicle targeting might not be focused or might lose polarity entirely. To test this, we analyzed the deposition of pectin under three conditions: with BFA, with KCN, and without an inhibitor. To quantify deposition, the PI signal from each pixel along the edge of the Figure 5. PI continues to accumulate in the absence of apical extension. A, KCN (200 mM) was added to the growing pollen tube at 30 s. At 150 s, the KCN washout was begun. The images show PI fluorescence. Bar = 10 mm. B, Data drawn from the experiment shown in A are graphically displayed. The blue trace represents the growth rate of the pollen tube tip, whereas the red line represents the PI fluorescence at the apex of the cell. Note that both the growth rate and PI fluorescence oscillate before and after, but not during, growth inhibition. A.U., Arbitrary units. pollen tube was measured using a custom graphic analysis program created with the R statistical analysis package (see "Materials and Methods"). These data were then compiled in a matrix consisting of the edge values for each frame in the time series. The matrix can then be visualized as a kymograph where the y axis represents all points along the edge in a given frame and the x axis shows time (Fig. 7A). The pollen tube's apex is set to 0 mm along the y axis. Therefore, the signal along one edge of the pollen tube is a positive distance from the apex and that along the other edge is negative. A kymograph of the PI signal for a single control pollen tube is shown in Figure 7A. The pollen tube's growth can be visualized by noting how the top and bottom edges trend upward or downward as one's eye moves from left to right along the x axis. This reflects the fact that the pollen tube gets longer and, therefore, the edge is longer. One can also see that the majority of deposition occurs near the apex, with the highest signal centering at 0 mm throughout the kymograph. Furthermore, very little change in the signal can be observed 5 to 8 mm away from the apex.
When the edge values for a single time point are plotted, with distance from the apex on the x axis and signal intensity on the y axis, a Gaussian distribution emerges (Fig. 7B). The highest signal is at the apex, which drops quickly farther from the apex. The point with the highest signal represents the area of maximal pectin deposition (McKenna et al., 2009;Rounds et al., 2011a). We fit the PI data to a Gaussian function using a nonlinear least-squares fitting algorithm within R. From the Gaussian fit, we derived an SD, s, thus providing a statistical description of the dispersion of the PI signal around the tube tip. If we model the pollen tube apex as a circle, 2 3 s describes an arc of the highest signal along this circle (Fig. 7, B and C; see "Materials and Methods"). This arc represents the majority of the PI signal at the apex; that is, it is the area of greatest deposition. As pollen tubes vary in diameter quite substantially, we chose to quantify PI dispersion by calculating the subtended angle of this arc (Q; Fig. 7C). We calculated Q for each frame of the pollen tube used for Figure 6A and plotted it against the time (Fig. 7D). This allows us to compare Q with the amplitude of PI fluorescence at the apex (Fig. 7D). Notice in particular that the amplitude of the PI signal oscillates as expected (McKenna et al., 2009;Rounds et al., 2011a), as does Q. In untreated pollen tubes, Q oscillates around 62°( 63°SE; n = 10). We also analyzed the change in Q and, therefore, the arc of pectin deposition using BFA and KCN. We reasoned that because BFA abolishes all exocytosis, Q should go from the control value of approximately 62°to 0°or an undefined angle. If KCN causes a loss of control over deposition, then the angle should increase dramatically. We conducted trials using 10 control pollen tubes, 11 treated with KCN, and five treated with BFA and then analyzed the results as with the control described above.
The BFA treatment shows a dramatic reduction in deposition (Fig. 8A). When quantified along with the signal at the apex, it is clear that deposition has slowed dramatically or stopped (Fig. 8, A and B), in agreement with our initial observations (Fig. 2). Our analysis of five pollen tubes confirms this observation and reveals that Q, which begins at 45°to 50°, cannot be calculated after treatment with BFA because the PI signal is low and uniform across the tip.
In the KCN-treated pollen tube shown (Fig. 8, C and D), the deposition initially falls off as with the BFA treatment, suggesting that, at first, deposition at the apex halts. Then, quite suddenly, deposition begins again, but Q is changed dramatically. After first stopping, Q quickly rises to a plateau significantly above the preinhibitor Q (Fig. 7D). Our analysis showed that the Q rose from 67°6 3°(n = 11) to 136°6 18°(n = 11; Fig. 9A), or roughly twice as high (P , 0.001 with Student's t test). These data suggest that in the presence of KCN deposition, polarity has been altered dramatically, perhaps even lost entirely.

DISCUSSION AND CONCLUSION
The results of this study indicate that the cortical actin fringe plays a major role in controlling the polarized Figure 6. The actin fringe dissipates during KCN inhibition. KCN was added to a growing pollen tube expressing Lifeact-mEGFP at 49 s, and the washout was begun at 126 s. A, The top image shows the pollen tube at the beginning of the experiment before KCN addition. A prominent apical fringe of actin can be seen. The middle image shows the pollen tube during inhibition, when the cell is not growing. Longitudinal filaments are still present, but the fringe is absent. The bottom image shows the pollen tube after growth recovery; note that the fringe has reappeared. Scale bar 5 10 mM. B, A kymograph of a five-pixelwide band down the center of the pollen tube shows the fate of the actin fringe in response to KCN. When the fringe is present, it appears as a diagonal line immediately behind the tip. Arrows denote where the fringe disappears along with growth cessation and then reappears with the resumption of growth. The asterisk shows the area where the fringe is completely absent. deposition of pectin, and therefore growth, in the lily pollen tube. In the first experimental approach, we used BFA, a well-established inhibitor of the exocytotic pathway. Here, we show that the deposition of pectin, as indicated by extracellular PI fluorescence, declines markedly in the absence of exocytosis. Simultaneously, we observe that the actin fringe dissipates and that a starshaped aggregate of Lifeact-mEGFP appears in the clear zone behind the tube apex. These observations suggest to us that the process of exocytosis and the presence of a normal cortical actin fringe are connected.
The second experimental approach involves the use of LatB as a sensitive inhibitor of actin turnover and dynamics. By binding to actin monomers, LatB prevents the formation of F-actin, thus leading to its degradation, especially of those filaments that are rapidly turning over. The apical actin fringe indeed turns over quickly and keeps strict pace with the rapidly growing pollen tube, as Figure 7. Deposition of pectin is restricted to a distinct zone at the pollen tube apex. A, The PI signal along the edge of a growing pollen tube was collected at each time point in a series. The apex is assigned 0 mm on the y axis. The PI intensity at each point moving away from the tip along the pollen tube's edge is then plotted on the y axis. The x axis represents time. This kymograph shows the changes in fluorescence intensity along the pollen tube's edge over time. B, The PI signal at a single time point (60 s) is plotted, with the distance from the apex on the x axis and the intensity on the y axis. A Gaussian function was fit to these data, and the s was calculated. The dashed blue line shows the apex, and the dotted blue line shows the distance indicated by s. C, Schematic shows the relationship between s, which defines an arc at the apex, and Q, the angle subtended by twice this arc. D, Data drawn from the same pollen tube show changes in the amplitude of the apical PI signal (blue) and Q (red). A.U., Arbitrary units.
shown in Figure 1. Therefore, it is not surprising that this array preferentially degrades in the presence of low concentrations of LatB. While 2 nM LatB degrades actin structure, nevertheless a cortical array of actin appears in the apex and, as shown by Vidali et al. (2009), is expressed as a rim of actin around the perimeter of the apical dome. A through-focus series reveals that at least some of the Lifeact-mEGFP fluorescence derives from randomly oriented microfilaments in the plane of the plasma membrane .
In lily pollen tubes treated with LatB, it is particularly interesting that pectin deposition continues but remains focused to definite hot spots. However, these focal locations change relatively quickly, where they define new axes of pollen tube growth. It is possible that a similar process occurs in tobacco pollen tubes expressing an RNA interference or antisense transgene against formin FH5, given their modified structure of apical actin and their meandering growth pattern (Cheung et al., 2010). For lily pollen tubes, which change growth direction when treated with LatB, it seems plausible that local exocytosis occurs in regions where there are breaks in the cortical rim of actin and that these momentarily become the focused deposition sites we observe.
The third experimental approach used KCN as a reversible inhibitor of the electron transport chain and pollen tube growth. Shortly after the addition of KCN, pollen tube growth stops abruptly. Concomitantly, the apical fringe dissipates. At this juncture, pectin deposition as measured by PI fluorescence at the apex drops dramatically, although only temporarily. Almost immediately, Figure 8. KCN and BFA alter the deposition of pectin at the pollen tube apex. A and B, Results from a single experiment in which BFA was added to a pollen tube. The kymograph and Q graphs were created as described in Figure 7. Deposition halts just after 300 s for BFA, and Q becomes undefined. C and D, Treating pollen tubes with KCN at first causes an apparent loss in deposition, as seen in the BFA experiment, but then results in a continuous increase in deposition over a broader section of the apex, as shown in D. A.U., Arbitrary units. deposition begins again, but with a significant change. No longer is it focused on the apex; instead, deposition occurs over a broad arc at the pollen tube tip. We have measured the angle subtended by this arc (Q) and show that it doubles in the absence of the apical actin fringe (Fig. 8, C and D). Given the absence of the apical actin fringe under these experimental conditions, it is attractive to conclude that this unique cytoskeletal structure contributes to the polarized wall deposition. Thus, during normal growth when the cortical fringe is present, moving vesicles to the extreme apex might prevent their docking and fusion along the side of the pollen tube, ensuring that these events will be focused at the tube tip (Fig. 8D). However, when the fringe is degraded (e.g. during KCN treatment), exocytosis takes place more widely, accounting for the loss of polarity in wall deposition. In addition, the more widely spread infusion of methoxylated pectins will broadly weaken the wall, allowing the cell apex to swell (Fayant et al., 2010); although some swelling is usually observed, it is not as dramatic as the depolarization in pectin deposition (Fig. 8D).
We fully recognize that many factors, including physiological processes and signaling pathways, contribute to the formation and activity of the actin fringe as well as to membrane trafficking. Among the players, the GTPase ROP has often been characterized as a master regulator (Qin and Yang, 2011). Through its effector proteins Ric3, Ric4, and ICR1 Interactor of Constitutive Active Rho-of-Plants1 (ICR1), ROP is thought to control Ca 2+ influx, actin dynamics, and exocytosis (Fu et al., 2001;Gu et al., 2005;Hwang et al., 2005;Lavy et al., 2007;Lee et al., 2008). In this scenario, active ROP through Ric4 initiates actin polymerization and the accumulation of vesicles at the tip. Through Ric3, active ROP initiates the subsequent Ca 2+ influx and actin depolymerization. Finally, through ICR1, ROP leads to the recruitment of the exocyst and vesicle tethering (Lavy et al., 2007;Qin and Yang, 2011). These results are also consistent with the idea that apical actin is critical for exocytosis.
While a ROP-dependent sequence may play an important regulatory role, it is not the only pathway that deserves attention. Given the tip-focused Ca 2+ gradient, which will have profound effects on the apical actin, possibly through specific actin-binding proteins such as villin/gelsolin Zhang et al., 2010a;Qu et al., 2013), it becomes apparent that any mechanism that modulates intracellular Ca 2+ could induce a cascade of events crucial to the targeted delivery of vesicles and polarized growth. Candidate Ca 2+ channels include those that are activated by mechanical forces or stretch (Dutta and Robinson, 2004), by changes in the membrane potential (Wang et al., 2004;Shang et al., 2005;Qu et al., 2007;Wu et al., 2007Wu et al., , 2010, or by the presence of a ligand Figure 9. KCN and BFA affect both cell wall deposition and actin structure. A, Q in pollen tubes before KCN inhibition (fringe present), during growth stoppage (fringe absent), and under control conditions. Error bars represent SE. Student's t test between tubes before and after KCN treatment yielded P , 0.001. There was no significant difference between pretreatment and control tubes. B, Gray arrows in control (center) or KCN-treated pollen tubes show the direction of vesicles leading to the deposition of wall material. The deposition is represented by the gray-to-black gradient shown at the apex in the control and spreading backward in the KCN-treated pollen tube. In the image representing the control pollen tube, the actin fringe is represented as black lines at the base of the clear zone. This fringe is missing from KCN-treated pollen tubes, whereas the actin is reorganized into a star-shaped aggregate in BFA-treated tubes.
such as reactive oxygen species (Potocký et al., 2007;Wu et al., 2010), cyclic nucleotides (Frietsch et al., 2007), or amino acids (Michard et al., 2011;for review, see Hepler et al., 2012). In brief, there is a rich array of factors and mechanisms that could modulate Ca 2+ and, by extension, the structure and organization of the cortical actin fringe.
Finally, the results presented in this study show that during normal pollen tube growth, deposition of PIstained pectin occurs primarily at the apex of the tube, with a dispersion of approximately 60°centered at the polar axis. These results are in close agreement with those obtained through the study of the growth-dependent displacement of fluorescent particles on the pollen tube surface (Rojas et al., 2011). These results may also include sufficient breadth to render them in agreement with studies using the membrane-marking FM dyes (Bove et al., 2008;Zonia and Munnik, 2008). Those authors have argued that exocytosis occurs in an annulus around the polar axis. However, given the rapid increase in area as one moves away from the polar axis, there will unavoidably be a great need for additional exocytosis from that seen strictly at the polar axis.
The results from this study provide support for the hypothesis that the cortical actin fringe plays a pivotal role in controlling the polarized growth of the pollen tube (Fig. 8B). During normal growth in the presence of the actin fringe, deposition is limited to the apical dome. However, when growth is blocked with BFA or KCN, the fringe degrades. BFA, because it inhibits membrane trafficking, prevents any further exocytosis. By contrast, KCN only stops pectin deposition momentarily. Even in the continued presence of the metabolic inhibitor, deposition resumes, but notably, it now occurs over the entire apical dome of the pollen tube. However, when the inhibitor is removed and normal growth resumes, the actin fringe reemerges, providing evidence for the close coupling between the actin fringe and polarized pectin deposition. Finally, with LatB, although the localized actin fringe degrades, a rim of cortical actin appears, which is spread across the dome of the pollen tube . Under these conditions, some pectin deposition continues, which is confined to small regions. When taken together, these results give support to the idea that cortical actin participates in moving vesicles to the preferred sites of fusion at the extreme tube apex.

Pollen Tube Growth Conditions
Pollen was grown as described previously (Rounds et al., 2010). Briefly, all pollen was from lily (Lilium formosanum) stocks stored at -80°C and germinated for 1 to 1.5 h on a rotator at room temperature in a standard growth medium (Lily Pollen tube Growth Medium [LPGM]: 7% (w/v) Suc, 1.6 mM H 3 BO 3 , 0.1 mM CaCl 2 , and 15 mM MES buffer adjusted to pH 5.7 with 10 N KOH (a final potassium concentration of approximately 6 mM); all reagents were from Fisher Scientific unless noted otherwise. For microscopic observations, pollen was prepared one of two ways. For wide-field epifluorescence microscopy, a pollen suspension was spread on custom-made well slides with a growth medium solution containing a final concentration of 0.7% (w/v) low-melting agarose (Sigma-Aldrich). The immobilized pollen was then covered with fresh growth medium for imaging. Cells were allowed to recover for at least 0.5 h before imaging or further manipulations. For spinning-disc confocal microscopy, cells were cultured in agarose as described above or on slides coated with high-molecular-mass (150-300 kD) poly-L-Lys. These slides were first cleaned in a Harrick Plasma cleaner (PDC-001) for 30 s. They were then coated in approximately 50 mL of 1 mg mL 21 poly-L-Lys/water solution and allowed to dry. The slides were thoroughly rinsed in water and then allowed to air dry before use.

PI Staining of Pollen Tubes
For PI staining of the primary cell wall, the growth medium covering the immobilized and recovered cells on microscope slides was replaced with LPGM supplemented with 20 mM PI (Sigma-Aldrich). Cells were imaged more than 10 min after addition of the stain.
Growth Inhibition with KCN, BFA, and LatB Cells were grown as described above. For inhibition with KCN, a two-tube peristaltic pump (Bio-Rad) was used to add 200 mM KCN in LPGM. One tube was used to remove the growth medium from the slide under analysis, while the other added medium with inhibitor. The rate was set to approximately 0.5 mL min 21 for the duration of the experiment. For experiments with either BFA or LatB, cells were immobilized in agarose or on poly-L-Lys, and then 200 mL of LPGM was added. Cells were allowed to recover, and then 200 mL of LPGM supplemented with either 30 mM BFA or 4 nM LatB was added to yield a final concentration of 15 mM BFA or 2 nM LatB.

DIC and Epifluorescence Microscopy and Imaging
PI, NADH, and DIC images were acquired using a CCD camera (Quantix Cool Snap HQ; Roper Scientific) attached to a Nikon TE300 inverted microscope (Nikon Instruments) with a 403/1.3 numerical aperture oil-immersion objective lens. All the equipment was operated with MetaMorph/MetaFluor (Molecular Devices) software. A filter wheel system (L10-2; Sutter Instruments), mounted immediately before the CCD camera, controlled the position of a polarizing filter for DIC or an emission filter for fluorescence imaging. Fluorescence excitation light was provided by a 175-W ozone-free xenon lamp in a DG-4 switching system (Sutter Instruments). Transmitted light was provided by a low-voltage halogen lamp. We used the following filter setup for PI imaging: excitation, 495 nm; a 565-nm dichroic long pass; and emission, 509 nm long pass (all filters were from Chroma). Exposure times varied but were generally approximately 10 ms for DIC and approximately 800 ms for PI unless stated otherwise. We used the following filter setup for NAD(P)H imaging: 360 nm (10-nm band pass) as excitation filter; 380-nm dichroic; and 400-nm long-pass emission filter (all filters were from Chroma). We employed an exposure time of 750 ms and binned the images using ImageJ before analysis.

Spinning-Disc Confocal Microscopy and Imaging
For imaging of Lifeact-mEGFP, slides were prepared as described above and then mounted on an inverted microscope (model Ti-E; Nikon) equipped with a spinning-disc head (model CSU-X1; Yokogawa Corporation of America) and a 512 3 512 electron-multiplying CCD camera (iXON; Andor Technology). Images were collected with a 1.4 numerical aperture/603 oil-immersion objective (Nikon) at room temperature. Fifty percent laser power was used with the 488-nm laser. The exposure time varied between 50 and 100 ms. The image-acquisition process was controlled by MetaMorph software (Molecular Devices). To create maximal projections, ImageJ (Abramoff et al., 2004) was used. Lifeact-mEGFP kymographs were prepared using the multikymograph plugin for ImageJ with a fiveor 11-pixel-wide scan along a hand-selected midline.

Growth Rate and NAD(P)H Fluorescence Measurements
Growth rate was measured using the tip-tracking feature of the MetaMorph software package (Molecular Devices). The average NAD(P)H fluorescence was measured in a 10-mm 2 box centered 5 mm from the pollen tube tip (Cárdenas et al., 2006) using a custom R script (Ihaka and Gentleman, 1996;Rounds et al., 2010).

Bombardment
Plasmid DNA for Lifeact-mEGFP was constructed as described . Plasmid DNA was prepared using alkaline lysis followed by precipitation with polyethylene glycol and extraction with phenol-chloroform. DNA was coated onto 1 to 3 mg of 1.1-mm-diameter tungsten particles (Bio-Rad) according to the manufacturer's instructions. The coated microprojectiles were divided into aliquots onto two macrocarriers (Bio-Rad). Pollen was allowed to hydrate in 1 mL of the appropriate growth medium (see below) before being placed on a 25-mm MF-Millipore membrane (Millipore), which in turn was set on Whatman paper moistened with pollen growth medium. The macrocarrier assembly was positioned in the top slot of the PDS-1000/He biolistic system and the sample assembly in the slot below (Bio-Rad). Pollen grains were bombarded twice (once with each aliquot) using a 1,100-p.s.i. rupture disc (Bio-Rad). After bombardment, pollen was transferred to a microcentrifuge tube and incubated for 2 h at room temperature with constant rotation. Cells were then immobilized on a microscope slide in growth medium as described above.

Edge/Midline Detection and PI Fluorescence along the Edge
For each tiff image in a sequence, approximate cell edges were first estimated by segmentation and Canny edge detection in ImageJ (Abramoff et al., 2004). The cell midline was drawn by eye using smoothing splines and then digitized. Approximated edges were digitized and then used as input to an R script that scanned the perimeter of each cell pixel by pixel, extending a sample line perpendicular to the approximate edge trace (Supplemental Protocols S1 and S2). Bicubic interpolation along the sample line provided pixel values that were used to determine the point of the maximum PI signal. The X and Y locations of the maximum PI signal were used as the first estimate of edge location. The center of the tip was calculated as the intersection of the midline and the edge trace. Then, distances from the tip along the cell edge (S) were calculated for each frame in a sequence as the geometric distance between pairs of (X,Y) points. X and Y values were parameterized as smooth splines versus S, and new X and Y coordinates for the smoothed edge were interpolated and used to determine the PI value at each point from the original images. The location of the edge and the value of the PI fluorescence at each location were saved for future analysis.

Kymographs
Kymographs were plotted using the kymo command in R.

Dispersion
For each image in a series, PI fluorescence versus S (with S = 0 as the cell tip) data were first smoothed, minimum and maximum values were determined, and the data were normalized to the maximum PI value. The resulting data were fit to a Gaussian function using the nls (for nonlinear least squares) function in R: PI norm ¼ e 2 ðS 2 mÞ 2 2s 2 where PI norm is the normalized PI fluorescence, S is the distance from the tip, m is the location of the peak, and s is the SD (e.g. the square root of the variance, a measure of the dispersion of the function and of the PI signal). To compare pollen tubes over time and treatment, we used s to calculate the angle Q (°) as an index of PI dispersion (Fig. 4) as follows:

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Protocol S1. R script for generating kymographic data.
Supplemental Protocol S2. R script for analyzing kymographic data to determine Q.
Supplemental Video S1. PI labeling of a lily pollen tube.
Supplemental Video S3. PI-labeled pollen tube treated with BFA.
Supplemental Video S5. The same pollen tube as in Supplemental Video S4 after BFA inhibition.
Supplemental Video S6. NAD(P)H autofluorescence in a lily pollen tube treated with KCN.
Supplemental Video S7. DIC image of a lily pollen tube treated with KCN.
Supplemental Video S8. PI-labeled lily pollen tube treated with KCN.