Author for correspondence:Peter K. HeplerTel. +1 413 545 2083Fax: +1 413 545 3243Email: firstname.lastname@example.org
I. Introduction 000
II. Ion gradients and flux patterns 000
III. Oscillations 000
IV. The need for a Ca2+ store 000
V. Intracellular targets for Ion activity 000
VI. Extracellular targets for ions: the cell wall 000
VII. Ions in navigation 000
VIII. Role of ions in self-incompatibility 000
IX. The plasma membrane; site of global coordination and control 000
X. A model for pollen tube growth 000
IX. Conclusions 000
Pollen tube growth attracts our attention as a model system for studying cell elongation in plants. The process is fast, it is confined to the tip of the tube, and it is crucial for sexual reproduction in plants. In the enclosed review we focus on the control of pollen tube growth, giving special attention to the role of ions, especially calcium and protons. During the last decade technical advances have made it possible to detect localized intracellular gradients, and extracellular fluxes of calcium and protons in the apical domain. Other ions, notably potassium and chloride, are also receiving attention. An important development has been the realization that pollen tube growth oscillates in rate; in addition, the ion gradients and fluxes oscillate in magnitude. Although all the ionic oscillations show the same period as that of the growth rate, with the exception of extracellular chloride efflux, they are not in phase with growth. Considerable effort is devoted to the elucidation of these different phase relationships, with the view that a hierarchical order may provide clues about those events that are primary vs. secondary in growth control. Attention is also given to the targets for the ions, for example, the secretory system, the cytoskeleton, the cell wall, in an attempt to provide a global understanding of pollen tube growth.
The pollen tube is a highly specialized cell type that delivers the sperm cells to the ovule for fertilization. Pollen tube growth is thus crucial for the process of plant sexual reproduction and food production; it is also of fundamental interest because it is one of the fastest growing plant cell types known. Understanding the mechanisms by which pollen tubes grow has the potential to be valuable both to plant breeders and cytologists alike. Among the essential underlying processes is the regulation of several different ions, notably calcium and protons, which contribute in a fundamental way to the control of pollen tube elongation, navigation and even in the self-incompatibility response of some species.
Of all the ions, calcium (Ca2+), in particular, has long been known to be necessary for pollen tube growth, and is an essential component of the culture medium (Brewbaker & Kwack, 1963; Picton & Steer, 1983; Steer & Steer, 1989). But it is also widely appreciated that protons (H+) are essential, with acidic conditions (pH 6.5 or less) favoring growth in many species. Over the last few years developments and improvements in ion detection techniques have enabled both intracellular and extracellular profiles of Ca2+ and H+ to be determined in growing pollen tubes. The emerging picture reveals the presence of intracellular Ca2+ and H+ gradients, and extracellular fluxes of these and other ions that are maintained as the tube grows (Fig. 1). These patterns, however, are not static but undergo periodic oscillations, as does the rate of pollen tube elongation. Studies of the interaction between fluctuations in growth rate, ion gradients, and ion currents are leading us closer to an understanding of the processes involved in tip growth in the pollen tube.
In this review, we address the role of ions in pollen tube growth, where we focus on the recent developments that reveal the oscillatory properties of ion expression. In addition, we attempt to place these ion oscillations within the context of growth, where we try to understand how the ions are controlled, and in turn what components or processes are controlled by the ions. An indication of the interest and excitement surrounding the issue of pollen tube growth and its control can be appreciated by the appearance of several recent reviews. Since the classic articles by Heslop-Harrison (1987) and by Steer and Steer (1989), which covered virtually all aspects of pollen tube growth, there are more recent articles that focus on different aspects of polarized growth (Mascarenhas, 1993; Derksen et al., 1995; Li et al., 1997; Taylor & Hepler, 1997; Malhó, 1998a; Franklin-Tong, 1999a, 1999b; Geitmann & Emons, 2000; Malhóet al., 2000; Feijóet al., 2001; Hepler et al., 2001), including in particular the role of Ca2+ in pollen tube elongation (Rudd & Franklin-Tong, 2001). Readers interested in more information on this topic are encouraged to consult these articles.
II. Ion gradients and flux patterns
Ratiometric ion imaging during the last 10–12 yr has revealed that growing pollen tubes possess a dramatic intracellular Ca2+i gradient that is tightly appressed to the apical plasma membrane (Rathore et al., 1991; Miller et al., 1992; Pierson et al., 1994, 1996); this has been referred to as a ‘tip-focused gradient’ and has been shown to be closely coupled to cell elongation. Disruption of the cytoplasmic Ca2+ concentration ([Ca2+]i) gradient invariably results in cessation of pollen tube extension, while inhibition of elongation by various means results in a concomitant dissipation of the gradient (Rathore et al., 1991; Pierson et al., 1994). When examined with fura-2 dextran, the results indicate that the [Ca2+]i gradient extends from around 1–5 µm at the apex of a growing tube, to basal values of 150–300 nm within 20 µm (Rathore et al., 1991; Miller et al., 1992; Pierson et al., 1994, 1996; Holdaway-Clarke et al., 1997) (Fig. 1a). However, the peak concentration is at the limit of the dynamic range of fura-2-dextran, which is one log unit either side of its pKa (0.57 µm). More recent work using the bioluminescent indicator, aequorin, which has a much greater dynamic range (Williamson & Ashley, 1982), has revealed that the peak [Ca2+]i at the tip is likely to be in the 3–10 µm range (Messerli et al., 2000), which is about two to three times that previously reported.
The ratiometric imaging of [Ca2+]i further reveals that the high point of the gradient is tightly appressed to the plasma membrane at the extreme apex of the growing pollen tube (Pierson et al., 1994, 1996; Malhóet al., 1995; Holdaway-Clarke et al., 1997) (Fig. 1a). In addition, the images show that the gradient drops off sharply, and reaches a basal level of 150–200 nm within approximately 20 µm from the apex (Pierson et al., 1996). Presumably elements of ER and mitochondria within the apical cytoplasm rapidly sequester the elevated Ca2+ and reduce its concentration to a basal level. It is worth noting here that nonratiometric techniques are completely inadequate at revealing the true spatial nature of the [Ca2+]i gradient, because the path length reduces sharply as the tip tapers toward the apex, creating a significant artifact, which causes the point of highest [Ca2+]i to appear to be located some distance back from the tip of the tube.
Membrane associated Ca2+ may also be high at the pollen tube apex based on the strong signal from cells stained with chlortetracycline (Malhóet al., 2000); however, the data have not been analyzed through ratiometric analysis, and thus they need further attention before definitive conclusions can be made. Nevertheless, when considered together with the tip-focused [Ca2+]i gradient, it is evident that the apex of the growing pollen tube is the site of localized Ca2+ expression.
Measurements of the extracellular flux of Ca2+ using an ion-selective vibrating electrode reveal a tip-directed Ca2+ influx in growing pollen tubes of up to 20 pmol cm2 s−1 (Kühtreiber & Jaffe, 1990; Holdaway-Clarke et al., 1997; Messerli et al., 1999), which is inhibited by treatments that stop growth and dissipate the [Ca2+]i gradient (Pierson et al., 1994). These studies thus establish that there is a close coupling between the intracellular tip-focused gradient, the extracellular tip-directed influx, and the elongation of the pollen tube. However, simple calculations suggest that the magnitude of the Ca2+ influx may be substantially greater than that needed to account for the observed [Ca2+]i gradient (3 µm) (Holdaway-Clarke et al., 1997); even if the high value of 10 µm is used (Messerli et al., 2000), the flux required (1–2 pmole cm−2 s−1) would appear to be much less than that measured using the Ca2+-selective vibrating probe (Fig. 2).
Source of Ca2+ Results from different avenues of endeavor support the idea that the intracellular Ca2+ gradient derives directly from influx of extracellular Ca2+ across the plasma membrane. Over 25 yr ago Jaffe et al. (1975) revealed that exogenous 45Ca2+ was rapidly taken up by lily pollen tubes, and initially concentrated in the apical region of the cell as indicated by autoradiography. A more recent study shows that manganese (Mn2+), applied to the extracellular medium, quickly quenches the fluorescence from indo-1, the indicator dye, at the tip (Malhóet al., 1995). The simplest interpretation is that Mn2+ passes through Ca2+ channels at the apex of the tube and binds to indo-1; under normal circumstances Ca2+ would pass through these channels (see section next section). The presence of a marked influx of extracellular Ca2+, which is specifically directed towards the apex of the growing tube, is also consistent with the idea that the intracellular gradient is derived from influx of extracellular Ca2+ (Fig. 2). Although it seems likely that the bulk of this flux is a result of binding within the cell wall itself, a fraction presumably crosses the plasma membrane and accounts for the observed intracellular gradient (Holdaway-Clarke et al., 1997) (see External stores: the cell wall as a Ca2+ sink/source). Finally, the spatial profile of the Ca2+ gradient, as revealed by ratiometric imaging, places its highest point immediately adjacent to the plasma membrane at the extreme apex of tube, again providing support for an influx of extracellular Ca2+. Further progress on this question might benefit from the use of an indicator dye (e.g. C18-Fura-2; Etter et al., 1994) that is targeted to the plasma membrane.
Despite these observations and arguments, we cannot rule out the release of Ca2+ from internal stores as a contributor to the tip-focused gradient. There are elements of the ER in the apical domain that could release Ca2+, possibly in response to inositol 1, 4, 5 trisphosphate (IP3) (Franklin-Tong et al., 1996; Malhó, 1998b); however, the ER distribution does not fit closely with the spatial profile of the gradient (Lancelle & Hepler, 1992; Lancelle et al., 1997). Perhaps a more likely candidate would be the secretory vesicles themselves. Some of these are closely appressed to the inner surface of the plasma membrane, and it is possible, as has been suggested in other systems (Mitchell et al., 2001), that they could be stimulated to release Ca2+ and contribute to the gradient, as well as to their own exocytosis. These issues deserve attention in future experimentation.
Ca2+ transport Although it is known that Ca2+ crosses the pollen tube plasma membrane at the tip (Malhóet al., 1995), the identity and characteristics of the channels responsible for this influx are not known (Fig. 1e). It has been suggested that stretch activated channels might be involved (Pierson et al., 1994; Feijóet al., 1995). The deformation of the apical plasma membrane, resulting from cell wall yielding and turgor driven expansion, would cause the opening of Ca2+ channels, especially at the apex where the mechanical deformation would be the greatest. Stretch-activated Ca2+ channels have been found in some plant cell systems (Cosgrove & Hedrich, 1991; Ding & Pickard, 1993), but perhaps the more pertinent ones are those mechanosensory Ca2+ channels discovered at the hyphal tips of the oomycete Saprolegnia ferax (Garrill et al., 1992, 1993), and rhizoid tips of the alga Fucus (Taylor et al., 1996). Through the production of spheroplasts correlating to their position in the apical region of the fungal hypha, Garrill et al. (1992, 1993) were able to show that stretch activated channels were most abundant on those plasma membranes derived from the apical region. These channels were Ca2+ permeable and could be blocked with Gd3+, which is known to inhibit the apical Ca2+ gradient that occurs in growing hyphae (Garrill et al., 1993). In Fucus rhizoids, Taylor et al. (1996) used laser microsurgery to produce plasma membrane-bound protoplasts, which when subjected to patch clamp analysis revealed the presence of stretch activated channels that conducted K+ outward and Ca2+ inward. In addition to being mechanosensitive, it is noteworthy that these channels are also voltage gated, with their open probability increasing upon depolarization of the membrane potential (Taylor et al., 1996). However, data on channel location fail to show an accumulation in the apical domain; rather, Taylor et al. (1996) argue that loosening of the cell wall at the rhizoid apex results in local activation, which accounts for the influx and marked intracellular gradient observed in these cells.
Root hairs are yet another tip growing cell with several parallels to pollen tubes. Patch clamp experiments on ‘spheroplasts’ formed from the membrane and cytoplasm of the apical portion of Arabidopsis root hair protoplasts have revealed a voltage-activated (but apparently not stretch activated) Ca2+ channel that is implicated in the tip-focused [Ca2+]i gradient (Véry & Davies, 2000). Véry and Davies (2000) in addition calculate that the macroscopic current through this channel is in agreement with Ca2+ fluxes measured in vivo. Notably, the Ca2+ flux at the tip of root hairs (3–5 pmol cm2 s−1) is similar to the flux that is thought to be required to produce the tip-focused [Ca2+]i gradient in growing lily pollen tubes (1.7 pmol cm2 s−1; Fig. 2) (Holdaway-Clarke et al., 1997).
The lanthanides, Gd3+ and La3+, which block Ca2+ channel permeability in fungal hyphae (Garrill et al., 1993) and Fucus rhizoids (Taylor et al., 1996), also exert an inhibitory effect on pollen tube growth (Malhóet al., 1994; Geitmann & Cresti, 1998; Franklin-Tong et al., 2002), and more specifically on intracellular Ca2+ gradients (Malhóet al., 1994) and extracellular Ca2+ fluxes (Franklin-Tong et al., 2002). However, these observations are only consistent with the presence of stretch activated channels; by themselves they do not prove their existence. Direct detection using patch clamp analysis is needed. Nevertheless, it should be realized that this method is extremely difficult to apply to the plasma membrane of growing, turgid cells, in particular those of the rapidly growing pollen tube. Efforts along these lines by Armstrong et al. (1999) on pollen tubes have thus far failed to detect a Ca2+ channel although they have identified a K+ channel (see K+ transport for more detail). It will be important therefore in future work to experimentally identify and characterize the Ca2+ channels, which must be present.
Although considerable attention has been given to Ca2+, it must be recognized that H+ may also serve important regulatory roles in the control of growth and development in plant cells, including pollen tubes (Felle, 2001). It is widely recognized that H+ pumps, through the generation of local pH gradients, ultimately power other transport processes in plant cells (Sze et al., 1999; Felle, 2001). However, owing to the greater mobility of H+, when compared to Ca2+, the detection of local pH gradients has been difficult, because the typical indicator dyes, especially at elevated concentrations will dissipate these presumptive gradients. Indeed, initial studies on growing pollen tubes failed to detect pH gradients (Fricker et al., 1997; Parton et al., 1997; Messerli & Robinson, 1998). However, when the high mobility of H+ is taken into account, and low concentrations of the indicator dye (e.g. BCECF-dextran) are used, pH gradients are observed. In the growing pollen tube these appear as an acidic domain at the extreme apex, and perhaps even more interestingly, a constitutive alkaline band that is positioned at the base of the clear zone (Feijóet al., 1999, Cárdenas, unpublished observations) (Fig. 1b).
Despite the detection of local pH gradients it is puzzle that pollen tube growth can continue in their absence. An explanation may derive from the limited resolution available with these various ion imaging light microscope systems (Feijóet al., 1999). Even under optimal resolution conditions a local pH gradient immediately adjacent to the plasma membrane might express a spatial width that is less than the 0.2 µm resolution limit of the microscope, and therefore be difficult to detect. In addition, the presence of buffers, such as the indicator dye, will make the gradient steeper (Stern, 1992), and even more spatially restricted to the point of H+ influx. When these issues are taken together with the findings that pH buffers indeed slow and eventually inhibit growth (Feijóet al., 1999), it seems likely that pH gradients are necessary for pollen tube elongation. It is of further interest that the acidic tip seems to be present only in growing tubes, while the alkaline band persists even in tubes that are not growing (Feijóet al., 1999).
As the H+ profile inside the cell differs from that of Ca2+, so does the pattern of extracellular H+ flux. The extracellular H+ fluxes measured around growing pollen tubes almost perfectly match the cytoplasmic H+ distribution. In lily there is a large H+ influx at the tip, which corresponds to the acidic domain at the apex, and a marked H+ efflux 20–40 µm back from the tip, which correlates with the base of the clear zone and the location of the alkaline band (Feijóet al., 1999) (Fig. 1b). H+, in contrast to Ca2+, thus exhibit a current loop within the apical domain of the pollen tube. It seems plausible that H+ pumps located on the plasma membrane towards the base of the clear zone are responsible for driving the efflux, while nonspecific cation channels at the apex permit H+ influx. Given the apical location of this current loop and the apparent activity of a H+ pump, which generates the alkaline band even when pollen tubes have stopped growing, there is speculation that these H+ currents may play a primary role in controlling pollen tube polarity (Feijóet al., 1999, 2001).
H+ transport As noted above H+ pumps are the primary energetic mechanism for membrane transport in most plant cells (Sze et al., 1999; Felle, 2001), and pollen tubes are no exception. Both a H+-ATPase and its cytochemical activity have been localized to the pollen grain and pollen tube (Feijóet al., 1995), although curiously the activity along the tube is relatively low (Obermeyer et al., 1992). Nevertheless, given the presence of the prominent alkaline band together with the corresponding H+ efflux pattern, it would seem evident that a H+-ATPase is present, and indeed localizes on the plasma membrane towards the base of the clear zone (Feijóet al., 1999). The activity of the H+-ATPase appears to be important in determining the growth rate of pollen tubes, as exposure to fusicoccin, which stimulates the H+- pump, increases the germination and growth rates of pollen tubes, while the ATPase inhibitors, azide, vanadate and N-ethylmaleimide, inhibit pollen tube growth (Rodriguez-Rosales et al., 1989; Feijóet al., 1992; Fricker et al., 1997; Pertl et al., 2001). Recent work by Pertl et al. (2001) further shows that fusicoccin increases the membrane association of the H+-ATPase, in a reaction that appears to involve a 14-3-3 protein. Of particular interest, the localized activity of the H+-ATPase may be controlled by the [Ca2+]i gradient (Feijóet al., 1999), because micromolar levels of Ca2+ reduce the activity of plasma membrane H+ pumps by Ca2+-dependent phosphorylation (Kinoshita et al., 1995; De Nisi et al., 1999). Within the pollen tube the high [Ca2+]i at the apex would inhibit H+ pumping, while further back from the tip where [Ca2+]i is at basal levels, the pump could function maximally, producing both the cytoplasmic alkaline band and the observed extracellular H+ efflux.
H+, like Ca2+, are much more concentrated outside than inside the cell, and hence there is a strong driving force, including both concentration and charge, that will tend to cause these ions to enter the cell. Although specific inward conduction paths have not yet been identified, it is possible for H+ to enter the cell through common cation selective ion channels (Hille, 2001). Thus, in Fig. 1(e), we suggest that both H+ and Ca2+ enter at the tip by the same nonspecific cation channel. In relationship to their oscillatory influx, having both ions pass through the same channel would not pose any special difficulty as their phase offsets with respect to growth are not significantly different (Feijó, 1999; Messerli et al., 1999).
Potassium and chloride
While Ca2+ and H+ are important regulatory ions, we must not ignore others, notably potassium (K+) and chloride (Cl−), which are present in relatively large amounts and whose movements control such basic cellular parameters as turgor and the membrane potential. Unfortunately there are no published accounts in which cytoplasmic Cl− or K+ have been imaged, however, extracellular fluxes of these species have been measured (Messerli et al., 1999; Zonia et al., 2001, 2002). Chloride fluxes are perhaps the more interesting, with efflux of the anion occurring at the tip and influx starting some 12 µm back from the tip in Nicotiana tabaccum. Cl−, like H+, thus form a current loop in the apical domain of the pollen tube (Zonia et al., 2001, 2002), as illustrated in Fig. 1(c). In addition, several different channel blockers, including inositol 3,4,5,6 tetrakisphosphate (IP4), a possible endogenous Cl− channel inhibitor, all inhibit pollen tube growth and lead to cell swelling (Zonia et al., 2002). The inference from these data is that Cl− fluxes regulate turgor and water uptake essential for normal pollen tube growth.
Measurable K+ influx occurs at the tip of most, but not all, longer growing pollen tubes (Messerli et al., 1999) (see Fig. 1d). Because of the relatively poor performance of the K+-selective electrodes, fluxes of this ion are not always detected. Nevertheless, K+ fluxes play a prominent role in establishing the membrane potential, and thus there may be substantial movement of these ions during pollen tube growth.
K+ Transport The only direct detection of channel activity in pollen grains and tubes involves those that participate in the transport of K+ (Fig. 1e) Considerable attention has been initially given either to quiescent grains (Obermeyer & Blatt, 1995) or to protoplasts derived from grains (Obermeyer & Kolb, 1993; Fan et al., 1999, 2001). In lily pollen protoplasts three different K+ channels have been observed; the most common has a conductance of 19 pS and shows an increase in activity with increased [Ca2+] (Obermeyer & Kolb, 1993). Using pollen protoplasts of Brassica, which have been examined using a whole-cell clamp analysis, Fan et al. (1999, 2001) have identified an inward K+ channel that is insensitive to intracellular Ca2+. By contrast this channel is sensitive to extracellular Ca2+. When the extracellular [Ca2+] is increased from 10 mm to 50 mm, the inward K+ currents are inhibited by 46%; whereas when the extracellular [Ca2+] is decreased from 10 mm to 1 mm, there is no effect (Fan et al., 2001). This inward K+ current is also sensitive to pH, being stimulated at 4.5 but markedly inhibited at 8.5 (Fan et al., 2001). The most recent advance has been the identification of an inward K+ channel in Arabidopsis, which is a member of the Shaker family (Mouline et al., 2002). Referred to as SPIK (Shaker Pollen Inward K+), this channel is expressed specifically in pollen, and is localized to the apical domain of the tube. When expressed heterologously in COS cells, and subsequently subjected to patch clamp analysis SPIK displayed a conductance of 14 pS in 50 mm K+ (Mouline et al., 2002), suggesting that it may be similar to the 19 pS inward K+ channel identified earlier by Obermeyer and Kolb (Obermeyer & Kolb, 1993). It is particularly noteworthy that mutants in SPIK had reduced K+ uptake and correspondingly impaired growth, thus supporting the conclusion that this channel plays a major role in K+ uptake during pollen tube growth (Mouline et al., 2002).
Cl− transport Inhibitor studies indicate that the channel through which Cl− leaves the pollen tube tip is similar to the ‘ubiquitous’ Ca2+-activated Cl− channel found in humans, which is negatively regulated by IP4 (Zonia et al., 2002) (Fig. 1e). The elevated [Ca2+]i at the tip coupled with a membrane potential more negative than the equilibrium potential for Cl− (approx. −60 mV), would thus account for the efflux of Cl− at the tube apex. A Cl−/HCO3− exchanger is an attractive candidate for the means by which Cl− crosses the plasma membrane at the shank of the pollen tube, and may also be involved in the regulation of cytoplasmic pH in plants (Trofimova & Molotkovskii, 1993).
One of the exciting findings during the last decade has been the discovery that the growing pollen tube is an endogenous oscillator (Weisenseel et al., 1975; Li et al., 1992, 1994; Pierson et al., 1995, 1996; Geitmann et al., 1996). Microanalysis of the growth rate of in vitro cultured lily pollen tubes reveals that it changes from 0.1 to 0.4 µm s−1 over an interval of 15–50 s (Pierson et al., 1996). Additionally, it is becoming apparent that many of the underlying physiological processes, including intracellular ion gradients and extracellular fluxes, also display a similar oscillatory pattern, in which the frequency or period is the same as that of growth, but which exhibit different phase relationships. The quest has been to define these various oscillatory patterns and to decipher their phase relationship to growth, with the view that a temporal hierarchy of the putative factors and/or processes controlling growth will provide clues about which are primary and which are secondary. The phase relationships of these oscillations are summarized in Fig. 3.
The [Ca2+]i gradient oscillates
Initial investigations of growing lily pollen tubes indicated that the [Ca2+]i at the apex oscillated markedly in magnitude (750–3000 nm) (Pierson et al., 1996); moreover, these changes were reported to be approximately in phase with growth rate oscillations (Holdaway-Clarke et al., 1997; Messerli & Robinson, 1997). Subsequent analysis, using a system with higher temporal resolution and the luminescent Ca2+ indicator aequorin, revealed that the peaks in the oscillation of the growth rate precede the peaks in the oscillation of [Ca2+]i at the tip by around 4 s, representing a phase difference of approximately 38° with respect to the average period of oscillation (Messerli et al., 2000) (Fig. 3). Despite the magnitude of the apical [Ca2+]i change under normal growth conditions, it does not propagate basipetally; rather the elevated region of Ca2+ remains restricted to the tube tip, with the highest point appressed to the plasma membrane (Holdaway-Clarke et al., 1997). A further interesting observation is that [Ca2+]i oscillations, in a limited number of instances, have been observed in the absence of growth thus raising important and unsolved issues regarding how these ion changes are generated (Messerli & Robinson, 2003).
It has long been recognized that the high amplitude of the Ca2+ signal within a localized gradient carries information that can be used to spatially confine key reactions. Thus Ca2+ binding proteins, such as calmodulin, or calmodulin-like domain kinase (CDPK), will be saturated and possibly maximally active where the gradient is high (tube apex) and inactive just a few microns away. With an oscillating gradient, the sequential repetition or frequency modulation of the amplitude signal can be used as further information to control developmental or physiological processes (Berridge et al., 2000; Evans et al., 2001; Feijóet al., 2001). In plants a particularly good example comes from studies on the role of [Ca2+]i in the control of stomatal aperture (Allen et al., 2001; Evans et al., 2001). Recent results, for example, reveal that repetitive transients in the [Ca2+]i are much more effective in reducing stomatal aperture than a single plateau of Ca2+ that achieves a similar magnitude of ion change (Allen et al., 2001). From an evolutionary point of view, a system that works through an oscillating [Ca2+] may hold a large advantage over one that simply generates a fixed high level of the ion. The presence of a fixed high level of Ca2+ could not be spatially confined because it would spread through diffusion. Moreover, it would exert a profound detrimental effect on phosphate-based energy production and ultimately on cell viability because of the formation of highly insoluble Ca2+ phosphate complexes. With an oscillating level of Ca2+, as for example in the pollen tube, the high points are tightly controlled in space and limited in time; as a consequence they would have a lessened negative effect on energy metabolism, while still being sufficient to activate the processes essential for maintaining the rapid rate of elongation. Although oscillations may be a natural consequence of a nonequilibrium system responding to various intra and/or extracellular perturbations, it is important to realize that they become a regulatory feature, whose spatial and temporal properties entrain and coordinate a host of underlying processes (Feijóet al., 2001).
Cytoplasmic pH oscillates
Oscillations in pHi have been observed in the apical region of the pollen tube (Feijóet al., 1999); specifically the alkaline band, which spatially correlates with the base of the clear zone, oscillates in intensity, with the band achieving maximal alkalinity (lowest [H+]) when the growth rate is slowest. The time resolution was not sufficient to enable precise correlation, but the general trend indicated that the intensity of the alkaline band is approximately 180° out of phase with the growth rate (Feijóet al., 1999). Messerli and Robinson (1998), while failing to observe a pH gradient, nevertheless detected an acidic wave that moved basipetally in an oscillatory manner. Correlation analysis revealed that the peak of the pHi pulses occur around 7.5 s, or 67.5° after the peak of growth pulses. Messerli and Robinson (1998) suggest that the oscillation of cytoplasmic acidity at the tip may control the growth oscillations by terminating the preceding growth rise.
The extracellular fluxes of Ca2+, H+, K+ and Cl− oscillate
In addition to the oscillations in the intracellular concentration of Ca2+ and H+ reported above, recent studies have shown that the extracellular fluxes of these ions as well as K+ and Cl− also oscillate in relationship to the growth rate (Fig. 3). The tip-directed influx of extracellular Ca2+ oscillates with a period that matches the growth rate; however, the peaks of Ca2+ influx follow peaks in growth rate by approx. 12 s, representing a phase difference of 136° (Holdaway-Clarke et al., 1997; Messerli et al., 1999). Likewise, influxes of extracellular H+ and K+ at the pollen tube apex oscillate with similar periods to growth, and with phase delays of 103° and 100°, respectively (Messerli et al., 1999). These phase delays are not significantly different from each other or from that of extracellular Ca2+, however, they are significantly different from the growth rate (Fig. 3).
Extracellular Cl−, in contrast to Ca2+, K+ or H+, exhibits an oscillatory efflux at the extreme apex of the growing pollen tube of Lilium longiflorum and Nicotiana tabaccum that is exactly in phase with oscillations in the growth rate (Zonia et al., 2002). Cl− thus stands apart from all other ions in its close phase relationship with growth (Fig. 3). It is additionally pertinent that the magnitude of the Cl− flux (up to 8000 pmole cm−2 s−1) is substantially larger than that of any other ion measured thus far. Zonia et al. (2002) have been able to detect oscillations in the bulk movement of vesicles into the apical dome, and find that these also occur in phase with Cl− efflux and thus with growth rate. Cl− movement thus emerges as a possible central regulator in pollen tube growth; further work on how Cl− itself is regulated and how it affects growth is clearly warranted.
Although they are not in phase with one another there are some interesting parallels between the extracellular fluxes of Cl− and H+. Whereas the efflux of Cl− and the influx of H+ at the tip both oscillate, and both ions form a current loop in the apical domain, neither the influx of Cl− nor the efflux of H+ along the shank of the tube oscillate. This is particularly curious for H+, because the intracellular alkaline band does oscillate (Feijóet al., 1999).
Within the context of all the ions noted thus far it is important to emphasize that the maximum flux or gradient high point, with the exception of extracellular Cl− efflux, occurs after the maximum growth rate. These conclusions have been established by different laboratories using cross-correlational analysis, in which irregular features occurring in both the growth and ion peaks establish that the ion changes follow growth and not the opposite (Holdaway-Clarke et al., 1997; Messerli et al., 1999) (Fig. 3).
Membrane potential oscillation
Given the oscillatory changes in net ions (Weisenseel et al., 1975), the conditions are established that might support an oscillation in the membrane potential. Oscillations in the membrane potential at the pollen tube tip could, in turn, drive oscillations in the [Ca2+]i gradient. Using conventional intracellular recording methods, Weisenseel and Wenisch (1980) observed membrane potentials of −90 to −130 mV in pollen grains of lily with short tubes, and −60 mV in grains with long pollen tubes. However, oscillations have not been reported. Despite these negative results, we emphasize that these studies are inconclusive. Given the localization of ion fluxes, especially H+ and Cl−, in the apical domain, it may be necessary to similarly restrict the measurement of the membrane potential to this region. In addition, it will be essential, at least with lily, to inspect pollen tubes that are exhibiting growth oscillations. However, this will be technically demanding because of the difficulty of impaling the apex while still maintaining tube elongation. Until more definitive studies have been conducted, perhaps with optical dyes, we are reluctant to rule out the presence of an oscillating membrane potential in the apical domain.
It is well established that elevated levels of Ca2+ stimulate exocytosis, and therefore it seems likely that the oscillating tip-focused gradient would lead to a similar oscillation in membrane fusion and secretion. Recently, Parton et al. (2001) report that the number of vesicles in the clear zone, which are labeled with a membrane dye, FM4-64, oscillate over time in a manner related to the growth rate oscillation. Although attempts were made to record the potential oscillation of fluorescence in the extreme apex, presumably where exocytosis occurs, periodic changes in this region were not detected. However, 3–5 µm back from the apex, fluctuations in the fluorescence were observed and these oscillated in intensity with the same period as growth, but with an apparent advance in phase by 5–10 s. As Parton et al. (2001) speculate, these changes in fluorescence, which become part of the retrograde flow observed within the vesicular cloud of the inverted cone, likely have more to do with the dynamics and replenishment of this pool than with the actual events of exocytosis. Still, the oscillatory data indicate that these vesicle dynamics are coupled to growth, and possibly to ion changes, although the relationships are not yet clear.
Cell wall thickness oscillations
Changes in wall thickness have been observed during pulsatile growth of Gasteria pollen tubes (Plyushch et al., 1995). A recent reinvestigation of this question indicates in lily pollen tubes that the wall thickness oscillates with the same frequency as growth, however, its phase is offset such that the wall is thinnest about 6 s. after the growth rate peaks, and thickest 6 s after the slowest growth point (McKenna et al., unpublished) (Fig. 3). While it is tempting to equate these changing thickness data directly to secretion, we currently have too little information to make a clear connection. There are several different, but overlapping, processes that need to be considered. For example, in addition to changes in secretion proper, there might also be an oscillation in the yielding properties of the cell wall at the tip. It is important here to develop a clear and independent measure for secretion in order to resolve the meaning of these thickness changes.
Given the oscillation in growth rate, and the realization that the underlying actin cytoskeleton must keep pace or even lead growth, it is reasonable to imagine that it also oscillates. Although there is some question about the structure and organization of the actin cytoskeleton at the extreme apex of the tube, in the shank of the tube, back from the apex, pollen tubes possess a rich array of actin microfilaments, and in addition it is these structures, together with myosin, that generate cytoplasmic streaming and transport the vesicles to the apical domain (Shimmen et al., 2000; Hepler et al., 2001). Using tobacco pollen tubes expressing a GFP-talin construct, which labels actin microfilaments, Fu et al. (2001) detected changes in apical fluorescence that were correlated with the oscillatory growth rate. Specifically they noted that the fluorescence from GFP-talin increased 15–30 s before the peak in growth rate, leading them to conclude that actin might be a regulator of oscillatory growth (Fu et al., 2001). These data are extremely interesting, however, they lack good time resolution, which would allow one to establish a more exact phase relationship. Also, while the increase in bulk fluorescence presumably means that more actin is moving into or becoming polymerized in the apical zone, it is not established how this translates into a subsequent increase in growth rate. This topic deserves additional attention.
IV. The need for a Ca2+ store
One of the exciting, but puzzling, observations about the oscillation of Ca2+, is that the intracellular gradient and the extracellular flux are not in phase with one another (Holdaway-Clarke et al., 1997; Messerli et al., 1999). The fact that [Ca2+]i peaks approximately 8 s before the maximum extracellular Ca2+ influx, indicates that the latter does not immediately give rise to the former. These data thus necessitate some kind of store, so that Ca2+ taken up in one cycle is ‘invisibly’ sequestered, but then ‘appears’ in the cytoplasm in the next cycle of oscillation. Two potential candidates for this role stand out: firstly, internal stores, that is, membrane delimited compartments within the cytoplasm (e.g. ER or vesicles), and secondly, external stores, that is, the cell wall.
Although we think it is most likely that the [Ca2+]i gradient is derived from influx of extracellular Ca2+, considerable interest surrounds the idea that internal compartments could be involved. To a certain extent these views gain prominence because they build upon a well-established body of information derived from studies of Ca2+ regulation in animal cells (Berridge et al., 2000; Putney et al., 2001). Thus, internal stores have been suggested as being involved in the production of the tip-focused [Ca2+]i gradient, with their continual filling and emptying resulting in the observed oscillations (Trewavas & Malhó, 1997). As the vacuole is located behind the clear zone, it is unlikely that this compartment contributes to the [Ca2+]i gradient. However, ER is present in the clear zone, and this has been hypothesized to act as a ‘capacitative’ Ca2+ store (Trewavas & Malhó, 1997). In animal cells Ca2+ entry across the plasma membrane may be regulated by the degree to which an intracellular Ca2+ store is depleted (Parekh & Penner, 1997); also a detectable [Ca2+]i gradient is not always seen in the cytoplasm as stores are filled (Mogami et al., 1997).
Clearly there are releasable internal Ca2+ stores in pollen tubes, evidenced by several observations of experimentally induced Ca2+ elevations and waves in the cytoplasm. Malhó and Trewavas (1996) observed a tip-directed Ca2+‘wave’ resulting from uncaging of Ca2+ in the region of the nucleus of Agapanthus pollen tubes, and single Ca2+ waves traveling basipetally from the apex of Lilium pollen tubes have also been observed upon cessation of growth caused by exposure to caffeine or injection of BAPTA buffer (Pierson et al., 1994, 1996). Franklin-Tong and coworkers (1996) observed an IP3-induced Ca2+ wave in tubes of Papaver rhoeas pollen and noted a rapid elevation in [Ca2+]i in the shank of the pollen tube in response to challenge by self-incompatability proteins, which was also propagated acropetally. Finally, we note that [Ca2+]i oscillations can occur in the complete absence of growth, which is interpreted as strong support for release from internal stores (Messerli & Robinson, 2003).
A prime candidate process for the release of Ca2+ from internal stores would be the activation of the IP3 receptor, and several lines of evidence show that this pathway is present in pollen tubes. Photolysis of caged IP3 causes an elevation of internal Ca2+ (Franklin-Tong et al., 1996; Malhó, 1998b), while the inhibition of IP3 production with mastoparan or the IP3 receptor with heparin, blocks pollen tube growth. Despite these positive indicators there are different aspects of the IP3-induced Ca2+ transients that raise questions about its role during normal pollen tube growth and development. Thus, the Ca2+ release is often more pronounced in the shank of the tube than in the apical domain, and in poppy pollen tubes it does not occur immediately but only after a delay of a few min (Franklin-Tong et al., 1996). In addition, IP3-mediated release often propagates a wave that is not seen in unperturbed tubes. Nevertheless, it is possible that the IP3 system participates in the regulation of internal stores; for example in the conformational coupling model, developed in animal systems, IP3 receptors located on the ER respond to low [Ca2+]i and transmit this information to the plasma membrane resulting in the influx of extracellular Ca2+ (Putney et al., 2001). A similar system operating in pollen tubes might account for the refilling of internal stores, and thus explain the difference in oscillatory phase between the gradient and the extracellular influx. These ideas need experimental verification.
The arguments above focus on the ER as the most obvious candidate for the internal Ca2+ store. However, its physical location is not consistent with the spatial profile of the tip-focused gradient (Lancelle & Hepler, 1992). While the gradient is appressed to the plasma membrane (Fig. 1), the ER is present throughout the pollen tube and is not particularly concentrated in the clear zone, or close to the plasma membrane at the tube apex (Lancelle & Hepler, 1992). Indeed, recent studies on living lily pollen tubes using a targeted GFP, indicate that the ER is reduced in amount in the region of the inverted cone when compared to other regions of the pollen tube apex (Parton, pers. comm.).
The population of vesicles, which occupies the clear zone, and occurs in an ‘inverted cone’ with the base of the cone at the tip of the pollen tube, is another possibility for a Ca2+ store that could be involved in supporting the tip-focused [Ca2+]i gradient. In neuroendocrine cells recent studies show that the dense core vesicles contain a pool of rapidly exchangeable Ca2+, which appears to be released by cyclic ADPribose, rather than IP3 (Mitchell et al., 2001). Because the ryanodine receptor is present on these vesicles, it is plausible that through Ca2+-induced Ca2+ release they could stimulate their own fusion and exocytosis. By comparison, it is possible that the vesicles in the inverted cone of the growing pollen tube also contain a store of Ca2+, which if released at the inner surface of the plasma membrane before fusion, could contribute to the Ca2+ profile observed.
External stores; the cell wall as an ion sink/source
Although often ignored, plant cell walls are a well-known Ca2+ sink (Fig. 4). The cell wall of the pollen tube is highly specialized at the apex, composed mostly of pectin, callose and arabinogalactan proteins (AGPs) with very few cellulose microfibrils (Heslop-Harrison, 1987; Derksen et al., 1995; Jauh & Lord, 1996). Immunocytochemical studies have shown that pectins located at the tip of the growing pollen tube are highly esterified, while de-esterified pectins are more concentrated in areas away from the tip (Jauh & Lord, 1996; Li et al., 1994). Pectins are secreted primarily as methylesters (Li et al., 1997, 2002; Lennon & Lord, 2000) and subsequently de-esterified by pectin methylesterase (PME), which has also been secreted into the apoplast (Li et al., 2002). De-esterification of pectin by PME permits Ca2+-induced gelation of pectate, in which Ca2+ cross-links pectin chains, increasing the rigidity of the wall. In vitro, the gel strength of isolated pectins is sharply reduced when esterification is increased from 70% to 80%, and in vivo maximal elongation of tobacco cells occurs when pectins are 80% esterified (McCann & Roberts, 1994). PME also alters the porosity and elasticity of pectin gels (Willats et al., 2001). In a number of species, increased esterification of pectins is correlated with cell elongation, and de-esterification is correlated with cessation of growth (Kim & Carpita, 1992; McCann et al., 1994). Indeed, preliminary studies indicate that in lily pollen tubes de-esterification of pectins by exogenously applied PME inhibits growth (Bosch et al., unpublished).
With Ca2+ continually binding to newly exposed sites on pectins, it is evident that the cell wall constitutes a significant Ca2+‘sink’. In other systems, for example, cell walls from mung bean hypocotyls, the Ca2+ content of pectins increases from 80 µmol g−1 d. wt of wall in young tissue to 122.5 µmol g−1 in mature tissue. The increase in Ca2+ content correlates with an increased degree of pectin de-esterification in mature walls (Goldberg et al., 1986).
Given an extensive binding of Ca2+ it seems reasonable to ask how this will affect ion influx. The calculations shown in Fig. 4 provide a rough estimate of the Ca2+ required by the ‘maturing’ pollen tube cell wall. A lily pollen tube with a diameter of 16 µm and wall thickness of 0.2 µm, growing at 10 µm min−1, produces cell wall at a rate of approximately 1.6 µm3 s−1. This is equivalent to 640 fg s−1 if it is assumed that the wall material has a specific gravity of 1 and that the wall is 60% water (Grignon & Sentenac, 1991). Assuming that Ca2+ movement were to occur through a circular patch of membrane at the apex with a diameter of 10 µm, then the influx would be approximately 35 pmol cm−2 s−1, which is in good agreement with the actual fluxes observed with the Ca2+-selective vibrating probe, for example, approx. 1–20 pmol cm−2 s−1, especially when it is realized that the vibrating electrode is only 50% efficient (Holdaway-Clarke et al., 1997; Messerli et al., 1999). Regardless of the specific details, these calculations indicate that the cell wall is a major extracellular Ca2+ store. In addition, we further emphasize that influxes, which are recorded in association with pollen tube growth, must first pass through the wall where there are numerous Ca2+ binding sites. A role for the cell wall in regulating both intracellular and extracellular Ca2+ seems inescapable.
V. Intracellular targets for ion activity
Ca2+ binding proteins
The foregoing establishes that there are gradients of ions, especially Ca2+ and H+, at the apex of the pollen tube. In order for these to affect growth they must bind to and either activate or inactive a target molecule, which in turn modulates downstream processes. For Ca2+ the obvious candidates that come to mind are calmodulin (Snedden & Fromm, 2001) and calmodulin-like domain protein kinases (CDPKs) (Harmon et al., 2000; Cheng et al., 2002). Calmodulin is present in the pollen tube, although recent work indicates that it is uniformly distributed throughout the cytoplasm and not concentrated at the tip (Moutinho et al., 1998b). While we still need to identify target molecules, at least one, villin, an actin binding protein, has been found that requires both Ca2+ and calmodulin to stimulate the unbundling of actin microfilaments (MFs) (Yokota et al., 2000). Although specific data are lacking as they pertain to pollen tube growth, it is reasonable to suggest that calmodulin, based on its activity in other systems, participates in the regulation of pumps, for example, on the plasma membrane, vacuole and/or endoplasmic reticulum, that are involved in the restoration of basal Ca2+ (Snedden & Fromm, 2001).
A pollen specific CDPK has been identified in maize, which has been suggested to be involved in the regulation of the actin cytoskeleton (Estruch et al., 1994). Mention is also made of a CDPK, derived from Nicotiana alata pollen tubes, that phosphorylates stylar RNAses, which are involved in the self incompatibility reaction (Kunz et al., 1996). More recent work on CDPKs has focused on the localization and activity of kinases in response to pollen tube reorientation (Moutinho et al., 1998a). The authors have used a novel ratiometric indicator, which couples the protein kinase inhibitor, bisindolylmaleimide, to a fluorophore, Bodipy. At low concentrations the complex binds to CDPKs, but does not apparently inhibit them, permitting their localization to be imaged. The results indicate that the kinase is distinctly in the apical domain, in a pattern that is similar to the intracellular Ca2+ gradient, and moreover it shifts its location in response to reorienting stimuli (Moutinho et al., 1998a). Its immediate targets however, are not known, although it might possibly be involved in exocytosis.
A paradigm that applies to both plant and animal systems is that elevated [Ca2+]i stimulates exocytosis (Zorec & Tester, 1992; Thiel & Battey, 1998; Battey et al., 1999). Studies of maize coleoptile protoplasts, for example, reveal an abrupt rise in electrical capacitance, and hence membrane area, when the [Ca2+] increases to 1 µm or higher (Thiel et al., 2000). Evidence is also emerging from studies of pollen tubes in support of a role for Ca2+ in secretion. Recent studies using FM1-43 as a membrane marker confirm that elevated [Ca2+]i, brought about by local photolysis of caged Ca2+, stimulates exocytosis, as evidenced by a decrease in FM1-43 fluorescence (Camacho & Malhó, 2003). These conditions, however, did not alter growth rates, indicating that secretion and growth might not be closely coupled. A further curious result came from studies using the nonhydrolyzable GTP analogue, GTPγS, which is known to activate G-proteins. Uncaging this agent led to an increase in exoctyosis, and an increase in growth, but unexpectedly, a small decrease in [Ca2+]i (Camacho & Malhó, 2003). While these last mentioned results may seem to negate the positive relationship between Ca2+ and secretion, it is possible that the tip focused gradient is well above that required to support maximal secretion, and that small decreases, as reported (Camacho & Malhó, 2003), will have little affect. Recent studies by Sutter et al. (2000) indicate that Ca2+-dependent exocytosis in maize coleoptile cells saturates above 1.5 µm with half-maximal stimulation of exocytosis occurring at ∼0.9 µm. If this is so in the pollen tube, then it is possible the [Ca2+]i in the tip focused gradient is above the threshold to support maximal secretion. These data also raise the possibility that the oscillations in [Ca2+]i at the tip do not necessarily result in oscillating rates of exocytosis at the apex.
Further support for an association between Ca2+ and secretion comes from studies of lily pollen tubes treated with the Yariv reagent, which inhibits the function of arabinogalactan proteins (AGPs) (Majewska-Sawka & Nothnagel, 2000). When lily pollen tubes are treated with Yariv reagent the tubes quickly stop growing. In marked contrast to other agents that block pollen tube growth with the concomitant dissipation of the tip-focused Ca2+ gradient, in cells treated with Yariv reagent the [Ca2+]i elevates, and spreads throughout the apical domain and even basipetally for 50–80 µm (Roy et al., 1999). Of interest secretion continues, although there are at least two major differences from controls as follows: firstly, because the high level of Ca2+ is spatially delocalized, so too is secretion; and secondly, the vesicular contents, once secreted fail to become properly incorporated into the cell wall. Thus secretion becomes uncoupled from polarized cell elongation. The observations are consistent with the possibility of a positive feedback between secretion and Ca2+, where the deformation of the plasma membrane, resulting from a failure of the vesicle contents to be incorporated into the wall, causes mechanical strain that opens channels. In this instance, the deformation will result in an inward, rather than outward, curvature of the plasma membrane, which nevertheless, can still induce a mechanical strain that opens channels. While these studies leave open the mechanism on how both the intracellular Ca2+ gradient and exocytosis remain spatially localized under normal growth conditions, they nevertheless underscore the close relationship between Ca2+ and secretion in pollen tube growth (Roy et al., 1999).
Although the proteins that might respond to Ca2+ and stimulate secretion in pollen tubes are unknown, some potential candidates include the annexins, which have been located to the apical domain of pollen tubes and which have been shown to participate in secretion (Blackbourn et al., 1992; Carroll et al., 1998). Another candidate is the CLB1 gene product, which contains a Ca2+ binding domain similar to synaptotagmin (Kiyosue & Ryan, 1997), a well-recognized membrane trafficking protein in animal cells (Goda & Sudhof, 1997; Mahal et al., 2002). This area deserves close attention; given the amplified nature of exocytosis during growth, the pollen tube could be an excellent object.
The prominent oscillatory efflux of Cl−, together with the known capacity of this ion to regulate salt extrusion and turgor in other plant cells, notably guard cells (Cosgrove & Hedrich, 1991), lead Zonia et al. (2002) to suggest that local turgor regulation is occurring. The question at this point is not so much about whether Cl− regulates turgor, but whether turgor regulates growth. In certain fungal hyphae, notably among the Oomycetes, evidence has emerged that tip growth can occur in the absence of measurable turgor (Harold et al. 1995; Harold et al. 1996; Money, 1997; Heath & Steinberg, 1999; Money, 2001). However, to a first approximation these ideas do not seem to hold for cells of higher plants, including pollen tubes; in brief turgor appears to be necessary for elongation (Benkert et al., 1997; Geitmann, 1999).
Messerli et al. (2000) argued that changes in turgor pressure might be a prime regulator of subsequent growth changes. However, these conclusions stand in marked contrast to an extensive literature on higher plants showing that growth is associated with an increase in the yielding of the cell wall rather than an increase in the turgor pressure (Cosgrove, 1986, 1993). In this scheme, pollen tube elongation occurs when the cell wall at the apex is sufficiently pliable so that it yields to the internal turgor pressure. In addition to the prevailing dogma, there is direct evidence that turgor changes do not regulate pollen tube growth. Benkert et al. (1997), using a pressure probe, could not detect any significant correlation between turgor pressure and growth rate. These authors also failed to detect any oscillation in turgor even though they examined long (> 1000 µm) pollen tubes of lily, which were presumably oscillating in growth rate. It is clear that turgor pressure plays a crucial role in pollen tube growth, which is well demonstrated in recent work showing that a sharp decrease in the osmolarity of the medium can bring about a brief increase in growth and [Ca2+]i (Messerli & Robinson, 2003). However, these studies fail to establish a connection between turgor and growth oscillation.
Despite the dogma and despite these arguments against oscillatory changes in turgor, it is nevertheless necessary to keep an open mind about the possibility of local changes that are restricted to the clear zone, which could alter the biophysical properties of the apical domain. Given the magnitude of the oscillatory Cl− efflux (approx. 8000 pmole cm−2 s−1) (Zonia et al., 2002), and its specific localization to the apical domain, it becomes plausible to suggest that there is an accompanying oscillation in the flow of water, which could affect cytoplasmic activity and structure, such as an actin gel (see Actin and tube growth below), which in turn might contribute to the forward extension of the tube.
Another widely appreciated generalization is that the cytoskeleton is regulated in part by ions, especially Ca2+, but also H+; these relationships also apply to the cytoskeleton in the growing pollen tube (Vidali et al., 2001b). The acto-myosin system in pollen tubes is best recognized as being responsible for cytoplasmic streaming (Shimmen et al., 2000). The pattern of flow, described as ‘reverse fountain’, is required to deliver vesicles to the tip so as to perpetuate growth of the cell by secretion of new membrane and wall material (Pierson et al., 1990). However, elevated [Ca2+]i, at levels expressed in the tip focused gradient, fragments F-actin in pollen tubes (Kohno & Shimmen, 1987), thus apparently destroying or disorganizing the apical most extension of the actin microfilament system. It seems plausible that F-actin fragmentation is carried out by villin, a Ca2+-calmodulin sensitive actin binding protein recently identified in lily pollen (Yokota et al., 1998, 2000; Vidali et al., 1999, 2001a, 2001b). Involvement of villin would account for the fact that the long actin cables become notably less organized in the clear zone at the pollen tube tip, where the [Ca2+]i gradient is present (Kost et al., 1998; Lancelle et al., 1987; Miller et al., 1996). However, in the shank of the pollen tube, where the [Ca2+]i is at basal levels, villin will no longer disorganize F-actin, but instead will cross-link and stabilize microfilament bundles (Yokota et al., 2000).
Yet another protein, which can respond to Ca2+, and which can modify the structure of actin is the small G-actin binding protein profilin. Kovar et al. (2000) have shown that profilin sequesters more actin and reduces the amount of F-actin in regions of high [Ca2+]i; indeed, their calculations show that in the apex of a pollen tube, where the [Ca2+]i might be 1 µm, the F-actin could be reduced by nearly one half.
Finally, the tip-focused [Ca2+]i gradient is expected to inhibit myosin motor activity, and block cytoplasmic streaming in that location. Yokota et al. (1995) have isolated a 170-kDa protein from lily pollen, which based on the partial cDNAs, is identified as a class XI myosin. Although the details of the mechanism are not entirely clear, it seems possible, as with class V myosins, that elevated Ca2+ binds to calmodulin and modifies its association with the myosin neck region, thus decreasing ATPase activity and motility (Yokota et al., 1999; Yokota, 2000). However, at basal Ca2+ levels, as exhibited throughout the shank of the pollen tube, the myosin will be fully active, and serve as the motor for cytoplasmic streaming.
H+ can also have a profound interaction with the actin cytoskeleton, and are expected to participate in the control microfilament turnover. Of particular note is the presence of actin depolymerizing factor (ADF)/cofilin (Lopez et al., 1996; Gungabissoon et al., 1998; Allwood et al., 2002; Chen et al., 2002), which is a pH sensitive actin binding protein found in pollen tubes. It has been noted that the pollen forms of ADF are less sensitive to pH than somatic/constitutive ADF (Smertenko et al., 2001). Nevertheless, there is a sensitivity (Chen et al., 2002), which given the substantial pH elevation evident in the alkaline band (Feijóet al., 1999), might contribute to the turnover of actin microfilaments. It should also be noted that Ca2+, in addition to H+, can affect the activity of ADF (Smertenko et al., 1998). By stimulating phosphorylation, Ca2+ down regulates the activity of ADF (Smertenko et al., 1998), although recent work questions the importance of this reaction as a regulatory mechanism, at least in lily pollen tubes (Allwood et al., 2002). Because of the current level of activity in uncovering new and different actin binding proteins, for example actin interacting protein (AIP; Allwood et al., 2002), some of which will be modulated by ions, we can anticipate that our understanding of the ways in which ions interact with microfilament formation and function will increase in the next few years.
An exciting and important development in our understanding of polarized growth has been the discovery of Rops, which are small-GTPases in the Rho family, that appear to play a pivotal role in the growth and development of pollen tubes, root hairs, and possibly in many other plant cell systems (Valster et al., 2000; Zheng & Yang, 2000a, 2000b; Fu & Yang, 2001; Yang, 2002). Although they show 70% sequence similarity to Racs, they are nevertheless a separate group, and for that reason will be referred to as Rops, based on their original identification. We will not attempt to review in detail the role of these proteins, as there are several authoritative current articles (Fu & Yang, 2001; Valster et al., 2000; Zheng & Yang, 2000a, 2000b; Yang, 2002). However, we will focus on their role in polarized growth, with specific reference to their possible interaction with ion gradients and fluxes.
It is clear that Rops are involved with polarized growth. Rop1 (Li et al., 1999) and Rop5 (Kost et al., 1999) are localized to the apical plasma membrane of pollen tubes. Dominant negative forms of the protein or function inhibiting antibodies block pollen tube growth, while overexpression or constitutively active forms cause the formation of balloons, in which cell expansion is delocalized. These observations are not restricted to pollen tubes because similar results have been obtained in studies of root hairs (Molendijk et al., 2001). In an effort to make a connection between Rop activity and the intracellular Ca2+ gradient Li et al. (Li et al., 1999) injected a function-inhibiting antibody into cells loaded with the Ca2+ indicator, fura-2 dextran. The results showed that within 1–2 min the antibody inhibited growth and led to a dissipation of the tip-focused Ca2+ gradient.
As these data are repeatedly referred to as providing evidence that Rops regulate the formation of the tip-focused Ca2+ gradient, it is important to look at the original results carefully. Firstly, there are several agents or conditions including high osmoticum, BAPTA buffers, brief thermal shock, and caffeine (Pierson et al., 1994, 1996), all of which block pollen tube growth and simultaneously lead to the dissipation of the tip-focused Ca2+ gradient (Pierson et al., 1994, 1996). These data unfortunately do not establish a hierarchy of activity; they only say that inhibition of growth correlates with the dissipation of the apical Ca2+ gradient. Therefore, while it is interesting that anti-Rop antibodies block growth (Li et al., 1999), those data do not establish that Rop activity directly regulates Ca2+. For example, the anti-Rop antibody might directly block secretion, which prevents extension of the cell wall, which in turn prevents the opening of stretch activated ion channels, and which finally prevents a [Ca2+]i increase. Secondly, Li et al. (Li et al., 1999), although showing an affect on the intracellular Ca2+ gradient, nevertheless repeatedly refer to this as an inhibition of ‘extracellular Ca2+ influx’. While it is true that the intracellular gradient may derive directly from extracellular influx, the release from internal stores has not been ruled out. Definitive statements about influx need to be supported by direct extracellular current measurements.
Additional observations however, help make a connection between Rops and Ca2+. Li et al. (Li et al., 1999) showed that dominant negative mutants of rop1at and antisense rop1at RNA are growth inhibited in low extracellular [Ca2+] (0.5 mm) where wild-type tubes grow well. However, high extracellular [Ca2+] (10–20 mm), which inhibits growth of wild-type tubes, partially rescues the dominant negative mutants, and completely reverses the effect of expression of antisense rop1at (Li et al., 1999). Although there are several uncertainties, the simplest interpretation of the results from the dominant negative mutants is that Rop1At increases the efficiency of the Ca2+ channels at the pollen tube tip such that they can grow in lower extracellular Ca2+. In order to dispel ambiguities, it is imperative in future work to directly determine both the intracellular [Ca2+], and the extracellular Ca2+ current under all pertinent experimental conditions.
A further interesting observation, which may relate to Ca2+ regulation, is the finding that Rop5 (At-Rac2) associates with PIP-Kinase in the apical plasma membrane of the growing pollen tube (Kost et al., 1999). This enzyme catalyzes the formation of phosphatidyl inositol 4, 5 bisphosphate (PIP2), which also localizes to the apical membrane, again in a pattern that is similar to Rop. The connection to Ca2+ stems from the ability of PIP2 to be hydrolyzed to diacylglycerol (DAG) and IP3. While DAG will remain in the plasma membrane, and possibly participate in various signaling events, for example, activation of protein kinase C, IP3 will diffuse into the cytoplasm where it can stimulate the release of Ca2+ from internal stores. Thus the spatial localization of PIP2 to the apical plasma membrane, which makes possible a scheme for generating localized IP3 production and localized Ca2+ release, provides a plausible mechanism in support of the idea that Rops might be associated with internal Ca2+ regulation and that internal Ca2+ release might contribute directly to the tip-focused gradient.
Although somewhat aside from ion regulation, it is important to emphasize that the Rop–PIP2 association could also have a connection to actin regulation. Many actin binding proteins including profilin, ADF/cofilin, and villin, which are present in pollen tubes, and which play key roles in the control of actin polymerization, are bound to and in part regulated by PIP2 (Martin, 1998). On the one hand PIP2 can dissociate the profilin/G-actin complex; it can remove villin from the F-actin barbed end, and promote rapid polymerization; and it can bind ADF/cofilin and block its association with F-actin (Vidali & Hepler, 2001). On the other hand profilin can bind to PIP2 and inhibit its hydrolysis by phospholipase C (Martin, 1998). In summary it should be clear that there are many reactions involving PIP2 and the actin cytoskeleton, which given the apical localization of PIP2 would be expected to spatially restrict cytoskeletal transformations. To the extent that the synthesis and hydrolysis of PIP2 might be regulated by Rops, we can appreciate the central role that this small G-protein could play in pollen tube growth.
VI. Extracellular targets for ions: the cell wall
In the External stores section above we argued that the cell wall might have a profound effect on the magnitude of the extracellular ion fluxes, especially Ca2+. Here we address an equally important issue concerning how extracellular ions affect the structure and yielding properties of the cell wall. The cell wall of the pollen tube, as with other plant cells, is that entity which surrounds the protoplast and resists the internal turgor pressure, thus preventing the cell from bursting (Derksen, 1996). It has many unique properties based on its composition and the way in which the components are deposited, and importantly it is the most crucial element in defining the morphology of the cell, including notably the highly polarized shape of the pollen tube. As noted earlier, the cell wall at the tip is mainly pectinaceous (Heslop-Harrison, 1987; Derksen et al., 1995; Li et al., 1997), in which the components are secreted primarily as methylesters, and subsequently de-esterified by pectin methylesterases. The de-esterification yields carboxyl residues that are available for cross-linking and rigidifying by Ca2+ (Jauh & Lord, 1996; Li et al., 1994). Here we restrict our discussion to structures and activities at the extreme tip of the tube, and to the changes associated with cell elongation. In particular we consider the interaction with ions, especially Ca2+ and H+, but also boron, which are almost certainly key factors in controlling the structural and yielding properties of the wall.
It seems increasingly clear that the tendency for pollen tubes to burst in low concentrations of Ca2+ (i.e. 10 µm or less) is caused by a limitation in the amount of ion cross-linking in the cell wall. Although Steer (Steer, 1989) argued that at 10 µm Ca2+ there should be enough ions to satisfy the necessary binding sites, his analysis fails to consider the temporal nature of the process. Thus the de-esterification of the secreted pectin methyl esters creates a need for Ca2+, which, if not quickly satisfied, will momentarily produce a cell wall that is weak and will burst given the substantial internal turgor pressure. The increased binding of Ca2+ will cross-link adjacent pectin residues and contribute markedly to the rigidity of the cell wall. Earlier (External stores) we argued that this binding of Ca2+ accounts for the large extracellular influx of Ca2+ that has been measured (Holdaway-Clarke et al., 1997). It will be further evident that the gradient in increased de-esterification is also expected to be a gradient in increased wall stiffness (Derksen, 1996). It seems plausible that this gradient in stiffness contributes to the cylindrical morphology of the pollen tube, because it will be most pliable and yielding at the extreme apex and become decreasingly so within a few microns from the tip. This situation alone will restrict cell elongation to the apex. Demethylation of pectin not only allows for the formation of Ca2+ bridges between pectin chains, but can also change the sensitivity to expansins (Carpita et al., 1996). Grass pollen actually secretes expansins, presumably to loosen cell-cell connections in the stigma and thus facilitate penetration by the pollen tube (Cosgrove, 2000).
Another target for Ca2+ in the cell wall might be extracellular calmodulin (Ma & Sun, 1997). Using pollen primarily from Hippeastrum, but also several other species, Ma and Sun (1997) show that pollen germination and tube growth are completely inhibited by the extracellular application of anticalmodulin serum, and the impermeable antagonist W-7, whereas germination and growth are stimulated by the exogenous application of calmodulin. Although it would seem that calmodulin should be fully saturated with Ca2+ in the extracellular space and thus unable to act as a signaling agent, Ma et al. (1999) make two important counter observations. Firstly, in environments of low pH, characterized by the cell wall space, the affinity of calmodulin for Ca2+ is reduced by as much as 10-fold, and secondly, that the [Ca2+] within the wall itself might be much lower than that in the medium. This second point is verified by studies with the extracellular vibrating electrode which directly demonstrate that the [Ca2+] in the cell wall is lower than in the surrounding medium (Kühtreiber & Jaffe, 1990). In studies that attempt to further characterize the effect of extracellular calmodulin, Ma et al. (1999) provide evidence that heterotrimeric G-proteins might be involved; they show that the inhibitory effect of anticalmodulin serum is reversed by the G-protein agonist, cholera toxin, and conversely that the stimulatory effect of exogenously applied calmodulin is blocked by pertussis toxin. Ma et al. (1999) conclude that the G-proteins may operate downstream from extracellular calmodulin.
While Ca2+ is generally considered a stiffening agent in cell walls, H+ are often associated with wall loosening (Cassab & Varner, 1988). PME activity in cell walls is down-regulated by increased acidity (Moustacas et al., 1986). Because the action of PME on the pectins generates H+ when methoxyl groups are converted to carboxyl groups, the resulting change in pH may then control the action of PME itself as well as that of other cell wall enzymes. We have previously suggested this as a possible feed-back mechanism for the regulation of PME activity at the tip of the growing pollen tube (Holdaway-Clarke et al., 1997), and for accounting for an oscillation in the yielding properties of the cell wall. However, low pH might enhance the activity of acidic isoforms of PME (Li et al., 2002), and if these act together with pectin hydrolases they can cause the degradation of pectin gels (Bordenave, 1996). Note is also made of two wall-bound exo-α-glucanases recently isolated from lily pollen tubes, which are highly pH-sensitive (Kotake et al., 2000). Because inhibition of these glucanases blocks pollen tube growth, it seems possible that these enzymes play an important role in the regulation of tip extension. With this enzyme, maintaining the optimal pH at 5.5 seems important because its activity drops off sharply either above or below that value (Kotake et al., 2000). By comparison with root hairs, where ratiometric imaging has revealed local changes in the apoplastic pH (Bibikova et al., 1998), it will be important in future work on pollen tubes to probe the pH within the cell wall in the apical domain of the growing, oscillating pollen tube.
Boron, added in the form of boric acid, is also essential for the in vitro culturing of pollen from most species; for example, it is well appreciated that elimination of boric acid from the culture medium often leads to tube bursting. Although boric acid might contribute to the control of pollen tube growth through its ability to stimulate the plasma membrane H+-ATPase (Obermeyer et al., 1996), most current lines of evidence emphasize its role in the formation of borate ester links that dimerize rhamnogalacturonan II, a subclass of pectins (O’Neill et al., 2001; Ridley et al., 2001; Brown et al., 2002). These dimers are also stabilized by Ca2+ (Matoh & Kobayashi, 1998) and may be crucial in wall rheology at the pollen tube tip, as boron-depleted cells of Chenopodium album are larger, show less cell adhesion and have more porous walls (Fleischer et al., 1998, 1999). Although it is clear that borate esters participate in cell wall structure, there is nevertheless a puzzle concerning how the borate anion arises because the acidic conditions in the wall (pH 5–7), would appear to be highly unfavorable for the ionization of boric acid (pKa = 9.24) (Brown et al., 2002). Perhaps a hint to the solution of this problem comes from the recent identification of a boron transporter in Arabidopsis (Takano et al., 2002). Boric acid can freely enter cells by passive diffusion (Brown et al., 2002; Dannel et al., 2002), but BOR1, a membrane protein with homology to bicarbonate transporters, is required for borate efflux (Takano et al., 2002). In pollen tubes, a plausible scheme could include first the passive uptake of boric acid, followed by its conversion to borate in the cytoplasm. It would then be actively exported to the cell wall where it participates in the formation of the 1, 2 diole borate ester linkages in rhamnogalacturonan II.
Effect of extracellular ions on growth oscillations
As the cell wall structure is affected by Ca2+, H+, and borate, it has seemed possible that modulation of the extracellular concentration of these ions might modify cell wall properties and even oscillatory growth. Accordingly two recent studies have addressed the role of varying extracellular ion concentration on growth oscillations in lily pollen tubes (Holdaway-Clarke et al., 2003; Messerli & Robinson, 2003). Holdaway-Clarke et al. (Holdaway-Clarke et al., 2003) show that Ca2+ and boric acid, but especially the latter, have the most profound effect on oscillatory growth. For example, for boric acid there is an optimum concentration for growth that is 3.2 mm. When the concentration is either less than or greater than 3.2 mm the pollen tubes grow more slowly and also exhibit longer periods and greater amplitudes of oscillation (Holdaway-Clarke et al., 2003). The results for elevated concentrations of boric acid, as well as Ca2+, can be explained by invoking the ability of both of these ions to cross-link pectin residues, for example, the homogalacturanans by Ca2+, and the rhamnogalacturonans by borate (Ridley et al., 2001) and Ca2+ (Matoh & Kobayashi, 1998). Stiffer cell walls yield less readily, which accounts for the slower growth, as well as the longer period of oscillation.
Focusing on Ca2+ and H+, Messerli and Robinson (2003) report that increasing the extracellular [Ca2+] from 0.13 mm to 1.3 mm decreased the peaks growth amplitudes but not the average growth. However, a further increase to 10 mm reduced growth by 43%. Holdaway-Clarke et al. (2003) also noted that 10 mm Ca2+ caused a marked reduction in average growth in oscillating, but not nonoscillating, tubes. In studies on H+, lowering the pH from 5.5 to 4.5 causes a marked reduction in growth, and the elimination of [Ca2+]i oscillations (Messerli & Robinson, 2003). Raising the pH to 6.5 also led to a reduction in growth, but in addition a great deal of tube bursting and cell death. A further study of considerable interest involved the use of MES buffer at high concentration (50 mm) to clamp the cell wall pH at the optimum level of 5.5 (Messerli & Robinson, 2003). These studies, which were designed to test if changes in cell wall pH were necessary for normal growth, revealed that the frequency of oscillations decreased by 75% and the average growth rate decreased by 30%. Although not embraced by the authors, these results can be readily interpreted as a need for local cell wall acidification below pH 5.5 in order to maintain normal rapidly oscillating pollen tube growth. These data are quite similar to those showing that increased concentrations of boric acid reduced the frequency of oscillations and the average rate of growth (Holdaway-Clarke et al., 2003). When taken together these studies further emphasize the potential importance of the cell wall as a primary regulator of the growth behavior of pollen tubes.
VII. Ions in navigation
Although the mechanisms by which pollen tubes navigate their way to the female gametes are not well understood (Wilhelmi & Preuss, 1997; Lush, 1999), it is clear that the position of the peak of the tip-focused [Ca2+]i gradient correlates with the position and direction of cell elongation (Hepler, 1997). Pollen tube growth can be induced to reorient by various ionic or ion-related treatments, such as exposure to electric fields (Malhóet al., 1992), iontophoretic injection (Malhóet al., 1994), uncaging of Ca2+ or a Ca2+-chelator (Malhó & Trewavas, 1996), and most recently by agents that increase cyclic AMP (Moutinho et al., 2001; Trewavas et al., 2002). It seems likely that in planta chemical directional clues alter ionic relations within the pollen tube, which then cause a change in the direction of growth.
Electric fields can effect the direction of pollen tube elongation when tubes are grown close to the cathode or anode, with these orientation effects being suppressed by omission of Ca2+, K+, Mg2+ or Cl− from the medium (Malhóet al., 1992). The presence of low concentrations of the Ca2+ channel blocker La3+ in the medium before exposure to electric fields inhibits the usual rise in [Ca2+]i and the reorientation of growth, indicating that an extracellular Ca2+ influx across the plasma membrane is involved in the process of growth reorientation (Malhóet al., 1994). Manganese (Mn2+) quench experiments, mentioned earlier, also confirm the presence of a substantial extracellular Ca2+ influx at the tip of growing, but not nongrowing, pollen tubes (Malhóet al., 1995). The quenching was almost completely inhibited by the presence of La3+ in the extracellular medium, further demonstrating that this ion indeed blocks extracellular Ca2+ influx. Notably, this effect was most pronounced in tubes exposed to either iontophoresis or electric fields indicating that additional extracellular influx of Ca2+ may be involved in producing the reorientation response to these stimuli. Both iontophoresis and exposure to electric fields cause membrane depolarization that the authors suggest may be involved in activating voltage activated Ca2+ channels in the pollen tube apical dome (Malhóet al., 1995).
In further studies, Malhó and Trewavas (1996) showed that localized elevations or decreases in [Ca2+]i in the apical dome of a growing pollen tube produce a predictable reorientation, dependent on the nature of the change of [Ca2+]i induced. Photoactivation of caged Ca2+ (nitr-5) or exposure to a localized extracellular gradient of the Ca2+ ionophore A-23187 caused tubes to grow toward the direction of elevated [Ca2+]i. By contrast, intracellular uncaging of the Ca2+ chelator diazo-2 or exposure to an extracellular gradient of Gd3+, resulted in tubes elongating away from the site of lowered [Ca2+]i, and away from the highest concentration of Gd3+. This is consistent with the hypothesis that an apical lead-point is determined by the gradient of Ca2+ channel activity in the tip. It is also possible to induce tube reorientation by manipulating [Ca2+]i further back from the tip using similar techniques, but the direction of the new orientation is random. In these instances the authors observed a [Ca2+]i wave moving slowly from the site of Ca2+ manipulation toward the tip. Ion imaging revealed that the ratio of Ca2+-dependent pixel intensities in the right and left hemispheres of the growing pollen tube were predictive of the direction in which the tube would turn. Further, the highest value of this ratio occurred not immediately after photoactivation of caged Ca2+, but when cell bending commenced, indicating that the release of caged Ca2+ into the cytoplasm was amplified by the cell, which then resulted in a change in the direction of growth (Malhóet al., 1995).
A recent study by (Moutinho et al., 2001) builds upon these already extensive results and shows that cAMP also appears to be involved in pollen tube navigation. Using forskolin and dideoxyadenosine, which stimulate adenylate cyclase activity, Moutinho et al. (2001) observe firstly that these agents cause a detectable increase in cAMP, and secondly that they also lead to changes in growth direction. They further find that local application of dibutyryl cAMP, a permeant form of this signaling agent, causes growth to reorient towards the source of dibutyryl cAMP (Moutinho et al., 2001). It is well appreciated that cAMP might be associated with changes in [Ca2+]i, and therefore Malhóet al. (2000) monitored for Ca2+ levels in pollen tubes in which the cAMP levels had been increased either by uncaging a caged form of this agent, by applying permeant cAMP and/or by addition of forskolin. Although the results suffer from using a single wavelength Ca2+ indicator, and thus cannot provide quantitative information, they nevertheless show in each instance that the pixel intensity increases in the apical domain, consistent with the conclusion that the local [Ca2+]i has become elevated.
The role of IP3 has been investigated by photoactivation of caged IP3 in the subapical or nuclear domain of pollen tubes; these treatments resulted in a transient increase in [Ca2+]i and reorientation of the direction of growth that could be inhibited by heparin (Malhó, 1998b). Pollen tubes loaded with heparin and exposed to an electric field, which would normally induce pollen tube bending, failed to change their direction of growth (Malhó, 1998b). These results are consistent with the idea that Ca2+ release from an IP3-dependent store back from the tip, such as the vacuole, is involved in the signal transduction pathway controlling pollen tube navigation. However, it is not clear how these conclusions fit with other studies showing that enhanced influx in the apical domain regulates the direction of pollen tube growth.
VIII. Role of ions in self-incompatibility
Self-incompatibility mechanisms are wide spread among angiosperms, and have been extensively reviewed (Rudd & Franklin-Tong, 2001; Nasrallah, 2002). Here we focus briefly on those instances in which a role for Ca2+ has been shown to be involved, although readers are encouraged to consult Rudd and Franklin-Tong (Rudd & Franklin-Tong, 2001) for a more complete discussion of this topic. A well studied example is the self-incompatibility (SI) response in Papaver rhoeas, which is mediated by a transient elevation in [Ca2+]i elicited by both incompatible stigmatic S glycoproteins, and purified S protein (Franklin-Tong et al., 1993, 1996). Ratio-imaging has revealed that application of incompatible S proteins produces a rapid increase in [Ca2+]i in the subapical region of the pollen tube lasting several minutes and that a ‘wave’ of elevated [Ca2+]i appears to originate in the shank of the pollen tube, and propagate to the tip where the tip focused [Ca2+]i gradient is disrupted and growth stops (Franklin-Tong et al., 1997). Recent investigations using the extracellular Ca2+-selective vibrating electrode have revealed that challenge with incompatible S protein also stimulates an extracellular Ca2+ influx in the subapical region of the pollen tube, in the region of the male germ unit (Franklin-Tong et al., 2002). Application of lanthanides partially blocks the SI-induced extracellular Ca2+ influx, and prevents the initiation of the [Ca2+]i‘wave’ originating back from the tip, indicating that the extracellular compartment is a source for the generation of this intracellular Ca2+ transient.
Ca2+-dependent reactions may be involved in other examples of self-incompatibility. For example, Kunz et al. (1996) find that the S-RNases, from at least two species of Nicotiana, are substrates for phosphorylation by a CDPK derived from pollen tubes of Nicotiana alata. The regulatory role of Ca2+ is supported by the activation of the CDPK in small elevations of [Ca2+] and its inhibition by La3+ (Kunz et al., 1996). However, it has not been established if the phosphorylation reaction controls the activity of the S-RNase. It seems likely that future work will identify additional connections between Ca2+ and the control of pollen tube growth in self-incompatibility reactions.
IX. Plasma membrane; site of global coordination and control
While it is clear that there are structures and activities in both the cytoplasm and the cell wall that are crucial to the process of growth, it is important here to focus briefly on the plasma membrane, the cellular component that rests at the interface between the cytoplasm and the cell wall. Although the general concept is obvious, we nevertheless need to remind ourselves of the central role that the plasma membrane plays in virtually all developmental processes (Kohorn, 2000; Harold, 2001; Heath, 2001; Brownlee, 2002). This may be particularly true for ionic events because it is within the plasma membrane that the specific channels are located, and regulated.
The plasma membrane is that entity which is simultaneously in contact with the cytosol, and the extracellular environment, and thus stands at the interface where it can receive information, and deliver appropriately encoded responses. In the growing pollen tube, if the cell wall yields through a slippage of cross links, the ensuing mechanical strain on the plasma membrane could open stretch activated channels, and allow ions, such as Ca2+ to enter, which then affect secretion and/or the structure and activity of the cytoskeleton. Or somewhat similarly, the binding of a small signal molecule to a receptor on the plasma membrane might stimulate the opening or closing of certain channels, thus changing the ion balance within the immediate vicinity of the apex, and directly modulating down stream reactions.
Changes of chemistry within the wall itself, whether it is the binding of extracellular Ca2+ or the production of H+, would also impinge on the plasma membrane, to which the latter can respond in ways that then dictate how the processes of secretion, and cytoskeletal activity will follow. In this light the plasma membrane becomes the single most central regulatory component responsible for the control of growth. Unfortunately our detailed understanding of its activities is limited. We certainly realize it to be the locus of a variety of ion channels, most of which have yet to be identified and characterized.
Beyond the immediate consideration of ions, it is important to bring in other factors, which together with ions could be part of the pathways that regulate growth. In a recent consideration of the plasma membrane/cell wall interaction Kohorn (2000) identifies three proteins, cellulose synthase, the AGPs, and wall associated kinases (WAKs), that bridge this interface and likely play crucial roles in the growth and development of the plant cell. With regard to the pollen tube we can for the moment dispense with cellulose synthase, as reports indicate that there is little or no cellulose in the apical cell wall domain (Heslop-Harrison, 1987; Derksen et al., 1995; Jauh & Lord, 1996). By contrast, AGPs and WAKs command our attention. We have noted earlier that inhibition of AGPs using the Yariv reagent causes the apical [Ca2+]i to rise dramatically, with the consequent occurrence of secretion that is no longer focused at the apex (Roy et al., 1999). AGPs, which are tethered to the plasma membrane by a GPI anchor (Oxley & Bacic, 1999; Majewska-Sawka & Nothnagel, 2000), appear to play a role at the cell surface in controlling both the process of secretion and the subsequent proper incorporation of the secreted components into the cell wall. These results emphasize the importance of the plasma membrane/cell wall interaction in the control of cell polarity.
The wall associated kinases (WAKs) also emerge as key factors in our attempts to understand the control of plant cell growth (Kohorn, 2000). These proteins, which include conventional WAKs and a much larger group of WAK-like kinases (Verica & He, 2002), are serine/threonine kinases that occur widely in plants, including in pollen tubes. It is particularly pertinent in this current discussion to note that WAKs extend across the plasma membrane and physically interact with carbohydrates, specifically pectins, in the cell wall (Kohorn, 2000). Thus far WAKs have been associated with plant defense responses, but also with cell elongation. Although the work needs to be done, it seems reasonable to predict that WAKs will play an important role in the control of pollen tube growth.
On the cytoplasmic side we draw attention to the plasma membrane as the likely locus where specific associations and attachments with the cytoskeleton occur, and of course, where vesicle exocytotic and endocytotic events reside. Again, we draw attention to the phosphoinositides, of which PIP2 is specifically localized along the plasma membrane in the apical domain of the pollen tube (Kost et al., 1999). Through its ability to bind cytoskeletal proteins (profilin, villin, ADF/cofilin), protein kinases (PLCã), membrane proteins (synaptotagmin), and GTPase and accessory proteins (Rac/Rho GTP-GDP exchange factors) (Martin, 1998), it is abundantly evident that PIP2 is poised to orchestrate several processes, including the ion fluxes that underlie motile and vesicle fusion events, which are essential for pollen tube growth. How the plasma membrane operates at this locus will tell us much about how the pollen tube grows.
X. A model for pollen tube growth
Although many aspects of pollen tube growth have been described and characterized, we are still far from understanding the underlying mechanisms that control growth. In this section we attempt to pull some of these different data together and generate fresh ideas, which may help us uncover the central regulators of growth. And while our focus is on pollen tubes, it is likely that the underlying mechanisms will also apply to other plant cells, certainly to tip growing cells, for example, root hairs, but even to cell expansion in general.
What is the Pacemaker?
Throughout this review we have focused on ions, and we can provide persuasive evidence that given certain ion gradients and fluxes, many of the other events, such as vesicle secretion, cytoskeletal growth and activity will follow. But what orchestrates ion activity? Through an analysis of the phase relationships between the activities of different components within an oscillatory system there is hope that the primary regulator or pacemaker will be identified. An obvious initial focus has been on Ca2+ because of the central role that this ion plays in a wide host of signal transduction events, including in pollen tubes where it is part of the growth regulatory and navigational system. However, it does not appear to be the pacemaker because its changes follow rather than anticipate a growth rate change (Messerli et al., 2000). Instead it is a change in growth rate that predicts the subsequent change in Ca2+. Nevertheless, it is still a puzzle how Ca2+ can determine growth direction, and yet be a follower in controlling the growth process. Clearly, the Ca2+/cell growth connection needs further elucidation.
H+ can also be considered as a potential primary regulator of oscillatory growth. The H+ pump is almost certainly the primary energetic system that ultimately powers other transport processes (Sze et al., 1999), and its apparent continued activity even in nongrowing pollen tubes indicates a potential for initiating events (Feijóet al., 1999, 2001). However, extracellular fluxes of H+, like those of Ca2+ and K+, follow rather than anticipate growth rate changes (Messerli et al., 1999). Conceivably changes in the intracellular pH could anticipate growth; unfortunately the data thus far are too imprecise to definitively establish the phase relationship (Feijóet al., 1999). Thus, while recognizing the central role that H+ gradients play in pollen tube energetics, at present we nevertheless lack sufficient evidence to promote the oscillatory expression of these gradients as the pacemaker.
Of all the ions analyzed thus far Cl− emerges as a possible prime regulator. Most notable is the finding that the oscillatory efflux of Cl− at the apex of the pollen tube is in phase with growth rate oscillations (Zonia et al., 2002). This observation, together with the well-known role of Cl− in salt extrusion and turgor regulation, demands that further attention be given to this ion as a prime regulator of pollen tube growth.
Actin and tube growth
The idea that actin polymerization is essential for pollen tube growth is both novel and exciting (Gibbon et al., 1999; Vidali et al., 2001a). The initial studies show that the process of growth can be divided into two components, one of which is a constant, basal rate of growth, with a second, oscillating component superimposed upon it (Vidali et al., 2001a). While turgor may participate in the basal elongation (approx. 0.1 µm−1 s−1), current evidence, presented below, supports a role for actin polymerization in the oscillatory component (Cárdenas et al., unpublished). When pollen tubes of lily, exhibiting growth rate oscillations, are cultured in the presence of low concentrations of latrunculin B (lat B; 2 nm), the oscillatory component is damped while the basal growth rate continues (Cárdenas et al., unpublished). Further, these low concentrations of lat B appear to affect only the subapical actin cytoskeleton. Thus, control cells have a ‘funnel’ of actin at the base of the clear zone, where cytoplasmic streaming reverses direction, while in cells treated with lat B the funnel is either absent or greatly reduced as is the extent of the clear zone (Vidali et al., 2001a). The tentative conclusion is that the actin ‘funnel’ at the base of the clear zone is involved in the oscillatory component of pollen tube growth in lily. Taken together these data focus our attention on the structural organization and dynamic transformations of the actin cytoskeleton, and their control by ions, especially Ca2+ and H+ (Fig. 5).
Considerable attention surrounds the role of actin polymerization in cell migration. Employing a Brownian ratchet mechanism (Peskin et al., 1993; Theriot, 2000), it is suggested that bundles of actin with their barbed ends proximal to the cell surface, through the action of thermal motions that displace the membrane, add new subunits to the existing MFs, and drive the leading edge forward (Theriot, 2000). In tip growing fungal hyphae, Money (Money, 1997) has calculated for a cylindrical cell 10 µm in diameter, which was packed with 1 × 106 parallel actin MFs, that a force equivalent to only 0.02 MPa of turgor would be generated. By comparison, the lily pollen tube generates a turgor pressure of 0.2 MPa (Benkert et al., 1997), which is 10-fold greater. In addition to the relatively weak force that would be generated by actin polymerization, the structural organization of actin microfilaments in the pollen tube apex is quite dissimilar to that at the leading edge of a migrating animal cell or at the cytoplasmic surface of an ingested Listeria bacterium (Theriot et al., 1992). Bundles of F-actin do not appear to extend into the growing apex of the tube. Moreover, the reverse fountain streaming pattern indicates that many actin microfilaments are oriented with their minus end forward, and presumably unavailable for protrusive growth. These marked differences make it difficult to promote the currently popular ‘Brownian ratchet’ model as a primary force generator for pollen tube growth.
Another model, which we think deserves consideration, involves actin gel/sol transformations and gel-osmotic expansion (Oster & Perelson, 1994). When an actin gel is solated its elastic modulus is reduced. Because of the fixed negative charges, mobile positive charges move into the gel together with water, and the gel swells (Oster & Perelson, 1994). The swelling is an endogenous property of the gel and is not dependent on a membrane barrier. Although the swelling occurs in all directions, it can become directional depending on the constraints of the cell. We think that the apical extension of the pollen tube is a candidate for regulation by gel-osmotic expansion. Given the tip focused Ca2+ gradient in the apex and alkaline band in the clear zone, the ionic conditions are favorable for the activation of villin, ADF/cofilin, and profilin, all of which might contribute to the solation of the presumed apical actin gel (Fig. 5). The ensuing gel expansion, which would be constrained by the relatively nonyielding cell wall along the shank of the tube, would occur in an axial direction and possibly contribute to the forward extension of the cell. It is even conceivable that the funnel of actin at the base of the clear zone would act as a barrier to bias the expansion in an anterograde, as opposed to retrograde, direction. The presumed severing of the F-actin would also create free barbed ends, which would stimulate formation of new filaments (Vidali & Hepler, 2001). As Ca2+ and pH decline, and severing is reduced, the new actin filaments would re-establish the gel structure.
An additional function of the network of actin filaments in the clear zone might derive from their physical linkage to and regulation of ion channels. As noted earlier, there are uncertainties about the structure of actin in the pollen tube apex, nevertheless, a recent study using phalloidin as a fixative and a stain reveals an extensive array of filaments of fragments close to the tip (Foissner et al., 2002). It is possible that these microfilaments might regulate apical ion channels, because it is known that Ca2+ currents in various fungal and mammalian cells are sensitive to drugs that modulate F-actin (Levina et al., 1995; Lader et al., 1999). Thus actin in the region of the clear zone may act dually: as a participant in the process of solation-expansion, and in regulating the flow of Ca2+ into the cytoplasm at the tip.
The cell wall
The cell wall is certainly a key player in the process of pollen tube growth, because the balance between cytoplasmic turgor and wall yielding determines whether the cell wall will stretch and the cell elongate (Cosgrove, 1993). Local changes in the Ca2+ and/or H+ content of the cell wall could generate changes in wall yielding, as would oscillations in the rate of exocytosis. Is it possible that there are factors, for example, expansin, enzymes (e.g. pectin hydrolase), which are released by the cytoplasm into the wall space, that increase the yielding of the cell wall at the tip and permit turgor driven wall stretching. Thus it is possible that oscillations in pollen tube growth rate are directly the result of an oscillation in the yielding properties of the cell wall.
Extracellular ion fluxes and intracellular ion gradients occupy a pivotal position in the control of pollen tube growth. In some instances we can appreciate a role for the ions, for example, Ca2+ and vesicle exocytosis, but in many instances we are unclear about their role. On top of this there is the intracellular/extracellular conundrum, which is probably most acute for Ca2+. Thus, within the cytoplasm Ca2+ regulates a variety of events in the 0.1–10 µm range, however, on the outside of the plasma membrane, in the cell wall, this same ion participates in the control of an entirely different spectrum of events, where it works in the 10 µM-10 mM range. These considerations underscore the importance in giving close attention to both the cytoplasm and the cell wall, not only for Ca2+, but for the other physiological ions. H+ deserve continued attention because the H+ pump almost certainly is the prime energetic mechanism for the pollen tube, as with other plant cells. But in addition, as with Ca2+, H+ in the apoplast can have a profound effect on cell wall structure. Cl− and K+ also command our attention; the recent results with Cl− (Zonia et al., 2002) are especially interesting and bring forth the realization that massive movements of this ion may underpin the control of turgor and water flux in the pollen tube.
It has been both exciting and informative to realize that nearly all the ion expressions that have been analyzed thus far exhibit periodic oscillation in activity, and that these oscillations are tightly coupled to growth. The opportunity thus still presents itself, that through phase analysis of these oscillations we can establish a hierarchical order of events that may help us discover the primary regulator or pacemaker, and thus to decipher the fundamental mechanism of growth.
We thank our colleagues, Maurice Bosch, Luis Cárdenas, Joseph Kunkel, Alenka Wheeler, and Kathleen Wilsen, from the University of Massachusetts, and Jose Feijó, from the Gulbenkian Institute and the University of Lisbon, Lisbon, Portugal, for many helpful discussions and specific criticisms on this manuscript. However, we alone assume responsibility for all statements. This work was supported by funds from the United States National Science Foundation (Grant No. MCB-0077599).