Tip growth in neuronal cells, plant cells, and fungal hyphae is known to require tip-localized Rho GTPase, calcium, and filamentous actin (F-actin), but how they interact with each other is unclear. The pollen tube is an exciting model to study spatiotemporal regulation of tip growth and F-actin dynamics. An Arabidopsis thaliana Rho family GTPase, ROP1, controls pollen tube growth by regulating apical F-actin dynamics. This paper shows that ROP1 activates two counteracting pathways involving the direct targets of tip-localized ROP1: RIC3 and RIC4. RIC4 promotes F-actin assembly, whereas RIC3 activates Ca2+ signaling that leads to F-actin disassembly. Overproduction or depletion of either RIC4 or RIC3 causes tip growth defects that are rescued by overproduction or depletion of RIC3 or RIC4, respectively. Thus, ROP1 controls actin dynamics and tip growth through a check and balance between the two pathways. The dual and antagonistic roles of this GTPase may provide a unifying mechanism by which Rho modulates various processes dependent on actin dynamics in eukaryotic cells.
Tip growth is fundamentally important for the development and morphogenesis of all eukaryotic organisms. This extremely polarized growth is dependent on targeted exocytosis in the apical region of the cell, allowing the generation of tubular structures such as neuronal axons in animals, hyphae in fungi and pollen tubes, and root hairs in higher plants. Several common features exist in tip growth among these different systems: (a) growth occurs rapidly over a long distance; (b) the cell apex is capable of sensing directional cues to redirect growth; and (c) the dynamics of F-actin and Ca2+ are required. Single-celled yeast, a model system for the study of polarized growth, shares some but not all of these features; but it is not clear whether or not the yeast paradigm for polarized growth is applicable to tip growth in higher multicellular organisms. Consequently, the mechanism for tip growth in higher eukaryotes is not well understood.
The pollen tube of higher plants is a fascinating model system for the study of tip growth and thus has attracted broad interest and intense scrutiny for many years (Franklin-Tong, 1999; Zheng and Yang, 2000; Hepler et al., 2001; Johnson and Preuss, 2002; Lord and Russell, 2002). During pollination, pollen grains are deposited on the surface of the stigma (pollen-receiving part of the female pistil in flowers) and communicate with the stigma to select a site for germination. Once germinated, pollen generates a tube that travels along a tortuous path to the ovule where sperm are delivered to the egg cell. This guided tip growth is analogous to neuronal axon targeting (Palanivelu and Preuss, 2000; Lord, 2003; Yang, 2003). Pollen tube growth and axon development and guidance apparently share remarkable similarities in the underlying molecular and cellular mechanisms. For example, Rho family GTPases, Ca2+ fluctuations, and actin dynamics all play important roles in these processes (Hepler et al., 2001; Fu et al.; 2001; Vidali et al., 2001; McKenna et al., 2004). However, the mechanistic linkage among these events is unknown in these systems.
ROP1, belonging to the Rho family of small GTPases including Rho, Cdc42, and Rac, is a central regulator of tip growth in pollen tubes (Lin and Yang, 1997; Kost et al., 1999; Li et al., 1999). Tip-localized ROPs are essential for tube elongation and the control of growth polarity. Pollen tube growth is blocked by depletion or inhibition of ROP1, whereas expression of constitutively active mutants of ROP1 induces growth depolarization (i.e., bulbous tubes or tip swelling; Lin and Yang, 1997; Kost et al., 1999; Li et al., 1999). ROP1 regulates the dynamics of F-actin in the tip and possibly Ca2+ signaling as well (Li et al., 1999; Fu et al., 2001). ROP1 inactivation eliminates a tip-focused Ca2+ gradient essential for pollen tube growth and alters pollen tube growth responses to extracellular Ca2+ (Pierson et al., 1994; Malho and Trewavas, 1996; Li et al., 1999). ROP1 inactivation also depletes apical F-actin, whereas ROP1 overexpression (OX) induces both the stabilization of this F-actin and the depolarization of growth. The ROP1 OX–induced growth depolarization is suppressed by latrunculin B (LatB), an actin-depolymerizing drug, indicating that ROP-dependent actin dynamics are critical for tip growth (Fu et al., 2001).
Actin dynamics play a fundamental role in many processes in all eukaryotic organisms; e.g., neuron development, growth, and guidance; cell polarity, growth, and movement; and gene expression (Dickson, 2001; Luo, 2002; Miralles et al., 2003; Pollard and Borisy, 2003). For example, the axonal growth cone motility that involves constant extrusion and retraction of filopodia and lamellipodia is controlled by the dynamics of cortical fine F-actin (Dickson, 2001). In most cases, actin dynamics involve Rho family GTPase-mediated actin assembly and its disassembly promoted by ADF, gesolin, and/or profilin (Etienne-Manneville and Hall, 2002; Pollard and Borisy, 2003).
In this paper, we demonstrate that ROP1 GTPase controls actin dynamics in pollen tip growth via two coordinately counteracting pathways controlled by the ROP1 targets RIC3 and RIC4. The RIC4 pathway promotes the assembly of the apical F-actin, whereas the RIC3 pathway promotes disassembly of this actin via a Ca2+-dependent process. Furthermore, these pathways regulate each other to control actin dynamics and tip growth. To our knowledge, this is the first demonstration of a Rho GTPase signaling network with two antagonistic pathways acting coordinately to control actin dynamics and tip growth, providing important new insights into the mechanism underlying these fundamental processes in higher eukaryotic organisms.
RIC3 and RIC4 are two distinct ROP1 target proteins
To understand how ROP1 regulates pollen tube growth, we investigated a class of putative ROP1 effector proteins from Arabidopsis thaliana, termed RICs (ROP-interactive CRIB-containing proteins; Wu et al., 2001). RICs are novel proteins that share a Cdc42/Rac interactive binding (CRIB) motif, which is present in many Cdc42/Rac effectors and responsible for their specific interaction with activated Cdc42/Rac (Burbelo et al., 1995; Aspenstrom, 1999). Among nine RICs expressed in A. thaliana pollen, only RIC3 or RIC4 OX induced depolarized growth in tobacco pollen tubes as did ROP1 OX, suggesting that these two RICs, but not other RICs, may be ROP1 targets to control polarized tip growth in pollen tubes (Wu et al., 2001).
If RIC3 and RIC4 are ROP1 targets, they are expected to bind active ROP1 and to display ROP1 activation-dependent localization and function. In vivo ROP1 interaction with RIC3 or RIC4 was tested using fluorescence resonance energy transfer (FRET) analysis of tobacco pollen tubes coexpressing CFP-ROP1 and YFP-RIC3 or YFP-RIC4 fusion genes. As shown in Fig. 1 A, YFP-RIC4 or -RIC3 signals were detected when tubes were excited with a laser that excites CFP but not YFP, indicating FRET was occurring. The same settings did not detect any YFP signals in tubes expressing CFP-ROP1/YFP, CFP/YFP-RIC4, or CFP/YFP-RIC3 (unpublished data). FRET between CFP-ROP1 and YFP-RIC3 or -RIC4 was further confirmed by acceptor photobleaching experiments, in which an increase in CFP-ROP1 signal was detected when YFP-RIC fluorescence was bleached (unpublished data). No FRET signals were detected when CFP-ROP1 was replaced with CFP-DN-rop1, a dominant-negative T20N mutant of ROP1 (Li et al., 1999; Fig. 1 A). Furthermore, no FRET occurred when CFP-ROP1 and YFP-RIC3 or -RIC4 were coexpressed with RopGAP1 (ROP GTPase-activating protein 1; unpublished data), which deactivates ROP1 (Wu et al., 2000). Finally, no FRET signals were detected between ROP1 and another plasma membrane (PM)–targeted ROP, ROP10, showing that FRET does not occur between two noninteracting PM-localized proteins (unpublished data). These results indicate that RIC3 and RIC4 specifically interact with active ROP1 in vivo.
The CFP-ROP1/YFP-RIC4 FRET signals were preferentially distributed to the apical region of the PM (Fig. 1 B), similar to the GFP-RIC4 localization (Wu et al., 2001). CFP-ROP1/YFP-RIC3 FRET signals were slightly stronger in the apical region of the PM compared with the cytoplasm (Fig. 1 B), whereas GFP-RIC3 is preferentially localized to the apical region of the cytoplasm (Wu et al., 2001). This difference could be explained by rapid dissociation of activated RIC3 from the PM-localized GTP-bound ROP1. Nonetheless, these results suggest that the apical localization of RIC3 and RIC4 involves their interaction with activated ROP1.
We further examined if the apical RIC localization requires its interaction with ROP1. Mutations in two conserved histidine residues in the CRIB motif (H37D/H40D) in RIC1 abolished its interaction with ROP1 in vitro (Wu et al., 2001). As shown in Fig. 1 B, similar mutations eliminated the preferential localization of GFP-RIC3 to the cytoplasmic region behind the apical dome as well as the PM localization of GFP-RIC4. Furthermore, these mutations abolished the ability of RIC3 or RIC4 to induce tip swelling (Fig. 1 B). Overexpressing either RopGAP1 or DN-rop1 effectively caused GFP-RIC3 or GFP-RIC4 to uniformly localize to the cytosol like soluble GFP alone (Fig. 1 B). Furthermore, RopGAP1 suppressed GFP-RIC3– or GFP-RIC4–induced growth depolarization by reducing the enhanced width of GFP-RIC3–overexpressing tubes (from 14.7 to 9.2 μm; n = 22) and GFP-RIC4–overexpressing tubes (from 18.4 to 8.9 μm; n = 27). DN-rop1 reduced the width of GFP-RIC3–overexpressing tubes from 14.7 to 12.3 μm (n = 25) and GFP-RIC4–overexpressing tubes from 18.4 to 13.8 μm (n = 23). All the differences were statistically significant (P < 0.05, Student's t test). Thus, ROP1 activation is required for RIC3 and RIC4 localization and function. Together, we conclude that RIC3 and RIC4 are both ROP1 targets in the control of pollen tube tip growth.
Because several other ROPs (ROP8–11) functionally distinct from ROP1 are also expressed in A. thaliana pollen, we tested whether RIC3 or RIC4 could act as targets of these ROPs (Fig. S1 and Table S1, available at http://www.jcb.org/cgi/content/full/jcb.200409140/DC1). Unlike ROP1, none of these ROPs altered GFP-RIC3 or -RIC4 localization when overexpressed, indicating that in cultured pollen tubes RIC3 and RIC4 are specific functional targets of ROP1 in the control of polarized tip growth but not of other pollen-expressed ROPs functionally distinct from ROP1.
RIC4 promotes the assembly of apical F-actin
Because tip-localized ROP1 regulates the dynamics of apical F-actin (Fu et al., 2001), we tested whether a RIC3 or RIC4 is involved in this aspect of ROP action. A GFP-tagged actin-binding domain of mouse talin (GFP-mTalin) decorated two distinct forms of F-actin in tobacco pollen tubes: axial cables that extend throughout the tube but end behind the tip of growing tubes and some form of dynamic tip F-actin (Fu et al., 2001). The GFP-mTalin–labeled apical structure seems to consist of short bundles at the extreme apex, apparently alternating with a collarlike structure a few micrometers back from the extreme apex (Fu et al., 2001; Fig. 2 A). The exact configuration of this tip actin structure remains to be determined because GFP-mTalin binding to F-actin might alter its configuration; so this form of F-actin is vaguely referred to as tip F-actin hereafter. Nonetheless, RIC4 and RIC3 OX induced distinct changes in apical GFP-mTalin labeling patterns in tobacco pollen tubes (Fig. 2 A vs. Fig. 4 A). RIC4 OX converted the dynamic tip F-actin into a dense F-actin network, but did not change the axial actin cables (Fig. 2 A; 90% tubes, n = 21). Quantitative analysis shows that RIC4 OX resulted in a significantly greater amount of F-actin at the tip than control tubes (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.300409140/DC1). The change in the tip F-actin was similar to that caused by ROP1 OX, except that RIC4 OX did not induce abnormal subapical actin hoops seen in ROP1-overexpressing tubes (Fu et al., 2001). Mutations in two conserved histidine residues in the CRIB motif (H107D/H110D) eliminate the ability of RIC4 to affect tip F-actin (Fig. 2 A), indicating that RIC4 interaction with active ROP1 is critical for its regulation of actin organization.
To test whether the RIC4-induced stable F-actin network was the effect or the cause of growth depolarization, RIC4-overexpressing tubes were treated with LatB, which partially suppresses the ROP1 OX phenotype (Fu et al., 2001). LatB recovered normal apical F-actin and tip growth in the majority (85%, n = 52) of RIC4-overexpressing tubes (Fig. 2, A–C; and Fig. S2). Similarly, cooverexpression of AtPFN3, a G-actin sequestering protein expressed in pollen, suppressed both depolarized growth and the apical actin network induced by either ROP1 (not depicted) or RIC4 OX (Fig. 2, B and C). Compared with LatB, the recovery of tip growth by AtPFN3 was only partial (80% tubes, n = 50), suggesting either that a factor that activates AtPFN3 is rate limiting or that another actin-depolymerizing factor coordinates with AtPFN3 in regulating actin dynamics.
Similar to ROP1 OX (Fu et al., 2001), RIC4 OX disrupted the oscillation of normal tip F-actin (Fu et al., 2001; Fig. 2 D). Furthermore, LatB recovered the oscillation of tip F-actin in RIC4-overexpressing tubes (Fig. 2 D; 92% tubes, n = 51). Together, our observations indicate that RIC4 promotes the assembly of tip F-actin in pollen tubes.
To further investigate the role of RIC4 in pollen tube growth, we isolated a loss-of-function (LOF) ric4 mutant in A. thaliana, ric4-1, which contained a T-DNA insertion in the second intron and exhibited reduced RIC4 mRNA level (Fig. 3 A). The ric4-1 tubes were significantly shorter and were more sensitive to growth inhibition caused by LatB (0.1–0.5 nM) than wild-type (WT) tubes (Fig. 3 C). Additional LOF ric4 lines were generated using an RNAi construct under the control of the pollen-specific LAT52 promoter (Eyal et al., 1995). RIC4 mRNA levels were dramatically reduced in several independent RIC4(RNAi) lines (Fig. 3 B). Pollen tube growth in these lines was inhibited and was hypersensitive to LatB as seen in ric4-1 (Fig. 3 D), indicating that the ric4-1 phenotype was not due to an extragenic second site mutation. These results show that RIC4 is important for pollen tube elongation.
ROP1 inactivation causes depletion of tip F-actin in tobacco pollen tubes (Fu et al., 2001). Similarly, ric4-1 tubes contained less tip F-actin than WT tubes (Fig. S3, available at http://www.jcb.org/cgi/content/full/jcb.200409140/DC1). However, WT A. thaliana pollen tubes did not contain as conspicuous F-actin at the tip as do tobacco pollen tubes. This finding could be explained by greater actin disassembly and/or less actin assembly in A. thaliana pollen tubes, which is consistent with their slower growth, as compared with tobacco pollen tubes. Indeed, RIC4 OX increased tip F-actin accumulation (Fig. S3), promoted pollen tube growth, and reduced LatB sensitivity (Fig. 3 D). These data suggest that actin assembly is a rate-limiting step in the elongation of WT A. thaliana pollen tubes.
RIC3 regulates calcium signaling
In contrast to RIC4 (Fig. 2 A), RIC3 OX did not convert the dynamic tip F-actin into a dense network in tobacco pollen tubes (Fig. 4 A). Instead, RIC3 OX caused loss of tip F-actin and protrusion of the axial actin cables toward the extreme apex of the tobacco pollen tube (Fig. 4 A; 85% tubes, n = 20). In WT tubes, the actin cables were absent in the apex (Fu et al., 2001). Furthermore, unlike RIC4 (Fig. 2 A), RIC3-induced tube swelling was not suppressed by either LatB (Fig. 4 A; 88% tubes, n = 50) or AtPFN3 (not depicted). Thus, RIC3 has a role distinct from RIC4.
The RIC3-induced actin changes were similar to actin reorganization induced by high [Ca2+]ex in WT pollen tubes (Fig. 4 A). We speculated that RIC3 might regulate Ca2+ signals in the tip. Indeed both EGTA (Ca2+-chelating agent) and LaCl3 (which blocks PM-localized inward Ca2+ channels) suppressed RIC3 OX induced growth inhibition and actin reorganization (Fig. 4, A and B; for LaCl3, 82% tubes, n = 50; and for EGTA, 86% tubes, n = 50, respectively).
To further investigate the connection between RIC3 and Ca2+ signaling, we examined Ca2+ responses of pollen tubes from A. thaliana RIC3(RNAi) and RIC3 OX lines. We obtained multiple independent homozygous lines having similar phenotypes and altered RIC3 transcript levels for each of the RIC3(RNAi) and RIC3 OX constructs (Fig. 4 C). WT A. thaliana pollen tubes require [Ca2+]ex for growth, likely due to the dependency of tip-focused cytosolic Ca2+ ([Ca2+]cyt) gradients on an influx of [Ca2+]ex (Obermeyer and Weisenseel, 1991; Pierson et al., 1994). The elongation of WT tubes showed a typical bell shape response to various [Ca2+]ex with 5 mM being optimal (Fig. 4 D). Growth inhibition by high [Ca2+]ex is thought to be due to a negative feedback regulation of Ca2+ influxes (Li et al., 1999), but might also involve increased rigidity of apical pectic wall and Ca2+-mediated inhibition of membrane permeability (Robinson, 1977; Holdaway-Clarke et al., 1997). Compared with WT control, RIC3(RNAi) inhibited tube growth at lower [Ca2+]ex (0.5 to 5 mM) but promoted growth at high [Ca2+]ex (10 mM), at which RIC3(RNAi) tubes elongated to the length of WT tubes grown under the optimal [Ca2+]ex condition (5 mM). RIC3 OX had a completely opposite effect; it promoted tube elongation at 0.5 mM [Ca2+]ex, but inhibited growth at 2 mM or higher [Ca2+]ex. These results cannot be explained by any role for RIC3 in mediating cell wall rigidity or general membrane leakiness, but can be satisfactorily explained by its promotion of calcium influxes, which are subject to a negative feedback regulation by high Ca2+ levels.
LaCl3 treatments were used to further test RIC3 regulation of [Ca2+]ex influxes. LaCl3 (100 μM) mimicked the effect of RIC3(RNAi) on the growth of WT tubes; i.e., optimal growth shifted from 5 to 10 mM of [Ca2+]ex (Fig. 4 E). LaCl3 reversed the effect of RIC3 OX; i.e., LaCl3-treated RIC3-overexpressing tubes mimicked the Ca2+ response of WT tubes in the absence of LaCl3 (Fig. 4 E). Furthermore, LaCl3 suppressed the ability of high [Ca2+]ex to rescue pollen tube elongation inhibited by RIC3(RNAi) (unpublished data). These results suggest that RIC3 modulates Ca2+ signaling by affecting [Ca2+]ex influxes.
RIC3 promotes the accumulation of tip-localized calcium
We next examined whether RIC3 actually influenced the level of cytosolic Ca2+ at the tip of the growing pollen tube. We used petunia pollen tubes for Indo-1–based ratio imaging of Ca2+ levels, as we have been unsuccessful in pressure microinjecting A. thaliana pollen tubes with Ca2+ imaging dyes, whereas microinjection of petunia tubes has a high success rate. To confirm that RIC3 likely acts in petunia tubes as predicted from A. thaliana, pollen bombarded with LAT52:RIC3 was cultured at different [Ca2+]ex levels. At 100 μM [Ca2+]ex, untransformed pollen did not germinate (0% germination, n > 200 grains, three separate experiments), whereas 100% of LAT52:RIC3 pollen grains germinated (n = 50, three separate experiments). At 1 mM [Ca2+]ex, 70% of untransformed pollen germinated (n = 620), but few LAT52:RIC3 pollen grains germinated (9%, n = 45). These results are consistent with the effect of RIC3 on Ca2+-dependent growth in both tobacco and A. thaliana pollen, although the optimal Ca2+ level for pollen tube germination and elongation differs among the three species. At 100 μM [Ca2+]ex, only GFP-RIC3–overexpressing tubes showed a tip-focused Ca2+ gradient similar to those of control tubes growing at 1 mM Ca2+ (Fig. 4 F). However, LAT52:RIC3 tubes at 1 mM [Ca2+]ex not only displayed the swollen tip phenotype as predicted from the effect on tobacco tubes, but also displayed an altered tip focused Ca2+ gradient, which extended throughout the swollen apex of the tube (Fig. 4 F). Thus, under conditions where control pollen tubes exhibited a normal tip-focused Ca2+ gradient, RIC3 OX led to a delocalized Ca2+ gradient and disrupted apical growth. Interestingly, the delocalized Ca2+ gradients were not observed in RIC4-overexpressing petunia pollen tubes, although they are also swollen at the tips like RIC3-overexpressing tubes. Instead, RIC4 OX led to an apparently reduced tip-focused Ca2+ gradient.
RIC3 and RIC4 counteract to control actin dynamics and apical growth
The results described above indicate that ROP1 activates two pathways: (1) a RIC4 pathway promoting apical F-actin assembly, and (2) a RIC3 pathway that modulates [Ca2+]ex influxes required to generate tip-focused [Ca2+]cyt gradients. The ROP-dependent tip F-actin and Ca2+ gradients oscillate with the same period but in the different phases (Fu et al., 2001), suggesting a possible functional coordination between the RIC3 and RIC4 pathways. To test this possibility, we cooverexpressed RIC3 and RIC4 in tobacco pollen tubes. Surprisingly, this resulted in the recovery of tip growth and normal dynamics of tip F-actin (Table I and Fig. 5 A). This antagonistic effect of RIC3 and RIC4 is unlikely to be due to a competition between these two proteins for binding to ROP1 because RIC3 and RIC4 do not show identical subcellular localization (Wu et al., 2001) and because GFP-RIC4 localization to the apical region of the PM was not altered by RIC3 OX and vice versa, whereas other PM-localized RICs (e.g., RIC1) can remove RIC4 from the PM (unpublished data). Importantly, the ability of RIC3 and RIC4 to counteract each other requires their respective downstream targets, Ca2+ and actin (see the next three sections).
A balance between RIC3 and RIC4 is critical for efficient tip growth
The counteraction of the RIC3 and RIC4 pathways suggests that a check and balance between them is important for ROP control of tip growth. Therefore, we next determined whether there was an optimal RIC3/RIC4 ratio for the most efficient polar growth. Varied amounts of LAT52:RIC3 and LAT52:RIC4 were coexpressed in tobacco pollen tubes. When either RIC3 or RIC4 was overexpressed alone, the tube length was only half that of WT (Fig. 5 B). With fixed RIC4 DNA amount and RIC3 DNA amount increasing, the tube length increased. With RIC3/RIC4 = 1:2, optimal tip growth occurred.
To test whether or not optimal tip growth requires a proper balance of endogenous RIC3 and RIC4, we took advantages of A. thaliana LOF ric3 and ric4 lines with reduced endogenous expression of respective RIC3 and RIC4 genes. Pollen tubes from each of these individual mutants were much shorter than WT tubes at 5 mM optimal [Ca2+]ex (Fig. 5, C and D). However, pollen tubes from the ric4-1/RIC3(RNAi)-1 double mutant grew almost as long as did WT tubes (Fig. 5, C and D). Furthermore, the elongation of ric4-1/RIC3(RNAi)-1 tubes showed similar responses to [Ca2+]ex levels as that of WT tubes (Fig. S4, available at http://www.jcb.org/cgi/content/full/jcb.200409140/DC1). Apical F-actin in ric4-1/RIC3(RNAi)-1 tubes was also normal (Fig. S3). These results clearly demonstrate that it is the balance of counteracting RIC3 and RIC4 that is critical for actin dynamics and tip growth in pollen tubes.
RIC3 acts through Ca2+ to promote the disassembly of RIC4-dependent F-actin
Next, we investigated the mechanisms by which the RIC3 and RIC4 pathways counteract and coordinate with each other. We suspected that RIC3-mediated Ca2+ signaling might be critical for the dynamics of tip F-actin. Ca2+ activates several actin disassembly factors, e.g., profilin and gelsolin-like proteins (Kovar et al., 2000; Huang et al., 2004), and elevation of intracellular Ca2+ levels causes dramatic actin disassembly in poppy pollen tubes (Geitmann et al., 2000). Interestingly, raising [Ca2+]ex from 0.01 to 1 mM effectively recovered normal tip growth (Table I) and the dynamics of apical F-actin (Fig. 5 A) in RIC4-overexpressing tobacco pollen tubes, demonstrating that Ca2+ antagonizes RIC4 action. To test if RIC3-induced Ca2+ elevation in the tip of pollen tubes caused the RIC4-induced disassembly of tip F-actin, we examined the effect of LaCl3 on the counteraction of RIC3 on RIC4. Indeed, LaCl3 treatments of tubes cooverexpressing RIC3 and RIC4 induced growth depolarization (Table II) and the formation of a stable apical F-actin network (not depicted) as found in untreated tubes overexpressing RIC4 alone. These results indicate that RIC3 acts through Ca2+ to promote actin dynamics and to antagonize RIC4.
Profilin may be one of the Ca2+-dependent actin disassembling factors linking RIC3-induced Ca2+ signals to the dynamics of RIC4-activated F-actin because it partially suppressed RIC4 OX phenotypes (Table III). Increasing the amount of PFN3 did not improve its ability to suppress the RIC4 OX phenotype, suggesting a potential rate-limiting factor required for the function of PFN3 (unpublished data). When RIC4 was coexpressed with both PFN3 and RIC3, normal pollen tube growth was completely recovered, although the same amount of RIC3 or PFN3 alone only partially suppressed the RIC4 OX phenotype (Fig. 2 and Table III). However, PFN3 OX did not enhance RIC3 OX phenotype, suggesting that RIC3-dependent Ca2+ may have additional cellular targets such as vesicle fusion. Together, our results indicate that RIC3-mediated [Ca2+]cyt, probably acting through profilin and possibly other Ca2+-mediated actin-binding proteins such as gelsolins (Huang et al., 2004), promotes the disassembly of RIC4-dependent apical F-actin, leading to its dynamics during tip growth.
RIC4 acts through F-actin to counteract the RIC3 pathway
RIC4 suppression of tip-localized Ca2+ accumulation (Fig. 4 F) suggests that RIC4 may counteract the RIC3-dependent pathway. To test whether or not the RIC4 counteraction of RIC3 is dependent on F-actin, we treated GFP-RIC3 and RIC4 cooverexpressing tubes with 5 nM LatB, a concentration that efficiently suppressed RIC4 OX–induced tube swelling (Fig. 2). GFP-RIC3 and RIC4 cooverexpressing tubes showed polarized growth, but LatB-treated tubes cooverexpressing GFP-RIC3 and RIC4 exhibited growth depolarization identical to that of GFP-RIC3–overexpressing tubes (Table IV). WT tubes treated with 5 nM LatB showed a significant reduction in tube elongation but only a slight growth depolarization (Fig. 2). Therefore, these results suggest that RIC4-mediated F-actin accumulation is required to counteract the RIC3 pathway.
In this paper, we have established a novel mechanism for Rho GTPase regulation of actin dynamics and polarized cell growth. ROP1 activates two counteracting pathways that regulate actin assembly and disassembly, respectively. The RIC4 pathway leads to the assembly of apical F-actin, whereas the RIC3-mediated Ca2+ signaling is required for the disassembly of this F-actin. A check and balance of these two downstream pathways is critical for efficient polarized growth of pollen tubes, not only because both pathways are required for correct control of actin dynamics but also because they coordinately regulate each other. This paper is also the first to demonstrate an interplay between a Rho GTPase and Ca2+ in the control of actin dynamics and to establish a molecular circuitry controlling tip growth in higher eukaryotic cells (Fig. 6).
ROP GTPase controls two pathways crucial for tip growth
Our earlier studies have implicated tip-localized ROP GTPases in the regulation of both the generation of tip-focused Ca2+ gradients and the assembly of dynamic tip F-actin (Li et al., 1999; Fu et al., 2001). In this paper, we provide evidence that ROPs may coordinate these two cellular events through the activation of the RIC3 and RIC4 pathways, respectively. The LOF mutants for either RIC3 or RIC4 exhibit impaired pollen tube elongation, suggesting that both pathways are required for pollen tube growth. This finding is consistent with the observation that ROP-dependent apical F-actin and tip-focused Ca2+ gradients are essential for pollen tube growth (Li et al., 1999; Fu et al., 2001; Hepler et al., 2001; Vidali et al., 2001).
Rho GTPases modulate actin assembly in all eukaryotic organisms, and many Rho effectors controlling this process are conserved in animals and yeast. However, RIC4 does not share sequence similarity to any known Rho GTPase effectors or any other conserved proteins that control actin assembly (e.g., WASP and SCA/WAVE). RIC4 also promotes fine cortical F-actin in A. thaliana leaf pavement cells (Fu et al., 2005). The mechanism by which RIC4 promotes actin assembly and the structural nature of RIC4-mediated F-actin remain to be investigated.
Based on the ability of the PM Ca2+ channel blocker (LaCl3) to block the effect of RIC3, we favor the hypothesis that RIC3 regulates the tip-localized influx of extracellular Ca2+, which subsequently modulates for the formation of the tip-focused Ca2+ gradient. Consistent with this hypothesis are recent studies of ROP2 signaling in the control of root hair tip growth (Molendijk et al., 2001; Jones et al., 2002). ROP2 may act upstream of RHD2, which encodes an NADPH oxidase implicated in the regulation of Ca2+ influxes in the control of root hair tip growth (Foreman et al., 2003). RIC3 might regulate NADPH oxidase-mediated Ca2+ influxes.
Checks and balances between the RIC3 and RIC4 pathways are critical for spatiotemporal control of tip growth
Our LOF and gain-of-function studies of RIC4 and RIC3 have demonstrated a critical role for a check and balance between their actions in the control of apical growth in pollen tubes (Fig. 5). The check and balance is necessary for the dynamics of apical F-actin, which is required for the spatial control of polarized growth in pollen tubes (Fu et al., 2001). Actin dynamics depends on both the assembly and the disassembly of an actin filament. Interestingly, the two ROP downstream pathways (RIC4 and RIC3) promote the assembly and the disassembly of apical F-actin, respectively (Fig. 6). RIC3 promotes the disassembly of F-actin through Ca2+, which may activate profilin and possibly gelsolins. Thus our results demonstrate for the first time that a Rho GTPase controls actin dynamics through two downstream pathways respectively activating actin assembly and disassembly.
The check and balance action of RIC3 and RIC4 may also involve a temporal coordination and regulation of their respective targets, tip Ca2+ and F-actin. These two ROP signaling targets coordinately oscillate with the same period but in the different phases during the growth of pollen tubes (Holdaway-Clarke et al., 1997; Messerli et al., 2000; Fu et al., 2001). Thus, the ROP GTPase-orchestrated check and balance between the RIC3 and RIC4 pathways seems to underscore the spatiotemporal control of rapid (up to 1 cm/h) and oscillatory polarized growth in pollen tubes. The check and balance between the RIC3 and RIC4 pathways also entails the antagonistic effect of the RIC4 pathway on the RIC3 pathway. In tobacco and petunia pollen tubes, RIC3 OX causes depolarized growth due to excess Ca2+ accumulation at the tip (Fig. 4 F). The suppression of the RIC3 OX phenotype by RIC4 is actin dependent, as LatB abolishes this suppression (i.e., RIC3- and RIC4-coexpressing pollen tubes treated with LatB show the RIC3 OX phenotype; Table IV). Similarly, WT pollen tubes and root hairs treated with either LatB or cytochalasin D also show tip swelling (Gibbon et al., 1999; Ketelaar et al., 2003). How does the RIC4–actin pathway counteract the RIC3–Ca2+ pathway? RIC4-mediated actin assembly appears to inhibit the accumulation of Ca2+ at the tip, as suggested by lower [Ca2+]cyt in the tip of RIC4-overexpressing tubes (Fig. 4 F). This observation is further supported by the observation that F-actin inhibits a Ca2+-permeable PM channel in A. thaliana pollen (Wang et al., 2004). Consequently, RIC4-mediated tip F-actin suppresses RIC3 OX–induced Ca2+ overload in the tip.
Rho GTPase activation of two counteracting pathways: a paradigm for the regulation of actin dynamics, polar cell growth, and cell migration?
It is known that Rho GTPases regulate actin dynamics mainly through the inactivation of actin disassembly and/or promotion of actin assembly in various systems (Ridley, 2001; Etienne-Manneville and Hall, 2002; Pollard and Borisy, 2003). Our observation that ROP GTPase controls actin dynamics by activating both actin assembly and Ca2+-mediated actin disassembly raises an important question: do these dual and antagonistic roles of a Rho GTPase provide a general mechanism for its control of actin dynamics and other cellular processes in various eukaryotic cells? Many observations are consistent with this notion. ROP regulation of root hair tip growth also involves both tip-focused Ca2+ gradients and tip-localized dynamic F-actin (Molendijk et al., 2001; Jones et al., 2002). Similarly, fungal hyphal growth is also mediated by Rho GTPases as well as tip-focused Ca2+ gradients and F-actin (Heath, 2001).
Rho GTPase modulates neuritogenesis, neurite outgrowth and axon guidance at least partly through the regulation of actin dynamics at the leading edge (Dickson, 2001; Luo, 2002). These processes require cell migration but are also dependent on tip growth as in pollen tubes (Pfenninger et al., 2003). Both of these cellular processes involve Cdc42/Rac promotion of actin assembly through the activation of the Arp2/3 complex and Ca2+ transients at the leading edge. Ca2+ has been shown to promote actin depolymerization by regulating α-actinin in neuroblast cells (Fukushima et al., 2002). Other actin-binding proteins such as profilin and gelsolin could also be potential targets of Ca2+ signals. Importantly, Rho family GTPases promote both entry of extracellular Ca2+ and release of intracellular Ca2+ stores in several animal cells (Hong-Geller and Cerione, 2000; Mehta et al., 2003). Thus, Cdc42/Rac likely regulates actin dynamics through Ca2+-mediated actin disassembly in neurons and other animal cells in addition to their activation of actin assembly.
Cortical F-actin has a dual and antagonistic role in exocytosis in mammalian cells (i.e., cortical actin targets vesicle but also acts as a barrier to vesicle fusion; Gasman et al., 2004). Cdc42/Rac stimulates exocytosis by activating actin assembly (Gasman et al., 2004) and the IP3–Ca2+ pathway in animal cells (Hong-Geller and Cerione, 2000). In addition to stimulation of vesicle fusion, Cdc42/Rac-mediated Ca2+ signals may also promote actin dynamics through their role in actin disassembly. Therefore, Cdc42/Rac stimulation of exocytosis seems to be analogous to ROP GTPase activation of tip growth in pollen tubes in that Rho GTPase signaling coordinately controls two antagonistic downstream pathways: (1) the assembly of cortical F-actin that is necessary for vesicle targeting but blocks vesicle fusion, and (2) Ca2+ signaling that may promote both actin disassembly and vesicle fusion.
In conclusion, we have shown that ROP GTPase modulates apical growth and actin dynamics in pollen tubes through two downstream counteracting pathways. This dual and antagonistic role of Rho GTPases may provide a unifying mechanism for their control of various actin dynamic–dependent processes in various eukaryotic systems (e.g., cell migration, neuritogenesis, neuron growth and guidance, and exocytosis).
Materials And Methods
Plant materials and growth conditions
A. thaliana ecotypes Columbia and Wassilewskija WT were grown at 22°C in growth rooms with a light regime of 16 h of light and 8 h of dark. Nicotiana tabacum plants were grown in growth chambers at 22°C under a light regime of 12 h of darkness and 12 h of light. Petunia inflata plants were grown under greenhouse conditions.
DNA manipulation and plasmid construction
For DNA manipulation and plasmid construction, see online supplemental material.
Particle bombardment-mediated transient expression in tobacco and petunia pollen
Mature pollen grains collected from tobacco plants were used for transient expression using a particle bombardment procedure as described previously (Fu et al., 2001). For tobacco, routinely 0.5 mg of gold particles were coated with 0.5 μg (2 μg for petunia) of LAT52:GFP-RIC DNA or a mixture of 0.5 μg (2 μg for petunia) of LAT52:GFP-RICs DNA with 0.5 μg of other fusion constructs. The pollen grains were incubated for 5 h before observation under an inverted microscope (model TE300; Nikon) as described previously (Fu et al., 2001) or confocal microscope as described in the section Analyses of RIC localization and OX phenotype and F-actin visualization in tobacco pollen tubes.
For LatB treatments, various concentrations of LatB were included in germination medium using a 5-mM stock solution in DMSO. For LaCl3 and EGTA treatments, a stock solution of 100 and 500 mM was added to the germination medium to a final concentration of 100 μM and 1 mM, respectively.
FRET analysis was performed using a confocal microscope (model TCS SP2; Leica) equipped with He-Cd laser (used for CFP excitation at 442 nm) and Argon laser (which provides excitation at 514 nm for YFP). The spectral photometric detection system of Leica confocal device allows freely adjusting collection bandwidth for emission signal to minimize bleedthrough signals. Settings for these experiments were as follows: (a) CFP-ROP1 or CFP-DN-rop1: excitation 442 nm, emission 450–490 nm; (b) YFP-RIC3 or YFP-RIC4: excitation 514 nm, emission 525–600 nm; and (c) FRET: excitation 442 nm, emission 560–638 nm. CFP and YFP images were acquired simultaneously and FRET images were acquired separately at the same plane. For photobleaching experiment, a half-dome region at the tip of the pollen tube was selected for scans using the 442-nm laser line. YFP images before and after scans were collected. The ratio of average intensity was calculated. Only cells expressing moderate levels of CFP-ROP and YFP-RIC were studied; high- and low-expressing cells were avoided.
Analyses of RIC localization and OX phenotype and F-actin visualization in tobacco pollen tubes
Approximately 5 h after bombardment, tubes expressing LAT52:GFP-RIC were identified using epifluorescence microscopy, imaged, and analyzed as described previously (Fu et al., 2001). The degree of depolarized growth was determined by measuring the diameter of the widest region of the tube and the degree of polar growth was determined by measuring the length of pollen tubes. The subcellular localization of GFP-tagged RIC3 and RIC4 in tobacco pollen tubes was analyzed as described previously (Wu et al., 2001). Transient expression of GFP-mTalin was used to visualize F-actin in tobacco pollen tubes as described previously (Fu et al., 2001). For time-sequence analyses, mid-plane sections of the tip were scanned every 15 s. All confocal images were analyzed using the Metamorph v4.5 and processed using Adobe Photoshop v5.5.
Generation of RIC3 and RIC4 RNAi and OX lines and isolation of T-DNA insertional mutants
To screen for RIC4 T-DNA knockout line from Wisconsin's BASTA collection (Weigel et al., 2000), the gene-specific primer MQKSF and T-DNA left border primer JL-202 were used. To generate RIC3 OX and RIC4 OX constructs, the full-length RIC4 or RIC3 cDNA sequence was cloned into pC1300LAT52 vector derived from pCAMBIA1300 (CAMBIA). To generate the RIC3 RNAi construct, a 407-bp cDNA fragment was amplified using primer 5dsRIC3 and 3dsRIC3. The PCR fragment was ligated into pFGC5941 (ChromDB) to generate a sense construct pFGC5941RIC3sense. The antisense fragment was introduced into pFGC5941RIC3sense. The CaMV 35S promoter was replaced with LAT52 after EcoRI–NcoI. RIC4 RNAi construct was generated similarly using a 360-bp cDNA fragment amplified using primer 5dsRIC4 and 3dsRIC4. The aforementioned constructs were introduced into Agrobacterium tumefaciens GV3101 by electroporation and transformed into A. thaliana ecotype Columbia. T2 homozygous plants were selected for analysis.
A. thaliana pollen tube growth measurement
Flowers collected from A. thaliana plants 2 wk after bolting were used for the examination of pollen tube phenotypes. Pollen was germinated on standard agar medium containing 5 mM Ca2+ (with an equal molar ratio of CaCl2 and Ca[NO3]2) or various concentrations of Ca2+ as described previously (Li et al., 1999). Approximately 9–12 h after germination, images of pollen tubes were recorded through a cooled CCD camera (model C4742-95; Hamamatsu) attached on an Eclipse inverted microscope (model TE300; Nikon). The images were analyzed using the MetaMorph v4.5 measurement function. For each transgenic plant, ∼100 pollen tubes were chosen randomly for length measurement.
Cytoplasmic Ca2+ concentration measurements
Pollen tubes were microinjected with the fluorescent Ca2+ indicating dye Indo-1-dextran, and [Ca2+] was monitored by ratio imaging using a confocal microscope (model LSM 410; Carl Zeiss MicroImaging, Inc.) according to Bibikova et al. (1997).
Online supplemental material
Time-lapse videos supplement Figs. 2 and 4 (Videos 1 and 2) and Fig. 5 (Video 3). Videos play at four frames per second. Table S1 shows quantitative analysis of the phenotype of pollen tubes coexpressing ROP1 with various other ROPs. Table S2 lists all primers used in this work. Fig. S1 shows the effects of various ROPs on the localization pattern of GFP-RIC3 or GFP-RIC4. Fig. S2 shows the quantitative analysis of average apical fluorescence intensity in RIC4 OX tubes. Fig. S3 shows alteration of fine F-actin at the tip of pollen tubes in A. thaliana RIC4 mutants. Fig. S4 shows that pollen tubes of the ric4-1/RIC3(RNAi)-1 double mutant have normal calcium responses as WT pollen tubes.
We thank Elizabeth Lord for critical reading of this manuscript, the Salk Institute Genomic Analysis Laboratory for providing the sequence-indexed A. thaliana T-DNA insertion mutants, David Carter for assistance on FRET analysis, John Fowler for providing pBS-CFP vector, and Albrecht Arnim for providing pBS35S:YFP vector.
This work is supported by National Science Foundation grant MCB0111082 to Z. Yand, a Predoctoral Fellowship for Students with Disabilities from the National Institutes of Health to P.E. Dowd, and grants from the U.S. Department of Agriculture and from The Pennsylvania State University Innovative Biotechnology Program to S. Gilroy.
Abbreviations used in this paper: CRIB, Cdc42/Rac interactive binding; FRET, fluorescence resonance energy transfer; LatB, latrunculin B; LOF, loss-of-function; OX, overexpression; PM, plasma membrane; WT, wild-type.