Mechanism of CAP1-mediated apical actin polymerization in pollen tubes

Significance

Actin polymerization drives rapid polarized pollen tube growth, but the mechanism underlying actin polymerization within the growth domain of pollen tubes remains incompletely understood. We here identify CAP1 as a major player in driving actin polymerization via recharging ADP-G-actin and facilitating the formation and maintenance of a pool of polymerization-competent actin monomers in a manner that involves the nucleotide exchange activity of CAP1 in pollen tubes. Our study directly links the actin nucleotide exchange activity of CAP1 to its role in promoting actin polymerization. Our study thus significantly enhances our understanding of the mechanism of actin polymerization in pollen tubes.

Abstract

Srv2p/CAP1 is an essential regulator of actin turnover, but its exact function in regulating actin polymerization, particularly the contribution of its actin nucleotide exchange activity, remains incompletely understood. We found that, although Arabidopsis CAP1 is distributed uniformly in the cytoplasm, its loss of function has differential effects on the actin cytoskeleton within different regions of the pollen tube. Specifically, the F-actin level increases in the shank but decreases in the apical region of cap1 pollen tubes. The reduction in apical F-actin results mainly from impaired polymerization of membrane-originated actin within cap1 pollen tubes. The actin nucleotide exchange activity of CAP1 is involved in apical actin polymerization. CAP1 acts synergistically with pollen ADF and profilin to promote actin turnover in vitro, and it can overcome the inhibitory effects of ADF and synergize with profilin to promote actin nucleotide exchange. Consistent with its role as a shuttle molecule between ADF and profilin, the cytosolic concentration of CAP1 is much lower than that of ADF and profilin in pollen. Thus, CAP1 synergizes with ADF and profilin to drive actin turnover in pollen and promote apical actin polymerization in pollen tubes in a manner that involves its actin nucleotide exchange activity.

The actin cytoskeleton has been implicated in numerous fundamental physiological cellular processes (1). Most actin-based functions, if not all, are carried out by the polymeric form of actin (2). Therefore, a central question in this field is how actin monomers can rapidly assemble into actin filaments and organize into distinct structures to meet the demands of various physiological and cellular processes. A pool of polymerization-competent actin monomers must be available to support rapid and sustainable actin polymerization. ADF/cofilin is a central player in driving actin assembly and disassembly, and it preferentially binds to ADP-G-actin with high affinity and inhibits its nucleotide exchange (3, 4). Consequently, the dissociated actin monomers should be in the form of ADP-G-actin-ADF. Therefore, dissociation of G-actin from ADF and reloading of G-actin with ATP is the key to drive actin assembly.

CAP, also known as Srv2p in budding yeast, has been emerging as an important player in this process. It was originally identified as an adenylyl cyclase-associated protein (5, 6). It was proposed that Srv2p/CAP releases actin monomers from the ADF-ADP-G-actin complex by transferring actin monomers to profilin and subsequently delivering actin subunits to the barbed end of actin filaments (7). To date, there is overwhelming evidence supporting the role of Srv2p/CAP in promoting actin turnover both in vitro and in vivo via coordination with ADF/cofilin in different organisms (8–14). The synergistic action between Srv2p/CAP1 and ADF/cofilin enables the release of actin monomers from ADF, which allows us to speculate that CAP1 will be required for actin polymerization in cells. Indeed, it was shown that loss of function of CAS-1, a Caenorhabditis elegans cyclase-associated protein, impaired sarcomeric actin assembly in striated muscle (15), but the underlying molecular details remain to be uncovered. To date, direct evidence supporting the role of CAP1 in promoting actin polymerization is rather limited in different organisms, which is likely because loss of function of CAP1 causes dramatic defects in actin turnover, which result in increased actin bundling and F-actin levels in cells. To some extent, this explains why the actin turnover function of CAP1 has been studied much more extensively than its polymerization function. However, some aspects of CAP1 function in actin polymerization—in particular, the contribution of its actin nucleotide exchange activity to this process—remain largely unknown.

Pollen tube growth depends on a dynamic actin cytoskeleton (16–19). Earlier studies suggest that there exists a population of dynamic actin filaments, and their polymerization is crucial for rapid pollen tube growth (20, 21). Live-cell imaging of actin dynamics within the growth domain of pollen tubes revealed that actin filaments polymerize actively from the apical membrane, and this process is required for and concurrent with pollen tube growth (22). Based on the observations that the formins are required for apical actin polymerization (22–24) and that loss of function of profilins impairs apical actin polymerization (25), it was proposed that formin/profilin module-mediated actin polymerization is one of the major actin polymerization pathways within the growth domain of pollen tubes. Given that actin is predicted to be buffered by an equal amount of profilin in pollen (16, 26), continuous generation and maintenance of a pool of polymerization-competent ATP-G-actin-profilin complexes within the cytoplasm at the pollen tube tip is central to this hypothesis. In contrast to nonplant profilins, which enhance nucleotide exchange on actin (27–29), plant profilins inhibit actin nucleotide exchange (30, 31) or weakly enhance actin nucleotide exchange (12). In this regard, recharging ADP-G-actin, presumably complexed with ADF, before transferring it to profilin to form an ATP-G-actin-profilin complex, is crucial for actin polymerization. Arabidopsis CAP1 is the likely candidate to take on this role since it enhances nucleotide exchange on actin (12). Several CAP1 proteins from other organisms also harbor nucleotide exchange activities for actin (32–34). To date, however, the linkage between the activity of CAP1/Srv2p in stimulating actin nucleotide exchange and actin polymerization is rarely established.

By analyzing the function of CAP1 in regulating actin dynamics in pollen tubes, we here demonstrate unambiguously that CAP1 is required for the polymerization of actin filaments from the plasma membrane (PM) within the growth domain of pollen tubes. Our findings reveal that the reduction in the amount of apical actin filaments in cap1 pollen tubes results mainly from impaired de novo actin polymerization and demonstrate that the activity of CAP1 in stimulating actin nucleotide exchange is required for its function in driving actin polymerization. We thus propose that CAP1 acts in concert with ADF and profilin to drive actin turnover in pollen and consequently facilitates the formation of ATP-G-actin-profilin complexes to feed membrane-anchored formins to drive apical actin polymerization and normal pollen tube growth.

Results

Arabidopsis CAP1 Promotes ADF-Mediated Actin Depolymerization and Severing via Its N Terminus in Vitro.

To understand the mechanism by which CAP1 regulates actin assembly and disassembly in pollen, we examined its coordination with pollen ADF and profilin. Sequence alignment showed that Arabidopsis CAP1 harbors the conserved N-terminal residues which are crucial for the coordination of CAP1/Srv2p with ADF/cofilin (35) (SI Appendix, Fig. S1). Consistent with previous findings that Arabidopsis CAP1 enhances actin nucleotide exchange (12), it also contains the conserved residues that are crucial for the actin nucleotide exchange activity (36) (SI Appendix, Fig. S1). We therefore generated three recombinant proteins, CAP1, CAP1-90, and CAP1-109 (Fig. 1 A and B), among which CAP1-90 (35) and CAP1-109 (36) have mutations in the residues that are crucial for ADF/cofilin interaction and the actin nucleotide exchange activity, respectively. We then examined whether CAP1 and CAP1-90 act synergistically with pollen ADFs. Using high-speed F-actin cosedimentation assays, we found that CAP1 promotes pollen-specific ADF7- and ADF10-mediated actin depolymerization whereas CAP1-90 has reduced activity (Fig. 1 C–E). Taking ADF10 as the representative pollen ADF, we demonstrated by total internal reflection fluorescence microscopy that Arabidopsis CAP1 promotes ADF10-mediated filament fragmentation; again CAP1-90 has reduced activity (Fig. 1 F and G). Our data thus suggest that Arabidopsis CAP1 is able to enhance pollen ADF-mediated actin depolymerization and severing via a conserved mechanism.

Fig. 1.

Arabidopsis CAP1 enhances pollen ADF-mediated actin depolymerization and severing in vitro via its N terminus. (A) Schematic domain organization of CAP1 and the mutant constructs used in this study. CC, coiled coil domain; HFD, helically folded domain; P1 and P2, proline-rich region 1 and 2; WH2, WASP-homology 2 domain. The red and blue dots in the HFD domain and β-sheet indicate the amino acid residues that were mutated in CAP1-90 and CAP1-109, respectively. (B) Purified Arabidopsis CAP1 recombinant protein and its variants. (C) SDS/PAGE gels of the high-speed F-actin cosedimentation experiments. F-actin, 3 µM; ADF7, 20 µM; CAP1 or CAP1-90, 3 µM. S, supernatant; P, pellet. Gels in B and C were stained with Coomassie Brilliant Blue. (D and E) Quantification of the amount of actin in the supernatant from the high-speed F-actin cosedimentation experiments. F-actin, 3 μM; ADF7 or ADF10, 20 μM; CAP1 or CAP1-90, 3 μM. Data are presented as mean ± SD from three independent experiments. Statistical comparisons were performed using ANOVA with a post hoc Tukey test; different letters indicate significant differences, P < 0.01. (F) Time-lapse images of actin filaments assembled from Oregon-green (OG) actin. OG-actin, 1 µM, 50% labeled; ADF10, 0.3 µM; CAP1 or CAP1-90, 1 µM. The pH of the buffer is 7.0. Red scissors indicate filament fragmentation events. (Scale bar, 5 μm.) (G) Quantification of actin filament-severing frequencies. The data are presented as mean ± SEM. More than 40 actin filaments were counted from three independent experiments. Statistical comparisons were performed using ANOVA with a post hoc Tukey test; different letters indicate significant differences, P < 0.01. For ADF10 vs. ADF10+CAP1-90, P = 0.85.

Synergy Between CAP1 and ADF Promotes Actin Turnover in Vivo.

To reveal the function of CAP1 in pollen, we characterized two loss-of-function T-DNA insertion mutants of CAP1 (8) SI Appendix, Fig. S2 A and B). Consistent with previous results (8), we found that pollen germination and pollen tube growth were significantly inhibited in cap1 mutants compared with WT (SI Appendix, Fig. S2 C–E). In addition, we found that the morphology of pollen tubes derived from cap1 mutants was dramatically different from WT (SI Appendix, Fig. S2C), as evidenced by significant increases in the width of pollen tubes (SI Appendix, Fig. S2F). The developmental defects and pollen tube growth defects in Arabidopsis can be rescued by transformation of CAP1 or CAP1-EGFP under the control of its own promoter (SI Appendix, Fig. S2 G–K). Interestingly, we found that overexpression of CAP1 (SI Appendix, Fig. S2L) slightly but significantly inhibited pollen germination (SI Appendix, Fig. S2M) whereas it enhanced pollen tube growth (SI Appendix, Fig. S2N). Thus, these data together suggest that misexpression of CAP1 impairs pollen germination and alters polarized pollen tube growth.

To determine the effect of misexpression of CAP1 on the actin cytoskeleton in vivo, we initially examined the actin cytoskeleton in pollen grains. We found that the amount of actin filaments and the width of filamentous structures increased in cap1 pollen grains (SI Appendix, Fig. S3 A–C) but decreased in CAP1 overexpression pollen grains (SI Appendix, Fig. S3 D–F). These data suggested that CAP1 promotes actin turnover in vivo. Strikingly, we found that introduction of CAP1-90 into cap1 mutants (SI Appendix, Fig. S3G) failed to complement the increased amount and bundling of actin filaments in cap1 mutant pollen grains (SI Appendix, Fig. S3 H–J) and the pollen tube growth phenotype in cap1 mutant pollen tubes (SI Appendix, Fig. S3K). These data together suggest that CAP1 promotes actin turnover in a manner that requires its synergy with ADF in vivo.

Loss of Function of CAP1 Has Differential Effects on the F-Actin Level Within Different Regions of the Pollen Tube.

Next, we examined the effect of the misexpression of CAP1 on the actin cytoskeleton in pollen tubes. We initially found that actin filaments became brighter and more heavily bundled and that the amount of actin filaments increased in the shank region of cap1 pollen tubes (Fig. 2 A and B). Accordingly, we found that overexpression of CAP1 causes a reduction in the amount of F-actin throughout the entire pollen tube (SI Appendix, Fig. S4). Surprisingly, we found that actin filaments became dimmer and that the amount of actin filaments decreased within the apical region of cap1 pollen tubes compared with WT pollen tubes (Fig. 2 A and B). The results suggest that loss of function of CAP1 has differential effects on actin dynamics within different regions of the pollen tube.

Fig. 2.

CAP1 is distributed throughout the cytoplasm of pollen cells whereas loss of function of CAP1 has differential effects on the actin cytoskeleton within different regions of pollen tubes. (A) Micrographs of actin filaments stained with Alexa-Fluor-488–phalloidin in pollen tubes. The projection image or selected optical sections are shown. The dashed red lines indicate the apical region (0–10 μm from the tip). (Lower) Transverse sections at the indicated positions from the tip of pollen tubes (the distance from the tip is indicated on each image). (Right) The 3D distribution of fluorescence intensity of actin filaments within the apical region, which was generated by ImageJ software with a 3D interactive “Surface Plot” function. Warm and cold colors indicate higher and lower fluorescence, respectively. (Scale bar, 5 µm.) (B) Quantification of the fluorescence intensity of Alexa-Fluor-488–phalloidin staining from the tip to the base along the growth axis of pollen tubes. The dashed red line indicates the apical region as shown in A. Data are presented as mean ± SEM. More than 15 pollen tubes were measured. (C) Subcellular localization of CAP1-EGFP in a pollen grain and pollen tubes of different lengths. Pollen derived from cap1 plants expressing CAP1pro:CAP1-EGFP was visualized. The projection images and selected optical sections are shown. (Scale bar, 10 µm.) (D) Plot of the fluorescence intensity of CAP1-EGFP and EGFP (as control) in the pollen tube. More than 15 pollen tubes were measured. Data are presented as mean ± SEM.

CAP1 Is Uniformly Distributed in Pollen Tubes.

We next asked whether the differential effect of CAP1 on actin dynamics within different regions of the pollen tube results from the distinct intracellular localization of CAP1 in the pollen tube. The above data show that introduction of a CAP1-EGFP fusion construct driven by the native CAP1 promoter into cap1 mutants fully rescued the developmental defects and pollen tube growth defects in cap1-1 mutants (SI Appendix, Fig. S2 I and K), which suggests that CAP1-EGFP can act as a faithful probe to indicate the intracellular localization of CAP1. We found that CAP1 has a uniform distribution and does not form obvious filamentous structures in the cytoplasm of pollen grains and pollen tubes (Fig. 2 C and D). This is consistent with the fact that CAP1 acts as an actin monomer-binding protein. Thus, our data suggest that CAP1 is distributed uniformly in the cytoplasm of the pollen tube.

Actin Dynamics Is Reduced in cap1 Pollen Tubes and Increased in CAP1 Overexpression Pollen Tubes.

To further reveal the defect of actin dynamics in cap1 pollen tubes, we performed real-time visualization of the dynamics of actin filaments decorated with Lifeact-eGFP as described previously (37, 38). Similar to the findings shown above (Fig. 2A), we found that actin filaments became heavily bundled in the shank region of cap1 pollen tubes compared with WT pollen tubes (Fig. 3A). We found that the dynamics of actin filaments are reduced substantially (SI Appendix, Fig. S5A) and that the filaments are arranged into heavy but curved bundle structures in the shank region of the cap1 pollen tube (SI Appendix, Fig. S5 A and B). This was also supported by measurements showing that the filament elongation rate, filament shortening rate, and filament severing frequency are significantly reduced in the shank region of cap1 mutant pollen tubes compared with WT (SI Appendix, Fig. S5 C–E). Next, we visualized the dynamics of actin filaments within the apical region of the pollen tube. As reported previously, we found that actin filaments continuously polymerize from the apical membrane in WT pollen tubes (Fig. 3 A, Upper, and Movie S1). By comparison, apical membrane-originated actin polymerization was severely compromised in cap1 pollen tubes, leading to impaired formation of the apical actin structure (Fig. 3 A, Lower, and Movie S2). After tracking individual membrane-originated actin filaments (Fig. 3B), we found that, consistent with the findings shown for shank-localized actin filaments (SI Appendix, Fig. S5), their depolymerization rates and severing frequencies are reduced, which consequently leads to an increase in the maximal filament lifetime in cap1 pollen tubes compared with WT pollen tubes (Fig. 3C). Thus, we showed that cap1 pollen tubes exhibit a similar reduction in actin dynamics within both apical and shank regions, but the outcomes are different in the two regions in terms of the amount of actin filaments. This allows us to speculate that the decrease in the amount of apical actin filaments is due to the defect in de novo actin polymerization within the growth domain of cap1 pollen tubes. In support of this speculation, we found that the elongation rate of actin filaments is reduced in cap1 pollen tubes compared with WT (Fig. 3C).

Fig. 3.

Actin polymerization is impaired within the apical region of cap1 pollen tubes. (A) Time-lapse images of actin filaments in growing pollen tubes of WT and cap1. Actin filaments were decorated with Lifeact-EGFP. See the entire series in Movies S1 and S2. The red boxes indicate the apical region of the pollen tube, and red brackets mark the region with bright apical actin filaments originating from the apical membrane. (Right) Kymograph analysis of the Left images: the fluorescence intensity of Lifeact-EGFP indicates the amount of actin filaments in the pollen tube (from tip to shank region along the longitudinal axis) during the growth process. Red arrows indicate the bright apical F-actin polymerized from the membrane. (Scale bars, 10 µm.) (B) Time-lapse images of actin filaments within the apical region of pollen tubes. The far Left panels show the pseudotricolored overlayered images of WT and cap1 pollen tubes taken at intervals of 6 s. (Right) The time-lapse images of actin filaments within the red-boxed apical region (Left) of pollen tubes. Membrane-originated actin filaments are labeled with different colored dots. The origination sites of actin filaments are indicated by yellow circles. Actin filament elongation, shortening, and severing events are indicated by blue, purple, and green arrowheads, respectively. (Scale bars, 5 µm.) (C) Dynamic parameters of actin filaments in the apical region (as shown in the red-boxed region of B). Measurements were taken from more than 15 pollen tubes for each genotype. Numbers of actin filaments measured are indicated in parentheses. Values are presented as mean ± SD ***P < 0.001; ND, no significant difference (Student’s t test).

Accordingly, we found that the rate of actin elongation increased in CAP1 overexpression pollen tubes compared with WT (SI Appendix, Fig. S6 A and B). However, the amount of actin filaments polymerized from the apical membrane appears less overall in CAP1 overexpression pollen tubes than in WT (SI Appendix, Fig. S6A and Movies S3 and S4). We showed above that CAP1 promotes ADF-mediated actin depolymerization in vitro (Fig. 1 C–G) and enhances actin turnover in vivo (SI Appendix, Fig. S3 D–F). Therefore, we speculated that the reduction in the amount of apical membrane-originated actin filaments in CAP1 overexpression pollen tubes is likely due to the up-regulation in actin depolymerization. In support of this speculation, we found that depolymerization rates and severing frequencies of membrane-originated actin filaments increased, leading to the reduced maximal filament lifetime and filament length in CAP1 overexpression pollen tubes compared with WT (SI Appendix, Fig. S6B). Our data suggest that misexpression of CAP1 affects both actin polymerization and depolymerization, but the final outcome in terms of the amount of actin filaments within cells depends on the relative change in actin polymerization and depolymerization.

CAP1 Is an Abundant Cellular Protein in Pollen That Synergizes with Pollen ADF and Profilin to Promote Actin Turnover and Actin Nucleotide Exchange in Vitro.

To gain insights into the biochemical basis of the action of CAP1 in pollen cells, we sought to determine the cellular concentration of CAP1 as well as its molar ratio relationship with actin and two other proteins involved in actin turnover, ADF and profilin. Consistent with previous measurements (20, 39, 40), we found that the molar ratio of total actin to profilin is roughly 1:1 in Arabidopsis pollen (Fig. 4A). The cellular concentration of Arabidopsis CAP1 is about 1.2 μM in pollen, which corresponds to a molar ratio of roughly 1:25 to actin or profilin (Fig. 4A). This suggests that the actin monomer-sequestering activity is contributed mainly by profilin in pollen. In addition, the concentration of CAP1 is about one-fourth that of ADF (Fig. 4A), which is actually consistent with the notion that CAP1 acts as an intermediate player by transferring actin monomers from ADF to profilin in the actin turnover machinery as proposed previously (7). Indeed, we found that CAP1 substantially enhances actin turnover in the presence of pollen-specific ADF10 (41) and PRF5 (25) (Fig. 4B). In addition, we found that CAP1 is able to overcome the inhibitory effect of ADF7 on ADP-G-actin nucleotide exchange (Fig. 4C). Strikingly, we found that the role of CAP1 in promoting ADP-G-actin nucleotide exchange is much more potent than that of profilin, as 50 nM CAP1 substantially enhanced the rate of ADP-G-actin nucleotide exchange, whereas 4 μM PRF5 only slightly enhanced it (Fig. 4D). Our study, along with previous observations (12, 30, 31), suggests that the role of recharging ADP-G-actin will be taken over by CAP1 rather than profilin in vivo. Interestingly, we found that profilin and CAP1 synergistically enhance ADP-G-actin nucleotide exchange (Fig. 4D). The synergy between profilin and CAP1 in promoting ADP-G-actin nucleotide exchange is consistent with the fact that CAP1 retains the conserved PP1 and PP2 motifs (SI Appendix, Fig. S1), as the PP1 motif of Srv2p was previously demonstrated to interact with profilin (42). Therefore, our results showed that CAP1 is an abundant cytosolic protein that is distributed uniformly in pollen cells and acts synergistically with pollen ADF and profilin.

Fig. 4.

CAP1 is an abundant protein in pollen that synergizes with pollen ADF and profilin in promoting actin turnover and stimulating actin nucleotide exchange. (A) Quantification of cytosolic concentrations of Actin, CAP1, ADF, and profilin in mature pollen of Arabidopsis thaliana. Measurements were made as described in Materials and Methods. Protein abundance (mean ± SD) was measured from at least three experiments. Recombinant Arabidopsis ADF, PRF, CAP1, and ACT1 were used as the loading controls. (B) CAP1 efficiently promotes actin turnover in the presence of ADF and profilin. All reactions contain 4 µM ε-ADP-F-actin in Buffer G supplemented with 1×KMEI and 1 μM ATP. ADF10, 4 µM; PRF5, 4 µM; CAP1, 1 µM. (C) CAP1 overcomes the inhibitory effect of ADF7 on actin nucleotide exchange. ADP-actin, 2 μM. The concentrations of actin-binding proteins are indicated. (D) CAP1 synergizes with PRF5 to enhance actin nucleotide exchange. ADP-actin, 2 μM. The concentrations of actin-binding proteins are indicated.

Both CAP1-109 and CAP1-90 Fail to Rescue the F-Actin Reduction Phenotype at cap1 Pollen Tube Tips.

We next introduced CAP1-109 (Fig. 1A) into cap1 mutants, as CAP1-109 was demonstrated to lack actin nucleotide exchange activity (36). We initially demonstrated that CAP1-109 retains the capability to enhance ADF-mediated actin depolymerization in vitro and promote actin turnover in pollen grains (SI Appendix, Fig. S7 A–F). We found that CAP1-109 retains the G-actin–binding activity, as evidenced by data showing that CAP1-109 inhibits spontaneous actin polymerization and binds to ADP-G-actin as effectively as CAP1 and CAP1-90 (SI Appendix, Fig. S8 A–C). However, we found that the ability of CAP1-109 to promote actin nucleotide exchange is substantially compromised when it is compared with CAP1 and CAP1-90 (SI Appendix, Fig. S8D). Consistent with this, we found that introduction of both CAP1 and CAP1-109 reduced the amount of actin filaments and bundling in the shank of cap1 pollen tubes (Fig. 5 A and B), which suggests that they retain the capability to enhance actin turnover in vivo. However, CAP1-109 cannot complement the reduction in the amount of apical actin filaments in cap1 pollen tubes (Fig. 5 A–C) or the pollen tube growth defect (SI Appendix, Fig. S7G). Further observations revealed that apical actin polymerization is impaired in CAP1-109 pollen tubes compared with WT pollen tubes (Fig. 5D and Movies S5 and S6), as the elongation rate of membrane-originated actin filaments is significantly reduced in CAP1-109 pollen tubes (Fig. 5E). Surprisingly, the depolymerization rate of apical actin filaments is reduced significantly in CAP1-109 pollen tubes compared with WT pollen tubes. This reduced depolymerization rate leads to an increase in the maximal filament length and lifetime (Fig. 5E). We do not currently know the reason for this. It might be due to a feedback regulatory mechanism to counteract the reduction in the amount of apical actin filaments. Nonetheless, our study suggests that the activity of CAP1 in stimulating actin nucleotide exchange is crucial for its role in regulating apical actin polymerization.

Fig. 5.

CAP1-109 cannot fully rescue the defective actin polymerization within the apical region of cap1 pollen tubes. (A) Micrographs of actin filaments stained with Alexa-Fluor-488–phalloidin in pollen tubes of WT and CAP1-109 transgenic lines. CAP1-109 is CAP1pro:CAP1-109;cap1-1. The Z-projection and selected optical sections are shown. (Scale bar, 5 µm.) The red boxes indicate the apical region (0–10 μm from the tip). (Lower) The distribution of fluorescence intensity of actin filaments in transverse sections at the indicated distances from the tip of pollen tubes. Pseudocolored transverse images are also shown. Warm and cold colors indicate higher and lower fluorescence, respectively. (Scale bar, 5 µm.) (B) Quantification of the fluorescence intensity of actin filaments from the extreme tip to the base along the growth axis of pollen tubes in WT, cap1-1, COM-1, and CAP1-109 lines. The dashed red line indicates the base of the red-boxed regions shown in the Upper of A. More than 20 pollen tubes were measured for each line. Data are presented as mean ± SEM. (C) Quantification of the mean fluorescence intensity of actin filaments within the red-boxed region of pollen tubes. More than 20 pollen tubes were measured for each genotype. Data are presented as mean ± SEM. Statistical comparisons were performed using ANOVA with a post hoc Tukey test; different letters indicate significant differences, P < 0.01. WT vs. COM-1, P = 0.99; WT or COM-1 vs. cap1-1, P < 0.0001; 1# vs. 2#, P = 0.98; 1# vs. cap1-1 P = 0.021; 2# vs. cap1-1, P = 0.03; WT vs. 1#, P = 0.032; WT vs. 2#, P = 0.02; COM-1 vs. 1#, P = 0.014; COM-1 vs. 2#, P = 0.009. (D) Time-lapse images of actin filaments in growing pollen tubes of WT and CAP1-109 plants. Actin filaments were revealed by decoration with Lifeact-EGFP. Red brackets indicate the apical actin filaments in the pollen tube. See the entire series in Movies S5 and S6. (Right) Pseudocolored kymograph analysis of apical actin filaments during pollen tube growth. Warm and cold colors indicate high and low fluorescence intensity, respectively. The fluorescence intensity of Lifeact-EGFP indicates the amount of actin filaments in the pollen tube (from tip to shank region along the longitudinal axis) during the growth process. The enlarged kymograph images show the pollen tube tip region (0–10 μm) in the Upper. (Scale bars, 10 µm.) (E) Dynamic parameters of actin filaments in the apical region of pollen tubes. Measurements were taken from more than 15 pollen tubes for each genotype. The numbers of actin filaments analyzed are shown in parentheses. Values are presented as mean ± SD ***P < 0.001; *P < 0.05; ND, no significant difference (Student’s t test).

To determine whether CAP1 mainly promotes the conversion of ADP-G-actin bound with ADF into ATP-G-actin to facilitate membrane-originated actin polymerization, we examined the actin cytoskeleton in pollen tubes derived from CAP1-90 plants. We found that apical actin polymerization is impaired in the CAP1-90 pollen tube, and this causes a reduction in the amount of apical actin filaments compared with WT (Fig. 6 A and B and Movies S7 and S8). This reduction was directly visualized in transverse sections derived from the apical region of the pollen tube (Fig. 6C). Surprisingly, we found that the amount of actin filaments increased in transverse sections derived from the shank region of the CAP1-90 pollen tube compared with WT (Fig. 6C). Notably, heavy actin cables are more prominent in the shank of the CAP1-90 pollen tube compared with WT (Fig. 6C). This is consistent with the above data showing the reduced rate of actin turnover in CAP1-90 pollen grains (SI Appendix, Fig. S3 H–J). Indeed, we found that actin depolymerization rate and severing frequency of membrane-originated actin filaments are reduced and that their lifetime is consequently increased in the CAP1-90 pollen tube compared with WT (Fig. 6D). These data suggest that the reduction in the amount of apical actin filaments in CAP1-90 pollen tubes results from the impaired actin polymerization. In support of this notion, we found that the elongation rate of membrane-originated actin filaments is reduced significantly in the CAP1-90 pollen tubes compared with WT (Fig. 6D). These data suggest that the coordination of CAP1 with ADF is crucial for CAP1 in promoting apical membrane-originated actin polymerization in pollen tubes, which implies that CAP1 mainly recharges ADP-G-actin bound with ADF in cells.

Fig. 6.

CAP1-90 cannot fully rescue the defective membrane-originated actin polymerization within the apical region of cap1 pollen tubes. (A) Time-lapse images of actin filaments in growing pollen tubes of WT and CAP1-90 plants. Actin filaments were revealed by decoration with Lifeact-EGFP. Red brackets indicate the apical actin filaments in the pollen tube. See the entire series in Movies S7 and S8. (Scale bar, 10 µm.) (B) Pseudocolored kymograph analysis of apical actin filaments during pollen tube growth. The fluorescence intensity of Lifeact-EGFP indicates the amount of actin filaments in the pollen tube (from the tip to shank region along the longitudinal axis) during the growth process. The green box indicates the red-bracketed region shown in A. Warm and cold colors indicate high and low fluorescence intensity, respectively. (Scale bar, 10 µm.) (C) Projection images showing the organization of actin filaments in WT and CAP1-90 pollen tubes. Pseudocolored images are shown at the Right. The (Lower) The distribution of fluorescence intensity of actin filaments in pseudocolored transverse sections at the indicated distances from the tip of pollen tubes. Warm and cold colors indicate higher and lower fluorescence, respectively. White arrowheads indicate heavy actin cables. (Scale bar, 10 µm.) (D) Dynamic parameters of actin filaments in the apical region of pollen tubes. Measurements were taken from more than 15 pollen tubes for each genotype. The numbers of actin filaments analyzed are shown in parentheses. Values are presented as mean ± SD ***P < 0.001; *P < 0.05; ND, no significant difference (Student’s t test).

Discussion

We initiated the functional characterization of Arabidopsis CAP1 in pollen with the aim of understanding the regulation of actin polymerization within the growth domain of pollen tubes. We found that actin is almost equimolar to profilin in pollen (Fig. 4A), consistent with previous measurements (20, 39, 40). These previous findings led to the conclusion that actin monomers exist mainly in the form of actin-profilin complexes (16, 26, 43). In addition, it was previously shown that the formin/profilin module plays an essential role in driving actin polymerization from the membrane within the growth domain of the pollen tube (22–25). This urges us to ask how a pool of polymerization-competent ATP-actin-profilin complexes is generated and maintained. In contrast to the situation in nonplant systems, where profilins can enhance actin nucleotide exchange (27–29), plant profilins lack or inhibit actin nucleotide exchange (30, 31). In this regard, recharging of ADP-G-actin before the formation of ATP-G-actin-profilin complexes should be critical. ADF is a very abundant protein in pollen (Fig. 4A), and pollen ADFs prefer ADP-G-actin, inhibit actin nucleotide exchange, and play an essential role in regulating actin turnover in pollen (41, 44). Dissociation of ADP-G-actin from ADF is a prerequisite for the reentry of actin into the actin polymerization cycle. CAP1 will be an extremely relevant player in this process as it synergizes with ADF and profilin to promote actin turnover as well as actin nucleotide exchange activity (12). However, the role of CAP1 in regulating actin polymerization, particularly the contribution of its actin nucleotide exchange activity, remains largely unknown.

By examining the effect of loss of function of CAP1 on the actin cytoskeleton in pollen tubes, we found that CAP1 has differential effects on the actin cytoskeleton within different regions of the pollen tube (Figs. 2A and 3A). Loss of function of CAP1 severely reduces the F-actin level within the growth domain but increases the F-actin level in the shank region (Figs. 2 A and B and 3A). The reduced F-actin level in the apical region of cap1 pollen tubes is due to impaired actin polymerization, as severing and depolymerization of apical actin filaments are reduced in cap1 pollen tubes compared with WT pollen tubes (Fig. 3C). This suggests that the formation of the apical actin structure is highly dependent on actin polymerization, as demonstrated recently in growing pollen tubes (22). The differential effect of the loss of function of the diffusely distributed CAP1 (Fig. 2C) on the actin cytoskeleton within different regions of the pollen tube is very likely due to the fact that formins efficiently nucleate actin assembly from the PM within apical and subapical regions (24). In addition, the differential effect of CAP1 loss of function on the actin cytoskeleton might also be due to the differential collaboration of CAP1 with ADFs within different regions of the pollen tube, as ADFs densely localize to the shank-localized actin cables but distribute comparatively diffusely within the cytoplasm of the pollen tube tip (41, 44). Furthermore, we directly link the actin nucleotide exchange activity of CAP1 to its function in promoting actin polymerization, as we found that the mutant protein CAP1-109, which binds G-actin and promotes actin turnover but has reduced actin nucleotide exchange activity (SI Appendix, Fig. S8), cannot fully rescue actin polymerization defects within the growth domain of cap1 pollen tubes (Fig. 5). Interestingly, we found that CAP1-90 fails to rescue the apical actin polymerization defect in pollen tubes (Fig. 6), which suggests that the conversion of ADP-G-actin bound with ADF is the rate-limiting step for the actin polymerization and depolymerization cycle in cells. These data together suggest that CAP1 acts to transfer actin monomers from ADF to profilin. It was reported that overexpression of profilin can suppress the defects caused by loss of the C-terminal domain of CAP (45), which suggests that the actin monomer-sequestering function of CAP might be critical in vivo. The actin monomer-sequestering function was previously proposed for plant CAP1 (46). It was shown that the binding affinity of CAP1 to plant actin is higher than that of ZmPRO5 (Zea mays profilin5), although both have roughly similar binding affinity to muscle actin (12). Similar results were obtained in our research for the binding of CAP1 to muscle actin (SI Appendix, Fig. S8C). Our finding that the cellular concentration of CAP1 is much lower than that of actin and profilin (Fig. 4A) suggests that the actin monomer sequestering function is achieved mainly by profilin rather than CAP1 in plants. In this regard, actin mainly forms complexes with profilin rather than CAP1 in plant cells. However, considering that plant profilins inhibit actin nucleotide exchange (30, 31), the role in promoting actin nucleotide exchange will be taken over by CAP1 in vivo, as we demonstrate in this study. Even in other systems where both profilin and CAP1 harbor actin nucleotide exchange activity, CAP1 appears to be better suited for this function since it binds to the substrate (ADP-G-actin) with much higher affinity than profilin (36, 47). In line with our findings, a recent report showed that deletion of four amino acids of CAP specifically affects its nucleotide exchange activity and impairs its role in promoting actin cable formation (34). Therefore, our results provide unambiguous evidence that CAP1 promotes apical actin polymerization in pollen tubes, and this function requires the ability of CAP1 to stimulate actin nucleotide exchange.

Previous characterizations support the function of Arabidopsis CAP1 in promoting actin turnover (8, 46). In particular, loss of function of CAP1 increases the amount and bundling of actin filaments in Arabidopsis root hairs (8). A similar phenomenon was noted in the shank region of the pollen tube and pollen grains (Figs. 2A and 3A and SI Appendix, Fig. S5). In support of the role of CAP1 in promoting actin turnover in vivo, it was previously shown that Arabidopsis CAP1 inhibits spontaneous actin polymerization and is capable of synergizing with ADF to promote actin depolymerization (12). However, the precise mechanism underlying the action of CAP1 is not well understood. We here demonstrate that Arabidopsis CAP1 enhances ADF-mediated actin severing and depolymerization via its N terminus (Fig. 1 C–G), which suggests that Arabidopsis CAP1 promotes actin turnover using the same mechanism as reported for CAP protein from other organisms (14). We previously demonstrated that ADF is an essential regulator of actin turnover in pollen (41, 44). Accordingly, we found here that ADF is very abundant, as its cytosolic concentration reaches about 5 μM in pollen (Fig. 4A). Although the molar ratio of ADF to actin is roughly 1:6 in Arabidopsis pollen (Fig. 4A), the percentage of F-actin is less than 10% of total actin in pollen (20, 39). Therefore, we speculate that ADF is sufficient to saturate F-actin in pollen cells. In partial support of this speculation, we found that both ADF7 and ADF10 form prominent filamentous actin structures in pollen cells (41, 44). Strikingly, we found that disruption of the interaction between the N terminus of CAP1 and ADF impairs its function in promoting actin turnover in pollen cells (Fig. 6D and SI Appendix, Fig. S3 H–J). This suggests that the role of CAP1 in driving actin turnover in vivo is exclusively through its interaction with ADF.

Based on the in vitro and in vivo data presented here, we propose that CAP1 synergizes with pollen ADF to enhance actin monomer dissociation and filament severing, which subsequently releases ADP-G-actin from ADF. After stimulating the nucleotide exchange of ADP-G-actin to convert it into ATP-G-actin, CAP1 subsequently transfers ATP-G-actin to profilin to form ATP-G-actin-profilin complexes to maintain a pool of polymerization-competent actin monomers. This will be utilized by membrane-localized formins to support actin polymerization from the PM within the growth domain of pollen tubes (Fig. 7). We thus identify CAP1 as a major player in driving actin polymerization via recharging ADP-G-actin and facilitating the formation and maintenance of a pool of polymerization-competent actin monomers. Our study significantly enhances our understanding of the mechanism of actin polymerization in pollen tubes as well as the mechanism of action of Srv2p/CAP1 in regulating actin assembly and disassembly in general.

Fig. 7.

Schematic model illustrating the action of CAP1 in promoting apical actin polymerization in the pollen tube. This simple schematic model shows CAP1 as an important player in coordinating with ADF and profilin to promote the turnover of membrane-originated actin filaments within the growth domain of the pollen tube. The diagram on the Left reflects the molar ratio of CAP1 with actin and several other actin associated proteins and their intracellular localization patterns, as shown by refs. 25, 41, 44, 51. (Right) An enlarged picture of the boxed apical region. The relative amount of CAP1 is increased to highlight the function of CAP1. Briefly, within the cytoplasm of the pollen tube, CAP1 coordinates via its N terminus with ADF to promote actin turnover. After releasing ADP-G-actin from ADF, CAP1 enhances actin nucleotide exchange via its C terminus to generate ATP-G-actin and subsequently transfers ATP-G-actin to profilin to form ATP-G-actin–profilin complexes. In this regard, CAP1 plays an important role in maintaining the pool of polymerization-competent actin monomers, which can be utilized by the membrane-anchored formins within the apical and subapical regions of pollen tubes as we demonstrated recently (24).

Materials and Methods

Plant Materials and Growth Conditions.

The information about two CAP1 T-DNA insertion lines, cap1-1 (Salk_112802) and cap1-2 (GK_453G08), had been described previously (8). Arabidopsis CAP1 overexpression and complementation plants were generated as described below. Arabidopsis Columbia-0 ecotype was used as WT in this study. Arabidopsis plants were grown in a culture room at 22 °C under a 16-h-light/8-h-dark photoperiod.

Complementation of cap1 Mutants and Visualization of the Intracellular Localization of CAP1.

To generate the CAP1 complementation construct, in which expression of CAP1 was driven by the CAP1 promoter, the genomic sequence of CAP1 was amplified with primers pgCAP1_For and pgCAP1_Rev+TAA (SI Appendix, Table S1) using Arabidopsis genomic DNA as the template and subsequently moved into pCAMBIA1301-NOS to generate pCAMBIA1301-CAP1pro:CAP1g-NOS plasmid. To determine the function of the interaction between CAP1 and ADF and the nucleotide exchange activity of CAP1 in vivo, we generated pCAMBIA1301-CAP1pro:CAP1-90g-NOS and pCAMBIA1301-CAP1pro:CAP1-109g-NOS plasmids using pCAMBIA1301-CAP1pro:CAP1g-NOS plasmid as the template with primer pairs pgCAP1-90_For/pgCAP1-90_Rev and CAP1-109_For/CAP1-109_Rev (SI Appendix, Table S1), respectively. They were transformed into cap1-1 to generate transgenic plants pCAMBIA1301-CAP1pro:CAP1g-NOS;cap1-1, pCAMBIA1301-CAP1pro:CAP1-90g-NOS; cap1-1, and pCAMBIA1301-CAP1pro:CAP1-109g-NOS; cap1-1, respectively. To generate the CAP1-EGFP fusion construct driven by the CAP1 promoter, the genomic sequence of CAP1 was initially amplified with primers pgCAP1_For and pgCAP1_Rev-TAA (SI Appendix, Table S1). EGFP was amplified with primers EGFP_For and EGFP_Rev (SI Appendix, Table S1). Both fragments were digested and subsequently moved into pCAMBIA1301-NOS to generate the final pCAMBIA1301-CAP1pro:CAP1g-EGFP-NOS plasmid. pCAMBIA1301-CAP1pro:CAP1g-EGFP-NOS construct was transformed into cap1-1 to generate the transgenic plant pCAMBIA1301-CAP1pro:CAP1g-EGFP-NOS;cap1. To visualize the intracellular localization of CAP1, pollen derived from pCAMBIA1301-CAP1pro:CAP1g-EGFP-NOS;cap1 plants was visualized under a fluorescence light microscope equipped with a 100× oil objective (1.46 numerical aperture HC PLAN), and images were captured by confocal laser scanning microscopy excited with the 488-nm line of an argon laser. Optical sections were collected with a step size of 0.5 μm for both pollen grains and pollen tubes.

F-Actin Staining and Quantification in Fixed Pollen Grains and Pollen Tubes.

Actin filaments in pollen grains and pollen tubes were stained with Alexa-488 phalloidin as previously described (48, 49). To reveal actin filaments in pollen grains and pollen tubes, pollen grains were subjected to staining with Alexa488 phalloidin after culturing for 10 min and 2 h at 28 °C. Briefly, 300 µM 3-maleimidobenzoic acid N-hydroxysuccinimide ester in liquid germination medium (GM) was added onto the surface of solid GM. After incubation for 1 h, the GM plate was washed three times with TBSS buffer [50 mM Tris⋅HCl, 200 mM NaCl, 400 mM sucrose and 0.05% (vol/vol) Nonidet P-40, pH 7.5] in GM. Finally, pollen grains or pollen tubes were incubated with 150 nM Alexa-488 phalloidin in TBSS buffer overnight at 4 °C. The samples were observed under a laser scanning confocal microscope (Olympus FV1000MPE) equipped with a 100× oil objective (numerical aperture of 1.4). The sample was excited with the 488-nm line of an argon laser, the emission was set in a range of 505–525 nm, and the Z-series images were collected with the Z-step set at 0.5 µm. Maximum intensity projection of Z-series and the optical section of both pollen grains and pollen tubes are displayed in the figures. The amount of actin filaments was quantified by determining the fluorescence intensity of Alexa-Fluor-488–phalloidin staining with ImageJ software.

Direct Visualization and Quantification of Actin Filament Dynamics in Living Pollen Tubes.

Actin filaments in pollen tubes were revealed by decoration with Lifeact-EGFP as described previously (37, 38). To reveal actin filaments in pollen tubes of cap1, CAP1 overexpression, and CAP1-90 and CAP1-109 plants, the marker Lifeact-EGFP was introduced by crossing them with WT Arabidopsis plants harboring Lat52:Lifeact-EGFP. After segregation, T3 homozygous plants harboring Lat52:Lifeact-EGFP were used for the analysis of actin filament dynamics in pollen tubes, and pollen tubes derived from sibling Arabidopsis plants harboring Lat52:Lifeact-EGFP were used as the control for comparison. Actin filaments were observed under a spinning disk confocal microscope equipped with a Yokogawa CSUX1 scanning head. The time-lapse Z-series images were acquired with an Andor iXon3 DU888 EMCCD camera at 2-s intervals, and the Z-step size was set at 0.7 μm. To reveal the overall dynamics of apical actin filaments, three consecutive images, pseudocolored differently, were merged as described previously (50). A kymograph along the growing direction at the center of the growing pollen tube was generated to analyze the intensity of F-actin. First, a 5-μm wide band was drawn along the growing axis of a pollen tube in the time-lapse stacks with the built-in tool “segmented line” in ImageJ. Then the plot profiles were analyzed with the macro “Stack profile data,” and the results were further imported into ImageJ to produce the kymograph. Each line along the growth axis represents the average plot profile along the 5-μm band in the pollen tube at one time point. To clearly observe the apical actin filaments, a grayscale kymograph image was applied to the color lookup tables in the submenu of ImageJ. The kymograph images of the pollen tube tip were displayed after pseudocolor processing. Warm and cold colors indicate high and low fluorescence intensity, respectively. The dynamic parameters of individual actin filaments were analyzed as described previously (38). Two additional parameters, bundling and debundling frequencies, were measured to reveal actin dynamics within the shank region of pollen tubes as described previously (44).

Generation of CAP1 overexpression transgenic plants, qRT-PCR analysis, observation and quantification of pollen germination and pollen tube growth, sequence alignment, protein production, production of polyclonal antisera, determination of cytosolic protein concentration by quantitative Western blotting assay, in vitro actin assembly and disassembly assays, actin monomer-binding assay, and nucleotide exchange assay are described in SI Appendix, Materials and Methods.