Regulation of the Pollen-Specific Actin-Depolymerizing Factor LlADF1

Pollen ADFs Have Conserved Sequences

Previously, four pollen-specific ADFs were identified, two from maize (ZmADF1 and ZmADF2), one from lily (LlADF1), and one from Brassica napus (BnADF) (Kim et al., 1993; Chung et al., 1995; Rozycka et al., 1995; Lopez et al., 1996; Smertenko et al., 2001). In previous articles, we have shown that these four pollen ADFs show a greater degree of identity with each other than with the maize vegetative tissue ADF, ZmADF3, suggesting a common role for an evolutionarily conserved ADF in pollen development and function (Maciver and Hussey, 2002). Here, we show that polyclonal antiserum raised against maize or lily pollen ADFs identifies only the pollen ADFs on protein gel blots (Figures 1A and 1B , respectively), whereas antiserum raised against ZmADF3 identifies ADF in all maize tissues examined with the exception of pollen. These results demonstrate structural distinctions between the pollen ADFs and the maize and lily vegetative ADFs as well as differential tissue localization. Moreover, in a phylogenetic analysis of all known plant ADFs (Maciver and Hussey, 2002), the four pollen ADFs fall in one clade, and no ADFs that are known to be expressed in vegetative tissues (i.e., ZmADF3, AtADF5, AtADF6, and PeADF) fall within this group (Mun et al., 2000; Dong et al., 2001a; Maciver and Hussey, 2002).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Protein Gel Blots of Plant Extracts.

(A) Protein gel blots of various maize total protein extracts probed with anti-ZmADF3 (vegetative ADF) or anti-ZmADF1 (pollen ADF). BMS, Black Mexican Sweetcorn culture cells.

(B) Coomassie blue–stained gel of various lily total protein extracts and corresponding protein gel blot probed with anti-LlADF1 (pollen ADF).

LlADF1 Is Not Regulated by Phosphorylation

To determine the biochemical characteristics of a pollen ADF, we used the lily pollen ADF, LlADF1, because it produces stable, soluble recombinant protein when expressed in, and purified from, bacteria (Smertenko et al., 2001). We first assessed whether LlADF1 could be phosphorylated by plant kinases and therefore whether LlADF1 was likely to be controlled by this mechanism. LlADF1 has the conserved Ser at position 6, which has been shown to be phosphorylated in ZmADF3. The CDPK known to phosphorylate ZmADF3 (Allwood et al., 2001) does not phosphorylate LlADF1 in in vitro phosphorylation assays, as shown in Figure 2A . To determine whether another kinase could phosphorylate this pollen ADF, we performed in-gel phosphorylation assays. In these assays, maize pollen, maize leaf, lily pollen, and lily leaf extracts all were shown to contain at least one protein kinase that was capable of phosphorylating ZmADF3, but neither homologous nor heterologous extracts contained a kinase that was capable of phosphorylating LlADF1 (Figure 2B). We conclude that lily pollen ADF is not controlled by reversible phosphorylation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

LlADF1 Is Not Phosphorylated.

(A) In vitro phosphorylation assays of recombinant ADF by CDPK. Autoradiograph and corresponding colloidal silver–stained gel of assays using ZmADF3 or LlADF1 in the presence of CDPK are shown.

(B) In-gel phosphorylation assays using either ZmADF3 or LlADF1 incorporated into the gel. Colloidal silver–stained blots showing equal loading of the various extracts onto the gel and corresponding autoradiographs are shown.

LlADF1 Is Regulated by pH

The activity of plant ADFs is affected by pH (Carlier et al., 1997; Gungabissoon et al., 1998). To determine whether this is the case for the lily pollen ADF as well, pure recombinant LlADF1 was incubated for 20 min at room temperature with F-actin over a range of pH values and centrifuged at 365,000g to pellet the actin. The amount of ADF that remained bound to F-actin and therefore was found in the pellet was determined by SDS-PAGE and densitometry and plotted graphically (Figure 3) . As with the ZmADF3 control, LlADF1 was more able to depolymerize F-actin at higher pH, releasing monomeric actin and ADF itself into the supernatant. At lower pH, a greater proportion of the LlADF1 remained in the pellet bound to the F-actin, indicating that at these pH values, the actin-depolymerizing activity of LlADF1 was low (Figure 3), although the effect was not as great as with ZmADF3. These data show that LlADF1 has pH-sensitive activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Effect of pH on LlADF1 Binding to F-Actin.

F-actin and LlADF1 (diamonds) or ZmADF3 (squares) cosedimentation assays were performed in the pH range 6.0 to 9.0. The percentage of ADF bound to F-actin and found in the pellet is plotted.

LlADF1 Is Inhibited by Phospholipids

Members of the ADF/cofilin group, including ADFs from plants, are known to bind specific phospholipids, and this interaction inhibits F-actin binding. Moreover, the presence of PIP₂ in membranes at the tips of pollen tubes (Dong et al., 2001b) is consistent with a regulatory role of the phosphoinositide signal transduction pathway in ADF activity and hence actin dynamics. To test for specific phosphoinositide binding, LlADF1 was preincubated with a range of phospholipids before being run on a nondenaturing gel (Figure 4A) . In the presence of phospholipids that are not able to bind, LlADF1 ran as a single distinct band similar to the control, but preincubation with PIP or PIP₂ caused it to form a smear on the gel, indicating binding between the phospholipid and ADF. These results are similar to those observed for ZmADF3. In addition, LlADF1 inhibited the activity of phospholipase C (Figure 4B), confirming that LlADF1 is able to sequester PIP₂ in vitro. Incubation of LlADF1 with either phosphatidylcholine or varying amounts of PIP₂ before performing a cosedimentation assay showed that PIP₂ inhibited the binding of LlADF1 to F-actin and that this inhibition increased between a 0- and 50-fold molar excess of PIP₂ over LlADF1. By contrast, a 50-fold molar excess of phosphatidylcholine had no effect on the binding of LlADF1 to F-actin (Figure 4C). These data indicate that LlADF1 activity is inhibited by PIP₂.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Regulation of Pollen ADF by Phospholipids.

(A) Native gels of either LlADF1 or ZmADF3 mixed with a 50-fold molar excess of purified phospholipids. (−)PL, control without phospholipid; PI, phosphatidylinositol; PIP, phosphatidylinositol 4-monophosphate; PIP₂, phosphatidylinositol 4,5-bisphosphate; PE, phosphatidylethanolamine; PC, phosphatidylcholine; PS, phosphatidylserine; IP₃, inositol 1,4,5-triphosphate; OAG 1-oleoyl-2-acetylglycerol.

(B) Effect of LlADF1 on plant phospholipase C activity. The activity of plant phospholipase C was measured in the presence of increasing concentrations (0 to 12.5 μM) of LlADF1 (diamonds) or ZmADF3 (squares).

(C) Inhibition of LlADF1 binding to F-actin by PIP₂. F-actin and LlADF1 cosedimentation assays were performed in the presence of a 50-fold molar excess of phosphatidylcholine (PC) as a control and a 50- to 0-fold molar excess of PIP₂. The percentage of LlADF1 remaining in the supernatant is plotted. The data shown are means of three replicate experiments, and the bars represent standard errors.

Identification of the Plant Homolog of Yeast AIP1p

So far, we have established that although the lily pollen ADF does not appear to be regulated by phosphorylation, it can be regulated by pH and may be inhibited by phospholipids. However, neither of these methods of regulation can explain how the relatively low levels of pollen ADF activity observed in vitro could account for the high actin turnover that would be required to extend a pollen tube if, as postulated for root hairs, active ADF is essential for tip growth (Jiang et al., 1997; Smertenko et al., 2001). Therefore, it seems likely that another method of regulating the pollen ADF occurs in vivo. One regulatory protein that enhances ADF activity is the WD-repeat protein AIP1. Screening of the Arabidopsis databases revealed a large number of WD-repeat proteins. The sequence that showed the highest identity with X. laevis AIP1 was isolated from an Arabidopsis flower cDNA library. This Arabidopsis sequence showed 34 and 31% identity with X. laevis and yeast AIP1, respectively. To determine whether this protein was AIP1 like, we performed two-hybrid assays in which the putative AIP1, an actin, and an ADF were inserted into both the activation and the DNA binding domains of yeast two-hybrid vectors, and the various combinations were cotransformed into yeast. The results show that the Arabidopsis AIP1-like protein was able to interact with actin and ADF but not with itself (Figure 5A) . Moreover, the WD-repeat structure was essential, because deletion of the seventh of the nine WD repeats to generate the mutant AIPΔ7 abolished any interaction with ADF or actin (Figure 5A). These results confirm that the WD-repeat protein cloned from the Arabidopsis cDNA library has AIP1-like activity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

AIP1-Like Proteins in Plants.

(A) Yeast two-hybrid analysis of the interactions between AtAIP1-1 or AIPΔ7 and actin or AtADF2. LamC was used as a negative control. NI, no insert.

(B) Phylogenetic tree of all known AIP1 sequences. Arabidopsis contains two AIP1 genes. Sequences represented are C. elegans AIP1-2 and AIP1-1, Drosophila melanogaster AIP1, X. laevis AIP1, Gallus gallus AIP1, Homo sapiens AIP1, Mus musculus AIP1, Saccharomyces cerevisiae AIP1, Schizosaccharomyces pombe AIP1, Neurospora crassa AIP1, Dictyostelium discoideum AIP1, Physarum polycephalum AIP1, Oryza sativa AIP1, and Arabidopsis AIP1-1 and AIP1-2. Accession numbers are given at the end of Methods.

(C) RNA gel blots of total RNAs from various Arabidopsis tissues probed with gene-specific probes generated from the two AtAIP1 sequences.

Reproductive and Vegetative/Constitutively Expressed AIP1 Genes

Screening of the completed Arabidopsis genome sequence revealed only one other AIP1-like gene, AtAIP1-2. The two sequences share 67% identity with each other at the amino acid level. A phylogenetic tree was generated of all of the known AIP1 sequences (Figure 5B). Interestingly, the sequences cluster according to their taxonomic groups. Moreover, to date, no more than one AIP1 homolog per organism has be found in the databases, with the exception of the nematode C. elegans and the plant Arabidopsis, for which two forms are found. Plant actin, profilin, and ADF isoforms can be divided into those expressed in reproductive tissues and those expressed constitutively and/or in vegetative tissues (Christensen et al., 1996; Huang et al., 1996; Lopez et al., 1996; Kandasamy and Meagher, 1999). RNA gel blot analysis showed that the AtAIP1-1 transcript was present in floral tissues but not in vegetative tissues, whereas the AtAIP1-2 transcript was found in all tissues examined (Figure 5C). Thus, there are reproductive and vegetative/constitutively expressed AIP1 genes in plants.

AtAIP1-1 Enhances the F-Actin Depolymerization Activity of LlADF1

The AtAIP1-1 cDNA was expressed in Escherichia coli, and the recombinant protein was purified. The effect of AtAIP1-1 and LlADF1 on F-actin was investigated using cosedimentation assays (Figure 6A) . As expected, the bulk of the F-actin control was found in the pellet, whereas only small amounts of ADF and AIP1-1 were seen in the pellets. When F-actin was mixed with AtAIP1 or LlADF1, there was a small increase in the amount of soluble actin in the supernatant, indicating that both the AIP1 and the ADF can solubilize F-actin. However, incubation of F-actin with both LlADF1 and AtAIP1-1 resulted in a massive 60% increase in the amount of actin found in the supernatant (Figure 6B), indicating that the AIP and the ADF cooperated to depolymerize the F-actin. This finding strongly supports the hypothesis that AIP1 may cooperate with ADF to increase actin dynamics in vivo.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Analysis of the Combined Activity of AtAIP1-1 and LlADF1 on F-Actin.

(A) Sedimentation assays analyzed on one-dimensional gels using F-actin, AtAIP1-1, or LlADF1 alone or as mixtures. p, pellet; s, supernatant.

(B) Histogram of the quantity of soluble actin as determined by densitometric scanning of the supernatant lanes on the gel shown in (A).

AIP1 and ADF Are Located Similarly in Pollen Grains and Pollen Tubes

For AIP1 to cooperate with pollen ADF, it should be expressed in pollen. Polyclonal antiserum raised against the AtAIP1-1 recombinant protein stained a band of ∼67 kD on one-dimensional protein gel blots of total protein extracts from Arabidopsis pollen and leaf and lily pollen and leaf (Figure 7A) . These data indicate that this antiserum detects both vegetative/constitutively expressed and reproductive AIP1 proteins and cross-reacts with AIPs from other plants.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Localization of Pollen AIP1.

Protein gel blots of Arabidopsis and lily pollen and leaf total protein extracts probed with the anti-AtAIP1-1 antibody. Immunofluorescence of Narcissus pollen grains ([B] to [E]) and pollen tubes ([F] to [L]) stained for actin using rhodamine-conjugated phalloidin ([B], [D], [F], and [H]) or anti-actin (J), for pollen ADF using anti-LlADF1 ([C], [G], and [K]), or for AIP1 using anti-AtAIP1-1 ([E], [I], and [L]). Pollen grains and pollen tubes shown in (B) to (I) were fixed directly using formaldehyde, and pollen tubes in (J) to (L) were prepared for fixation using the freeze-substitution method.

Figures 7B and 7C show that pollen ADF binds to bundles and aggregates of F-actin in pollen grains, and Figures 7F, 7G, 7J, and 7K show that when pollen grains exit dormancy and germinate, the ADF is located mainly in the cytoplasm. When these pollen grains and pollen tubes are stained with anti-AIP1 antiserum, a similar pattern of staining is observed (Figures 7D, 7E, 7H, 7I, and 7L). We used two fixation protocols (conventional formaldehyde fixation [Figures 7F to 7I] and freeze substitution [Figures 7J to 7L]) for the pollen tubes, and both gave similar results, although the freeze substitution gave a better preservation of actin filaments and pollen tubes (Cresti et al., 1987). This similar localization of pollen AIP1 and ADF, coupled with the observation that AIP1 enhances the activity of ADF, suggests that these proteins may work as a pair and cooperate in vivo.

Protein Extraction, Electrophoresis, and Protein Gel Blot Analysis

Protein extracts were prepared, fractionated on one-dimensional gels, and blotted onto nitrocellulose membranes as described by Hussey et al. (1988). The membrane was probed with mouse anti-pollen actin-depolymerizing factor (ADF) antiserum (ZmADF1 or LlADF1), rabbit anti-ZmADF3 diluted 1:1000, or mouse anti-AtAIP1-1 diluted 1:200 in TBST (10 mM Tris, pH 7.4, 140 mM NaCl, and 0.05% Tween 20) buffer including 1% (w/v) fat-free dried milk. Anti-mouse or anti-rabbit secondary antibodies conjugated to horseradish peroxidase (Amersham Pharmacia) were used at a dilution of 1:5000 in TBST-milk buffer. Protein bands cross-reacting with anti-ADF or anti-AIP1 were visualized using the enhanced chemiluminescence detection kit (Amersham Pharmacia). After enhanced chemiluminescence detection, nitrocellulose membranes were washed in water and then stained with colloidal silver.

Phylogenetic Tree

The phylogenetic tree was prepared using ClustalW to align sequences (http://www.ebi.ac.uk/clustalw/), followed by TreeView PPC version 1.5.3. (createdbyRodericD.M.Page,http://taxonomy.zoology.gla.ac.uk/rod/rod.html).

Protein Phosphorylation

In vitro phosphorylation was performed using partially purified calmodulin-like domain protein kinase, as described by Allwood et al. (2001). In-gel phosphorylation assays were performed as follows. To prepare extracts containing kinases, cells were harvested and homogenized in 50 mM Hepes-KOH, pH 7.6, 2 mM DTT, 2 mM EDTA, 20 mM β-glycerophosphate, 20% (v/v) glycerol, 1 mM Na₃VO₄, 1 mM NaF, and 50 μg/μL protease inhibitors (complete; Boehringer Mannheim). The extracts were centrifuged at 14,000g for 20 min, and a 40 to 70% ammonium sulfate fraction was prepared and then desalted using 2-mL Sephadex G25 coarse-spin columns. A 10% SDS-PAGE gel to which 7 μM ADF was added immediately before polymerization was prepared, and the extracts were run through the gel. After electrophoresis, the gel was washed in 25 mM Tris-HCl, pH 7.5, 0.5 mM DTT, 1 mM Na₃VO₄, 5 mM NaF, 0.5 mg/mL BSA, and 0.1% Triton X-100 with gentle shaking at room temperature for three separate 30-min washes. The gel then was equilibrated with 25 mM Tris-HCl, pH 7.5, 0.5 mM DTT, 1 mM Na₃VO₄, and 5 mM NaF at 4°C overnight to renature the proteins. After renaturing, the gel was incubated for 60 min at room temperature in reaction buffer but without ATP (25 mM Tris-HCl, pH 7.5, 0.5 mM DTT, 1 mM Na₃VO₄, 2 mM EDTA, 12 mM MgCl₂, and 1 mM CaCl₂). The gel was incubated in 15 mL of fresh reaction buffer, this time containing 50 μCi of γ-³²P-ATP, for 60 min with gentle shaking. The majority of the unincorporated γ-³²P was removed by rinsing in electrotransfer buffer several times for a few seconds each, and the gels were subjected to electroblotting for 3 h at 400 mA. Proteins from the extracts that were run through the gel were transferred onto nitrocellulose membranes in this way, whereas ADF that was incorporated into the gel before polymerization remained fixed on the gel. Gels were stained with Coomassie Brilliant Blue R250, destained, and after equilibration in 5% methanol and 20% glycerol, dried onto paper. Gels and blots were exposed to x-ray film overnight. After exposure, blots were stained with colloidal silver.

Actin Binding and Enzyme Assays

Cosedimentation assays, phospholipid binding assays, and assay of phospholipase C activity were performed as described by Gungabissoon et al. (1998). To measure the extent of ADF binding to F-actin over a range of pH values, F-actin and ADF, each at 10 μM, were mixed in buffer F (12 mM Bis-Tris-Propane, 50 mM KCl, 1 mM MgCl₂, 1 mM EGTA, 0.2 mM ATP, 0.5 mM DTT, 0.2 mM CaCl₂, 1 mM NaN₃) in the pH range 6.0 to 9.0 and centrifuged at 386,000g, and the proteins in the pellets and supernatants were separated by SDS-PAGE. The amounts of protein in the supernatants and pellets were quantified by densitometric scanning of the SDS-PAGE gels. In buffer F, Bis-Tris-Propane (Sigma) was used to buffer pH values at 6.0 to 9.0, making use of the fact that this buffer has two pK_a values, 6.8 and 9.0.

To assess the effect of phospholipids on ADF activity, similar cosedimentation experiments were performed in buffer F at pH 8.0 with the following modifications. ADF (2 μM) was incubated with 100 μM phosphatidylinositol, phosphatidylinositol 4-monophosphate, phosphatidylinositol 4,5-bisphosphate (PIP₂), phosphatidylethanolamine, phosphatidylcholine, phosphatidylserine, inositol 1,4,5-triphosphate, or 1-oleoyl-2-acetylglycerol for 20 min at room temperature. F-actin then was added at a final concentration of 5 μM and incubated for 40 min at room temperature. The amount of ADF in the supernatants after sedimentation was determined by densitometric scanning, as described above.

To assess the effect of actin-interacting protein (AIP) on the depolymerization of F-actin, cosedimentation assays were performed at pH 8.0 using 10 μM actin and 5 μM AIP and ADF. This experiment was repeated three times, and the results shown in Figure 6 are typical of the data obtained in each experiment.

To assess the binding of phospholipids to ADF, native gels were used as described by Gungabissoon et al. (1998), using a modified method based on that described by Maciver and Weeds (1994). A total of 2 μg of ADF was loaded with phosphatidylinositol, phosphatidylinositol 4-monophosphate, PIP₂, phosphatidylethanolamine, phosphatidylcholine, phosphatidylserine, inositol 1,4,5-triphosphate, or 1-oleoyl-2-acetylglycerol (Sigma) at a molar ratio of 1:5 in 5% Suc. Six percent gels were prepared in half-concentrations of gel buffer containing 0.2 mM ATP, 0.2 mM EGTA, and 0.5 mM DTT; the tank buffer also was at half concentration. These gels were run at 10 mA for 40 min.

ADF inhibition of phospholipase C activity was assayed as described by Gungabissoon et al. (1998). After temperature equilibrium, phospholipase C activity of plasma membranes (5 μg of protein) was assayed at 25°C in 50 μL of phospholipase C buffer (50 mM Tris/malate, pH 6.0, and 10 μM CaCl₂) containing a sonicated micellar suspension of PIP₂ (50 μM) spiked with 0.86 kBq ³H-PIP₂ (specific activity of 325.6 GBq/mmol). Various concentrations of ADF were preincubated with the PIP₂ micelles in phospholipase C buffer (10 min at 25°C), and the assays were started by the addition of plasma membranes. After 6 min of incubation, the reaction was stopped by the addition of 1 mL of chloroform:methanol (2:1, v/v), and tubes were placed on ice for 5 min. HCl (0.25 mL of 0.6 N) was added to facilitate phase separation, and the tubes were vortexed vigorously and centrifuged at 14,000g for 2 min. Aliquots of the aqueous top phase were removed from each tube, and radioactivity was determined by liquid scintillation spectrometry after the addition of the scintillation fluid (Hionic-Fluor; Hewlett-Packard, Bracknell, UK).

Cloning of AtAIP1-1

AtAIP1-1 was identified in the Arabidopsis thaliana genome database in a BLAST search using the Xenopus laevis AIP1 sequence. The predicted protein showed 34% identity at the amino acid level. An Arabidopsis cDNA library (CD4-6 flower cDNA library from the ABRC [Columbus, OH]) (Weigel et al., 1992) was used as a template for PCR to amplify the 5′ region of the coding sequence (245 bp). This PCR product was used to screen the same library, and a full-length clone was isolated. Standard molecular biological techniques were used for PCR and library screens (Ausubel et al., 1992).

Two-Hybrid Assays

PCR was used to generate the coding sequences of AtAIP1-1 and AtADF2 with NcoI-BamHI ends incorporated into the 5′ and 3′ primers, respectively. Products were subcloned initially into pCR2.1 (Invitrogen, Carlsbad, CA) and then transferred to the NcoI-BamHI sites of either the GAL4 activation domain vector pACT2 or the GAL4 DNA binding domain vector pAS2-1 (Clontech, Palo Alto, CA). The actin cDNA (mouse) in pACT2 and pAS2-1 was kindly provided by I. Hauser-Hahn (Bayer, Leverkusen, Germany). All vector constructions were checked by restriction analysis and DNA sequencing.

All assays were performed using the GAL4-based two-hybrid system (Clontech). For qualitative assays, pairwise combinations of vectors were cotransformed into yeast colonies (HF7c strain) and were plated on the same selective medium to test for protein–protein interaction. The test used a β-galactosidase activity assay on a filter with 5-bromo-4-chloro-3-indolyl-l-d-galactopyranoside or growth in the absence of His in medium containing 5 to 10 mM 3-aminotriazole.

RNA Gel Blot Analysis

Arabidopsis total RNA was prepared as described by Jordan et al. (1994), fractionated on a 1% formaldehyde-agarose gel, and transferred to a nylon membrane (Hybond-N; Amersham). Each lane on the gel contained 30 μg of RNA, as quantified by A₂₆₀ spectrophotometric assays supplemented by comparison of rRNAs on ethidium bromide–stained agarose gels. Gene-specific 3′ untranslated regions for both AIP1-1 and AIP1-2 were used to generate riboprobes with high specific activity (using the Promega Riboprobe In Vitro Transcription System).

Expression and Purification of AIP1-1

AtAIP1-1 was amplified by PCR using primers that incorporated 5′ NdeI and 3′ HindIII restriction sites and cloned into pET28a. The vector construct was checked by restriction digests and sequencing. The protein was expressed with a His tag in Escherichia coli strain BL21 (DE3). Luria-Bertani medium (1 L) was inoculated with 5 mL from an overnight culture of transformed cells, and the culture was grown to an OD₆₀₀ of 0.6 at 37°C. Then, the flasks were transferred to 10°C, and the temperature was allowed to equilibrate for 1 h. The cells were induced with 1 mM isopropylthio-β-galactoside for 24 h before being harvested. Cell extract was prepared, and recombinant epitope-tagged protein was purified using a His-Trap column (Amersham Pharmacia) according to the manufacturer's instructions. Briefly, the cells were pelleted after induction by centrifugation at 6000g for 5 min, and the pellet was resuspended in START buffer (10 mM imidazole, pH 7.4, 10 mM Na₂HPO₄, 10 mM NaH₂PO₄, and 400 mM NaCl). The cells were disrupted by sonication (20 × 2 s over 1 min), and the extract was centrifuged at 13.5g for 45 min to pellet the cell debris. Extract was filtered through a 0.2-μm filter, loaded onto a Pharmacia His-Trap column previously loaded with nickel, and then equilibrated with START buffer. The column was washed with START buffer and then with 5 mL each of a series of wash buffers containing START buffer and imidazole concentrations of 40, 60, and 100 mM. Recombinant protein was eluted in elution buffer (START buffer containing 500 mM imidazole). Soluble recombinant protein was eluted in the first 2 mL of elution buffer. For biochemical assays, this protein was used within 60 min.

Immunofluorescence Microscopy

Narcissus pollen grains were germinated in vitro as described by Heslop-Harrison and Heslop-Harrison (1992). The pollen grains and pollen tubes were prepared for indirect immunofluorescence microscopy as described by Smertenko et al. (2001). The anti-LlADF1 and anti-AtAIP1-1 antisera were diluted 1:500 in PBS containing 0.5% BSA. The anti-mouse IgG fluorescein isothiocyanate conjugate (Sigma) was used as a secondary antibody in a dilution of 1:100. Actin filaments were stained with 0.66 μM rhodamine-phalloidin in PBS (Molecular Probes, Leiden, The Netherlands) for 20 min immediately before microscopy, as described by Smertenko et al. (2001) (for Figures 7B and 7D to 7F). Alternatively, anti-actin antibodies (C4 mouse monoclonal antibody) were used as described by Jiang et al. (1997) (for Figure 7J). It should be noted that no staining was observed with either the preimmune mouse serum or the second antibody alone. Also, the single staining of actin on the red channel generated no signal on the green channel and vice versa.

For freeze-substitution fixation, daffodil pollen grains were attached to 3-mm² pieces of dialysis tubing placed on the surface of solidified germination medium. After 3 h of germination, the material was immersed rapidly in liquid propane and prepared for immunofluorescence microscopy as described by Cresti et al. (1987). Rhodamine-phalloidin did not decorate actin filaments after freeze-substitution fixation, so mouse monoclonal anti-actin IgG clone C4 (ICN, Basingstoke, UK) antibody was used (at a dilution of 1:200). The secondary antibody was anti-mouse IgG fluorescein isothiocyanate conjugate.

All fluorescently stained samples were mounted in Vectashield (Vector Laboratories, Burlingame, CA) and examined with a Bio-Rad Radiance 2000 laser confocal microscope. The images digitally captured were processed with Adobe Photoshop 3 software (Mountain View, CA).

Upon request, all novel materials described in this article will be made available in a timely manner for noncommercial research purposes. No restrictions or conditions will be placed on the use of any materials described in this article that would limit their use for noncommercial research purposes.

Accession Numbers

The accession numbers for AtADF2 are AL162459 (EMBL) and At3g46000 (TAIR). Other accession numbers (NCBI unless indicated otherwise) are as follows: C. elegans AIP1-2, AAK11613.1; C. elegans AIP1-1, CAB03187.1; Drosophila melanogaster AIP1, AAF49822.1; X. laevis AIP1, AAD22062; Gallus gallus AIP1, AAD05042.1; Homo sapiens AIP1, AAH00201.1; Mus musculus AIP1, AAD05043.1; Saccharomyces cerevisiae AIP1, P46680; Schizosaccharomyces pombe AIP1, CAB11489.1; Neurospora crassa AIP1, T49741; Dictyostelium discoideum AIP1, AAB05588.1; Physarum polycephalum AIP1, AAC26321.1; Oryza sativa AIP1, BAB21186; Arabidopsis AIP1-1, AC006200 (GB) and At2g01330 (TAIR); and Arabidopsis AIP1-2, AB020749 (EMBL) and At3g18060 (TAIR).