Geminating Pollen Has Tubular Vacuoles, Displays Highly Dynamic Vacuole Biogenesis, and Requires VACUOLESS1 for Proper Function

VCL1 Is Expressed in Mature Pollen

To determine if VCL1 was expressed in pollen, total RNA from mature pollen of Arabidopsis plants was examined for the presence of VCL1 mRNA using reverse transcriptase (RT)-PCR. Primers were designed to detect the VCL1 mRNA specifically. A major gel band was obtained that corresponded to the expected length of the RT-PCR product from the mature mRNA (Fig. 1, RNA lane, arrow). The product was confirmed by DNA sequencing to correspond to the predicted region of the VCL1 mRNA (data not shown). For comparison, a PCR product was obtained from Arabidopsis genomic DNA that corresponded to the expected product length that included several introns (Fig. 1, gDNA lane). These results indicated that VCL1 mRNA, and probably protein, was expressed from the gene in mature pollen.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

The VCL1 gene is expressed in mature pollen. RNA lane, Major 444-bp product of RT-PCR was from the VCL1 mRNA. Several minor PCR products were also detected in pollen RNA. Note that the bands appear more abundant upon duplication of the image. The largest product was confirmed by sequencing to be derived from genomic DNA contamination (data not shown). The shorter product was presumed to be nonspecific. gDNA lane, 722-bp PCR product from Arabidopsis genomic DNA using same primers as in RNA lane. MW, Lambda HindII, EcoRI molecular mass marker. Masses as follows from top to bottom: 2.0 kb, 1.6 kb, 1.4 kb, 950 bp, 830 bp, and 560 bp (faint).

δ-TIP::GFP as a Reporter for Vacuole Biogenesis in Pollen

Given that VCL1 was expressed in mature pollen, a key question was whether or not the presence of the vcl1 T-DNA gene knockout would affect vacuole morphology in germinating pollen. To address this question, the vcl1 mutant was transformed with a construct in which the tonoplast-specific marker δ-TIP was fused to GFP. The reporter was under the control of the pollen-specific promoter LAT52, which displays enhanced expression in mature pollen and throughout germination (Eady et al., 1995; Eyal et al., 1995). Self-crosses led to the identification of T₃ lines that were: (a) homozygous for both δ-TIP::GFP and functional VCL1 as a control (VCL1/VCL1) or (b) homozygous for δ-TIP::GFP and heterozygous for vcl1 (vcl1/VCL1). Because the homozygous defect leads to embryo lethality, vcl1 was maintained as a heterozygote. To study the effects of vcl1 on the morphology and dynamics of vacuoles, the most straightforward approach was to examine pollen collected from mutant and control lines germinated in vitro.

We germinated our pollen in 30-μL drops directly on coated microscope slides that were then inverted. The pollen sank to the liquid-air interface, providing sufficient gas exchange to support germination (Fig. 2, A and B). The germination rates were highly variable between flowers. However, pollen in different stages of germination could be prepared easily for microscopy by placing a microscope coverslip over the drop. The method was useful for other reasons as well (see “Discussion”). When the VCL1/VCL1 control line was viewed by LSCM, numerous intracellular structures were highlighted in grains and elongating pollen tubes, indicating that the δ-TIP::GFP reporter was expressed in pollen (Fig. 2, C–E). No fluorescence was detected in control pollen not expressing δ-TIP::GFP using the same instrument settings (data not shown). These results indicated that germination in hanging drops was a useful approach and that the δ-TIP::GFP tonoplast reporter was functioning in pollen.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Germination of δ-TIP::GFP expressing pollen in vitro. A, Pollen germinated in an inverted 30-μL drop of germination medium on a coated microscope slide. Bar = 320 μm. B, Pollen tubes of different lengths are present in germinating pollen. Bar = 193 μm. C, Control pollen viewed by laser scanning confocal microscopy (LSCM). Vesicles and vacuoles are apparent. Bar = 20 μm. D, Corresponding transmitted image to C shown for comparison. 1, Hydrated pollen grain; 2, early pollen germination; 3, pollen tube in mid growth. Arrow, Position of sperm cells. Bar = 20 μm. E, Control pollen viewed by LSCM. Bar = 40 μm. F, Corresponding transmitted image to E. Pollen is at a more mature stage (4). Bar = 40 μm. All pollen viewed after 3 h of incubation in germination medium.

Vacuole Morphology in Germinating Pollen

Germinating pollen from the control line was examined to understand the basic morphology and dynamics of the tonoplast and vacuoles and to serve as a baseline for comparison to pollen from the vcl1/VCL1 line. Because the dynamics of vacuole biogenesis in living germinating pollen are poorly characterized, germination in vitro was divided for convenience into several stages as follows (see Fig. 2, D–F): (a) hydrated pollen grains (no visible tube), (b) early germination (tube length < 20 μm), (c) tubes in early growth (tube length 20–100 μm; occasional callose plug), and (d) tubes in more advanced growth that had at least one callose plug (typically >100 μm in length).

Hydrated pollen (stage 1) was first examined for δ-TIP::GFP by LSCM and found to have a granular appearance suggestive of the presence of dispersed vacuoles (Figs. 2, C and D, and 3, A and B). To define these structures more carefully, higher resolution images were obtained. The images clearly described the vesicular morphology within the hydrated pollen grain (Fig. 3C), and this was consistent with electron microscopy (EM) images of both mature and hydrated Arabidopsis pollen showing the presence of small dispersed vacuoles (Van Aelst et al., 1993; Kuang and Musgrave, 1996; see also Park et al., 1998). Other aspects of pollen morphology were apparent. For example, sperm cells could be seen in transmitted images (Fig. 2D, arrow), and dark regions presumed to be sperm cells were seen within hydrated pollen grains (Fig. 3A). The presence of chromatin within the sperm cell nuclei was confirmed by staining with the DNA dye DAPI and viewed in Z sections along with δ-TIP::GFP. In Figure 3D, both dispersed vacuoles and two sperm cell nuclei are visible in overlays of confocal (Fig. 3D) and transmitted (Fig. 3E) images.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Vacuole biogenesis in Arabidopsis pollen germinated in vitro. A, Optical sections showing the granular appearance of the δ-TIP::GFP. The dark regions (arrow) correspond to the two sperm cell nuclei in the tricellular pollen grain. Bar = 20 μm. B, Corresponding transmitted image for comparison with A. Bar = 20 μm. C, Higher resolution image of pollen grain showing the presence of dispersed vacuoles. A maximum intensity series using four-line averaging and 60× water immersion objective is shown. Bar = 8.0 μm. D, Sequential scanning of δ-TIP::GFP (green) and 4′,6-diamidino-2 phenylindole (DAPI; red) in overlay showing the presence of two sperm cell nuclei. Note that the vegetative nucleus was detected with DAPI but was at a different optical plane and not visible in this image. Bar = 8.0 μm. E, δ-TIP::GFP and DAPI images from D layered onto their corresponding transmitted image. Bar = 8.0 μm. F, Optical section through a pollen grain in early germination highlighting the presence of larger vacuoles and extended or tube-like tonoplast. Bar = 8.0 μm. G, Transmitted image corresponding to F for comparison. Bar = 8.0 μm. H, Optical section through several pollen grains during early pollen tube growth showing the presence of small vacuoles. Dark region of tube on left was due to curvature of the tube out of the focal plane. Bar = 20 μm. I, Transmitted image corresponding to H. Bar = 20 μm. J, Optical section of a pollen grain showing increased vacuolation in the grain. Note the tubular appearance of tonoplast in the growing tube (arrow) and the lack of GFP signal in the tip region. The background contains a more mature pollen tube that is highly vacuolated. Bar = 20 μm. K, Transmitted image corresponding to J. Note the lack of material in the tip region (arrow). The background contains a more mature pollen tube that is highly vacuolated. Bar = 20 μm. L, Higher resolution image of a pollen tube in early growth showing the presence tubular vacuoles. The pollen tube tip (not visible) is toward the upper left of the image. Bar = 8.0 μm. M, Optical section of pollen at a more advance stage of tube growth showing increased vacuolation of the grain and tube and tubular vacuoles near the tip. Bar = 40 μm. Inset, Magnification of the grain and plug. Arrow, Bright δ-TIP::GFP structures (see text). Bar = 8.0 μm. N, Transmitted image corresponding to M. The sperm cell nuclei are visible within the pollen tube (arrow). Bar = 40 μm. Inset, Magnified overlay image of the grain and plug. Bar = 8.0 μm. The plug is clearly visible between areas of δ-TIP::GFP fluorescence. O, Overlay image of δ-TIP::GFP (green) and DAPI (red). Overlap of GFP and DAPI signals results in the orange appearance of the nuclei (arrow). Bar = 40 μm. P, Transmitted image corresponding to O. Bar = 40 μm. All images shown are of pollen incubated for 3 h in germination medium before viewing by LSCM.

In pollen at early germination (stage 2), somewhat larger vacuoles could be detected frequently in addition to the tonoplast, which had an extended or tubular morphology (Fig. 3, F and G). This morphology was noted later in germination (see below) and, along with vacuoles, developed presumably via the coalescence of the smaller dispersed vacuoles. As germination progressed into early growth of the pollen tube (stage 3), numerous small vacuoles could be seen in both the pollen grain and in the elongating tube, again resulting in a granular appearance (Fig. 3, H and I). In general, as pollen tube growth proceeded, larger vacuoles were observed, particularly in the pollen grain (Fig. 3, J and K). In addition, the vacuoles in the tube at times appeared to be extended in appearance. This is apparent in Figure 3J (arrowhead). To investigate the unusual vacuoles in more detail, higher resolution maximum intensity images of a Z series were obtained, and they indicated that the vacuoles assumed a tubular morphology (Fig. 3L). It has been observed generally for pollen tubes that the tip region is devoid of vacuoles (Lord and Russell, 2002). This is apparent in transmitted images that lack vesicular material in this area (Fig. 3K, asterisk) and is further evidenced by a corresponding lack of δ-TIP::GFP fluorescence at the tip (Fig. 3J).

Pollen at a more advanced stage of germination (stage 4) was characterized by the presence of a large central vacuole in the grain and more extensive vacuolation in the tube, except in the region toward the tip that frequently contained dispersed or tubular vacuoles presumably necessary for growth (Fig. 3, M and N). Another feature was callose plugs. The plugs formed at the base of the pollen tube and were apparent in LSCM as a region devoid of δ-TIP::GFP fluorescence (Fig. 3M). The plug was also visible in the corresponding transmitted image (Fig. 3N) and an overlay image (Fig. 3N, inset). Additional callose plugs were observed at more distal regions of the pollen tube (data not shown). Sperm cells migrating toward the growing tip could be seen in transmitted images (Fig. 3N, arrow), indicating that pollen germinated in vitro retained this function. To confirm the presence of chromatin within the sperm cell nuclei, germinating pollen at a similar stage was stained with DAPI and viewed in Z sections along with δ-TIP::GFP. In Figure 3O (arrow), sperm cell DNA was visualized in the pollen tube.

These results indicated that vacuole morphology was ordered throughout germination. Germination rates were highly variable between grains; however, there was a general progression from dispersed vacuoles in hydrated grains and tubes in early germination to more extensive vacuoles in more mature pollen tubes. During this transition, an unusual feature was observed: the presence of tubular vacuoles in early germination in regions of the pollen tube proximal to the tip. This region was presumed to be important for rapid elongation. The range of morphologies observed suggested that vacuoles in germinating pollen were undergoing rapid biogenesis.

The Vacuoles in Germinating Pollen Are Highly Dynamic

Having observed vacuole morphology in static images, we investigated the dynamics of vacuole biogenesis during germination. Pollen with tubes in early growth (stage 3) before callose plug formation was examined over time after 3 h of incubation in germination medium. The morphology of the vacuoles changed significantly even within 1 h; there were altered arrangements of large vacuoles in the grain and smaller vacuoles in the pollen tube. Tubular appearing vacuoles were also present. To capture these dynamics, a timed series of images were produced over a period of 1.3 h. Several remarkable aspects of germination were captured in the movie (supplemental Fig. S1, available in the online version of this article at http://www.plantphysiol.org). These included: (a) the rapid movement of large vacuoles in the pollen grain with the streaming of tonoplast toward the growing tube, (b) the rapid movement of tonoplast toward and away from the region of the tube tip in a reverse fountain pattern, (c) a lack of tonoplast flow into the extreme apex of the tube, (d) the presence of sperm cell nuclei in the tube and their movement, (e) the presence of mobile strand-like (tubular) vacuoles, and (f) the presence of unusual bright regions of δ-TIP::GFP fluorescence in the tonoplast of large vacuoles in the grain and similar bright structures within the pollen tube that were highly mobile. Some of these features were apparent in a static image from the timed series (Fig. 4A). Both large vacuoles (arrow) and tubular vacuoles (arrowhead) were present in the grain and throughout the tube, respectively. A dark region near the tip (bracket) corresponded to the sperm cells. As described above, the tip region is mostly devoid of tonoplast, and this is particularly apparent when compared with the corresponding transmitted image (Fig. 4B, asterisk). These data indicated that the germinating pollen possessed vacuoles that were extremely dynamic and formed structures that were unusual in morphology and fluorescence.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

The vacuoles in germinating pollen are highly dynamic. A, Image of pollen grain in early growth from a time series (Fig. 1S). The arrow indicates a large vacuole and the position of an associated region of bright δ-TIP::GFP fluorescence. Note that bright structures are found throughout the pollen tube. The arrowhead indicates the position of some tubular vacuoles. The dark region near the tip (bracket) indicates the position of the sperm cells. Bar = 16 μm. B, Transmitted image corresponding to C. Arrow indicates the tip regions lacking δ-TIP::GFP fluorescence. Bar = 16 μm.

Germinating Pollen Contains Mobile Cytoplasmic Invaginations

The unusual bright regions in the tonoplast of large vacuoles were obvious in many pollen tubes, particularly in the pollen grain shown (Fig. 4A, arrow), as were other bright structures that were widely distributed throughout the tube. There were several possible explanations for these regions (see “Discussion”). To determine the nature of the structures, additional pollen tubes were viewed in more detail in images from a time series over a period of 30 min. Selected images showing details of interest are presented (Fig. 5). In Figure 5A (circled area), a bright ring structure and a bright semicircle were captured. In progressive images, the ring appeared to fuse with another region of tonoplast, and the semicircle folds inward, forming a nearly complete circle (Fig. 5B, circled area). The nearly complete circle then reopened (Fig. 5C, circle). Similar structures formed in other regions of the pollen tube (Fig. 5D, circled area), at times forming what appeared to be complete circles. Similar structures also were seen in other regions of the images. The highly fluorescent structures described including lines, arcs, and circles are very similar to double tonoplast cytoplasmic invaginations into vacuoles that have been described in leaves and cotyledons of Arabidopsis (Saito et al., 2002) and are termed bulbs. The complete time series over a period of 30 min is shown as a movie in supplemental Figure S2 and demonstrates the extreme activity of vacuoles and cytoplasmic invaginations in the pollen tube. An examination of pollen at different stages of germination revealed the presence of localized bright regions that were probably cytoplasmic invaginations (see Figs. 2C; 3, F, J, and M, inset, arrow). These results provided the reasonable explanation that the highly fluorescent structures were rapidly changing cytoplasmic invaginations. Overall, our examination of vacuole morphology in control pollen provided details of vacuole dynamics and a solid basis for comparison to pollen containing the defective vcl1 gene.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Brightly fluorescent regions are cytoplasmic invaginations. Selected images are from a time series of a pollen tube at a more advanced stage in germination. The complete series (Fig. 2S) was produced over a period of 30 min. The images were produced at the following times after the beginning of the series: A, 90 s; B, 100 s; C, 420 s; and D, 1,640 s. Circled areas indicate the positions of several semicircles and rings of tonoplast. Pollen tube tip (not visible) is toward the top of the image. All bars shown = 8.0 μm.

Effect of vcl1 on Pollen Morphology

To ask if VCL1 was necessary for vacuole biogenesis or pollen germination, we investigated pollen from a heterozygous vcl1/VCL1 line. Because vcl1 is embryo lethal when homozygous, pollen containing the vcl1 gene knockout could be collected only from influorescences of vcl1/VCL1 plants. One-half of the pollen from heterozygotes was predicted to contain vcl1 in the haploid complement and would not express the gene. Given the essential nature of VCL1, lack of expression in even a proportion of the pollen would be expected to lead to a population that is distinct and defective in vacuole morphology or other aspect of germination. Pollen from vcl1/VCL1 plants was examined thoroughly at all stages of germination using the approaches described for the control pollen. In all, four independent experiments were done, totaling a minimum of 1,600 pollen grains examined from vcl1/VCL1 plants at all stages described for the control pollen. In addition, at least 650 freshly collected pollen grains from vcl1/VCL1 plants was examined for altered vacuole morphology (data not shown). Overall, our detailed analysis detected no defects in vacuole biogenesis in pollen from vcl1/VCL1 plants compared with pollen from control plants (data not shown).

Because in principle 50% of pollen from vcl1/VCL1 plants harbored vcl1, it was possible that the mutant pollen germinated at a lower frequency than VCL1 pollen. To address this possibility, pollen from vcl1/VCL1 and control plants was germinated in vitro and quantified. The frequencies for pollen germinated in liquid were statistically the same for the mutant and control. Pollen from vcl1/VCL1 yielded a frequency of 21.0% ± 1.3% (2,739 grains scored) compared with the control frequency of 22.9% ± 1.1% (2,618 grains scored). We also compared the germination frequency of pollen on solid medium (Fan et al., 2001), which resulted in a greater overall rate of germination. Pollen from vcl1/VCL1 yielded a frequency of 50.9% ± 5.2% (3,846 grains scored) compared with the control frequency of 43.5% ± 5.2% (3,655 grains scored). Again, this indicated no statistical difference in germination frequencies. Within the limits of our assays, which may not detect subtle differences in germination, our results indicated that vcl1 does not significantly reduce germination and supports the conclusion that vcl1 and VCL1 pollen share similar vacuole morphologies.

The lack of obvious defects in vacuole biogenesis in vcl1 pollen was surprising because VCL1 is essential and there are no known homologs in the Arabidopsis genome (Rojo et al., 2002). This raised the interesting question of how vacuole biogenesis occurred at all in pollen harboring the vcl1 defect (see “Discussion”). The effect of vcl1 on pollen vacuoles was either too subtle to detect using our methods, which was unlikely in the case of vcl1, or was at a different stage of development or fertilization. Thus, rather than focus further on morphology, we decided to ask if vcl1 pollen was affected functionally in vivo.

Male and female vcl1 Gametophytes Are Defective

As a functional assay, we used genetics to examine the fertility of vcl1 gametophytes. This approach was both sensitive and quantitative compared with the morphological approach. First, progeny of self-crosses of vcl1/VCL1 T₃ plants were evaluated for transmission of the vcl1 defect. Seedling DNA from the progeny was examined for the presence of the vcl1 defect via PCR to detect the characterized single T-DNA insertion within the VCL1 gene (Rojo et al., 2002). Because vcl1 homozygous embryos did not survive, the expected ratio of vcl1/VCL1 to VCL1/VCL1 progeny was 2 to 1. The observed ratio was 1.1 to 1. This was a highly significant result, having a χ² P value of 0.0031 (Table I) and indicated reduced transmission of the vcl1 gene to the progeny through the male or female gametophytes, or both. To determine which gametophytes were affected by vcl1, reciprocal backcrosses were done between vcl1/VCL1 and VCL1/VCL1 plants, and the progeny were scored by PCR. In these crosses, the expected ratio of vcl1/VCL1 to VCL1/VCL1 progeny was 1 to 1. When the pollen donor was vcl1/VCL1, the observed ratio was 0.45 to 1, which was highly significant by χ² test having a P value of 0.0093 (Table I). This strongly indicated that vcl1 transmission was suppressed in male gametophytes. When the pollen donor was VCL1/VCL1, the observed ratio of vcl1/VCL1 to VCL1/VCL1 progeny was only 0.53 to 1. Again, this was a significant result having a P value of 0.0218 (Table I). Interestingly, this indicated that vcl1 also affected transmission through the female, indicating that VCL1 is important for the function of male and female gametophytes. The results of the genetics also indicate that homozygous vcl1 is lethal at the sporophyte stage; however, at the haploid gametophyte stage, it is not fully expressive.

Nucleic Acid Preparation and RT-PCR

The RNA from mature pollen was a generous gift from Drs. Mark Johnson and Daphne Preuss (University of Chicago). Pollen was collected (Mayfield et al., 2001) from the Arabidopsis qrt (quartet) mutant, which has no detectable impact on pollen function (Preuss et al., 1994). Message-specific primers were designed to complement sequences within exons 11 and 14 (Rojo et al., 2001). The intervening gene sequences included exons 12 and 13 and introns 12 through 14. A gel band of 444 bp was obtained that corresponded to the expected length of the RT-PCR product from the mature mRNA minus introns 12 through 14. For comparison, a PCR product of 722 bp was obtained from Arabidopsis genomic DNA that corresponded to the expected product length, which included the introns. RT-PCR was performed using the Superscript II Reverse Transcriptase System (Invitrogen, Carlsbad, CA) according to the manufacturer. Genomic DNA was extracted from Arabidopsis Columbia-0 using the cetyl-trimethylammonium bromide method (Rogers and Bendich, 1985). The following gene-specific primers were used to amplify a region of the VCL1 gene diagnostic for either the mRNA or the gene containing introns 5′ CATGGGATATGGGGAAAAA 3′ and CAGATGAGAAAGCGGAAGCT 3′. The conditions for PCR were as follows: 94°C for 5 min followed by 30 cycles of amplification (94°C for 1 min, 61°C for 1 min, and 72°C for 1 min). Products were fractionated by 1% (w/v) agarose gel electrophoresis.

Construction and Transformation of δ-TIP::GFP Reporter

For construction of the LAT52 promoter-δ-TIP::GFP, the 35S promoter was removed from pEGAD (Cutler et al., 2000) as an SstI/Age I fragment. This was replaced by blunt-end ligation with an XbaI/HindIII fragment from pUC19-LAT52::GFP (a generous gift from Zhenbiao Yang, University of California, Riverside) containing the LAT52 promoter. The LAT52 promoter was originally from the laboratory of Sheila McCormick (Plant Gene Expression Center, University of California, Berkeley). The resulting construct pLAT52-δ-TIP::GFP was transformed into Argobacterium tumefaciens and introduced into plants heterozygous for vcl1 by floral dipping (Clough and Bent, 1998). The vcl1 mutant (Rojo et al., 2001) has a single T-DNA insertion in the VCL1 gene in Arabidopsis Columbia-0. Plants harboring pLAT52-δ-TIP::GFP were selected on BASTA, and heterozygous (vcl1/VCL1) plants were identified by PCR using NPTII gene-specific primers (see below) to identify the T-DNA insertion. Several resistant vcl1/VCL1 T₂ plants were self-crossed, and progeny were analyzed for the vcl1 defect. Among the genotyped T₃ progeny were plants homozygous for δ-TIP::GFP and heterozygous for the vcl1 defect. One such line, LGT4-28, was used in this study. This line will be available from the Arabidopsis Biological Resource Center (Ohio State University, Columbus).

Genotyping and Pollen Germination in Vitro

Seeds from self crosses of LGT4-28 were sterilized in 70% (v/v) ethanol and 0.05% (v/v) Triton X-100 and germinated on Murashige and Skoog Minimal Organics medium (Life Technologies/Gibco-BRL, Cleveland), then transferred to soil after the appearance of four true leaves. After about 2 weeks, one or several rosette leaves were removed for the preparation of DNA using Plant DNAzol Reagent (Invitrogen) according to the manufacturer's instructions. The final DNA product was dissolved in 15 to 30 μL of water for PCR. Because homozygous vcl1 is embryo lethal, it was maintained as a heterozygote in LGT 4-28. For all genotyping, vcl1/VCL1 plants were identified from those that were VCL1/VCL1 by using NPTII-specific primers to identify the T-DNA insertion in vcl1 (forward, 5′ CCGGTACCTGCCCATTC 3′; and reverse, 5′ GCGATAGAAGGCGATGCG 3′). The conditions for PCR were as follows: 94°C for 3 min, 28 to 30 cycles of amplification (94°C for 45 s, 45°C for 30 s, and 72°C for 2 min). The NPTII product of approximately 400 bp was fractionated by 1.2% (w/v) agarose gel electrophoresis. As a positive control, all PCR reactions included primers to the gene for the large subunit ADP Glc pyrophosphorylase (APL), which resulted in a product of approximately 700 bp. For germination in drops in vitro, pollen from vcl1/VCL1 or VCL1/VCL1 control plants was collected from flowers that were fully or nearly fully open with anthers that were freshly dehiscent. Flowers were air-dried for 2 to 4 h, then gently dabbed onto the surface of a 30-μL drop of Germination Medium 1C [10% (w/v) Suc, 0.01% (w/v) boric acid, 1 mm CaCl₂, 1 mm Ca(NO₃)₂, and 1 mm MgSO4, pH adjusted to 6.5 with KOH) on a coated microscope slide (Cytoslide, Thermo Shandon, Pittsburgh, PA). Pollen from one flower was typically examined per drop; however, pollen from multiple flowers could be pooled and germinated. The slide was then inverted and incubated for 3 to 15 h at 23 C and 100% humidity. For germination on agar, 2 mL of Germination Medium (Fan et al., 2001) was applied to the surface of a microscope slide and allowed to solidify. Pollen from individual flowers was then dabbed onto the agar, which was then incubated for 12 to 15 h at 100% humidity. To quantify germination in drops or on agar medium, pollen from individual flowers was germinated for 12 to 15 h and 50 to 150 (drops) or 100 to 200 (agar) grains were scored per flower. A minimum of 27 vcl1/VCL1 or VCL1/VCL1 flowers each were scored in liquid or on solid medium. In all cases, the frequencies reported plus or minus se were derived from three independent experiments. For examination by LSCM, a 24-× 60-mm coverslip was gently lowered onto the drop containing the germinating pollen. The DNA dye DAPI was added directly to the drop before coverslip addition at a final concentration of 2 μg/mL.

Microscopy

Germinating pollen was viewed using a Leica TCS SP2/UV Confocal Microscope (Leica Microsystems, Wetzlar, Germany) and either 20× or 63× water immersion objectives. For GFP and DAPI viewing, manufacturer settings were used. The sequential scanning function was used to eliminate the detection of δ-TIP::GFP in the DAPI channel and vice versa when producing merged images. For all static images, 2× line averaging was used. For detailed images as indicated, maximum intensity images were produced from a Z series with 4× line averaging using Leica Lite software (Leica Microsystems). For movies, images were taken at the rates indicated with no line averaging. To further reduce bleaching of δ-TIP::GFP fluorescence due to repeated scanning, the pinhole was widened to 150 μm from the usual setting of 100 μm used for static images.

Genetics

For segregation analysis of heterozygotes, LGT 4-28 plants were used that were allowed to self-pollinate. The seed was collected, sterilized, germinated on plates, and then transferred to soil. After about 2 weeks, DNA was prepared from rosette leaves, and plants were genotyped as described above. For analysis of reciprocal back crosses of vcl1/VCL1 plants by VCL1/VCL1 plants, approximately 30 independent reciprocal crosses were done. Seeds were collected and pooled for VCL1/VCL1 (female) by vcl1/VCL1 (male) crosses and for vcl1/VCL1 (female) by VCL1/VC1 (male) crosses. After sterilization and germination on plates, DNA was extracted as above at the appearance of the fourth true leaf. Progeny were genotyped for the presence of the vcl1 by PCR. χ² P and other values are indicated. P values were calculated using GraphPad Quickcalcs (free on-line at GraphPad.com).