Distinct short-range ovule signals attract or repel Arabidopsis thaliana pollen tubes in vitro

Background Pollen tubes deliver sperm after navigating through flower tissues in response to attractive and repulsive cues. Genetic analyses in maize and Arabidopsis thaliana and cell ablation studies in Torenia fournieri have shown that the female gametophyte (the 7-celled haploid embryo sac within an ovule) and surrounding diploid tissues are essential for guiding pollen tubes to ovules. The variety and inaccessibility of these cells and tissues has made it challenging to characterize the sources of guidance signals and the dynamic responses they elicit in the pollen tubes. Results Here we developed an in vitro assay to study pollen tube guidance to excised A. thaliana ovules. Using this assay we discerned the temporal and spatial regulation and species-specificity of late stage guidance signals and characterized the dynamics of pollen tube responses. We established that unfertilized A. thaliana ovules emit diffusible, developmentally regulated, species-specific attractants, and demonstrated that ovules penetrated by pollen tubes rapidly release diffusible pollen tube repellents. Conclusion These results demonstrate that in vitro pollen tube guidance to excised A. thaliana ovules efficiently recapitulates much of in vivo pollen tube behaviour during the final stages of pollen tube growth. This assay will aid in confirming the roles of candidate guidance molecules, exploring the phenotypes of A. thaliana pollen tube guidance mutants and characterizing interspecies pollination interactions.


Background
After a pollen grain lands on the surface of the pistil, it absorbs water from the stigma and forms a pollen tubea long polar process that transports all of the cellular contents, including the sperm [1]. Pollen tubes invade the pistil and migrate past several different cell types, growing between the walls of the stigma cells, travelling through the extracellular matrix of the transmitting tissue, and finally arriving at the ovary, where they migrate up the funiculus (a stalk that supports the ovule), and enter the micropyle to deliver the two sperm cells-one fertilizes an egg and other the central cell (Fig. 1a, 1b) [2]. Typically, only one pollen tube enters the ovule through an opening called the micropyle, terminates its journey within a synergid cell, and bursts to release sperm cells-a process defined as pollen tube reception [3].
A combination of genetic and in vitro assays has defined signals that contribute to the early stages of pollen tube guidance. Chemocyanin, a small basic protein from lily stigmas, attracts lily pollen tubes in vitro [4], and in A. thaliana, wild type pollen guidance was abnormal when grown on stigmas over expressing the A. thaliana chemocyanin homolog [5]. Other signals are active in the nutrient-rich extracellular matrix secreted by the female transmitting tissue. A pectin that may promote guidance by mediating adhesion of pollen tubes to this matrix has been identified in lily [6]. Glycoproteins that likely contribute to guidance have also been described: in lily, a lipid transfer protein that contributes to adhesion [6], and in tobacco, two glycoproteins (TTS1 and TTS2) that provide nutritional and guidance cues are known [7,8]. Although potential homologs of these proteins exist in A. thaliana, their role in pollen tube growth is yet to be determined [6].
After emerging from the transmitting tract, pollen tubes approach the ovule micropyle with remarkable precision. Mutants defective in pollen tube guidance have demonstrated that this process is controlled by a series of molecular signals that involve pollen tubes, ovule tissues, and female gametophytes [1]. The A. thaliana mutants, ino [9] and pop2 [10] point to a role for diploid ovule tissue in pollen tube guidance; these have aberrant interactions between pollen tubes and diploid ovule cells, yet their female gametophytes appear normal, and in the case of pop2, can be fertilized with wild-type pollen [11]. Pollen tubes fail to either reach or enter the micropyle in A. thaliana mutants with nonviable or aberrant female gametophytes yet apparently normal diploid ovule tissue, providing strong support for the role of the haploid germ unit in promoting growth to the micropyle [12,13]. Based on these studies, it was proposed that final stages of pollen tube growth can be divided into two distinct phases: funicular guidance, in which pollen tubes adhere to and grow up the funiculus, and micropylar guidance, where pollen tubes enter the micropyle to deliver sperm to the female gametophyte [13]. Micropylar guidance signals originate at least in part from the two synergid cells contained within the female gametophyte; pollen tubes do not enter ovules in which synergid cells were either laser ablated [14] or defective due to a lesion in A. thaliana MYB98 gene [15]. The maize EA1 protein, which is exclusively expressed in the egg and synergids of unfertilized female gametophytes, may specify a role for these cells in regulating micropylar guidance. Plants expressing EA1 RNAi or antisense constructs produced significantly fewer seeds than wild type, and wild type pollen tubes failed to enter mutant ovules [16].
In vitro assays have been used to characterize intracellular cues such as a Ca 2+ gradient at the tip of pollen tubes that is critical for growth. Disrupting this gradient by iontophoretic microinjection or by incubation with Ca 2+ channel blockers can change the direction of tube growth [17]. The Ca 2+ gradient in pollen tubes is controlled by Rho GTPases; injection of antibodies against these proteins into pollen tubes, or expression of dominant-negative forms of RhoGTPase, causes the tip-focused Ca 2+ gradient to diffuse and eliminates tube growth [18], presumably by disrupting F-actin assembly [19]. These pollen tube growth defects can be partially alleviated by adding high concentrations of extracellular Ca 2+ [18].
In vitro grown pollen tubes also reorient their growth in response to certain extracellular cues; lily pollen tubes are attracted to chemocyanin [4] and repelled by a point source of nitric oxide [20]. In addition, in vitro grown pearl millet pollen tubes are attracted to ovary extracts [21]. For T. fournieri pollen tube guidance across a simple medium and into the ovule was achieved after pollen Pollen tube targeting in vitro tubes were grown through a stigma and style [22]. In this species, the female gametophyte protrudes from the ovule, and pollen tubes enter the micropyle without interacting with funiculus [22]. Thus, the T. fournieri in vitro guidance system serves as a model for the micropylar, but not the funicular guidance phase of pollen tube growth to ovules [23]. Here, we describe an A. thaliana in vitro guidance assay that recapitulates both funicular and micropylar guidance, serving as a model for ovules with encased female gametophytes, an arrangement that is more common among flowering plants. With the sequenced A. thaliana genome, the large collection of mutants affecting reproductive functions, and comparative genomic resources, this assay will greatly facilitate identifying genes that mediate the final phases of pollen tube guidance.

Results
The A. thaliana stigma and style confer pollen tube targeting competence Previous studies indicated that pollen tubes germinated in a simple growth medium cannot be guided to the micropyle [14,22]. Consistent with these observations, when such assays are performed for A. thaliana, few ovules are targeted (~3%, Table 1). Hence, we instead removed the upper portion of the pistil (the stigma and style [4,10,14,24]), deposited pollen on the stigma surface and showed that A. thaliana pollen tubes emerged from the style, travelled across an agarose medium to excised ovules and successfully entered the micropyle (Fig. 1a,  1b). To facilitate pollen tube observation, especially after they enter the micropyle and are obscured by the opaque ovule integument cells, we transformed plants with a GFP reporter under the control of the pollen-specific LAT52 promoter [25], and identified GFP-expressing lines with fully functional pollen tubes. Upon reaching the female gametophyte, these tubes burst and release a large spot of GFP (Fig. 1c, 1d, arrowheads and see Additional files 1 and 2), conveniently marking targeted ovules. Pollen tubes that grew within ~100 µm of an unfertilized ovule often made a sharp turn toward an ovule; of the tubes that grew within this range, ~50% successfully entered the micropyle (Table 1). This targeting efficiency is significantly higher than that of tubes germinated on agarose (Table 1). Thus, pollen tubes acquire the ability from pistil tissue to perceive ovule guidance signals, perhaps by absorbing essential nutrients or undergoing critical developmental transitions; a similar phenomenon was reported in T. fournieri [22]. In some cases, pollen grains that germinated on the stigma formed tubes that grew onto the medium, rather than penetrating the pistil and growing through the style. Nonetheless, these tubes successfully targeted excised ovules, suggesting that interaction with the stigma alone is sufficient to confer pollen tube guidance competence (Table 1); targeting efficiency was significantly reduced, however, suggesting that direct contact with female cells, rather than exposure to diffusible factors in the medium, is most important.

Characterization of A. thaliana pollen tube-ovule interactions
As tubes left the style, they dispersed ( Fig. 1e and see Additional File 1), growing at 2.5 ± 1.0 (s.d) µm/min (n = 20) and up to 3 mm before reaching an ovule ( Fig. 2a-f). Near a virgin ovule, however, growth rates decreased (1.2 ± 0.59 (s.d) µm/min (n = 20; Fig. 2a-f) and the tubes often made sharp turns (Fig. 1d, arrows) within 33 ± 20 (s.d) mm of a micropyle (avg. 60 ± 38° (s.d); n = 60). Pollen tube guidance to ovules was abolished when fertilized or heat-treated ovules were used (Table 1), indicating they release a diffusible, heat-labile attractant prior to fertilization. Tube entry into the ovule appears to be not influenced by the number of tubes near a micropyle; targeting was achieved regardless of whether one or multiple tubes were in the vicinity of an ovule (see Additional files 1, 2, 3, 4). When approaching an ovule, the in vitro grown pollen tubes did not always migrate up a funiculus, Fig. 1c) before entering it. To directly test the role of this tissue, we removed funiculi from ovules, revealing a small but significant decrease in targeting efficiency (Table 1). These results indicate that an interaction between the pollen tube and funiculus is not essential, yet this interaction enhances successful entry into the ovule, perhaps by i) providing a physical support for pollen tubes to reach the micropyle, ii) aiding in the generation and maintenance of a signal gradient, or iii) enhancing the availability of ovule-derived guidance signals.  Fig. 2g-l and see Additional files 1 and 2). While previous work showed that pollen tube growth arrests only after reaching the female gametophyte [9,10], our observations point to additional signals that slow growth within the ovule prior to this arrest. This delay coincides with meandering pollen tube navigation past the integument and nucellus cells (Fig. 2g-k and see Additional files 1 and 2), potentially reflecting guidance by these cells.

Short-range guidance signals from A. thaliana ovules are developmentally regulated
The data presented above indicate that contact with A. thaliana stigmas and styles enables pollen tubes to respond to diffusible ovule signals. To understand the nature and source of these signals, we examined their activity during ovule development. Previously it was shown that pollen tubes grow randomly or fail to elongate in immature A. thaliana pistils [26]. However, it was impossible to distinguish the contributions of distinct tissues in these experiments. Here, we exploited the modular nature of the in vitro system, varying the age of stigmas, styles, and ovules. While mature flower parts (stage 14 [27] were optimal, the stage of ovule development was critical, with guidance factors completely absent at ~32 hrs (stage 12a) and lower at ~16-24 hrs (stages 12b-c) before flowers mature ( Table 2, upper panel). This pattern correlates with synergid development, the suggested source of pollen tube attractants; these cells form after stage 12a [27,28]. Even so, immature ovules promote better pollen tube guidance than heat-treated or fertilized ovules, suggesting that a basal signalling capability is established early and increases as the female gametophyte differentiates. In contrast, the developmental stage of the stigma and style did not significantly alter targeting (χ 2; P > 0.1; Table 2, bottom panel), indicating that the signals Pollen tube navigation time course that confer targeting competence to pollen tubes do not vary over the course of pistil maturation (stages 12a-18) and that they emerge as early as stage 12a.

Short-range guidance signals from ovules are highly species-specific
Like many traits that mediate reproduction [29,30], pollen tube guidance signals diverge rapidly -crosses between A. thaliana and its relatives show random or arrested pollen tube growth, even among species separated by <25 million years, MY [13,31]. Because these interspecies crosses utilized intact pistils, it has been impossible to discern the roles of individual tissues; moreover, early blocks in pollen tube migration have often made it difficult to assess interactions at downstream stages, including those near ovules. Here, we examined whether the stigma, style and ovule-derived signalling interactions are shared among A. thaliana relatives separated by ~5, 10, or 20 MY [31] ( Table 3). The ability of the stigma and style to promote pollen tube competence was highly conserved (χ 2 ; P > 0.05; Table 3, bottom panel), while the ovule-derived attractant diverged rapidly (χ 2 ; P < 0.01; Table 3, upper panel). For example, A. thaliana pollen tubes inefficiently target ovules from Arabidopsis arenosa (separated by 5 MY from A. thaliana), rarely target Olimarabidopsis pumila ovules (10 MY), and fail to target Capsella rubella or Sysimbrium irio ovules (10 and 20 MY, respectively, Table 3, upper panel). Because C. rubella and S. irio are challenging to transform, it was not possible to test whether ovules from these two species are able to guide self-pollen expressing GFP under our assay conditions. Nonetheless, the ability of A. thaliana pollen to target ovules correlated with phylogenetic separation [30], suggesting that A. thaliana pollen tubes are sufficiently diverged that they fail to recognize attractants from C. rubella and S. irio ovules. Unlike the calcium signals that emanate from synergids [32,33], the proposed source of micropylar guidance signals, our results point to a diffusible, heat-labile ovule-derived signal that is sufficiently complex for rapid divergence -criteria that are most consistent with a protein-based signal. Pollen tubes perceive this signal at a distance of ~100µm from ovules after a 5 hour incubation in the assay. To estimate the molecular weight of this signal, we measured the diffusion rates of fluorescently-labelled dextran molecules under the same conditions in which the in vitro assay was performed, and calculated that the ovule-derived signal could measure up to approximately 85 kD (see methods).

Targeted A. thaliana ovules repel supernumerary pollen tubes in vitro
Interestingly, while the ovule-derived attractant in the in vitro assay acted to guide multiple pollen tubes toward ovules, only one pollen tube gained access to each micropyle. This is reminiscent of polyspermy blocks in vivo, where only one tube generally migrates up the funiculus and into the ovule [13]. While the mechanisms that prevent multiple tubes from even approaching an ovule are highly efficient, it is nonetheless possible for more than one pollen tube to enter a micropyle. In wild-type maize, heterofertilization results when the egg and central cell are fertilized by different pollen tubes at a frequency of ~1/50 [34] and in A. thaliana, ~1% (wild type) and ~10% (feronia) of ovules are penetrated by multiple pollen tubes [3]. When we performed the in vitro assay with fertilized ovules, many tubes grew within 100 µm, but none entered ( Table 1), suggesting that the release of the ovule attractant terminates after fertilization, or alternatively, that a new signal repels additional pollen tubes. To distinguish between these possibilities, we used time-lapse imaging analysis ( Fig. 3 and see Additional files 3 and 4). While 44% of targeted ovules (n = 143) were approached by additional pollen tubes, in every instance, these tubes did not enter the micropyle. Repelled tubes either stalled near the micropyle or turned sharply away from the targeted ovule (84 ± 42°; n = 61, Fig. 2e-h and see Additional files 3 and 4), a response that was observed as early as 10 min after a successful targeting event. This effect was fairly short-range; only tubes that approached within 27 ± 22 (s.d) µm were repelled. The diffusion rate of a series of dextran molecules through the medium used in this assay allowed us to estimate that a repellent measuring <10kD could diffuse ~27 µm in 10 minutes. Such abrupt turning behaviours were not observed when a single tube approached a virgin ovule. Instead, these tubes changed their growth direction by 60 ± 38° (s.d; n = 60), migrating toward, and not away, from the micropyle.
In the A. thaliana female sterile feronia and sirene mutants, wild type pollen tubes enter the mutant ovules but fail to cease growth or burst. In addition to this defect, multiple pollen tubes gain access to feronia and sirene mutant ovules [3,35]. Based on these results, it was suggested that repulsion of supernumerary tubes does not initiate until pollen tube reception occurs. Our observations with wild type pollen tubes in this in vitro assay indicate this is not the case -repulsion responses occurred well before tube growth terminated or tubes released their cytoplasm (n = 50, Fig. 3j, 3n and see Additional files 3 and 4). Moreover, while previous work suggested that female gametophyte cells release an inhibitory signal [3,35], our results show that repulsion initiated soon after the pollen tubes entered the micropyle and long before they reached the female gametophyte ( Fig. 3 and see Additional files 3 and 4). Thus, this work points to a diffusible repulsive signal that is sufficient to override the ovule attractant. This signal may be derived directly from the diploid cells that surround the micropyle, from the female gametophyte, or from the successfully targeted, but unlysed, pollen tubes.

Conclusion
Based on the results described here, we have defined three signaling events that regulate pollen tube guidance in A. thaliana: i) contact-mediated competence conferred by the stigma and style, ii) diffusible ovule-derived attractants and iii) repellents exuded from recently-targeted ovules.
Pollen tubes avoid targeted ovules The species specificity and diffusion properties of the ovule attractant are consistent with a protein signal, while the abrupt transmission and response to the repellent suggests the activity of a small molecule, a peptide, or posttranslational modifications to signals present before fertilization. This investigation also provides a platform to confirm the roles of candidate guidance molecules and to explore the phenotypes of A. thaliana mutants, including those that affect the development of diploid [9,10] and haploid [3,16,35] female tissues or pollen tubes [36]. The ability to characterize interspecies pollination interactions with this assay could lead to improvements in generating novel plant hybrids, a process that often requires in vitro manipulations [37]. To consistently isolate pistils of varying developmental stages, we correlated the initial day of flowering of our plant population with previously defined floral development stages [26]. First, we confirmed that the youngest open flower is similar to stage 14 [26]. In A. thaliana, flowers continuously arise at the floral apex and are arranged in a spiral, with the younger buds on the inside. This predictable pattern allowed us to select stage 14 flowers as a starting point and identify older flowers (up to stage 18) and younger buds (up to 12a).

In vitro pollen tube guidance assay
Growth medium for in vitro manipulations of pollen tubes [10] was determined to be optimal for also growing pollen tubes through a cut pistil. For the in vitro assays described here, pollen growth medium (3 ml) was poured into a 35 mm petri dish (Fisher Scientific, Hampton, USA). This volume of medium was ideal both for pollen tube growth and for microscopically viewing the interactions between pollen tubes and ovules. Excised pistils were pollinated under a dissection microscope (Zeiss Stemi 2000), cut with surgical scissors at the junction between the style and ovary (World Precision Instruments, Sarasota, USA), and placed horizontally on pollen growth medium. Pollen tubes emerged from the pistil ~3 hours after pollination and dispersed along the agarose surface for up to ~3 mm from the pistil.
Unlike previous reports [12,18], ovules were excised dry under a dissection microscope with a 27.5 gauge needle, from pistils that were held horizontally on double-sided tape (Scotch brand, 3M, St. Paul, USA). Excised ovules were immediately placed on the pollen growth medium, 2 mm from the pistil, a distance that was typically accessible by the emerging pollen tubes. To maximize pollen tube-ovule interactions, 8-10 ovules were placed at the base of a pistil as shown in Fig. 1d. Because pollen tubes tend to disburse and grow randomly after leaving the style, not all ovules, particularly those placed near the cut pistil, are visited by a pollen tube (Fig. 1e).
For time-lapse imaging, ovules were placed with their micropylar end closest to the pistil excision site. Although not essential for targeting, ovules were oriented in this manner to reduce the time elapsed before targeting was achieved. In vitro assays were typically performed by completely coating stigmas of cut pistils with pollen (>100 grains per stigma); in contrast, for the repulsion assays only 20-30 pollen grains were deposited per stigma, making it possible to clearly observe individual tube behaviour. Based on experiments with limited amounts of pollen, we typically observed 50-80% of the pollen grains produced tubes that emerged from the style.

LAT52:GFP transgenic plants
A HindIII fragment encoding GFP expressed from a postmeiotic, pollen-specific LAT52 promoter [25], was cloned into PBI121 (Clonetech, CA) and introduced into A. thaliana (Columbia) plants by Agrobacterium-mediated transformation. Kanamycin-resistant transgenic plants were selected, and a line containing a single transgene insertion, based on segregation of kanamycin resistance and GFP, was chosen for this study. This line had no detectable reproductive defects.

Microscopy
Ovules targeted by pollen tubes in the in vitro assay were counted under a Zeiss fluorescent dissecting stereoscope. For calculating targeting efficiencies, we included only ovule micropyles within 100µm of a growing pollen tube; within this range pollen tubes exhibited responses that are typical for cells undergoing attraction: significant reorientation of growth towards the signal source, followed by a steady advance towards the target. For time-lapse fluorescent microscopy, GFP-labelled pollen tubes were observed using a Zeiss Axiovert 100 fitted with an automated shutter, motorized stage and CCD camera (Cool-SNAP fxHQ, Roper Scientific, Inc Tucson, AZ). Images were captured at 10-minute intervals, converted to a TIFF format per the manufacturer's instructions using Slidebook (Intelligent Innovations Imaging, Santa Monica, CA). Pollen tube behaviours (growth rate, angle of turning, distance from micropyle) were measured and images were assembled into movies using ImageJ image analysis software (http://rsb.info.nih.gov/ij/download.html).

Diffusion rates
To estimate the size of pollen tube signalling molecules, we performed time-lapse imaging of diffusion of a series of fluorescein-conjugated dextrans (Invitrogen, Carlsbad, CA) ranging in molecular weight from 3 to 70 kD on the pollen growth medium used for performing the in vitro pollen tube guidance assay. We dissolved each dextran compound in pollen growth medium, and spotted 2 ul each of 10 ng/ul and 100 ng/ul onto pollen tube guidance assay plates. Time-lapse imaging was performed as described for in vitro pollen tube behaviours except that images were captured once in 30 minutes. The rates of diffusion (3 kD = 5.5µm/min; 10 kD = 2.73 µm/min; 40 kD = 2.43 µm/min and 70 kD = 1.01 µm/min) were measured from these images using ImageJ software. Specifically, the fluorescent intensities along a line drawn from the centre of a dextran spot to the diffused periphery were calculated for an entire time-lapse series. The data was imported into Microsoft Excel and regression analyses were performed. Extrapolating from these values, molecular weights were estimated for attractants that would diffuse 33 µm and repellents that would diffuse 27 µm in 300 minutes (the time required for a typical pollen tube to reach an ovule in the in vitro assay).

Statistical analysis
To measure the significance of the differences among observed ovule targeting efficiencies, we employed a χ 2 test for consistency in observed frequency distributions with a dichotomous classification and variable sample size [38].