Arabidopsis thaliana ecotype Columbia-0 (Col-0) and mutant eru (SALK_083442C) seeds were obtained from the Nottingham Arabidopsis stock center. Plants homozygous for the eru T-DNA insert were selected by PCR using T-DNA and gene-specific primers (Supplementary Table 1), backcrossed to the Col-0 background twice and reselected for the eru T-DNA insert.

WT and eru plants were grown in soil (Tref substrate) and kept in a growth room at 21°C in a 16/8 h light period under a light intensity of 70–90 μmol m^-2s^-1.

Pollen from stage 13–15 flowers were germinated in the dark at 21–23°C in a closed Petri dish on a thin cellophane membrane overlaying solid pollen growth medium in accordance with Rodriguez-Enriquez et al. (2012). The medium was supplemented with 0.5, 1, or 2 mM CaCl₂ or 1 mM (NH₄)H₂PO₄ when appropriate. The solution was heated in a microwave for the agarose to dissolve and cooled down to 50–60°C for the pH to be readjusted. A 0.5 cm × 0.5 cm cellophane membrane (325P cellulose; AA Packaging Limited, Preston, United Kingdom) was placed on top of the medium, and pollen were applied directly onto the membrane.

The putative 1500 bp promoter region of ERULUS was examined for cis-elements using Place¹. Public transcriptomics data was consulted using the eFP browser (Winter et al., 2007) and Genevestigator (Hruz et al., 2008).

Constructs were generated using the Gateway Cloning system² (Life Technologies). Genomic DNA was extracted from WT seedlings using phenol extraction. For the promoter::reporter gene analysis a 619bp sequence upstream of the start codon of ERULUS was PCR amplified from Col-0 genomic DNA using Platinum high fidelity DNA polymerase (Life Technologies). The following primers were used: 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTC-GCTTTGAGGTCATTTTT-3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTAATATCCGGCGAGGTTTTG-A-3′. PCR products were subcloned into pDONR207 and the sequence was verified by DNA sequencing. The promoter region was subsequently fused to the β-glucuronidase reporter using the LR reaction and the pGWB3 destination vector (Nakagawa et al., 2007). The construct was electroporated into Agrobacterium tumefaciens C58. Plant transformation was done by the flower dip method (Clough and Bent, 1998). Seeds were selected on half strength MS medium containing 50 μg ml^-1 Kanamycin or 25 μg ml^-1 Hygromycin B. Homozygous T₃ progeny was used for further analysis. Eight independent homozygous lines were examined for GUS expression.

GUS activity staining was performed in WT plants homozygous for the proERU::GUS insert according to a modified protocol of Jefferson et al. (1987).

To determine pollen germination, eru and Col-0 pollen from stage 15 flowers were grown on pollen medium and the percentage of germinated pollen was counted 16 and 24 h after pollination.

To quantify PT growth dynamics, eru and Col-0 pollen were grown on pollen medium and gently overlaid by a cover glass directly onto the medium. Time-lapse movies were collected for approximately 1 h with a framerate of 20 s using transmitted light on a Zeiss Axioplan microscope using a Zeiss Achroplan 100× (na 1.25) oil immersion Ph3 objective. The PT length gain was measured for each frame. The data for individual PTs was aligned based on their initial length, averaged, and plotted against time.

Siliques of the T₃ progeny derived from three self-pollinated heterozygous eru T₂ plants were investigated. Fifteen siliques from each plant were collected and cut in the middle to generate a top half and a bottom half as described previously (Schiefelbein et al., 1993). Due to the fact that the T-DNA-specific Kanamycin resistance gene was found to be silenced in the eru mutant and that this mutation is recessive, seeds were sown on solid RH medium and the percentage of mutants was calculated by visual screening of the obvious eru RH phenotype (Bai et al., 2014a; Haruta et al., 2014).

Stage 12 WT flowers were emasculated and stigma were hand-pollinated with WT or eru pollen from stage 13–15 flowers. Twenty-four hours after pollination, the pistils were fixed for 2 h in 1:3 acetic acid:ethanol solution, softened overnight in 8 M NaOH, cleared with distilled water and subsequently stained with decolorized aniline blue (0.1% filtered aniline blue in 0.1 M K₂HPO₄-KOH buffer, pH 11) for 2 h in the dark (Mori et al., 2006). PTs were subsequently visualized by exciting DAB-stained callose using UV lighting on a Zeiss Axioplan fluorescence microscope. More than 300 PTs originating from 10 (eru pollinated) or 13 (WT pollinated) ovaries were classified according to their targeting phenotype. The length of these PTs was quantified as the distance between the stigma’s surface and the tip of the PT.

Agrobacterium GV3101 containing the NES-YC3.6 harboring pTKan vector with UBQ10 promoter (Krebs et al., 2012) was used to transform eru and WT plants by the floral dip method (Clough and Bent, 1998). Positive transformants were selected on MS plates containing 50 mg L^-1 Kanamycin and transferred to soil. Their progeny was screened for fluorescence using a Nikon AZ100 macroscope coupled to a fluorescence unit, and seeds were collected from plants showing a clear cpVenus-based fluorescence. Note that throughout the text, for simplicity, we will use YFP when referring to cpVenus.

Pollen from plants grown in long day cycles were used for subsequent in vitro germination and calcium imaging. A setup was devised to visualize growing PTs on an inverted Nikon Eclipse Ti-E. A small cover glass was covered by a thin film of pollen growth medium (prepared fresh each day), which was overlaid with a 0.5 cm × 0.5 cm square of cellophane membrane. Agarose was chosen as a gelling agent for its lack of ‘contaminating’ calcium compared to other gelling agents, and the fact that it does not rely on the presence of bivalent cations for matrix establishment. In addition, this specific solid medium was used to more closely resemble Arabidopsis dry stigma conditions and it was confirmed that its use leads to in vivo-like pollen germination and PT growth (Rodriguez-Enriquez et al., 2012). Finally, the use of solid medium prevented sample drifting during prolonged imaging. Pollen were applied on the membrane and the cover glass was flipped on another cover glass that was attached to an opening at the bottom of a small Petri dish. Water was applied on the inner sides of the Petri dish to maintain high humidity and the Petri dish was closed using parafilm. The pollen grains were germinated in a climate controlled room at 23°C, in the dark, and visualized approximately 2 h after pollination.

YC3.6 positive PTs were visualized at the tip using a 60× oil immersion objective and excited using the Prior Lumen 200 PRO fluorescent lamp (Prior Scientific) at 440 nm (436/20 nm). Images were collected with a Hamamatsu Dual CCD Camera ORCA-D2. CFP (465–500 nm) and FRET-induced YFP (520–570 nm) emission were detected simultaneously every 5 s with an exposure time of 300–400 ms using a beam splitter (A11400-03 optical block; 483/32 nm for CFP; 542/27 nm for YFP) and Hamamatsu dichroic 510 nm mirror. Images were collected simultaneously for each channel using binning 4 × 4.

Image analysis was performed in Fiji ImageJ (Schneider et al., 2012). For both channels, all frames were aligned with respect to the PT tip using the stackreg plugin and the rigid body detection algorithm. The YFP and CFP fluorescence intensities were extracted from the very tip of the growing PT, as well as the background signal intensity. The background intensity value was subtracted independently from both YFP and CFP gray values before YFP/CFP ratio calculation. The change in the ratio (ΔR) was normalized to the average ratio of the series (<R>) and plotted versus time (ΔR/<R>). The Ratio Plus Plugin for Fiji was used to visualize the calcium oscillations. Oscillograms were analyzed by Fourier transformation, in order to isolate the main oscillation frequency.

Statistics were performed using the R statistics platform (R Core Team, 2008). Significance (α = 0.05) was assessed by two-way analysis of variance (ANOVA; parametric) using linear mixed-effects models followed by a TukeyHSD (for pairwise statistical analysis), or Kruskal–Wallis tests (non-parametric).

Recent data showed that ERULUS, a RLK from the Arabidopsis thaliana subfamily of Catharanthus roseus RECEPTOR-LIKE KINASE 1-LIKE proteins (CrRLK1Ls), is expressed in trichoblast cells and that it controls tip growth of RHs in Arabidopsis (Bai et al., 2014a; Haruta et al., 2014). In higher plants, tip growth is restricted to RHs and PTs. Publicly available micro-array data indicated that ERULUS is transcribed in pollen too (Figure 1A). We performed an in silico analysis of the 1500 bp upstream promoter region and found that the ERU promoter contains two auxin response elements (AuxREs; TGTCTC), putative binding sites for auxin response factors (ARFs) which regulate auxin-induced transcription (Ulmasov et al., 1997), and (in addition to two RH-specific cis-elements), six pollen-specific cis-elements (Supplementary Table 1).

We visualized ERU expression by GUS staining in stably transformed WT x proERU::GUS plants and found that, in the areal parts, ERU transcription is restricted to mature pollen grains and in vitro growing PTs (Figure 1B). GUS staining was not detected in any other flower part, even after prolonged staining.

Given the apparent tip growth phenotype of eru RHs and the possibility of a similar defect in growing PTs, we performed an in vitro phenotyping of eru pollen and PTs. We compared pollen viability by Alexander staining and found no difference in viability between WT and eru mature pollen grains (Supplementary Figure 1A). Next, we germinated mutant pollen under standard growth conditions. Both WT and eru pollen germinated and produced normal looking PTs (Supplementary Figure 1B). No significant difference between eru and WT was detected with respect to both pollen germination (Supplementary Figure 1C) and PT length (Supplementary Figure 1D). In addition, callose plug formation, a prerequisite for normal sperm migration in the PT (Qin et al., 2012), was unaffected in eru PTs (Supplementary Figure 1E). Bai et al. (2014a) suggested that ERULUS plays a role in ammonium homeostasis in RHs. We therefore investigated the effect of ammonium supplementation on in vitro eru and WT pollen germination and PT growth. NH₄⁺ was supplemented to the medium and the pollen germination percentage (Supplementary Figures 2A,B) and PT length (Supplementary Figures 2A,C) were quantified 2.5 h after pollen application, when PTs are still in the process of growing. We observed no significant difference between WT and eru pollen in both control and NH₄⁺ supplemented conditions, implying that excess NH₄⁺ does not affect pollen germination (Supplementary Figure 2B), PT length (Supplementary Figure 2C), and PT morphology (Supplementary Figure 2A).

Given the importance of calcium signaling in PT growth and the role of ERU in establishing a functional calcium gradient during RH tip growth (Bai et al., 2014a), we grew mutant pollen on solid medium containing 1 mM (control), 0.5 mM (low) or 2 mM CaCl₂ (high). In agreement with our previous observations (Supplementary Figure 1) we observed no significant differences with respect to average eru and WT PT growth velocities when grown under standard conditions (Figure 2A). Strikingly though, in response to 0.5 mM CaCl₂ WT PTs grew faster than in the control condition (Figure 2B; 0.70 ± 0.04 vs. 0.57 ± 0.02 μm min^-1) whereas eru PTs grew slower on lower [Ca²⁺]_ext (0.44 ± 0.03 vs. 0.70 ± 0.04 μm min^-1). As a result, eru PTs grew 37% slower compared to WT PTs at low [Ca²⁺]_ext conditions. When grown on medium with increased [Ca²⁺]_ext (2 mM CaCl₂), both WT and eru PTs grew faster than in control conditions but no relative difference in average growth rate was seen (Figure 2C). Twenty minutes tracking of individual PT growth confirmed a consistently lower growth rate throughout the acquisition time (Figures 2D,E). Together, these findings indicate that ERU has a role in regulating PT growth velocity in response to [Ca²⁺]_ext.

To further dissect ERU-mediated signal transduction in response to [Ca²⁺]_ext, we visualized the tip-focused [Ca²⁺]_cyt gradient in growing eru and WT PTs. The cytosol-targeted Yellow Cameleon 3.6 (NES-YC3.6) calcium sensor was introduced in eru and WT plants by Agrobacterium-mediated transformation and [Ca²⁺]_cyt oscillations were visualized in in vitro growing PTs under low (0.5 mM), normal (1 mM), and high (2 mM) [Ca²⁺]_ext (Figure 3). Kymograms of the recorded cytoplasmic calcium dynamics indicated that in all conditions oscillating tip-focused calcium gradients were present (Figure 3A). Fourier analysis of the acquired oscillograms revealed that WT and eru PTs showed similar oscillation frequencies (0.042 ± 0.002 and 0.039 ± 0.002 Hz, respectively) and amplitude (5.05 ± 0.12 and 4.96 ± 0.10, respectively) under control conditions (Figures 3B–E). In the presence of high [Ca²⁺]_ext (2 mM CaCl₂), oscillations occurred so rapidly that image acquisition every 5 sec could not resolve separate peaks (Figure 3A at the bottom). However, the average amplitude (5.48 ± 0.10 for WT PTs and 5.46 ± 0.10 for eru PTs) remained unaffected (Figure 3E). Strikingly, however, WT PTs responded to low [Ca²⁺]_ext (0.5 mM CaCl₂) by increasing their main oscillation frequency from 0.042 ± 0.002 to 0.048 ± 0.003 Hz (Figures 3B–D). Contrastingly, and in direct agreement with the average growth speed response, the main oscillation frequency in eru PTs decreased to 0.033 ± 0.002 Hz when subjected to low [Ca²⁺]_ext (Figures 3B–D). Together, these results show that ERU is involved in regulating apical [Ca²⁺]_cyt oscillations in response to [Ca²⁺]_ext.

Given the role of calcium-mediated cross-talk during the PTs journey through the transmitting tract toward the ovule, we investigated in vivo PT targeting of WT and eru pollen. We emasculated non-pollinated WT flowers, hand-pollinated them with WT or eru pollen and visualized in vivo grown PTs using aniline blue staining. ERU loss-of-function PTs grew through the WT transmitting tract in a WT-like manner. However, compared to WT PTs, eru PTs were more likely to show aberrant ovule targeting (Figures 4A,B). More specifically, proportionally more mutant PTs grew around the funiculus or made multiple turns before targeting the micropyle (Figure 4A). Defective targeting was independent of the PT length (Figure 4C). Hence, aberrant targeting was not skewed toward ovules that were either situated at the top or bottom of the ovary.

To understand the biological relevance of this phenotype, we quantified the transmission efficiency of the eru mutant allele versus that of the WT ERU allele. Plants that were heterozygous for the eru T-DNA insertion were allowed to self-pollinate and their progeny was screened toward the presence of the eru short RH phenotype. In addition, to gain insight into the distribution of mutant seeds along the silique, each silique was cut in half and both parts were examined separately. A smaller proportion of eru homozygotes in the bottom part of a silique would indicate slower growth of mutant PTs in vivo (Schiefelbein et al., 1993). We observed a strong deviation from the expected Mendelian inheritance ratio (25% eru^HOM_, 50% eru^HET, and 25% WT) and found consistently more seedlings with a normal RH phenotype (86.1 ± 1.4% eru^HET+ WT), and thus less homozygous eru progeny (13.9 ± 1.4% eru^HOM) in both silique halves (Figure 4D), illustrating decreased fertilization efficiency of eru pollen. We observed no difference in inheritance ratio between the top and bottom silique parts. Given the complete lack of ovular ERU transcription (shown by public transcriptomics data and proERU::GUS expression), the altered eru transmission efficiency is ought to be specifically due to a PT defect (Figures 1A,B).

Together, these data show that ERU regulates PT targeting in vivo and is an important component of the fertilization pathway in Arabidopsis thaliana.