A potential role for protein O-fucosylation during pollen-pistil interactions

ABSTRACT

Putative protein O-fucosyltransferases (POFTs) represent a large family of Glycosyl Transferase family 65 domain-containing proteins in land plants, with at least 39 proposed members in the Arabidopsis thaliana genome alone. We recently identified a member of this family, AtOFT1 (At3g05320), in which loss-of-function mutants display impaired sexual reproduction that was linked to a defective male gamete. Specifically, oft1 mutant pollen tubes are ineffective at penetrating the stigma-style interface leading to a drastic reduction in seed set and a nearly 2000-fold reduction in pollen transmission. Our findings establish that AtOFT1 plays a critical role in pollen tube penetration through the stigma/style in Arabidopsis and further suggest an important role for protein O-glycosylation events that potentially influence pollen tube mechanical strength or the ability to respond to positional guidance cues during the process of tube growth and fertilization.

Keywords:

• pollen-pistil interactions
• pollen
• biomechanics
• O-glycosylation
• protein O-fucosyltransferase

Angiosperm sexual reproduction requires cooperative interactions between the male and female tissues to facilitate a process known as double-fertilization, in which the male gamete (pollen grain) must adhere to the stigmatic papillae, hydrate to produce a pollen tube (PT), and directionally grow downward into the female pistil tissues, where the PT must locate an unfertilized ovule and deliver its two sperm cells.^1-5 The pistil provides guidance cues to lead the PT along its path to the distant ovule and supplies essential nutrients that sustain rapid PT growth through the pistil tissues.^6-10 Additionally, the PT must circumvent numerous mechanical barriers by physically penetrating through cells and tissues, such as the papillar cell cuticle, stigma, style, and the micropyle.^3-5 Recently, we characterized loss-of-function mutations in a putative protein O-fucosyltransferase (At3g05320; AtOFT1) from Arabidopsis thaliana. These mutants displayed a near-sterile phenotype, with an approximately 10-fold reduction in seed set.¹¹ This physiological defect suggested a critical role for AtOFT1 in sexual reproduction.

While manual outcrosses with wild-type Col-0 pollen on oft1 stigmas did not exhibit segregation distortion for T-DNA transmission efficiency, when oft1^± pollen was used to fertilize Col-0 stigmas, only a single transmission event of the T-DNA allele could be detected out of 1,872 progeny, which strongly suggested that the impaired fertility of oft1 plants was due to a defective male gamete. Mutant oft1 PTs displayed no overt morphological or growth defects compared to Col-0 during in vitro germination. However, after manual pollination and imaging of analine blue stained tubes growing into the pistils of a male sterile 1 (ms1) host, we observed that oft1 PTs were largely retained in the style 24 hours after pollination (HAP), while Col-0 PTs had emerged into the ovary cavity. Under semi-in vivo (SIV) conditions, oft1 PTs exhibited at least a 1-hour delay in exiting the transmitting tract (TT) of ms1 pistils compared to Col-0. After 8 hours, the number of mutant PTs exiting the style was still 3-fold lower that WT. These results indicate that the stigma and style pose a significant reproductive barrier for oft1 mutant pollen and suggest that AtOFT1 is important for facilitating PT penetration through these female tissues. We further tested this hypothesis through a stigma/style-decapitation assay in which the stigma and style were dissected from the pistils, and pollen grains were germinated on the cut surface and allowed to grow into the ovary. When this assay was performed using pollen from a heterozygous plant that was segregating wild-type and oft1 mutant pollen, a 29.1% transmission of the mutant allele was observed. This dramatic improvement compared to the 0.1% transmission observed when pollinating an intact pistil further supported the hypothesis that the stigma and style critically interfere with the reproductive success of oft1 pollen. However, the observation that there is still a reduced transmission efficiency (29% instead of an expected 50% transmission of the mutant allele) even when the stigma/style barrier is removed, suggests that oft1 pollen are still impaired in penetrating additional physical barriers or responding to chemoattractants, as discussed below.

Protein O-fucosyltransferases (POFTs) are relatively understudied enzymes in plant systems. The Arabidopsis genome contains 39 putative POFT-like genes,^12,13 and emerging evidence suggests an important role for these enzymes in plant systems, yet characterization of the biochemical activity of even a single family member has remained elusive.^14,15 The Carbohydrate Active Enzyme Database (CAZy) classifies POFTs as GT65-domain containing proteins, and in metazoan systems, the GT65 class of enzymes is best represented by the enzyme POFT1.¹³ These enzymes utilize GDP-fucose to transfer L-fucose onto serine or threonine residues of target proteins typically within defined consensus sequences residing in Epidermal Growth Factor (EGF) or Thrombospondin Repeat (TSR) domains.^16,17 These post-translational modifications impart critical regulatory functions to fundamental biological pathways, such as the Notch signaling pathway, where mono-O-fucosylation of the Notch extracellular domain is essential for potentiating Notch interactions with Delta, Serrate, and Jagged ligands.^18-22 Phylogenetic analysis of the putative Arabidopsis POFT family, revealed that AtOFT1 shares close sequence homology to metazoan POFT1s. Moreover, the crystal structure of Caenorhabditis elegans POFT1 was recently determined, which defined specific amino acid residues that participate in metazoan POFT1 catalysis,¹⁷ and protein sequence alignment with AtOFT1 showed conservation in 4 of these 5 critical residues. We mutagenized these residues in AtOFT1 and assessed each mutant’s ability to complement the seed set phenotype of the oft1 background. Each of these conserved residues was essential for AtOFT1 in vivo function. Importantly, AtOFT1 H54 aligned to C. elegans POFT1 N43. When AtOFT1 H54N was retransformed into the oft1 mutant background, seed set was restored to that of wildtype. Overall, these results indicate that catalytically important residues for metazoan POFT1s are also functionally important in AtOFT1, supporting a working model that AtOFT1 might have a similar catalytic mechanism to metazoan POFTs.

Among the Arabidopsis putative POFT family, 7 members have been previously phenotypically characterized, implicating these enzymes in maintaining proper cell wall architecture and/or cell adhesion.^{14,15,23–26} The central function of cell adhesion underlying the PT’s journey through the female tissues has long been established,^27,28 and although AtOFT1’s particular role in this process remains unclear, we propose three hypotheses for the male-specific reproductive deficiency of oft1 pollen (Figure 1). First, we postulate that AtOFT1 could help facilitate degradation of the female tissue cell wall by post-translationally activating enzymes involved in this process or potentiating their secretion (Figure 1B). Cell wall loosening molecules involved in fertilization play a role in PT invasion of female tissues in other species, such as Nicotiana tabacum and Brassica napus, suggesting that a similar degradative mechanism may be required in Arabidopsis.^3,29–32 Thus, oft1 PTs may fail to penetrate through the stigma and style due to their inability to degrade the pistil cell wall ECM. Second, AtOFT1 could conceivably participate in the synthesis of an unidentified pollen cell-wall component that facilitates PT penetration through the stigma/style (Figure 1C). Although oft1 pollen were indistinguishable from wild-type pollen in vitro, it is reasonable to speculate that a currently undefined PT cell wall rigidifying component is necessary to forcibly invade through the pistil tissues. Finally, AtOFT1 may post-translationally modify and potentiate a receptor that is necessary for responses to positional guidance cues within the stigma/style and TT (Figure 1D). Supporting this hypothesis is the similar defective penetration behavior observed in Arabidopsis VACUOLAR SORTING PROTEIN41 (VPS41) mutants, in which vsp41 PTs are proposed to be unable to recycle guidance cue receptors, resulting in arrested PT growth in the style³³. Furthermore, PTs are known to become “capacitated” during invasion through the style by undefined mechanisms that alter the PT transcriptome, induce rapid polarized tip growth, and facilitate PT capacity to locate ovules^34,35. Impairing this process may additionally explain the “unmotivated” penetration behavior of oft1 PTs.

Figure 1. Potential models for AtOFT1 behavior during pollen tube penetration.

Illustration of proposed mechanisms for the oft1 mutant pollen tube penetration defect. A, Col-0 (WT) PT rapidly traversing the transmitting tract towards the ovary, while oft1 mutant PTs (B-D) exhibit slower penetration rates through these tissues, which most often results in arrested growth. B, The oft1 PT may fail to secrete or activate female tissue cell wall degrading enzymes, thus making passage through these tissues arduous. C, oft1 PT potentially lacks a cell wall rigidifying component that aids in penetrating though the female tissue layers during fertilization. D, oft1 PT potentially lacks a functional receptor that is necessary for the PT to respond to positional guidance cues within the stylar transmitting tract (TT) or to metabolize the nutrients supplied by the TT ECM. Borders around cells are cuticle (purple), cell wall (green), and plasma membrane (orange). The location of the papillar cells and the stylar transmitting tract are indicated.

Acknowledgments

This work was supported in part by the National Science Foundation EAGER Award to ISW (Award number IOS 1449068), a National Science Foundation Award to JFH (IOS 1056774) and a National Science Foundation Graduate Research Fellowship Award to DKS.

Additional information

Funding

This work was supported by the National Science Foundation Division of Organismal Systems [1449068] and [1056774]. Additionally, this work was supported by the National Science Foundation Graduate Research Fellowship Program.