In most complex eukaryotes, development starts with the establishment of cell polarity determining the first axis of the body plan. This polarity axis is established by the asymmetrical distribution of intrinsic factors1,
Cell polarization, defined as asymmetry within a cell9, is the process in which a side or a cortical region of the cell is determined and predestined for a certain cell function or fate. During the establishment of cell polarity, intrinsic factors such as ions10, proteins11, organelles12 or maternal RNAs5 are distributed unequally and thereby determine the polarization vector (composed of a direction and a sense) in a single phase. The resulting cellular asymmetry specifies two functional cell poles that ultimately result in two different cell fates. Here we demonstrate an alternative two-phased situation, which involves distribution of intrinsic factors along a previously fixed bidirectional axis representing the predetermined direction of the polarization vector. While predetermined polarity is often the rule in animal and plant zygotes, zygotes of oogamous brown algae such as Fucus (and the related genus Silvetia) serve as excellent models for the study of polarity acquisition7. To gain further insights in symmetry breaking mechanisms, we took advantage of Dictyota (Dictyotales, Phaeophyceae), a parenchymatous brown alga, only distantly related to Fucus, with poorly described embryology.
On fertilization, the spherical Dictyota eggs change shape to a prolate spheroid in approximately 90 seconds (Fig. 1a,b,e). At the population level, the process is highly synchronized by sperm entry with all zygotes being elongated within 4 minutes after fertilization (Fig. 1e). After 8 hours, rhizoids have developed at one of the two poles of the spheroid (Fig. 1c), implicating that the direction of the polarization vector is determined by the elongation axis. The length of the zygote remains stable during the first 4 hours after elongation until the onset of tip growth (Fig. 1c,d,f).
A first cell division perpendicular to the elongation direction, 7–8 hours after fertilization, results in two cells, one developing into a thallus pole, the other becoming the rhizoid (Fig. 1c). This asymmetric cell division gives rise to unequally sized daughter cells, where the thallus cell is the largest (Fig. 1c). The young embryo will then divide perpendicular to the elongation axis (Fig. 1d). Eggs develop parthenogenetically when left unfertilized, during which they remain spherical and produce a cell wall but only rarely elongate (Supplementary Fig. 1a–e).
We also considered other external factors that might influence the polarization vector. Because Fucus zygotes determine the polarization vector according to the light direction13, we tested whether the direction of the polarization vector in Dictyota is light responsive. Thereto, eggs were fertilized under unilateral light. The direction of the polarization vector of elongated zygotes was scored and shown to be independent of the direction of the light (Fig. 2a). Likewise, parthenogenetic germlings, which tend to remain spherical, do not align the direction of their polarization vector to the light direction (Fig. 2b). Because zygotes only adhere to the substrate after they start to elongate, any environmental vector is a priori unlikely to provide a stable cue to suspended eggs and zygotes. In addition, since parthenogenetic eggs are capable of elongating without any paternal contribution (Supplementary Fig. 1f) the sperm entry site can be ruled out as the determining factor.
The cytoplasm of Fucus eggs shows no polar organization14,15. In contrast, confocal microscopy of Dictyota egg cells showed two polar regions of more intense cortical autofluorescence and a hyaline equatorial zone of less intense autofluorescence in between (Fig. 2c,d). The regions of more intense autofluorescence overlapped with dark regions observed under light microscopy. Furthermore, kernel density maps based on transmission electron microscopy (TEM) sections showed two aggregations of chloroplasts surrounding the central nucleus separated by a small region of low density in between (Fig. 2c; Supplementary Fig. 2). In all sectioned eggs (N = 6), a significant deviation from a random distribution was observed (Supplementary Table 1), implicating that chloroplasts are distributed in a bimodal pattern along the polarity axis present in egg cells. Furthermore, in vivo time-lapse microscopy confirmed that on fertilization, elongation occurred according to this preformed axis (Fig. 2d).
Although the direction of the polarization vector is independent of the direction of unilateral light (Fig. 2a), the zygotes and parthenotes of Dictyota develop their rhizoid at the pole that was oriented away from the constant unilateral light source (Fig. 3a,b). Taking the results from Fig. 2a into account, it follows that the decision on the sense of the polarization vector is not determined in the egg/early zygote but in the elongated zygote, and that both poles of the young zygote are still capable of developing into the rhizoid. If also the sense of the polarization vector would have been maternally predetermined, a pattern where the rhizoid developed in 50% of the cases towards the light vector would be expected. Moreover, when the light direction is reoriented by 180° (Fig. 3c), the rhizoid pole may still switch until approximately 5.5 hours after fertilization indicative for early zygotes being capable to produce rhizoids from both poles. In Fig. 3c different populations of cells were reoriented by 180° at different time points, allowing us to determine the timing of the permanent fixation of the sense of the polarization vector. The timing was also assessed using toluidine blue O in Dictyota zygotes, a marker for fixation of positional information in the cell wall matrix in Fucus16. One of the poles became toluidine blue O positive (Fig. 3d) and in young germlings, the cell walls at the rhizoidal pole maintained the blue/purple coloration which disappears in the distal end of the rhizoid (Fig. 3e), confirming a link between the toluidine blue O patch and the process of cell polarization rather than cell fate specification. The time point when half of the cells show a toluidine blue O patch (Fig. 3e) and incipient tip growth (Fig. 1f) coincides with the moment zygotes committed a rhizoid pole according to the first light vector (Fig. 3c). These events represent the irreversible determination of the sense of the polarization vector that occurs much later than the elongation of the zygote.
We further tested whether the determination of the sense of the polarization vector is under maternal or zygotic control by RNA-seq of the poly (A)+ RNA isolated from eggs (15 minutes after release), sperm cells (1 hour after release), zygotes (1 hour after fertilization) and asymmetrically divided embryos (8 hour after fertilization) followed by de novo transcriptome assembly and functional annotation (Supplementary Notes 1–3 and Supplementary Figs 3,4). The fractions of differentially expressed genes (Fig. 3f), ontology groups (Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and gene ontology (GO)) (Supplementary Table 4) and the shift in average expression values indicate de novo transcription starting within 1 hour after fertilization with an increasing percentage of transcribed transcripts over time (Fig. 3g). Comparison with the sperm transcriptome shows that upregulation of transcripts in zygotes relative to eggs and the increased fraction of expressed genes cannot be explained by mere introduction of transcripts by the male gametes (Supplementary Fig. 5 and Supplementary Note 5). Pairwise comparisons show that some of the upregulated ontology groups are associated with pathways such as ‘ribosome’ (ko03010), ‘spliceosome’ (ko03040) (Fig. 3h) and a diverse range of transcription and translation related GO terms are significantly upregulated by 1 hour after fertilization and/or 8 hours after fertilization (P < 0.05) (Supplementary Table 4). Despite the here reported fundamental differences with Fucus the timing of the fixation of the sense of the polarization vector relative to the length of the first cell cycle, the dependency on unilateral light and toluidine blue O staining at the rhizoid pole suggest that the polarization process of Fucus7,13,16 may closely resemble the second phase of polarization in Dictyota. The last may also consist of an axis selection and axis amplification processes that ultimately maturate in axis fixation and germination7. Moreover, the nature of the ontology classes and upregulated transcripts during polarization suggests a conserved role for classes such as calcium signalling, phosphatidylinositol signalling and small GTPases known to be main players in cellular polarization in a broad range of species17,18 (Fig. 3h; see Supplementary Note 4 and Supplementary Table 4). A putative rhodopsin photoreceptor, beta-adrenergic receptor kinase and G-protein-coupled receptor subunits show enrichment in the egg, suggesting a role in light-stimulated egg release instead of photopolarization. Instead, putative orthologues of cytochrome 1 and aureochrome 2 were upregulated in zygotes and by the end of the cell cycle respectively (see Supplementary Note 4 and Supplementary Table 4).
Maternal polarization is common among organisms; however, where present (for example, Sargassaceae, ascidians, Xenopus, angiosperm sporophyte and male gametophyte), both the sense and the direction of the polarization vector are determined maternally, not only the direction5,19,
In most animal systems early embryo development relies on parental (mainly maternal) transcripts already present in the gamete whereas the inheritance of developmental control by the zygotic nuclear genome, termed the maternal-to-zygotic transition (MZT)8, is delayed24. There is accumulating evidence for the existence of parent-of-origin genes transcripts in the embryo of land plants, but the contribution of a delayed zygotic genome activation is unclear25,
Dictyota provides a good experimental system for testing the timing of the MZT without reliance on invasive micro-dissection techniques potentially causing transcriptomic responses28,29 and the risk of seed-coat contaminations26. In addition, inhibition of transcription using 30 µg ml–1 actinomycin D demonstrated that rhizoid development is dependent on early zygotic transcription between 0 and 45 minutes after fertilization (Fig. 3i) and transcription in parthenotes (Supplementary Fig. 1g) and corroborates that MZT occurs within 1 hour after fertilization as suggested by the transcriptional changes (Fig. 3f,g; Supplementary Fig. 5 and Supplementary Table 4). The zygotic genome acquires further control over development during the later stages, as suggested by further transcriptomic changes. The cells have acquired enough mRNA of each gene by 1 hour after fertilization to acquire the capacity of rhizoid development that is robust against further transcription inhibition. However, it cannot be excluded that actinomycin D does not function during the latter stage for unknown reasons or that the zygotic cell wall becomes less permeable for actinomycin.
We identified a novel cell polarization process that can be divided in two steps relying on different informational cues and occurring during distinct life cycle generations of the organism. The direction of the polarization vector is determined under maternal control in the oogonium (gametophyte), but the sense is determined using environmental cues under zygotic control (sporophyte). The phased polarization process in Dictyota therefore represents a unique model system to unravel the multiple mechanisms that generate the axis of the future body plan in multicellular organisms.
Assaying of polarization cues
Male and female gametophytes of Dictyota dichotoma were cultured and triggered to release gametes by exposure to daylight. Photopolarization of parthenotes and zygotes (Figs 2a,b and 3a,b) was tested using unilateral light and scoring the direction of elongation or rhizoid emergence using a cross on a reticle in the eyepiece of a light microscope. For the 180° reorientation experiments, the zygotes were plated on a Petri dish and exposed to cool white unidirectional fluorescent light at approximately 60 µmol photons m−2 s−1. Petri dishes were reoriented by 180° at 1 h intervals during the first cell division. Images of egg sections were taken using TEM microscopy, photographs were loaded in QGIS and centroids of chloroplasts sectional views were marked and kernel density maps were computed. To test the hypothesis that chloroplasts are diametrically bimodally distributed, the angles between the chloroplast centroid, nuclear centroid and the x-axis of an arbitrary coordinate system were calculated. The angles were transformed to a circular scale of 180° and analysed using a Rayleigh test. Eggs were fertilized and chloroplast autofluorescence was imaged on a laser confocal imaging system (model TCS SP5, Leica) at excitation 405 nm and emission wavelengths between 555 and 625 nm.
Staining of TBO patch
Fixation of the rhizoid pole of zygotes was assayed by toluidine blue O. Zygotes or germlings were stained for 15 min with 0.1% Toluidine Blue O artificial seawater at pH 1.5. Slides were rinsed in 99% ethanol three times for approximately 5 min and once for 1 h before being mounted in tap water.
De novo assembly of transcriptome and assessment of differential expression
mRNA from three biological replicates of eggs (15 min after release), sperm cells (1 h after release), zygotes (1 h after release) and embryos (8 h after fertilization) and additional tissue (vegetative, gamete and embryonic) was extracted, cDNA libraries were sequenced using Illumina HiSeq and Roche 454 and assembled de novo with Trinity and Mira, respectively. Assemblies were joined using CD-HIT, rRNA was filtered out and redundancy was reduced, TGICL and orthology guided assembly30. Three biological replicates of >14 million reads 75 bp reads (HiSeq/NextSeq) of eggs, zygotes and embryos were mapped to the assembled transcriptome using CLC Genomics Workbench version 7.5 and transcriptome library expression values were compared using false discovery rate (FDR) corrected (α = 0.05) Baggerley tests. Enrichment for functional classes was tested using Gene Set Enrichment Analysis. Full methods and any associated references are available in Supplementary Information 1.
Raw sequence files are available under NCBI BioProject PRJNA356500.
How to cite this article: Bogaert, K. A., Beeckman, T. & De Clerck, O. Two-step cell polarization in algal zygotes. Nat. Plants 3, 16221 (2017).
The authors are indebted to the Research Foundation Flanders (FWO) (PhD fellowship to K.A.B.) and ASSEMBLE (grant agreement no. 227799). We would like to thank S.M. Coelho, M. Claeys, K. Pauly, T. Motomura, J.H. Bothwell, C. Nagasato, A. Lipinska, F. Steen, G. Zuccarello and C. Katsaros for stimulating conversations and M. Claeys and S. D'hondt for practical assistance.