The baseless mutant links protein phosphatase 2A with basal cell identity in the brown alga Ectocarpus Open Access

In most animals and plants, key events during the first cell division lead to the establishment of one or more major body axes, providing the foundation for the deployment of the subsequent steps of multicellular developmental programmes (reviewed by Radoeva et al., 2019; Rose and Gönczy, 2014). In animals, partitioning defective (PAR) proteins play a key role in establishing the anterior/posterior body axis, and a number of pathways are involved in establishing the dorsoventral axis (e.g. Mongera et al., 2019). In the land plant Arabidopsis, both the apical/basal axis and apical/basal cell identities are established at the time of the first cell division, and genetic analysis has identified two main pathways involved in this process (reviewed by Bayer et al., 2017; Ueda and Berger, 2019). The first pathway involves SHORT SUSPENSOR (an interleukin-1 receptor-associated kinase/Pelle-like kinase), YODA (a MAP kinase kinase), MPK3 and MPK6 (MAP kinases) and downstream transcription factors, which may include WRKY2 and GROUNDED/RKD4. This pathway may also be influenced maternally through secreted peptide factors, such as EMBRYO SURROUNDING FACTOR 1 and CLV3/ESR-RELATED8. The second pathway involves auxin and consists of PIN-FORMED 7 (auxin efflux regulator), MONOPTEROS/AUXIN RESPONSE FACTOR 5 (transcription factor) and BODENLOS/INDOLE-3-ACETIC-ACID 12 (auxin response inhibitor). The establishment of zygote polarity is a pre-requisite of the asymmetrical first cell division. This process involves movement of the nucleus and other organelles, enlargement of the vacuole and reorganisation of microtubules (Kimata et al., 2016; Ueda and Berger, 2019).

The mechanisms underlying the onset of early development from an initial cell in multicellular plants and animals are relatively well understood, but research has lagged behind for the third most complex group of multicellular eukaryotes, the brown algae. The brown algae offer an interesting contrast to animals and plants because of their phylogenetic position and the fact that they evolved complex multicellularity independently from those groups. Moreover, in brown algae, the two generations are independent and often very distinct morphologically. The same genome, therefore, regulates the set-up of two independent and distinct developmental programmes from two different types of initial cells, opening interesting questions about the molecular control of alternation of generations (Arun et al., 2019; Cock et al., 2014; Coelho et al., 2007, 2011). Furthermore, gametophytes and sporophytes of brown algae develop from single cells outside the parent organism, indicating that they likely establish polarity in a cell-autonomous manner, without the involvement of factors delivered from the parental tissues, simplifying the study of polarity, axis establishment and pattern formation.

Investigations using the brown alga Fucus have shown that asymmetrical first cell division is driven by apical-basal polarity established within the zygotic cell (Brownlee and Bouget, 1998; Goodner and Quatrano, 1993; reviewed by Bogaert et al., 2023). The two daughter cells of this first division divide to produce the apical and basal systems of the alga: the thallus and the rhizoid, respectively (Brownlee and Bouget, 1998). Studies using Fucus zygotes have underlined the role of Ca²⁺ asymmetries, mRNA distribution and position-dependent information from the cell wall (involving an unknown diffusible apoplastic factor) in basal and apical system fate determination (Berger et al., 1994; Bogaert et al., 2023; Bouget et al., 1996; Brownlee and Bouget, 1998).

More recently, Ectocarpus has emerged as a suitable model in which to investigate the molecular mechanisms underlying initial cell division and cell fate determination in brown algae (Bogaert et al., 2023). This model has the advantage of availability of a range of genetic and genomic tools (Avia et al., 2017; Badis et al., 2021; Bourdareau et al., 2021; Cock et al., 2017; Coelho et al., 2020; Cormier et al., 2017; Umen and Coelho, 2019), but also, importantly, the regularity of the first cell division that characterises the early stages of development of both the gametophyte and sporophyte generations. Ectocarpus, as many brown algae, alternates between a gametophyte and a sporophyte generation, both being multicellular and independent (Fig. 1). Male and female gametophytes produce male and female gametes by mitosis, which fuse to produce the (diploid) sporophyte. In absence of fusion with gametes of the opposite sex, germinating parthenogenetic gametes deploy the sporophyte programme, despite being haploid, to produce partheno-sporophyte algae that are indistinguishable from diploid sporophytes in terms of morphology. Life cycle generation (i.e. deployment of a gametophyte or a sporophyte body plan) is therefore not determined by ploidy (haploid or diploid phase) and these two features of the life cycle may be uncoupled (Bothwell et al., 2010a; Coelho et al., 2020). In Ectocarpus, the gametophyte generation exhibits an asymmetrical initial cell division that produces a basal rhizoid cell and an apical cell, the latter further dividing to form the apical system of upright filaments. The upright filaments bear the reproductive structures (plurilocular gametangia, which produce the gametes by mitosis). In contrast, the initial cell of the sporophyte generation undergoes a symmetrical initial cell division to produce two daughter cells with similar fates, both being components of the basal system (Godfroy et al., 2017; Peters et al., 2008) (Fig. 1). The apical system of the sporophyte (upright filaments) is produced later, after an extensive system of basal filaments has been established. Reproductive structures (unilocular and plurilocular sporangia) are produced on the upright filaments.

Earlier work identified an Ectocarpus mutant, distag (dis), with abnormal cellular features during the first cell division, and that was unable to develop basal systems (rhizoids in the gametophyte, prostrate filaments in the sporophyte) (Godfroy et al., 2017). dis mutant alleles exhibit a strong phenotype in the initial cell, with disordered microtubule networks, larger cell size, altered Golgi structure and mispositioned nuclei and centrioles (Godfroy et al., 2017). The cell division plane, however, is normal and the cellular defects are only observed during the division of the initial cell, suggesting that DIS function may be specific to this cell. DIS encodes a Tubulin Binding Cofactor C (TBCC) domain protein of the TBCCd1 class, with conserved roles in animals and plants (André et al., 2013; Feldman and Marshall, 2009).

Here, we report the identification of the BASELESS (BAS) locus in Ectocarpus. bas mutants exhibit phenotypes that closely resemble those of dis mutants, including an atypical initial cell division that leads to failure to deploy a basal system in the adult organism, increased cell size, abnormal cellular features, such as disorganised microtubule cytoskeleton, loss of bipolar germination and more extensive Golgi apparatus compared with wild-type cells. These phenotypic features are associated with modified patterns of gene expression during early stages of development. BAS encodes a protein phosphatase 2A regulatory subunit type B″ with EF-hand domains. Together, our results are consistent with BAS being involved in a pathway that plays a key role in initial cell division and basal cell fate determination during both the gametophyte and sporophyte generations of the Ectocarpus life cycle.

During the Ectocarpus gametophyte generation, the two cells derived from the division of the initial cell develop as two germ tubes, and establish a rhizoid (a basal, root-like organ) and an upright filamentous thallus (an apical, shoot-like organ) (Fig. 2A,B). The first cell division of the sporophyte generation, in contrast, produces two daughter cells with similar morphology and equivalent cell fates (Fig. 2C). These two cells then divide to produce a prostrate filament, which branches to establish the basal system, and the apical system differentiates later in development, growing up into the medium from the extensive, prostrate basal system. Reproductive structures are produced on the apical system (Fig. 2D,E).

A screen of a large population of individuals mutagenised by ultraviolet (UV) irradiation identified two mutant strains (Ec800 and Ec801; Table S1) that failed to develop any of the basal structures normally observed during either the gametophyte or the sporophyte generation of wild-type strains (Fig. 2). The screen used gametes, which, in absence of fertilisation, develop into partheno-sporophytes, being thus initial cells of the sporophyte generation (Fig. 1) (Coelho et al., 2011; Godfroy et al., 2015, 2017; Peters et al., 2008). Initial cells of Ec800 and Ec801 gametophytes immediately developed as apical upright filament cells and no rhizoid cells were produced. Similarly, during the sporophyte generation, neither mutant strain produced the network of prostrate basal filaments typical of the wild-type sporophyte and, instead, the first divisions of the initial cell directly produced an upright filament (Fig. 2).

In wild-type Ectocarpus, secondary rhizoids, which are analogous to the adventitious roots produced from the stems of some land plants (Atkinson et al., 2014), are produced from apical upright filament cells at a late stage of development (Fig. 2E) (Peters et al., 2008). The Ec800 and Ec801 mutants did not produce secondary rhizoids (Fig. 2, Fig. S1). Hence, production of all basal attachment structures, both primary and secondary, was abolished in these mutants. Taking into account these phenotypes, the Ec800 and Ec801 mutants were named baseless-1 (bas-1) and baseless-2 (bas-2), respectively.

The establishment of reproductive structures on apical systems in both the gametophyte and sporophyte generations was unaffected in the bas-1 mutant, which was fully fertile after 3 weeks in culture (Fig. 2F-S). In the bas-2 mutant, the formation of the plurilocular sporangia (which contain mito-spores) was unaffected, whereas unilocular sporangia (where meiosis takes place) aborted and no functional meio-spores were produced, preventing the generation of gametophytes (Fig. 2T).

In summary, we identified two Ectocarpus lines showing abnormalities in the production of basal structures during both sporophyte and gametophyte generation. One of the lines presented, in addition, meiotic defects.

In wild-type Ectocarpus, the majority of the initial cells (91%) exhibited a bipolar pattern of germination, with two germ tubes emerging from opposite poles of the initial cell (Fig. 2U; Peters et al., 2008). In contrast, only 21% of the initial cells of bas-1 partheno-sporophytes exhibited this bipolar pattern of germination, the remaining 79% undergoing unipolar germination (Fig. 2U). A proportion of the bas-1 partheno-sporophytes that exhibited a bipolar germination pattern produced one or more enlarged and abnormally shaped cells at the extremity where the second germ tube would normally emerge, possibly corresponding to an aborted germ tube (Fig. 2O). Similar phenotypes were observed for bas-2 partheno-sporophytes (Table S2). Overall, thus, both bas-1 and bas-2 presented defects in germination patterns.

Mutations at the DIS locus strongly affect the organisation of the microtubule cytoskeleton (Godfroy et al., 2017). Because of the similarity between the morphological phenotypes of bas and dis mutants, we investigated the distribution of the microtubule network during early development of bas mutants compared with wild-type germlings (Fig. 3A,B). The microtubule cytoskeleton was markedly disorganised in the bas mutants, with supernumerary microtubule filaments and a highly disordered network (Fig. 3C). This microtubule phenotype is reminiscent of that of the dis mutant (Godfroy et al., 2017). Also, similarly to the dis mutant, we did not detect any abnormalities in the positioning of the cell division plane during early development; all bas initial cells produced a cell division plane perpendicular to the growing axis (Fig. 2U). Thus, the microtubule cytoskeleton was highly affected both in bas and dis mutants, but the initial cell division plane remained unaffected.

Transmission electron microscopy (TEM) and focused ion beam–scanning electron microscopy (FIB-SEM) were used to characterise further the cellular architecture of bas mutants. We focused on the early development of the sporophyte generation, i.e. when unfused gametes had started developing into partheno-sporophytes (two to five cells), which is the stage at which conspicuous morphological differences were observed (Fig. 2). This is also the stage when dis mutants exhibit altered subcellular phenotypes, including significantly more abundant cisternae and more fragmented Golgi compared with wild-type cells (Godfroy et al., 2017).

Germinating initial cells of bas exhibited increased cell area compared with wild-type initial cells (Fig. S2), a phenotype similar to that of dis mutants (Godfroy et al., 2017). Morphometric analyses of the subcellular structures in the bas mutant and in wild type indicated that the Golgi apparatus were about twice as numerous in bas as in wild type at the same developmental stage, but this difference was not significant (Fig. 4A-E, Fig. S3, Table S3). We also noticed that the number of cisternae per Golgi tended to be slightly higher in bas (Fig. 4F, Table S3, Fig. S3), yet bas mutant cells did not exhibit Golgi cisternae fragmentation (Wilcoxon test, P=0.99202; Table S3, Fig. S4), in contrast to dis mutants (Godfroy et al., 2017). Therefore, bas exhibited a Golgi defect, but this defect appeared to be less conspicuous than that of dis mutants, in which Golgi was strongly fragmented. TEM imaging did not reveal any other evident organellar defects in bas cells, such as abnormal structure or position of the nucleus or centrioles (Fig. 4G, Fig. S5). Note that dis mutants show abnormal nuclear and centriolar position during germination of the initial cell (Godfroy et al., 2017). Moreover, no visible defect, in particular at the Golgi, could be observed in the gametes prior to their release from the plurilocular gametangia (Fig. 4G). These observations indicate that bas cellular defects are detectable only once the initial cell initiates germination. A summary scheme illustrating the development of wild type versus bas mutants in the gametophyte and sporophyte generation is presented in Fig. S6.

A male bas-1 gametophyte (Ec800) was crossed with a wild-type female gametophyte (strain Ec25; Table S1, Fig. S7). The resulting sporophyte (Ec805) exhibited a wild-type pattern of development, indicating that the bas-1 mutation was recessive. A segregating family of 38 partheno-sporophyte individuals derived from this cross consisted of 16 and 22 phenotypically wild-type and mutant individuals, respectively, consistent with a 1:1 segregation ratio and Mendelian inheritance of a single-locus recessive mutation (χ² test with Yates correction=0.4767, P=0.4899; Table S5). The bas-1 mutation segregated with the phenotype in the 38 individuals used for the phenotype segregation analysis (Table S4).

The phenotypes of bas-1 and bas-2 strongly resembled that of the dis mutant (Godfroy et al., 2017). The dis mutant also fails to produce any basal structures, during both the sporophyte and gametophyte generations, and lacks secondary rhizoids, again during both generations. Sporophytes resulting from crosses between the bas-1 strain Ec800 and strains carrying either the dis-1 or the dis-2 allele resembled wild-type sporophytes (Fig. S8, Table S1), indicating complementation, and therefore that the DIS gene was not mutated in the bas-1 mutants. However, although heterozygous sporophytes had a wild-type phenotype during vegetative growth, none of the heterozygous sporophytes (Ec826, Ec827, Ec828, Ec829, Ec830 and Ec831) produced unilocular sporangia at maturity, suggesting a meiotic defect. Note that in Ec826, Ec827, Ec828 and Ec829, bas was the female partner, whereas in Ec830 and Ec831 it was the male that was bas. Therefore, the sexual background of the mutant did not change the meiotic defect. The lack of meiosis in heterozygous sporophytes precluded the analysis of any meiotic offspring to further investigate pleiotropy or additivity between DIS and BAS.

Genome resequencing identified a candidate locus on chromosome 21 for the location of the bas-1 and bas-2 mutations (Fig. 5). Whole genome resequencing (WGS) was carried out for the Ec800 (bas-1) and Ec801 (bas-2) mutant strains and the data compared with the wild-type Ectocarpus sp. strain Ec32 reference genome. More than 41,000 putative variants were detected for each strain. Those variants were compared with a list of 567,532 variants called during the analysis of 14 other mutant lines that showed a range of different phenotypes (Table S5) and shared variants were eliminated. This approach allowed the identification of 827 and 769 variants that were unique to the Ec800 and Ec801 mutants, respectively. Quality filtering of those variants (see Materials and Methods for details) resulted in 118 and 67 putative mutations for the Ec800 and Ec801 strains, respectively, corresponding to one mutation every 1.7-3.0 Mb of genome. Of these 185 putative mutations, 26 and 15 were in coding regions (CDS) (22%) in Ec800 and Ec801, respectively. Only one gene (locus ID: Ec-21_001770) contained a CDS mutation in both strains. A single nucleotide transition from T to C, at position 2,806,985 was identified in the bas-1 mutant (strain Ec800) and a G-to-A transition at position 2,807,321 in the bas-2 mutant (strain Ec801) (Fig. 5). The bas-1 mutation segregated with the phenotype in the 38 individuals used for the phenotype segregation analysis (Table S6).

The Ec-21_001770 gene encodes a protein of 646 amino-acids similar to protein phosphatase 2A regulatory subunit type B″ proteins. This polypeptide contains three predicted functional domains: a disordered region between positions 50 and 185 and two EF-hand domains at positions 280-370 and 380-550. The bas-1 mutation affects the first EF-hand, replacing a positively charged lysine residue with a negatively charged glutamic acid (K302E). This modification of electric charge may disrupt domain folding and/or function at least for the first EF-hand. It is possible that the bas-1 mutation leads to the production of a protein that is partially active. The bas-2 mutant carries a non-sense mutation that creates a premature stop codon at position 190 of the protein. This mutation is predicted to result in the production of a truncated protein that lacks both EF-hand domains (Fig. 5A).

BAS is predicted to encode a protein phosphatase 2A regulatory B″ subunit. PP2A phosphatases are protein complexes usually composed of three subunits: a catalytic C subunit, a scaffolding A subunit and a regulatory B subunit (Wlodarchak and Xing, 2016). Most species have multiple forms of each subunit and there are four distinct classes of the B subunit (B/B55/PR55, B′/B56/PR61, B″/PR72/PR130 and B‴/Striatin), which are unrelated at the sequence level. An analysis of the Ectocarpus sp. genome revealed that it encodes B, B′ and B″ subunits, but not B‴/Striatin (Fig. 5B). The BAS protein is predicted to belong to the B″ class, homologous to the human protein PPP2R3A (also known as PR130 or PR72) (Fig. 5B).

RNA-sequencing (RNA-seq) data (Bourdareau, 2018; Coelho et al., 2011; Godfroy et al., 2017; Gueno et al., 2022; Macaisne et al., 2017) were analysed to investigate BAS gene expression during the Ectocarpus life cycle. BAS transcripts were detected throughout development, during both the gametophyte and sporophyte generations (Fig. 6A). The BAS transcript was most abundant at the gamete stage (after the release from plurilocular gametangia) and had decreased in abundance about 1 week after germination, at the two- to five-cell stage (Fig. 6A; Table S7). This pattern of expression is consistent with a role of BAS in the early divisions of the initial cells of the partheno-sporophyte generation, provided that the transcript and/or protein persists in the initial cell during the first cell division. Interestingly, during the life cycle, the abundance of the BAS transcript was inversely related to the abundance of the DIS transcript, which was at a very low level in gametes but increased in abundance at later stages of development (Fig. 6, Tables S7, S8).

The similar phenotypes of dis and bas mutants suggest that the products of the two genes may play roles in common cellular processes. We investigated the expression of DIS in a bas background, and, conversely, the expression of BAS in a dis background. We noticed that in a bas background, DIS expression in early stages (two to five cells) was significantly higher (P=4.13E−02 and P=7.66E−02, respectively, for bas-1 and bas-2 comparison with wild type at the two- to five-cell stage, Fig. 6B; Table S7), whereas no difference was observed in the levels of expression of BAS in the absence of dis gene product. Taken together, these analyses suggest that BAS expression levels are not affected by DIS, but, conversely, DIS gene expression is disturbed by mutations at the BAS locus. This observation, however, is unlikely to explain the phenotypic similarity between bas and dis mutants.

To characterise the bas phenotype further, an RNA-seq approach was employed to study gene expression in the bas-1 and bas-2 mutants compared with wild type. Because of the severe impact of the bas mutations on the general thallus architecture at the adult stage (lack of basal structures) and the prominence of the bas phenotype during the early stages of development, we focused on the two- to five-cell stage during germination of the initial cell. Given the subtle differences in phenotype between bas-1 and bas-2 (in particular the meiotic defect in bas-2), we also focused on comparing bas-1 and bas-2 transcriptional patterns.

Overall, gene expression patterns in bas-1 and bas-2 were comparable, and both were different from the wild-type samples (Figs S9-S11, Tables S7-S9). However, bas-2 exhibited more differences compared with wild-type samples (Fig. S11): 40% of the transcriptome was differentially regulated in the comparison bas-2 versus wild type, whereas 20% of the genes were differentially expressed (DE) between bas-1 and wild type (Fig. S11, Table S8; see Materials and Methods for thresholds). Most genes were expressed under all conditions; only about 4% of the genes were not expressed in any of the samples.

DE genes exhibited very similar patterns of up- and downregulation in the two bas mutants compared with wild type. Only 11 genes exhibited divergent expression patterns, i.e. upregulated in bas-1 and downregulated in bas-2, and six genes were upregulated in bas-2 and downregulated in bas-1. The vast majority of the DE genes in bas-1 versus wild type were also differentially expressed in the bas-2 versus wild type comparison (76.69% of the downregulated genes and 76.73% of the upregulated genes) (Fig. 7A, Table S9).

Downregulated genes in bas-1 compared with wild type had, globally, lower expression levels in bas-2, and upregulated genes in bas-1 were slightly more expressed in bas-2 (Fig. 7B). Conversely, downregulated genes in bas-2 compared with wild type had expression levels in bas-1 similar to that of wild-type samples, whereas upregulated genes in bas-2 had intermediate expression levels compared with wild type and bas-2 (Fig. 7B, Table S9). Comparison of the fold-change distribution of the DE genes from bas-1 versus wild type with those of the DE genes from bas-2 versus wild type showed that DE genes exhibited greater fold changes in bas-2 than in bas-1 (Fig. 7C,D).

Overall, these observations indicate similar changes in transcriptome of both bas mutants, with a stronger effect in bas-2 than in bas-1 compared with wild type. Those results are in agreement with the predicted effects of the two mutations on the BAS protein. The bas-1 mutation may result in the production of a protein partially active, whereas the bas-2 mutation leads to a truncated protein with probably no activity.

The more severe phenotype of bas-2, compared with that of bas-1, could be due to additional mutations affecting other genomic loci and not a direct consequence of the bas mutation per se. In order to test this possibility, we analysed the genomes of bas-1 and bas-2 and compared the numbers of mutations in the two mutant genomes. We found that the bas-2 mutant had fewer mutations (65 in total, 14 in CDSs) than bas-1 (118 total, 27 in CDSs) (Table S5). In terms of the type of mutation, the proportion was similar between the two mutants: about 21-23% of the mutations affect CDSs, of which 40-50% were silent and 50-60% were missense (Table S5). The only non-sense mutation that was identified in bas-2 was the one that affected the BAS locus. Five genes had missense mutations, but none had a function that could be associated with meiosis. Furthermore, the functional annotation of those five genes with InterProScan is not modified by the mutations, suggesting that none of those mutations directly affect a key amino-acid residue. In conclusion, our data are consistent with the hypothesis that the phenotypic severity of bas-2 strains compared with bas-1 strains is due to the relative severity of the bas mutations caused by each allele, but cannot rule out mutations at other loci contributing to the bas-2 phenotype.

Gene Ontology (GO) term enrichment analyses of DE genes revealed that functions related to photosynthesis and metabolism were downregulated in both bas mutants, whereas upregulated genes in mutants were associated with intracellular protein transport transcription and protein synthesis (Fig. S12, Table S10). The downregulation of photosynthesis-related genes may reflect both the impact of the mutation on the algae photosynthetic metabolism, but could also reflect the lack of the basal, round cells, which are typically more pigmented than apical cells and rhizoids (Charrier et al., 2008 and references therein). It is interesting to note that some of the intracellular protein transport transcription and protein synthesis functions can be linked with the Golgi apparatus, which appears to be affected by the bas mutation (see above). Using the HECTAR predictor (Gschloessl et al., 2008), we examined the enrichment of DE genes in particular subcellular localisations. In coherence with GO term enrichment analyses, we observed an enrichment in ‘chloroplast’ localisation of the downregulated genes in both bas mutants. We also observed a slight, but strongly significant, enrichment in proteins with a ‘signal peptide’ suggesting an impact of the bas mutation on the production of secreted proteins (Table S11), which, again, may be consistent with defects observed in bas mutants at the level of the Golgi. Finally, we noted that 13 out of the 26 predicted BASIC LEUCINE-ZIPPER (bZIP) transcription factors in Ectocarpus are differentially expressed in the bas-2 mutant compared with wild type (Table S9).

We identified 26 and 41 genes that were exclusively expressed in the bas-1 and bas-2, respectively, compared with the wild type (i.e. genes that had 0 transcripts per million in wild-type samples). Seventeen genes were silenced in the bas-1 and 29 in bas-2 compared with wild type (Table S12, S13; see Materials and Methods for thresholds). Most of those genes have unknown functions and, owing to the low numbers of genes, no significant functional enrichment could be established. However, it is worth noting several transcription factors, which may be potential effector genes: Ec-14_003940 and Ec-16_000350, which are silenced in bas-1 and/or bas-2; Ec-00_004890, which is activated in both mutants compared with the wild type; and numerous genes associated with ‘Cellular regulation and signalling’ or ‘Membrane function and transport’. Looking at the putative localisation of those proteins, we observed that about 30% of the genes specifically silenced in bas-1 and/or bas-2 have ‘signal peptide’ or ‘signal anchor’ prediction, suggesting a membrane or cell-wall localisation, whereas only 16% of the total predicted proteins in the genome of Ectocarpus presented this signature (Table S12). This two-fold enrichment is even higher than the ‘signal peptide’ enrichment found in downregulated genes (Table S11).

The two Ectocarpus mutant alleles identified in this study, bas-1 and bas-2, lack basal structures during both the gametophyte and the sporophyte generations of the life cycle. Analysis of the initial cells of the bas mutant showed that the morphological phenotype was associated with several cellular anomalies during germination and the first cell division, including disorganisation of the microtubule network, an increase in the number of microtubule bundles and Golgi apparatus, and unipolar, rather than bi-polar, germination patterns. These observations highlight a key role for the BAS protein during the development of Ectocarpus. In particular, the BAS protein appears to operate during key cell divisions during development: the (first) initial cell division and, at later stages of development, during meiosis. However, no cellular defect was observed prior to initial cell divisions, suggesting that BAS operates only after release of the initial cells from the reproductive structures. The meiotic defect observed in the bas-2 mutant and in heterozygous sporophytes produced from dis×bas-2 mutants suggests that the BAS protein may have also a role during meiotic cell division. The absence of this meiotic defect in the bas-1 may be explained by the production of a protein with sufficient activity in bas-1 to ensure its role during meiosis but not during the first cell division. Higher penetrance of the bas-2 mutation is also indicated by the greater proportion of differentially expressed genes in bas-2 compared with bas-1. Indeed, about 40% of the genome was significantly mis-regulated in the bas-2 mutant during early stages of development, demonstrating that mutations at the BAS locus can lead to broad, large-scale modifications to the transcriptome. It is interesting to note that half of the bZIP transcription factors of Ectocarpus are differentially expressed in bas-2 compared with wild type. This observation suggests that BAS may be part of a pathway involving additional regulatory proteins that drives the establishment of the apical-basal axis during the early development of Ectocarpus.

Ectocarpus BAS is predicted to encode a PP2A phosphatase regulatory B″ subunit. In animals, PP2A phosphatases are involved in diverse cellular processes and constitute a major component of cellular serine/threonine phosphatase activity, dephosphorylating several hundred cellular substrates (Reynhout and Janssens, 2019; Wlodarchak and Xing, 2016). PP2A has been implicated in the reorganisation of several cellular structures, playing key roles in nuclear envelope breakdown during mitosis, and in chromosome segregation via effects on assembly of the mitotic spindle and attachment of cytoplasmic microtubules to kinetochores. It is also involved in rearrangement of endoplasmic reticulum and the Golgi apparatus (reviewed by Wlodarchak and Xing, 2016; Wurzenberger and Gerlich, 2011). In particular, the PP2A B″ subunit PR130 associates with CG-NAP (AKAP9), which localises to centrosomes and the Golgi apparatus (Takahashi et al., 1999), and restacking of newly formed Golgi cisternae requires dephosphorylation of Golgi stacking proteins by PP2A (Tang et al., 2008). More broadly, the animal PP2A B″ subunit is crucial for cell–cell communication, cell adhesion, migration, proliferation and differentiation during animal development (Creyghton et al., 2005; Zwaenepoel et al., 2010). Altogether, these observations link animal PP2A to roles in cell division, subcellular features and developmental pattern formation. In the multicellular brown alga Ectocarpus, our results are consistent with a similar role for PP2A in subcellular organisation, cell division and cell fate determination. Our results therefore provide an example of protein functional conservation across eukaryotes that evolved multicellularity independently, suggesting that the role of PP2A in multicellular development has likely been preserved across very divergent lineages.

Consistent with a conserved role for PP2A in development across multicellular eukaryotes, the Arabidopsis B″-δ/ε subunit of PP2A (At5g28900/At5g28850) interacts with bZIP transcription factors, and is implicated in leaf and root development as well as mechanical stress response (Tsugama et al., 2019; Van Leene et al., 2016). The PP2A regulatory B″ subunit FASS/TON2, is essential for the reorganisation of cortical microtubular arrays into a dense preprophase band preceding cell division. FASS-containing PP2A complexes are targeted to microtubules through an association with TONNEAU1 (TON1) and TON1-recruiting motif protein (TRM) (Spinner et al., 2010). PP2A interacts with and dephosphorylates KATANIN, an evolutionarily conserved microtubule-severing enzyme, to promote the formation of circumferential cortical microtubule arrays in Arabidopsis (Ren et al., 2022). The Ectocarpus BAS protein is related to the Arabidopsis FASS/TON2 protein but is orthologous to the AtB″-δ/ε protein mentioned above (At5g28900/At5g28850). The microtubule phenotype we describe here, in which PP2A B″ disruption in brown algae is associated with microtubule disorganisation, may further underline a conserved role for PP2A across plants and brown algae.

Golgi vesicle transport has been shown to play an important role in the establishment of the Fucus zygote polar axis prior to the first cell division (reviewed by Bogaert et al., 2023; Shaw and Quatrano, 1996). It has been proposed that the selective targeting of Golgi vesicles to the plasma membrane locally modifies the cell wall, participating in the establishment of the cell wall asymmetry required for rhizoid differentiation (Bogaert et al., 2013, 2023; Goodner and Quatrano, 1993). In this model, the cell wall plays a key role in the establishment of the initial cell asymmetry, which is interesting with regard to the dramatic changes in expression of numerous genes with putative ‘Membrane and transport’ functions in bas mutants.

In animals, the B″ class of PP2A B subunit is thought to be involved in Ca²⁺ signalling through the presence of multiple EF-hand domains (Xu et al., 2010). Note that two EF-hand domains are predicted in the BAS protein, and these EF-hands are absent or non-functional in bas mutants, suggesting that Ca²⁺ binding may be disturbed. In the brown alga Fucus, Ca²⁺ gradients have been shown to have a crucial function in zygote polarisation and in establishment of the apical/basal axis during the first cell division (reviewed by Bogaert et al., 2023; Brownlee and Bouget, 1998). In particular, cytosolic free Ca²⁺ accumulates on the side of the growing (basal) rhizoid, directly linking intracellular Ca²⁺ signalling and acquisition of basal cell identity. In this context, we speculate that the BAS protein may be a potential sensor of a pre-germination Ca²⁺ signal.

Similar morphological and cellular phenotypes were observed for bas and dis mutant strains, suggesting that BAS and DIS may be involved in similar cellular processes. However, some differences between the mutants are also apparent, such as the less marked Golgi fragmentation and no effect on nuclear positioning in early stages of development in bas, and also a meiotic defect was only observed in bas-2.

DIS is predicted to encode a TBCCd1 protein. This protein shares similarity with TBCC, which is a component of the complex (TBCA to TBCE) that mediates dimerisation of α and β tubulin subunits to form microtubules (Nithianantham et al., 2015; Tian et al., 1996). However, TBCCd1 lacks a conserved arginine residue that is essential for TBCC activity and is unable to complement TBCC in yeast, indicating that the two proteins may have different biochemical functions (Goncalves et al., 2010). TBCCd1 has been localised to both the centrosome and the Golgi in humans, Chlamydomonas and trypanosomes, and there is evidence that TBCCd1 plays important roles in positioning organelles within the cells of these diverse organisms (André et al., 2013; Feldman and Marshall, 2009; Goncalves et al., 2010). However, the molecular mechanisms underlying these cellular phenotypes are unclear and they may not involve direct effects on microtubule assembly (Goncalves et al., 2010).

Interestingly, analysis of human protein phosphatase interactions revealed that TBCCd1 is a partner of PP2A regulatory subunit B″ (Huttlin et al., 2015, 2017; Yadav et al., 2017). If this interaction also occurs in brown algae, it would provide a mechanism whereby DIS and BAS could act within the same pathway involved in cell fate determination, with BAS regulating the DIS protein. Further analysis of the biochemical functions of BAS and DIS will be necessary to test this hypothesis.

To summarise, both TBCCd1 and PP2A have been linked to cytoskeleton and Golgi functions and both proteins have been shown to play important roles in the regulation of cellular architecture in diverse eukaryotic systems. These observations are consistent with the pleiotropic cellular phenotypes of both the dis and bas mutants. We suggest that the observed morphological and cell fate (loss of basal cells) phenotypes of the bas and dis mutants are a consequence of cellular defects during the first cell division, perhaps through disruption of the distribution of hypothetical cell fate-determining factors during this critical step of development (see model proposed by Godfroy et al., 2017). Combining information about the Ectocarpus DIS and BAS proteins with observations of Ca²⁺ waves in the Fucus embryo, we speculate that BAS may be involved in sensing an intracellular Ca²⁺ signal that participates in the distribution of a cell fate-determining factor through reorganisation of cytoskeleton and Golgi function involving DIS.

Strain cultivation, genetic crosses, raising of sporophytes from zygotes, and isolation of meiotic families were performed as described previously (Coelho et al., 2012a,b, 2020; Godfroy et al., 2017). Ectocarpus sp. (species 7; Montecinos et al., 2017) gametes are able to develop parthenogenically to produce haploid partheno-sporophytes, which are identical morphologically to the sporophytes that develop from diploid zygotes (Bogaert et al., 2023; Bothwell et al., 2010b; Peters et al., 2008). This phenomenon was exploited to screen directly, in a haploid population, for mutants affected in early sporophyte development. UV mutagenesis of gametes was performed as described previously (Coelho et al., 2011; Godfroy et al., 2015, 2017) and mutant partheno-sporophytes lacking basal structures were identified by visual screening under a light microscope. Table S1 describes the strains used in this study.

Genetic crosses were performed as described by Coelho et al. (2012a). The bas-1 mutant (Ec800) was crossed with the outcrossing line Ec568 to generate a segregating population of 38 individuals. Each of the 38 individuals was derived from a different unilocular sporangium (each unilocular sporangium contains 50-100 meio-spores, derived from a single meiosis followed by at least five mitotic divisions). The meio-spores germinated to produce gametophytes, which were isolated and allowed to produce gametes that germinated parthenogenically. The resulting partheno-sporophytes were then observed under a light microscope to determine whether they exhibited the bas phenotype. The presence of the bas-1 mutation was determined by Sanger sequencing of PCR products (forward: TGACGAATGATGCTAAACTGGA; reverse: GACAACGGAGCAGACGAAC) for each of the 38 individuals. The bas-2 mutant (Ec801) was not usable for crosses because it did not form functional unilocular sporangia, which are required for gametophyte production.

Genomic DNA from Ec800 and Ec801 strains was sequenced on an Illumina HiSeq4000 platform (2×150 nt paired end; 8.50 and 7.95 Gbp of data, respectively; Fasteris, Switzerland). After quality cleaning using Trimmomatic (Bolger et al., 2014), the reads were mapped onto the Ectocarpus sp. reference genome (Cormier et al., 2017) using Bowtie2 (Langmead and Salzberg, 2012). Coverage depth and breadth were, respectively, 34× and 96.83% for Ec800 and 32× and 96.81% for Ec801. Variants were called and counted using bcftools mpileup (http://samtools.github.io/). These variants were compared with a list of variants identified in genome sequence data for 14 other Ectocarpus mutant lines in order to remove false-positive mutations caused by, for example, errors in the reference genome sequence.

Variants unique to strains Ec800 and Ec801 were quality filtered based on coverage depth (±50% of the genome mean), mapping quality (>20), variant quality (>50), variant frequency (>0.9) and variant support in both sequencing directions. A custom Python script allowed the identification of variants in CDSs and the effect of each CDS mutation on the predicted protein was accessed manually (Table S5). A scheme of the approach is depicted in Fig. S13.

Ectocarpus samples were processed as described (Coelho et al., 2012c) using a protocol adapted from Bisgrove and Kropf (1998). Briefly, Ectocarpus cells were settled on coverslips and at appropriate times after settlement were rapidly frozen in liquid nitrogen and fixed in 2.5% glutaraldehyde and 3.2% paraformaldehyde for 1 h, then washed in PBS and treated with 5% Triton X-100 overnight. Samples were then rinsed in PBS and 100 mM NaBH₄ was added for 4 h. Cell walls were degraded with cellulase (1% w/v) and hemicellulase (4% w/v) for 1 h, and the preparation was then rinsed with PBS and blocked overnight in 2.5% non-fat dry milk in PBS. Samples were treated with an anti-tubulin antibody (1/200; DM1A, Sigma-Aldrich) at 20°C overnight and then treated with the secondary antibody (Alexa Fluor 488-conjugated goat anti-mouse IgG; 62197, Sigma-Aldrich; 1:1000 in PBS) at 20°C overnight. The preparation was rinsed with PBS and blocked overnight in 2.5% non-fat dry milk in PBS and then treated with an anti-centrin antibody (1/1000; anti-centrin 1 ab11257, Abcam) at 20°C overnight, followed by the secondary antibody (1/1000; Alexa Fluor 555-conjugated goat anti-rabbit IgG; SAB4600068, Sigma-Aldrich) for 8 h. Samples were stained with 4′,6-diamidino-2-phenylindole (DAPI; 0.5 µg/ml in PBS) for 10 min at room temperature and mounted in ProLong Gold (Invitrogen).

Confocal microscopy was conducted using an inverted SP8 laser scanning confocal microscope (Leica Microsystems) equipped with a compact supply unit which integrates a LIAchroic scan head, several laser lines (405 and 488 nm), and standard photomultiplier tube detectors. We used the oil immersion lens HC PL APO 63×/1.40 OIL CS2. The scanning speed was set at 400 Hz unidirectional. The pinhole was adjusted to one Airy unit for all channels. The spatial sampling rate was optimised according to Niquist criteria, generating a 0.058×0.058×0.299-µm voxel size (xyz). The z-stack height fitted the specimen thickness. A two-step sequential acquisition was designed to collect the signal from three or four channels. The first step recorded the anti-tubulin fluorescence signal (excitation, 488 nm; emission, 530 nm) and the transmitted light. The second step acquired the DAPI fluorescence signal (excitation, 405 nm; emission, 415-480 nm). Signal intensity was averaged three times. Fiji software was used to optimise the raw images, including maximum intensity projection and de-noising (3×3 median filter). For any given data, both wild-type and mutant images were analysed simultaneously with similar settings.

Tracking of microtubule bundles was performed on maximum intensity projections of z-planes covering the whole thickness of the cells. We drew a line transversely, perpendicular to the growth axis of the cell and crossing the nucleus. Peaks corresponding to the microtubule bundles were then identified in plots of intensity profiles and counted, in order to estimate the number of microtubule bundles in each cell. Note that in the bas mutants, owing to the disorganised nature of the microtubule network, average bundle numbers may be underestimated. This is because this method is well adapted for tracking microtubule bundles oriented with their long axis parallel to the image plane, but we may have missed bundles that were perpendicular to the plane of the transection.

Medium containing mature Ectocarpus material was pipetted onto a plastic film (gel support films; ATTO). The film was cut into <1-cm side triangles, and these were attached to Petri dishes by adhesive tape. Plurilocular structures, which were attached to the triangles, were rapidly immersed in liquid propane cooled to −180°C by liquid nitrogen, and immediately transferred into liquid nitrogen. The samples were submerged in substitution solution containing 2% osmium tetroxide with acetone at −80°C for 2 days, −40°C for 2 h and 4°C for 2 h. Finally, the temperature of the samples was gradually allowed to rise to room temperature, and they were then washed with acetone several times. The gel support films were infiltrated and embedded in Spurr's low-viscosity resin (Polysciences) on aluminium foil dishes. The films with the samples were turned inside out on the upper surface of the resin. Serial sections were cut with a diamond knife using an Ultracut ultramicrotome (Reichert-Jung) and mounted on Formvar-coated slot grids. Sections were stained with TI blue (Nisshin EM) and lead citrate and observed using an electron microscope (JEM-1011; JEOL).

For TEM coupled to focused ion beam-scanning electron microscopy (FIB-SEM), freshly released gametes were collected in cellulose capillaries and cultivated in the capillaries in Provasoli-enriched seawater for 3-5 days at 14°C in the dark or approximately 1 day at 14°C in a 12 h light/12 h dark regime to produce two- to five-cell stage partheno-sporophytes. Cells in capillaries were frozen at high-pressure (HPF Compact 03, Engineering Office M. Wohlwend), freeze-substituted (AFS2, Leica Microsystems) with 0.2% OsO₄ and 0.1% uranyl acetate in acetone containing 5% H₂O as substitution medium (Read et al., 2021) and embedded in Epon. Ultrathin sections were stained with uranyl acetate and lead citrate and analysed with a Tecnai Spirit (Thermo Fisher Scientific) operated at 120 kV.

In order to identify a region of the sample containing algae at high density for FIB-SEM data acquisition, a 3D X-ray tomogram of the resin block was acquired with a Bruker Skyscan 1272. The region of interest was exposed using a 90° diamond trimming knife (Cryotrim 90, Diatome) mounted on a Leica UC7 ultramicrotome. The sample was then attached to a stub using conductive silver epoxy resin (Ted Pella) and gold sputter coated (Quorum Q150RS).

FIB-SEM imaging was performed with a Zeiss Crossbeam 540 or a Crossbeam 550, using Atlas 3D (FIBICS, Carl Zeiss Microscopy) for sample preparation and acquisition. After the deposition of a protective platinum coat on the surface above the region of interest, a 60-µm-wide trench was opened to identify and image several cells in parallel. During the stack acquisition, FIB slicing was performed at 30 kV and 700 pA current. The datasets were acquired at 8 nm isotropic voxel size with the SEM at 1.5 kV and 700 pA current, using a back-scattered electron detector (ESB). After acquisition, the image stacks were acquired using the Fiji plugin ‘Linear stack alignment with SIFT’ (Lowe, 2004). We acquired images for a total of five wild-type cells and five bas-2 cells, and chose a representative image to present in Fig. 4. Images containing Golgi stacks were retrieved from aligned image stacks and cropped to reduce the image dimensions for further segmentation with the IMOD software package (https://bio3d.colorado.edu/imod/), MIB (Belevich et al., 2016) and AMIRA (Thermo Fisher Scientific).

Multiple alignments were generated with Muscle in MEGA7 (Kumar et al., 2016). Phylogenetic trees were then generated with RAxML (Stamatakis, 2014) using 1000 bootstrap replicates and the most appropriate model.

Expression levels of the BAS and DIS genes were investigated using TPM values obtained after kallisto pseudo-mapping and calculation of the lengthScaledTPM using the tximport package in R (see ‘Comparative Transcriptome Analyses’ section for details). Previously generated RNA-seq data for wild-type and dis samples (Table S7) was used for comparisons.

RNA-seq analysis was performed to compare the abundances of gene transcripts in the mutants bas-1 and bas-2 and wild-type sporophytes. RNA-seq datasets were generated from triplicate samples of each genotype, and individuals were grown synchronously as described previously in standard culture conditions (Coelho et al., 2012b; Cossard et al., 2022). Germlings were filtered through a nylon mesh to recover only thalli at the two- to five-cell stage. Each replicate contained between 10⁴ and 10⁶ individual germlings. Tissue samples were rapidly frozen in liquid nitrogen and processed for RNA extraction. Total RNA was extracted from each sample using the QIAGEN Mini kit as previously described (Lipinska et al., 2015). For each replicate, cDNA was produced by oligo(dT) priming, fragmented, and prepared for stranded 2×150-bp paired-end sequencing on an Illumina HiSeq 3000 platform.

Raw and cleaned read quality was evaluated using fastQC (v0.11.9) and mutliQC (v. 1.9). Raw reads were trimmed and filtered based on quality score, and adapter sequences were removed using Trimmomatic (v0.39). Transcript abundance was evaluate using kallisto (v0.46.2) via pseudo mapping on CDS features. The matrix of counts and TPMs for all samples and all replicates were generated in R using the tximport package.

A gene was considered to be expressed if the TPMmean was above the 5th percentile (as in Cossard et al., 2022; Lipinska et al., 2019). About 4% of the genes in each sample had TPMmean values under this threshold and were considered not to be expressed in our samples (Table S9). Differential gene expression was analysed using DESeq2 (Love et al., 2014). Genes were considered to be DE when the log2 fold change was below or equal to −1 (downregulated, at least twice as weakly expressed) or above or equal to 1 (upregulated, at least twice as strongly expressed) and the adjusted P-value below or equal to 0.01. Genes were considered to be exclusively (uniquely) expressed in bas mutants when they were significantly upregulated compared with wild type and the wild-type mean TPM was equal to 0. Conversely, genes were considered to be specifically silenced in bas mutants when they were significantly downregulated compared with wild type and their TPM means were equal to 0.

GO term and HECTAR localisation enrichment were carried out in R using the ‘clusterProfiler’ package (Yu et al., 2012).

We thank Nana Kinoshita-Terauchi (Shimoda Marine Research Center, University of Tsukuba) and Iris Kock (MPI Tübingen) for helpful discussions and for images of the mutants; and Viktoria Buhanets for help with segmentation of FIB-SEM images. We acknowledge the Electron Microscopy Core Facility at MPI for Biology Tubingen. S.M.C is the recipient of a Bettencourt Schueller Foundation Award, which supported imaging facilities at the Station Biologique in Roscoff.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201283.reviewer-comments.pdf