Arabidopsis JANUS Regulates Embryonic Pattern Formation through Pol II-Mediated Transcription of WOX2 and PIN7.

Embryonic pattern formation relies on positional coordination of cell division and specification. Early axis formation during Arabidopsis embryogenesis requires WUSCHEL RELATED HOMEOBOX (WOX)-mediated transcription activation and PIN-FORMED7 (PIN7)-mediated auxin asymmetry. How these events are regulated is obscure. We report that Arabidopsis JANUS, a putative subunit of spliceosome, is essential for embryonic pattern formation. Significantly reduced transcription but not mRNA processing of WOX2 and PIN7 in janus suggested its role in transcriptional regulation. JANUS interacts with RNA polymerase II (Pol II) through a region outside of its spliceosome-association domain. We further show that Pol II mediates the transcription of WOX2 and PIN7 in a JANUS-dependent way and is essential for embryonic pattern formation. These findings reveal that JANUS recruits Pol II for the activation of two parallel pathways to ensure proper pattern formation during embryogenesis.

Introduction

Development of a specific body plan during embryogenesis requires precise cell fate determination based on the position of cells along the embryo axes. In Arabidopsis, the crucial cell types are established extremely early during embryogenesis as reflected by the stereotypic sequence of oriented cell divisions (Lau et al., 2012, ten Hove et al., 2015). Two pathways have been linked to the establishment of apical-basal axis and cell specification after zygotic division. One involves the transcription factors WUSCHEL RELATED HOMEOBOX2 (WOX2), WOX8, and WOX9, whereas the other depends on auxin, whose asymmetry is maintained by auxin efflux carrier PIN-FORMED7 (PIN7) (Lau et al., 2012). WOX2 and WOX8 are initially co-expressed in the zygote (Haecker et al., 2004). After zygotic division, WOX2 and WOX8 are restricted to the apical and basal cell lineage to control the following cell specification, respectively (Breuninger et al., 2008, Haecker et al., 2004). PIN7 is polarly localized to the apical plasma membrane (PM) of the basal cell, where it provides maternal auxin to the apical cell (Friml et al., 2003, Robert et al., 2018). The polar distribution of PIN7 ensures auxin maximum in the apical cell, which generates the proembryo and all apical structures of the plant. Functional loss of WOX2 or PIN7 compromised the formation of apical-basal axis during early embryogenesis. However, their defects at early embryonic pattern formation are later recovered (Friml et al., 2003, Robert et al., 2018). Whether these two pathways play redundant roles in embryogenesis and how their specific expression is controlled are unclear.

RNA polymerase II (Pol II) plays a pivotal role in regulating gene expression (Thomas and Chiang, 2006). Pol II in Arabidopsis consists of 12 core subunits (Ream et al., 2009), in which Nuclear RNA Polymerase B1 (NRPB1) and NRPB2 interact to form the catalytic center for RNA synthesis, whereas other subunits play structural and regulatory roles in transcription initiation, elongation, termination, or RNA processing (Cramer et al., 2008, Werner and Grohmann, 2011). Functional studies of genes encoding for Pol II subunits suggested its role in embryogenesis such that no homozygous mutants could be obtained for functional loss of NRPB2, as well as that of NRPB9 and NRPB11, genes encoding for the noncatalytic subunits of Pol II (Onodera et al., 2008, Ream et al., 2009, Tan et al., 2012). However, whether and how functional loss of Pol II affects embryogenesis is unclear.

Spliceosomes are large RNA-protein complexes mainly involved in pre-mRNA splicing. Nevertheless, subunits of spliceosomes also participate in many other processes, including mRNA export (Howard and Sanford, 2015, Muller-McNicoll et al., 2016), the maintenance of genome stability (Li and Manley, 2005, Xiao et al., 2007), and microRNA processing (Ben Chaabane et al., 2013, Wu et al., 2010). Interestingly, reports both in metazoans and in plants showed that a subunit of spliceosome, Serine/arginine-rich splicing factor 35 (SC35), interacts with a Pol II subunit and is required for the transcription of genes, in addition to its role in pre-mRNA splicing (Lin et al., 2008, Yan et al., 2017). A role of the spliceosome subunit in regulating Pol II-mediated transcription was proposed (Yan et al., 2017).

In this study, we report that the Arabidopsis homolog of human SPLICEOSOME-ASSOCIATED PROTEIN 49 (SAP49) and yeast Hsh49p, a subunit of spliceosome, is essential for pattern formation during early embryogenesis. We named it JANUS as it represents the god of new beginnings in ancient Rome and associates with the first steps of a journey. Functional loss of JANUS resulted in complete embryo lethality due to abnormal cell division immediately after the first zygotic division. The specific expression of PIN7 was disrupted in janus, resulting in defective auxin signaling. On the other hand, WOX2 was also transcriptionally downregulated in janus. Consistently, the disruption of both WOX2- and PIN7-dependent pathways resembled pattern formation defects of janus during early embryogenesis. We further showed that JANUS interacts with Pol II subunits independent of its role as a splicing factor and is required for Pol II-dependent transcription of WOX2 and PIN7. Indeed, functional loss of Pol II subunits mimicked embryonic defects of janus. Taken together, our findings demonstrate that JANUS recruits Pol II to transcriptionally activate WOX2- and PIN7-mediated pathways for pattern formation during early embryogenesis.

Results

JANUS Is Essential for Pattern Formation during Embryogenesis

JANUS was isolated for characterization because of the complete embryo lethality of its mutant emb2444 (Meinke et al., 2008). JANUS contains two RNA recognition motifs (RRMs) and is homologous to a subunit of the splicesome (Figures S1A, S1E, and S1F). Segregation ratio from reciprocal crosses between wild-type and janus/+ indicated that gametophytic transmission of the mutant was not affected (Table S1). The heterozygous mutant showed one-fourth of seed abortion (Figure 1A). Transcript abundance of JANUS was significantly reduced in janus/+ compared with that in the wild-type (Figure S1). In addition, the genomic fragment of JANUS with a GFP reporter gene in the control of its native promoter was introduced into janus/+. Transgenic lines of JANUSg-GFP;janus were obtained, and all showed no seed abortion (Figures 1A and S1), indicating that JANUS is the causal gene for seed abortion of janus.

Figure 1

JANUS Is Essential for Pattern Formation during Embryogenesis

(A) Seed set of different genotypes. Results are means ± standard deviation (SD, n = 8). Seed set of janus/+ is significantly different from others (Tukey's multiple comparison test, p < 0.05). Scale bars, 1 mm.

(B and C) Wild-type (WT) and janus embryo development by ovule clearing. DAP indicates days after pollination. Because janus embryos are much delayed in development, wild-type embryos and janus embryos are shown in pairs according to their developmental stages but not to the same DAP. Dotted lines in (C) indicate division planes. Scale bars, 20 μM.

(D and E) Confocal laser scanning microscopy (CLSM) of a Pro:GFP embryo (D) or a Pro:GFP;janus embryo (E). Images shown are merges of the GFP channel and RFP channel (propidium iodide [PI] staining in magenta).

(F) Schematic illustration of wild-type or janus embryogenesis. Arrowheads point at aborted seeds in (A). The arrowhead points at the Quiescent Center labeled by GFP in (D) but its absence in (E).

To determine at which stage developing seeds started to show defects in janus, we examined self-fertilized janus/+ by whole-mount clearing. Embryos developing within a single silique are approximately at the same developmental stage (Breuninger et al., 2008), which enabled an estimate of janus embryos, which are much delayed compared with their wild-type siblings. After the first zygotic division, one-fourth of embryos from janus/+ plants displayed an asymmetric oblique division (Figures 1C and 1F) rather than symmetric vertical division as observed in three-fourth of the other embryos as well as in embryos of the wild-type (Figures 1B and 1F). The division pattern followed was impaired in these presumably janus embryos (Figures 1C and 1F), which showed severe morphological defects at the early globular stage and were eventually arrested at the late globular stage (Figures 1C and 1F). In the arrested embryos, the outer walls of protoderm cells were distended, producing an uneven surface on the embryo proper (Figures 1C and 1F). Abnormal divisions occurred both in the apical and the basal lineages (Figures 1C and 1F). Furthermore, the formation and specification of quiescent center (QC) was also compromised judged by the irregularly oblique divisions in hypophysis and by the absence of GFP signals in Pro:GFP;janus (Figures 1D and 1E), which specifies the QC (Blilou et al., 2005). These results demonstrated that JANUS is an essential gene for early embryonic pattern formation and cell fate specification. Consistent with its role in embryogenesis, JANUS is highly expressed in developing embryos from the zygotic stage to the cotyledon stage (Figure S1).

JANUS Mediates the Expression of WOX2 and PIN7

JANUS is nuclear localized (Figure S1), suggesting a potential role in gene expression. We thus tested whether the two major pathways in controlling early embryogenesis, i.e., WOX-mediated transcriptional pathway and PIN7-mediated auxin signaling pathway, were affected in janus. First, we analyzed the transcription activity of WOX2 and WOX8 by introducing Pro:DsRed2/Pro:NLS-vYFP (Yu et al., 2016) in janus/+. WOX2 and WOX8 were transcriptionally activated in the apical and basal cells after the zygotic division in wild-type, respectively (Figure 2A), as reported (Breuninger et al., 2008, Haecker et al., 2004). By contrast, WOX2 showed a substantially reduced transcriptional activity in the apical cell lineage of one-fourth embryos in janus/+, whereas WOX8 was not affected (Figures 2A and 2B). The reduced transcription of WOX2 but not WOX8 was confirmed by additional reporter line (Figure S2) or native transcript levels (Figure S3). These results suggested that the transcription activity of WOX2 but not WOX8 in early embryogenesis depended on JANUS.

Figure 2

JANUS Mediates the Expression of WOX2 and PIN7

(A) CLSM of a Pro:DsRed2/Pro:NLS-vYFP3 zygote or embryo at early developmental stages either in the wild-type or in janus background. Images shown are merges of the RFP (magenta for dsRed) and the GFP (green for NLS-vYFP3) channels. Dotted circles indicate regions of interest (ROI). Scale bars, 20 μM.

(B) Fluorescence intensity of dsRed2 (for Pro) NLS-vYFP3 (for Pro). A.u. represents arbitrary fluorescence unit. Results shown are average fluorescence intensities within an ROI. Substantially reduced dsRed2 intensity is indicated in pink, presumably of the janus genotype.

(C and D) CLSM of PIN7g:GFP (C) or Pro:NLS-YFP (D) embryos in wild-type or in janus. Dotted lines in (C) indicate the silhouettes of the embryo. Scale bars, 20 μM.

(E) CLSM of DR5:GFP embryos in wild-type or in janus. Arrowheads point at cells with auxin maximum.

(F–K) WT (F), wox2-4 (G), pin7-1 (H), wox2-4;pin7-1 (I), janus/+ (J), or Pro:PIN7;janus/+ (K) embryo development by ovule clearing. Embryos are shown according to their developmental stages but not to the same DAP. Dotted lines indicate division planes. Results are means ± SD (n = 10). The arrowhead in (K) indicates the appearance of the Quiescent Center. In total, 147–404 embryos were examined. Scale bars, 20 μM.

Second, we examined whether PIN7 was transcriptionally affected by functional loss of JANUS. PIN7 (Blilou et al., 2005) was polarly localized to the apical PM of the basal cells in wild-type globular embryos (Figure 2C), as reported (Friml et al., 2003). In one-fourth embryos from PIN7:GFP;janus/+, presumably of the janus genotype, GFP signals were significantly reduced or even undetectable (Figure 2C). Because JANUS is homologous to spliceosome subunits (Figure S1), we examined the transcription activity of PIN7 in Pro:NLS-YFP;janus/+, which was reflected by fluorescence intensity, to exclude the possibility that mRNA splicing of PIN7 contributed to the expression difference. Consistent with the results obtained from PIN7:GFP, YFP signals were dramatically reduced or undetectable in the basal cells of one-fourth embryos in Pro:NLS-YFP;janus/+ (Figure 2D). These results implied that JANUS is crucial for the transcription of PIN7 during early embryogenesis. Indeed, no splicing defects of PIN7 as well as of WOX2 were detected in the siliques of janus/+ (Figure S2), confirming a role of JANUS in the transcription rather than RNA processing of the two genes. Because PIN7 is critical for auxin polar transport during early embryogenesis (Friml et al., 2003), the reduced PIN7 in janus would have severely compromised auxin signaling. It was indeed the case. By examining DR5:GFP, a synthetic auxin maximum reporter (Friml et al., 2003), we confirmed that GFP was accumulated in the apical cell lineage immediately after zygotic division and then restricted to the hypophysis of the globular embryos in wild-type plants (Figure 2E), as reported (Friml et al., 2003). By contrast, GFP signals were restricted to the basal cells and failed to establish an apical-basal gradient in one-fourth of embryos in DR5:GFP;janus/+ (Figure 2E). Similar defects were reported in embryos defective in PIN7 or treated with auxin efflux inhibitors (Friml et al., 2003).

Finally, we generated a double mutant of WOX2 and PIN7 under the rationale that if JANUS mediates the transcriptional activity of WOX2 and PIN7 during early embryogenesis, functional loss of both pathways would at least partially mimic the defects of janus. Compared with wild-type, the single mutant of PIN7 or WOX2, i.e., pin7-1 or wox2-4 (Friml et al., 2003, Zhang et al., 2017), showed a slight defect of apical-axis patterning in early stages during embryogenesis but finally recovered at the globular stage (Figures 2G and 2H), consistent with previous reports (Breuninger et al., 2008, Friml et al., 2003, Haecker et al., 2004). In comparison, the proportion of wox2-4;pin7-1 embryos with abnormal asymmetric division in early stages substantially increased (Figure 2I). In the wild-type, the first division of the apical cell was vertical and symmetric (Figure 2F), as reported (Breuninger et al., 2008). By contrast, in wox2-4;pin7-1, the divisions were asymmetric, either oblique or horizontal in both embryo proper and hypophysis (Figure 2I), which largely resembled those of janus during early embryogenesis (Figure 2J). Instead of a full recovery at late stages as seen in each single mutant (Breuninger et al., 2008, Friml et al., 2003, Haecker et al., 2004), over 30% embryos of wox2-4;pin7-1 arrested immediately after the first zygotic division (Figure 2I). By contrast, enhanced expression of PIN7 partially complemented the defects of janus such that early embryonic pattern formation was largely normal in Pro:PIN7;janus/+ (Figure 2K). These results suggested that both WOX2- and PIN7-dependent pathways may be transcriptionally regulated by JANUS and play redundant roles in apical-basal patterning during early embryogenesis.

JANUS Interacts with Pol II Whose Functional Loss Resulted in Embryo Lethality

To determine how JANUS affected the transcription activity of WOX2 and PIN7 during embryonic pattern formation, we tested whether JANUS interacted with components of Pol II since Pol II is responsible for the transcription of most mRNAs in eukaryotes (Thomas and Chiang, 2006) and was shown to interact with other components of the spliceosome (Yan et al., 2017). Of 11 Pol II subunits tested, NRPB7, NRPB10, and NRPB11 showed interaction with JANUS by yeast two hybrid (Y2H) and bimolecular fluorescence complementation (BiFC) assays (Figures 3A–3C and S4). Interestingly, the interaction is evolutionarily conserved: JANUS homolog also interacts with Pol II components in mice (Figure S4).

Figure 3

JANUS Interacts with Pol II whose Functional Loss Resulted in Embryo Lethality

(A) Bimolecular complementation (BiFC) assays demonstrating the interaction between three Pol II subunits and JANUS. YFP signals are shown in green. U1-70k is used as a nuclear marker (magenta). Scale bars, 20 μM.

(B) Quantification of BiFC signals based on fluorescence intensity within nuclei. Results are means ± SD.

(C) Yeast two hybrid (Y2H) assays showing the interaction between three Pol II subunits and JANUS. Diploid yeast strains are grown on medium lacking Trp and Leu (-WL). Positive interactions are determined by growth on medium lacking Trp, Leu, His, Ade, and supplemented with X-α-Gal (-WLHA + X-α-Gal). Results are representative of three biological replicates.

(D) WT, nrpb2-1, nrpb2-2, or nrpb11 embryo development by ovule clearing. Embryos are shown according to their developmental stages but not to the same DAP. Dotted lines indicate division planes. Scale bars, 50 μM.

(E) Schematic illustration of wild-type, nrpb2, or nrpb11 embryogenesis.

The physical interaction between JANUS and Pol II suggested that they might function in the same pathway. Indeed, all three genes are highly expressed during embryogenesis based on reporter analysis (Figure S4). We thus hypothesized that functional loss of Pol II would also result in defective embryonic pattern formation. To test this hypothesis, we characterized mutants of NRPB10 and NRPB11, both of which interact with JANUS (Figure S4). Although NRPB7 also interacts with JANUS, the lack of its mutants in all stock centers prevented further analysis. Likely due to the redundancy with NRPB10-like genes (Ream et al., 2015), functional loss of NRPB10 did not show seed set reduction and embryo lethality (Figure S4). In comparison, functional loss of NRPB11, a single gene in Arabidopsis (Ream et al., 2015), resulted in embryo lethality (Figure S4), similar to nrpb2 (Onodera et al., 2008). Because both genes are required for Pol II activity (Kershnar et al., 1998, Ream et al., 2009), these results hinted at an essential role of Pol II in embryogenesis. To identify the defects of embryogenesis due to Pol II loss of function, we examined developing embryos in nrpb2-1/+, nrpb2-2/+, and nrpb11-1/+. The apical-basal patterning during early embryogenesis was significantly affected in one-fourth of nrpb11-1/+ as well as in 29% of the nrpb2/+ mutants (Figures 3D and 3E). Similar to that in janus, the first division of the apical cell, i.e., symmetric and horizontal in wild-type, was asymmetric and oblique in nrpb11 and nrpb2 (Figures 3D and 3E). Irregular division patterns resulted in abnormal formation of embryo proper and hypophysis (Figures 3D and 3E). In most severe cases, embryos were arrested at as early as the two-cell stage (Figures 3D and 3E). Thus, these results demonstrated that Pol II is essential for early embryonic pattern formation and likely functions in the same genetic pathway as JANUS.

Pol II-Mediated WOX2 and PIN7 Transcription Depends on JANUS

That JANUS interacts with Pol II and affects the transcription activity of WOX2 and PIN7 suggested that Pol II might mediate WOX2 and PIN7 transcription during embryonic pattern formation in a JANUS-dependent way. We thus hypothesized that functional loss of Pol II would compromise WOX2 and PIN7 expression during early embryonic pattern formation. To test this hypothesis, we introduced Pro:GFP and PIN7:GFP into nrpb2-1/+ and examined the fluorescence distribution during embryogenesis. Indeed, the expression of WOX2 and PIN7 was severely reduced or even undetectable in one-fourth embryos of the nrpb2-1/+ plants (Figures 4A and 4B), reminiscent of that in janus/+. Auxin signaling was also compromised during early embryogenesis by functional loss of NRPB2 such that ectopic GFP signals were detected in the basal cells of one-fourth DR5:GFP;nrpb2-1/+ embryos (Figure 4C), similar to that by functional loss of JANUS. These results suggested that Pol II mediates the transcription of WOX2 and PIN7 during early embryonic pattern formation.

Figure 4

Pol II-mediated WOX2 and PIN7 Transcription Depends on JANUS

(A) CLSM of Pro:DsRed2/Pro:NLS-vYFP3 embryos at early developmental stages either in the wild-type or in nrpb2-1 background. Images shown on top are merges of the RFP (magenta for dsRed) and the GFP (green for NLS-vYFP) channels; on bottom are merges of the RFP, GFP, and transmission channels. Dotted lines indicate the silhouettes of the embryo. Scale bars, 50 μM.

(B and C) CLSM of PIN7g:GFP (B) or DR5:GFP (C) embryos in wild-type or in nrpb2-1. Dotted lines indicate the silhouettes of the embryo. Scale bars, 50 μM.

(D and E) CLSM of Pro:H2B-GFP (D) or PIN7g:GFP (E) in wild-type or in two lines of Pro:JANUS-RNAi. Dotted circles indicate ROI. Scale bars, 20 μM.

(F) Schematic representation of the structure of PIN7. The letters indicate the positions of primer pairs used for ChIP-PCR.

(G) Quantification data of the ChIP results. ChIP-PCRs were used to analyze the Pol II enrichment at PIN7, which is presented as ratio of (Pol II PIN7/input PIN7) to (Pol II Actin/input Actin). Results are means ± standard error (SEM) from three technical repeats. ChIP assays were repeated three times with similar results. Asterisks indicate significant difference (t test, p < 0.05).

(H and I) Intensity of H2B-GFP (for Pro) (H) or PIN7g:GFP (I). Results shown are average fluorescence intensities within an ROI, shown in (D) and (E), respectively. Different letters indicate significantly different groups (Tukey's multiple comparison test, p < 0.05).

(J and K) Relative transcript abundance of WOX2 (J) or PIN7 (K). Results shown are means ± SD (n = 3). Different letters indicate significantly different groups (Tukey's multiple comparison test, p < 0.05). Three biological replicates were examined with similar results.

To test whether Pol II-mediated transcription of WOX2 and PIN7 depended on JANUS, we examined Pol II occupancy at the promoter regions of WOX2 and PIN7 by chromatin immunoprecipitation (ChIP) using an antibody against NRPB2 (Yan et al., 2017). Because no homozygous janus plants could be obtained, we generated Pro:JANUS-RNAi transgenic plants. Transcript analysis as well as fluorescence quantification supported a significant reduction of JANUS by RNAi in Pro:JANUS-RNAi transgenic plants (Figure S5). ChIP assays indicated that Pol II occupancy was significantly reduced at the promoter regions of PIN7 in Pro:JANUS-RNAi lines (Figures 4F and 4G) as compared with that in the wild-type (Figures 4F and 4G). Consistent with the reduced Pol II occupancy, PIN7 was transcriptionally downregulated in Pro:JANUS-RNAi lines, either by quantitative PCRs for the endogenous genes (Figures 4J and 4K) or by fluorescence quantification of the PIN7:GFP transgenic plants (Figures 4D, 4E, 4H, and 4I). Except for embryos, WOX2 was weakly expressed in a few cells at root maturation zone (Figure 4D), which hindered the application of ChIP assays on its promoter regions. However, by analyzing the fluorescence in the Pro:JANUS-RNAi;Pro:H2B-GFP or Pro:DsRed2;janus/+ plants, we confirmed that JANUS-RNAi or janus significantly reduced the transcription activity of Pro (Figures 2A and 4D). These results suggested that Pol II-mediated transcription of WOX2 and PIN7 depends on JANUS. By contrast, the transcription of WOX8 and SCARECROW (SCR), a critical embryonic gene (Wysocka-Diller et al., 2000), depends on Pol II but not JANUS (Figure S3), suggesting that JANUS did not affect the transcriptional activity of Pol II in general.

Discussion

Early embryogenesis is the critical developmental phase during which the basic features of the plant body are established. Although distinct expression domains of WOX family transcription factors as well as directional auxin transport are known to be involved in early apical-basal patterning, their upstream regulators and potential interactions were obscure. Our results demonstrated that the two pathways control early embryonic pattern formation in a parallel way and both mediated by JANUS. First, functional loss of WOX2 and PIN7 caused a severe defect immediately after the zygotic division, which cannot be restored at late stages, unlike the disruption of each pathway. Second, both WOX2 and PIN7 were transcriptionally downregulated by JANUS loss of function. The effect of JANUS on PIN7 and WOX2 is specific because WOX8, another embryonic gene, requires RNA Pol II but not JANUS for its expression during early embryogenesis. Third, wox2-4;pin7-1 resembled janus in early embryogenesis. Finally, the accumulation of Pol II at PIN7 was significantly reduced by JANUS-RNAi, suggesting JANUS-dependent recruitment of the transcriptional machinery at PIN7.

Although transcription and RNA processing are coupled in vivo (Bentley, 2002, Hirose and Manley, 2000, Lee and Tarn, 2013, Proudfoot et al., 2002, Yan et al., 2017), studies have proven that splicing factors may interact with transcription machinery to directly influence gene expression in metazoans (Braunschweig et al., 2013, Das et al., 2007). Although we cannot exclude a role of JANUS in RNA splicing owing to its homology to the component of the spliceosome during embryogenesis, results presented here strongly suggested its role through transcriptional regulation. First, JANUS is critical for the transcription but not RNA splicing of WOX2 and PIN7, two genes critical for early embryonic pattern formation. Second, the RRM1 domain of yeast SAP49 or metazoan Hsh49p, homolog of Arabidopsis JANUS, is responsible for RNA binding and spliceosome association and is essential for spliceosome-mediated RNA splicing (Igel et al., 1998, Kuwasako et al., 2017, Pauling et al., 2000, van Roon et al., 2017). However, we found that JANUS directly interacts with Pol II through its RRM2 but not RRM1 domain, suggesting that its interaction with Pol II does not require its association with the spliceosome. Third, yeast and metazoan homologs of JANUS are part of an SF3b complex within the spliceosome (Kuwasako et al., 2017, van Roon et al., 2017). However, although other components of the SF3b complex in Arabidopsis have been functionally characterized (Aki et al., 2011, Wang and Brendel, 2006), none of the related mutants showed embryo lethality.

Eukaryotes decode their genomes using three essential nuclear DNA-dependent RNA polymerases (Cramer et al., 2008, Werner and Grohmann, 2011). In Arabidopsis, several subunits of Pol II, including NRPB2, NRPB5, NRPB9, and NRPB11, have been reported to affect plant viability (Onodera et al., 2008, Ream et al., 2009, Tan et al., 2012). However, whether and how Pol II regulates embryogenesis was unclear. We showed here that mutations at a few Pol II components resulted in complete embryo lethality (Figures 3D and S3I). A few components of Pol II are shared by two plant-specific RNA polymerase complexes, Pol IV and Pol V (Ream et al., 2009). However, we believe that Pol II is key for JANUS-mediated transcription of WOX2 and PIN7 because the JANUS-interacting component NRPB7 is specific for Pol II (Figures 3A and S3B) and only Pol II was reported to be essential for viability (Herr et al., 2005, Kanno et al., 2005, Onodera et al., 2005, Pontier et al., 2005, Ream et al., 2009).

Data presented here suggested that JANUS is important for the Pol II-mediated transcription of WOX2 and PIN7 during early embryogenesis. However, JANUS has no recognizable DNA-binding domains. It is unclear how JANUS determines the selectivity of Pol II on target genes such as WOX2 and PIN7. In eukaryotes, mediators, multi-subunit complexes, bridge transcription activators with Pol II at specific cis-elements for transcription initiation (Dolan and Chapple, 2017). Whether JANUS fulfills a role of mediator, recruiting DNA-binding transcription factors together with Pol II, to the promoter regions of WOX2 and PIN7, is an interesting scenario worthy of further investigation.

Limitations of the Study

We would like to note that, because both janus and Pol II mutants are embryo lethal, some molecular and biochemical experiments can only be carried out in Pro:JANUS-RNAi plants. Future efforts will be dedicated to generate weak mutant alleles of JANUS, for which homozygous plants can be used for further mechanistic analysis. We also would like to point out that, although JANUS is important for Pol II-mediated transcription of WOX2 and PIN7 during early embryogenesis, it is unclear how JANUS, a protein containing no DNA-binding domains, determines target selectivity for Pol II. Whether JANUS contains untraditional DNA-binding domains or interacts with transcription factors for its function will be interesting scenario for future studies.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.