Expression of PRX36, PMEI6 and SBT1.7 is controlled by complex transcription factor regulatory networks for proper seed coat mucilage extrusion

Abstract

Mucilage secretory cells (MSC) form an intriguing cell layer important for seed germination. In Arabidopsis thaliana, several master transcription factors (TFs) and “actor” proteins have already been identified as key players for seed coat differentiation including epidermal cell formation, mucilage production and extrusion. The regulation of the genes coding for MSC cell wall “actor” proteins by TFs needs to be better established. Here, the expression and the regulation of 3 known actors (PRX36, PMEI6, SBT1.7) and 2 additional putative actors (PRX56, DIR12) have been analyzed in T-DNA mutants affected in master TFs (ap2, egl3/gl3, gl2, myb5, tt8, ttg1, ttg2 and luh1/mum1). Genes with somehow similar function are differentially regulated and conversely, genes with different functions are regulated in similar manner.

Abbreviations

MSC

mucilage secretory cells

AP2

APETALA2

EGL3

ENHANCER OF GLABRA3

GL2

GLABRA2

TT8

TRANSPARENT TESTA8

TTG1

TRANSPARENT TESTA GLABRA1

TTG2

TRANSPARENT TESTA GLABRA2

LUH/MUM1

LEUNIG HOMOLOG1/ MODIFIED MUCILAGE1

During seed coat formation of numerous myxospermous plant species including Arabidopsis thaliana, large amounts of pectins are accumulated as a mucilage within a specialized apoplastic zone of outermost cell layer, either called mucilage secretory cells (MSC) or outer integument 2 (oi2). In parallel, a polarized non-lignified secondary cell wall synthesis occurs within these cells, between the deposited mucilage and cytoplasm forming a “volcano shape” structure named colum-ella.^1-3 Upon imbibition of dry seeds, the hydrated mucilage expands and the polarized rupture of the tangential outer primary cell wall allows the release of the viscous mucilage surrounding the seed. This extracellular polysaccharidic matrix acts as an hydrogel that could have a role for seed rehydration, germination in drought conditions and protection during early steps of germination.^4-6 The physiology of MSC and the polysaccharidic composition of the mucilage is now well reviewed,^2,7 but less is known about the gene regulation network leading to this complex cell differentiation and synthesis events. A partial regulation network has been drawn, mainly based on master regulatory genes and with a few genes coding for “actor” proteins with enzymatic activity and involved in MSC dynamism.^8-10

Several pleiotropic transcription factors (TFs) necessary for seed coat differentiation but also for normal epidermal cell morphology and mucilage production have been identified (APETALA2 (AP2), ENHANCER OF GLABRA3 (EGL3), GLABRA2 (GL2), MYB5, TRANSPARENT TESTA8 (TT8), TRANSPARENT TESTA GLABRA1 (TTG1), and TRANSPARENT TESTA GLABRA2 (TTG2)).^11-17 Other transcriptional regulators such as LEUNIG HOMOLOG1 (LUH1)/MUM1 have been shown to only regulate gene necessary for mucilage extrusion.^18-20 T-DNA mutants corresponding to these TFS are available and their phenotypes have already been described. Among sometimes other pleiotropic effects, they present either lack, or reduction of mucilage release. The phenotypes observed can be due to a default of mucilage synthesis or release.²¹ Despite the lack of mucilage phenotype description for the gl3 mutant, GLABRA3 (GL3) could also be implicated in regulation of mucilage synthesis and extrusion through its interaction with EGL3 demonstrated during trichome initiation.²²

Recently, 3 cell wall proteins: a subtilisin protease (SBT1.7), a pectin methyl esterase inhibitor (PMEI6) and a class III peroxidase (PRX36) have been identified as major proteins necessary for a normal A. thaliana seed mucilage release with a somehow common phenotype of no release of mucilage caused by the non-fragmentation of the outer cell wall.^10,23,24 SBT1.7, PMEI6 and PRX36 are expressed in MSC within a restricted temporal frame of seed coat development arguing for accurate and putative common transcriptional regulation process. The exact function of these 3 cell wall proteins is not yet exactly determined but the similar phenotype of the respective corresponding T-DNA mutants suggests a putative role in a similar polarized cell wall modifying process. Expression of PRX36, PMEI6 and SBT1.7 has been demonstrated to be controlled by NARS1 for PRX36,²⁴ by MUM1 and GL2 for PMEI6¹⁰ and by MUM1 for SBT1.7.¹⁰ The regulation of these 3 genes by the other masters TFs described above has not yet been shown and need to be established in order to support their previously described similar expression patterns and phenotypes.

In parallel, availability of a large amount of transcriptomic data (RNA seq and DNA microarrays) allows to produce gene co-expression network in order to identify other potential genes with related function. Using PRX36 as bait, we performed co-expression analysis with ATTED-II database.²⁵ Among the numerous genes identified as members of the same network, 2 additional putative candidates have been chosen: another class III peroxidase, PRX56; a dirigent protein, DIR12 (Table 1). While PRX56 role has not been previously studied, DIR12, also named DP1, belongs to a class of genes necessary to control stereochemical phenoxy radical coupling during natural product synthesis^26,27 and has been shown to be specifically expressed in the outer-seed-coat²⁸ with a role in production of a specific neolignan.²⁹ Here, the expression of genes coding for the 3 cell wall proteins, PRX36/PMEI6/SBT1.7, and the 2 candidates genes, PRX56/DIR12, will be studied in a panel of T-DNA mutants affected in master TFs described above (ap2, egl3/gl3, gl2, gl3, myb5, tt8, ttg1, ttg2 and luh1/mum1). The convenient ruthenium red test will be used to evaluate putative mucilage release phenotype in T-DNA mutants corresponding to the 2 new candidates.

Table 1.

Detail of the genes analyzed for their expression level within the mutants listed in Table 2

Accession #	Locus	Function
At3g50990	PRX36	Class III Peroxidase
At5g15180	PRX56	Class III Peroxidase
At2g47670	PMEI6	Pectin Methylesterase Inhibitor
At4g11180	DIR12/DP1	Dirigent Protein
At5g67360	SBT1.7	Subtilisin-like Serine protease

To analyze the expression and the regulation of PRX36, PRX56, PMEI6, DIR12 and SBT1.7 (Fig. 1), mRNA have been extracted from siliques of 2-month-old wild type and master TF mutant plants (Table 2) and semi-quantitative RT-PCR have been performed. PRX36, PRX56, PMEI6, DIR12 and SBT1.7 expression levels in the wild-type control plants were consistent with publicly Arabidopsis transcriptomic data available eFP Browser and Genevestigator.^30,31 The level of PRX56 expression is not modified in any of TFs mutants (Fig. 1B) and the level of SBT1.7 expression is only slightly repressed in the mum1, myb5, tt8 and ttg1 mutants (Fig. 1E). Interestingly, PRX36 and DIR12 expression levels are highly affected and present a similar pattern in 8 out the 9 tested master TF mutant lines (Fig. 1A and C). They are both not expressed in the gl2, tt8, ttg1 and ttg2 mutants and normally expressed in ap2, gl3, gl3×egl3 and myb5. The only observed difference is the additional PRX36 repression in a mum1 line. Finally, PMEI6 is not expressed in the gl2, gl3, gl3×egl3 and mum1 mutants, and partially repressed in the myb5, tt8 and ttg1 mutants (Fig. 1D). Notably, GL3 regulates only PMEI6 and none of the TF studied here appears as repressor for the 3 actors and the 2 new candidates since overexpression was never observed in the TF mutants for any of the tested genes.

Figure 1.

Regulation of PRX36, PRX56, DIR12, PMEI6 and SBT1.7 expression. RT-PCR was carried out on siliques of 2-month-old wild type and mutant plants (lines used are detailed in Table 1) grown in a culture chamber (16 h light, 200 μmol photons m−1 sec−1, 22°C, 70% RH). Total RNA was extracted with the Tri-reagent solution (Sigma-Aldrich) according to the manufacturer's instructions and treated with the RQ1 RNase-free DNase I (Promega). cDNA were generated using the MMLV-RT (Promega). The amount of cDNA template in each RT-PCR reaction was normalized to the signal from the actin-encoding ACT2 gene 37, and primers were designed to rule out the amplification of genomic DNA. Histogram bars represent the expression levels normalized versus ACT2 expression levels. Mean ± SD, n = 3. (A) Prx36 primers: 5’-CAAAGCACGAGCTCTGTAGC-3’ and 5’-ATCGGTGAGATCAAGTCCTTG-3’. (B) Prx56 primers: 5’-AGACCTCGTTGTGCTTTCAG-3’ and 5’-TCATAAAACAATCTCACAGACATA-3’. (C) DIR12 primers: 5’-GGGATCATAGTTTTGGGGAGTTAATCAT-3’ and 5’-CTTACTGCCTTGAGTGAGTTTGGTCTTG-3’. (D) PMEI6 primers: 5’-CCTCCAATAAAGAACATGACTTC-3’ and 5’-TTACAAGCCATCAAAAGCAA-3’. (E) SBT1.7 primers: 5’-ATGTCTTCTTCGTTTCTCTCCTC-3’ and 5’-TGCAGCTAAAATATCTGAGCTG-3’

Table 2.

Detail of the mutant lines used for the expression level analysis of the genes listed in Table 1

Accession #	Locus	Seed stock #
At4g36920	APETALA 2 (AP2)	CS3084 (ap2–9) in Ler
At1g79840	GLABRA 2 (GL2)	CS65 (gl2–1) in Ler
At5g41315	GLABRA 3 (GL3)	CS66 (gl3–1) in Ler
At1g63650	ENHANCER OF GLABRA 3 (EGL3)	CS6516 in gl3 background
At2g32700	MUCILAGE-MODIFIED 1 (MUM1)	CS3903 (mum1–1) in Col
At3g13540	MYB DOMAIN PROTEIN 5 (MYB5)	SALK_030942 (myb5–1) in Col
At4g09820	TRANSPARENT TESTA 8 (TT8)	CS111 (tt8–1) in Ler
At5g24520	TRANSPARENT TESTA GLABRA 1 (TTG1)	CS89 (ttg1–1) in Ler
At2g37260	TRANSPARENT TESTA GLABRA 2 (TTG2)	CS277 (ttg2–1) in Ler

Based on the RT-PCR results and the literature,^{8,13,16,13-16,19,20,32-34} a working model of the regulation network controlling seed mucilage extrusion of the 5 genes has been drawn including the ruthenium red phenotypes (Fig. 2). TFs have been regrouped following their described protein-protein interactions. GL3, known to interact with EGL3 during trichome initiation, has been positioned separately due to its low regulation function. Interestingly, ruthenium red staining test performed on T-DNA mutants of the 2 candidate genes shows that prx56 has a defect in mucilage release whereas dir12 shows a wild type phenotype (Fig. 2).

Figure 2.

Working model of the regulation network controlling seed mucilage extrusion. Gray-lined arrows correspond to already published relationships 8–10, 13–16, 19, 20, 32–34. White-lined arrows are deduced from Figure 1 data. Lines width is proportional to the level of activation. Nodes diameter reflects the weight of the corresponding transcription factor in the regulation network. Pictures of seed mucilage extrusion are from published articles 10, 23, 24 for prx36, pmei6 and sbt1.7. For prx56 and dir12 (mutant lines SALK_010145 and SAIL_60_D04, respectively, obtained from the Nottingham Arabidopsis Stock Center), mucilage extrusion experiments were performed using the ruthenium red protocol described in the aforementioned articles.

Despite the previously reported similar phenotype of prx36, pmei6 and sbt1.7, the corresponding genes appear to be differentially regulated. Indeed SBT1.7 is only slightly activated by a subset of the studied TFs, whereas PRX36 and PMEI6, which are both strongly regulated by GL2 and MUM1, are also triggered by a specific subset of TFs. In addition, PRX36 and DIR12 expression level is controlled by 4 common TFs even though the corresponding mutants display different phenotypes (Fig. 2). Finally, PRX36 and PRX56 are not regulated by similar TFs among those tested in this study while the corresponding mutants present a delay of mucilage release.

Seed mucilage contains a part of homogalacturonan pectin which can be differentially methylesterified, modifying its physicochemical properties. PMEI6 and SBT1.7 could control the degree of pectin methylesterification through different pathways thanks to the regulation of specific PME activities, respectively.¹⁰ This hypothesis could be confirmed by reverse genetic analysis of the different PME mutants. The class III Peroxidases could favor cell wall polarized weakening enabling proper mucilage release upon imbibition.^24,35 Finally, the nearly identical regulatory pattern of PRX36 and DIR12 is in good agreement with the very high score of DIR12 within the PRX36 ATTED-II network and their very similar spatiotemporal expression patterns.^24,28 However, this does not preclude the existence of a common mucilage release phenotype as demonstrated here. It could be hypothesized that MSC are not only linked to abundant polysaccharidic secretion but also harbors specialized natural products for defense purpose such as neolignans.²⁹

The synthesis of mucilage and of the columella, and the release of mucilage are complex processes which need to be controlled in a spatio-temporal manner. Due to the complexity of these processes, different pathways and hierarchy of regulation, including numerous master TFs and their target genes, are not surprising. We can imagine that the NARS1 and 2³⁶ and other unknown partners could also regulate the expression of target cell wall proteins linked to mucilage production, modification and extrusion but also putatively linked to the synthesis of natural products such as neolignans.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors are grateful to Helen North for providing seeds of various mutants tested in this study.

Funding

The authors are thankful to the Université Paul Sabatier Toulouse 3 and to the Center National de la Recherche Scientifique (CNRS) for granting their work.