The Plant Journal (2011) 67, 395–405
doi: 10.1111/j.1365-313X.2011.04602.x
Control of root hair development in Arabidopsis thaliana by an endoplasmic reticulum anchored member of the R2R3-MYB transcription factor family Erin Slabaugh, Michael Held and Federica Brandizzi* Plant Research Laboratory, Department of Energy, Michigan State University, East Lansing, MI 48824, USA Received 2 March 2011; accepted 5 April 2011; published online 20 May 2011. * For correspondence (fax +1 517 353 9168; e-mail
[email protected]).
SUMMARY The evolution of roots and root hairs was a crucial innovation that contributed to the adaptation of plants to a terrestrial environment. Initiation of root hairs involves transcriptional cues that in part determine cell patterning of the root epidermis. Once root hair initiation has occurred, elongation of the root hair takes place. Although many genes have been identified as being involved in root hair development, many contributors remain uncharacterized. In this study we report on the involvement of a member (here dubbed maMYB) of the plant-specific R2R3-MYB family of transcription factors in root hair elongation in Arabidopsis. We show that maMYB is associated with the endoplasmic reticulum membrane with the transcription factor domain exposed to the cytosol, suggesting that it may function as a membrane-tethered transcription factor. We demonstrate that a truncated form of maMYB (maMYB84–309), which contains the R2R3-MYB transcription factor domain, is localized and retained in the nucleus, where it regulates gene expression. Silencing of maMyb resulted in plants with significantly shorter root hairs but similar root hair density compared with wild type, implying a role of the protein in root hair elongation. 2,4-D (2,4-dichlorophenoxyacetic acid), an exogenous auxin analog that promotes root hair elongation, rescued the short root hair phenotype and maMyb mRNA was induced in the presence of 2,4-D and IAA (indole-3-acetic acid). These results indicate a functional role of maMYB, which is integrated with auxin, in root hair elongation in Arabidopsis. Keywords: root hair, endoplasmic reticulum, auxin, MYB-transcription factors.
INTRODUCTION The first land plants faced harsh environmental conditions and developed innovative ways to survive (Renzaglia et al., 2000; Heckman et al., 2001). Major evolutionary events occurred when plants began to colonize land; these included the development of complex life cycles, new organs and tissues and morphological adaptations such as leaves, stems, trichomes, stomata and roots (Kenrick and Crane, 1997). The emergence of roots provided plants with anchorage as well as water and nutrient acquisition, thus allowing plants to better colonize a terrestrial environment (Schiefelbein and Benfey, 1991; Raven and Edwards, 2001; Ishida et al., 2008). Three developmental zones are clearly defined in roots: the meristematic zone where cell division occurs, the elongation zone where cells elongate and the differentiation zone where root hairs and lateral roots form (Schiefelbein and Benfey, 1991). Root hairs aid the plant by providing an increased absorptive surface for nutrient and ª 2011 The Authors The Plant Journal ª 2011 Blackwell Publishing Ltd
water uptake (Peterson and Farquhar, 1996; Raven and Edwards, 2001). The determination of trichoblast (hair forming or H-cell) and atrichoblast (non-hair forming or N-cell) cell fate is established by cell positioning in relation to the underlying cortical cells and by transcriptional cues (Dolan et al., 1994; Masucci and Schiefelbein, 1996; Lee and Schiefelbein, 2002; Ringli et al., 2005). Once root hair cell fate has been determined, root hair elongation occurs. Root hairs elongate via tip growth, in a similar fashion to pollen tubes (Kost et al., 1999). During this process, many intracellular changes occur including those related to the pH in the cytoplasm and apoplast, delivery of cell wall material, formation of a calcium gradient, as well as levels of reactive oxygen species at the tip of the root hair (Jones et al., 2006). The cytoskeleton is also involved in root hair growth as the presence of microtubules and actin are necessary for proper root hair elongation (Kost et al., 1999). The plant hormones 395
396 Erin Slabaugh et al. auxin and ethylene have also been shown to positively regulate root hair elongation and specifically act downstream of transcriptional regulators (Masucci and Schiefelbein, 1996; Pitts et al., 1998). In addition, jasmonic acid and methyl jasmonate induce root hair formation through interaction with ethylene (Zhu et al., 2006). Environmental cues have also been shown to affect root hair development as iron and phosphate deficiency induces root hair formation (Rubio et al., 2009). A transcriptional regulatory network is in place for the formation of H- or N-cells. The most recent views support that a LRR receptor-like kinase SCRAMBLED (SCM) is responsible for position-dependent cell patterning and regulates the expression of a downstream transcriptional regulator WEREWOLF (WER) (Schiefelbein and Lee, 2006; Kwak and Schiefelbein, 2007). WER, in turn, induces a lateral-inhibition pathway in neighboring cells involving CAPRICE (CPC) and TRIPTYCHON (TRY) and ultimately regulates the expression of GLABRA2 (GL2) which leads to the suppression or induction of root hairs (Schiefelbein and Lee, 2006). Although multiple genes have been identified that are involved in root hair formation and elongation, these processes remain to be fully elucidated (Jones et al., 2006; Won et al., 2009). Considering the multitude of developmental, biological and environmental factors that influence the process of root hair formation, it is plausible to hypothesize that several other transcriptional regulators, in addition to the proteins already known, may influence these processes. Here, we show that a member of the R2R3-MYB transcription factor family, maMYB, is associated with the endoplasmic reticulum (ER) membrane and influences the elongation of root hairs but not their density. We demonstrate that such a phenotype can be rescued by exogenous 2,4-D and that maMyb mRNA is induced by 2,4-D and IAA. These data indicate that maMYB has a functional role in plant-specific processes related to root hair elongation and auxin signaling, possibly functioning as an ER membrane-tethered transcription factor. RESULTS
Figure 1. maMYB contains transmembrane domains, a putative R2R3 MYB domain and is localized to the endoplasmic reticulum. (a) A schematic diagram of maMYB depicts two predicted transmembrane domains at the N-terminus of the protein and a putative R2R3-MYB transcription factor domain at the C-terminus. (b) Membrane fractionation experiments show maMYB–YFP (expected molecular weight of approximately 60 kDa; top panel, arrow) is associated mostly with the pellet fraction (P) rather than in the soluble fraction (S). Controls include ST–GFP (33 kDa, lower panel, arrow) as a membrane associated protein, and YFP as a soluble protein (approximately 27 kDa; lower panel, arrow). Accordingly, ST–GFP signal is predominantly in the pellet while YFP is in the soluble fraction. (c) maMYB–YFP localizes to the ER as shown by co-localization with the ER marker ssGFP–HDEL in tobacco epidermal cells. Arrows indicate fluorescence of the ER network (top) and the nuclear envelope (bottom). Scale bars in the top and bottom panels represent 20 and 10 lm, respectively.
Membrane-anchored MYB (maMYB) is localized to the ER In an earlier proteomic analysis of the ER in Arabidopsis thaliana, a protein (here dubbed membrane-anchored MYB, maMYB) containing an R2R3-MYB transcription factor domain (Figure 1a), was identified (Dunkley et al., 2006). The subfamily of R2R3-type MYB-proteins is characteristic of plants and makes up the largest subfamily of MYB transcription factors in Arabidopsis with 126 members (Stracke et al., 2001; Dubos et al., 2010). These proteins contain two imperfect MYB repeats that form helix–turn–helix structures that bind DNA. Bioinformatics analyses suggest that maMYB is the only membrane-anchored protein in the R2R3-MYB
family (Kim et al., 2010), and that it is conserved through the plant lineage with homologs identified in Physcomitrella patens but not in Chlamydomonas reinhardtii (Altschul et al., 1997). To test if maMYB is indeed membrane associated, we performed a membrane fractionation experiment with maMYB–YFP using ST-GFP (Boevink et al., 1998) and free YFP as membrane associated and soluble controls, respectively. The results of which showed that maMYB is mostly associated with the membrane fraction (Figure 1b), validating the bioinformatic predictions (Kim et al., 2010).
ª 2011 The Authors The Plant Journal ª 2011 Blackwell Publishing Ltd, The Plant Journal, (2011), 67, 395–405
Root hair development in Arabidopsis thaliana 397 Once we established that maMYB is membrane associated, we next wanted to determine the subcellular localization of maMYB. Confocal microscopy imaging of a YFP fusion to the C-terminus of maMYB (maMYB–YFP) in cells co-expressing the known ER marker ssGFP–HDEL (Brandizzi et al., 2003) showed a distribution of maMYB at the ER and nuclear envelope (Figure 1c). Together these data provide empirical validation of the earlier bioinformatic (Kim et al., 2010) and proteomic (Dunkley et al., 2006) analyses that maMYB is an ER membrane associated MYB protein. The N and C termini of maMYB face the cytosol Once the ER localization of maMYB was established, we wanted to determine its topology in the membrane. maMYB has two predicted TMDs (ARAMEMNON, http://aramemnon.botanik.uni-koeln.de/) (Figure 1a). We hypothesized that the putative transcription factor domain of maMYB would be exposed to the cytosol; given the presence of two putative TMDs, we expected that also the N-terminus of maMYB would be exposed in the cytosol. To test this, we performed a bimolecular fluorescence complementation (BiFC) experiment (Hu et al., 2002; Zamyatnin et al., 2006). Not only is this method widely used to establish protein– protein interactions, but it can also be used to analyze the topology of membrane proteins (Zamyatnin et al., 2006). This approach is based on the fusion of the membrane protein to a half of YFP and co-expression of such fusion with the remaining YFP half distributed either in the cytosol or in the lumen of the organelle where the membrane protein resides. In protein topology experiments, generation of fluorescence signal is possible only in cells where the split YFP of the membrane protein-half YFP fusion and the complementary half of YFP reside in the same compartment. To proceed with this experiment, we split YFP into two protein halves, nYFP (1–154 aa) and cYFP (155–239 aa). When each of these halves was expressed in tobacco leaf epidermal cells individually, no fluorescence was detected (Figure S1). However, when nYFP and cYFP were co-expressed, YFP fluorescence was restored due to the spontaneous interaction of the two protein halves in the cytosol (Figure 2a), as expected (Zamyatnin et al., 2006). We used an ER resident protein, PVA12, whose C-terminus is known to reside in the ER lumen (Saravanan et al., 2009) as a negative control. When the PVA12–nYFP fusion, with nYFP in the ER lumen, was co-expressed with cytosolic cYFP, no fluorescence was detected (Figure 2b). Having validated the experimental system, we next aimed to determine the topology of maMYB in cells co-expressing maMYB fusions with split YFP with the remaining cytosolic YFP fragment. First, nYFP was fused to the C-terminus of maMYB (maMYB–nYFP) and co-expressed with cytosolic cYFP (Figure 2c). Then, nYFP was fused to the N-terminus of maMYB (nYFP–maMYB) and co-expressed with cytosolic cYFP (Figure 2d). In both cases, fluorescence was restored, indicating that nYFP could interact with cYFP
Figure 2. The C-terminus of maMYB, which contains the putative transcription factor domain, as well as the N-terminus, faces the cytosol. Transient transformation of tobacco leaf epidermal cells shows (a) restoration of fluorescence of the cytosol when co-expressing nYFP and cYFP as a positive control. (b) Fluorescence is not restored when co-expressing PVA12–nYFP and cYFP due to physical separation of the two protein halves of YFP. (c) Restoration of fluorescence of the ER when co-expressing maMYB–nYFP with cYFP. (d) Restoration of fluorescence of the ER when co-expressing nYFP–maMYB with cYFP. (e) Schematic of the topology of maMYB in the membrane of the ER. Scale bar: 20 lm.
when fused to either terminus of maMYB. Consistent with our hypothesis, these data indicate that both maMYB termini are exposed to the cytosol (Figure 2e). A truncated form of maMYB (maMYB84–309), containing the putative transcription factor domain, is localized to and retained in the nucleus We next wanted to gain functional insights into the activity of maMYB. R2R3-MYB factors are mainly involved in plantspecific processes and can be activators or repressors of gene expression (Meissner et al., 1999; Stracke et al., 2001). To test whether maMYB could have transcription factor activity, we generated a truncated form of the protein, maMYB84–309, which lacked the transmembrane domains but contained the putative transcription factor domains (aa 84–309; Figure 1a). We first aimed to establish the subcellular localization of maMYB84–309 by fusing YFP to the N-terminus of maMYB84–309 (YFP–maMYB84–309) for live cell analyses. We hypothesized that the protein would be targeted to the nucleus, as would be expected for active transcription factors (Reich and Liu, 2006). Confocal imaging of cells expressing the construct showed the protein in
ª 2011 The Authors The Plant Journal ª 2011 Blackwell Publishing Ltd, The Plant Journal, (2011), 67, 395–405
398 Erin Slabaugh et al. the nucleus with some labeling of the cytosol (Figure 3a). However, it is well known that untargeted fluorescent protein fusions can diffuse into and out of the nucleus if their molecular weight is lower than that of the molecular sieve of the nuclear pores (i.e.
