Control of root hair development in Arabidopsis thaliana by an endoplasmic reticulum anchored member of the R2R3-MYB transcription factor family

Authors


(fax +1 517 353 9168; e-mail fb@msu.edu).

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.

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 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 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

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).

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 μm, respectively.

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 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).

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 μm.

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 plant-specific 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 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. <50–70 kDa; Nigg, 1997). Because the predicted molecular weight of YFP–maMYB84–309 is approximately 52 kDa, we wanted to ensure that its targeting to the nucleus did not occur simply by diffusion. To do so, we performed fluorescence recovery after photobleaching (FRAP) analyses to monitor the kinetics of fluorescence recovery of YFP–maMYB84–309 upon bleaching (Cole et al., 1996). We expected that if the movement of maMYB84–309 into and out of the nucleus was selective, and/or if maMYB84–309 would bind DNA, as it is generally expected for transcription factors (Reich and Liu, 2006), then YFP–maMYB84–309 would have different dynamics compared with a fluorescent fusion that moved into and out of the nucleus by simple diffusion. Therefore, as a reference for simple diffusion dynamics, we used cytosolic YFP, which is free to move into and out of the nucleus due to its size of approximately 27 kDa (Reits and Neefjes, 2001). To proceed with this experiment, identical areas within the nuclei of cells, either expressing YFP–maMYB84–309 or cytosolic YFP were photobleached and then monitored for recovery of fluorescence (Figure 3a). We noticed that, differently from YFP–maMYB84–309, cytosolic YFP fluorescence recovered in the nuclei at a relatively fast rate. To quantify this phenomenon, we measured the mobile fraction, which estimates the available pool of protein to recover in a bleached area (Reits and Neefjes, 2001) and found that the mobile fraction for YFP–maMYB84–309 (median = 20%) was significantly lower (P = 1.73E−06) than that of cytosolic YFP (median = 65%) at 300 sec post-bleach (Figure 3b,c), indicating that YFP–maMYB84–309 movement into and out of the nucleus does not occur by simple diffusion. These data indicate that YFP–maMYB84–309 is localized to the nucleus where it is largely immobilized, likely due to binding of DNA in the nucleus as R2R3-MYB proteins are well known to bind DNA (Stracke et al., 2001). These observations leave open the possibility that this form of maMYB may function to regulate gene expression.

Figure 3.

 The putative transcription factor domain of maMYB is specifically targeted to the nucleus as shown by FRAP.
(a) Selected frames of a time lapse FRAP experiment (pre-bleach, bleach and recovery) in tobacco leaf epidermal cells expressing either cytosolic YFP or YFP–maMYB84–309. Time of individual frames is indicated at the left hand corner in seconds (sec). After photobleaching, cytosolic YFP, which can freely diffuse into and out of the nucleus, shows recovery of fluorescence compared with YFP–maMYB84–309. Scale bars: 10 μm.
(b) A plot of the mobile fraction values (%) for cytosolic YFP and YFP–maMYB84–309 shows that the mobile fraction of YFP–maMYB84–309 is substantially less than that of cytosolic YFP after 300 sec of recovery after photobleaching.
(c) The median mobile fraction for cytosolic YFP after 300 sec was 65% while the median mobile fraction for YFP–maMYB84–309 was 20%, indicating that maMYB84–309 is specifically targeted to the nucleus.

maMYB84–309 has transcriptional activity

To establish whether maMYB84–309 could have transcriptional activity in Arabidopsis, we tested whether its expression could influence gene expression. To test this, we generated stable Arabidopsis transformants expressing YFP–maMYB84–309 under the control of the CaMV 35S promoter (Figure S2a) and carried out a preliminary microarray experiment with Col-0 and YFP–maMYB84–309 expressing lines (E. Slabaugh and F. Brandizzi, unpublished data). We then performed quantitative RT-PCR (qRT-PCR) on two genes that were shown to be differentially regulated. We selected RHS14 (ROOT HAIR-SPECIFIC 14) and PDI9 (PROTEIN DISULFIDE ISOMERASE 9) that encode a member of the pectate lyase family of proteins (Won et al., 2009) and a protein associated with ER stress, respectively (Lu and Christopher, 2008). These genes were selected as, on the basis of publicly available microarray data (eFP Browser; Winter et al., 2007), they are highly expressed in roots, similarly to maMYB (Figure S3). Compared with the controls (Col-0 and Col-0 expressing free YFP), we found that RHS14 was significantly downregulated in YFP–maMYB84–309 expressing lines, and PDI9 was significantly upregulated in the YFP–maMYB84–309 lines (Figure S2b,c). Although we cannot establish whether these genes are direct targets of maMYB, these results indicate that maMYB84–309 can regulate gene expression in Arabidopsis suggesting that this portion of maMYB may function at as a transcription factor.

Silencing of maMyb results in a reduction of root hair length in Arabidopsis thaliana

To gain insight about maMYB’s function we used a gene silencing approach. As we were not able to identify T-DNA mutants with reduced or absent maMyb transcript from public germplasm banks, we generated knockdown lines using an RNAi approach. We identified a region of 400 bp of maMyb that appears to be unique in the Arabidopsis genome through the dsCheck website (Naito et al., 2005) and BLAST (Altschul et al., 1997) analyses and targeted this region to achieve specific silencing of maMyb. Arabidopsis Columbia-0 (Col-0) plants were transformed with either the pFGC5941 binary vector carrying the cDNA sequence for maMyb suppression, or the empty pFGC5941 vector as a control. To confirm effective silencing of maMyb, qRT-PCR was performed on independent and homozygous RNAi transformant lines. As a result, we selected and worked with two independent lines whose maMyb transcript levels were found to be 60–80% reduced compared with wild type or lines expressing the empty vector for RNAi silencing (Figure 4). We next observed the knockdown lines for phenotypic effects and discovered that the length of root hairs in the RNAi lines were severely reduced (Figure 5a–d). Root hair lengths were quantified and were shown to be significantly shorter in the maMyb RNAi lines compared with wild type (Figure 5e), while root hair density and primary root lengths remained unaffected (Figure 5f,g). Other phenotypic aspects were also analyzed including fresh weight of plants, rosette size and trichomes, no differences were observed (Figure S4), further implying a specific role for maMYB in root hair elongation.

Figure 4.

 Generation of knockdown lines of maMyb. RNAi lines were constructed that effectively silenced the expression of maMyb by approximately 60–80% as shown by qRT-PCR. Data are expressed relative to Col-0 and represent three experimental replicates. Error bars represent standard error. Asterisks represent a P-value of <0.01.

Figure 5.

 Silencing of maMyb results in a reduction of root hair length in Arabidopsis thaliana.
(a–d) Root hairs of two independent RNAi lines (RNAi-1 and RNAi-2) showed a reduction of root hair length compared with that of wild-type (Col-0) and empty vector lines. Data are representative of five experimental replicates. Scale bar: 0.5 mm.
(e) Root hair lengths were quantified at approximately 5 mm above the root tip, within the differentiation zone.
(f) Root hair density and (g) primary root lengths were not affected between RNAi lines and wild type or empty vector lines suggesting that maMyb specifically functions in root hair elongation. Asterisks represent a P-value of <0.01.

Exogenous auxin rescues the short root hair phenotype of maMyb suppressed lines

There is a tight connection between root hair elongation and hormonal cues. For example, the plant hormone auxin has been shown to promote root hair elongation (Masucci and Schiefelbein, 1996). Accordingly, the exogenous application of an auxin analog results in the increase of root hair length in wild type Arabidopsis seedlings (Pitts et al., 1998). As plants with a reduced abundance of maMyb transcript showed shorter root hairs compared with wild type and auxin can induce root hair elongation, we wanted to test if maMYB may function as part of auxin signaling pathways. To test this, we looked at maMyb mRNA levels in wild-type and empty vector lines in response to the auxin analogs, 2,4-D and IAA, both of which are known to induce root hair elongation (Pitts et al., 1998). Seven-day-old seedlings were treated with either 0.1 μm 2,4-D, 10 μm IAA or a mock treatment and mRNA levels were analyzed after 48 h. As controls, we looked at the mRNA of two genes whose expression changes in response to auxin: SCPL-14 (At3g12230), which is downregulated by auxin, and IAA2 (At3g23030), which is upregulated in response to auxin (Figure S5; Hagen and Guilfoyle, 2002; Nemhauser et al., 2004). After confirming that the hormone treatments were effective, we looked at the transcript levels of maMyb. The expression of maMyb mRNA increased by approximately 40% in the Col-0 and empty vector lines in response to 2,4-D and IAA compared with mock treatment (Figure 6). These results suggest that auxin may play a role in the regulation of maMyb at a transcriptional level.

Figure 6.

 Expression of maMyb mRNA is induced by a auxin analogs. mRNA levels of maMyb were measured in Col-0 and empty vector lines that were treated for 48 h with either (a) 0.1 μm 2,4-D, (b) 10 μm IAA or a mock treatment. Seedlings treated with 2,4-D and IAA showed a marked increase of maMyb mRNA in the Col-0 and empty vector lines compared with the mock treatment. In contrast, maMyb mRNA was not increased in the maMyb RNAi lines in response to treatment with 2,4-D or IAA compared with the mock treatment. Data was compiled from three experimental replicates. Error bars represent standard error. Asterisks represent a P-value of <0.05.

We next checked the influence of the auxin on maMyb mRNA levels in the RNAi lines treated with the auxin analogs and verified a small but not significant increase (Figure 6). In light of these results, we tested if the exogenous application of 2,4-D could rescue the root hair phenotype in the maMyb RNAi lines. To do so, we followed a standard protocol for measuring root hair elongation in the presence of hormones (see Experimental procedures) (Pitts et al., 1998), where wild type and maMyb RNAi lines were grown for 7 days on control medium and then transferred to media containing either 0.1 μm 2,4-D, or a mock treatment (Pitts et al., 1998). Root hair measurements were taken after 48 h of hormone treatment. As a result of exogenous application of 2,4-D, the root hair phenotype of the maMyb RNAi lines were fully recovered and root hair lengths were similar to that of wild type (Figure 7), demonstrating that the maMyb RNAi lines are sensitive to auxin and leaving the possibility open that the rescue of the root hair length phenotype by auxin may occur via a maMYB-independent mechanism.

Figure 7.

 Short root hair phenotype of maMyb silenced lines can be rescued by an exogenous auxin analog. Exogenous 2,4-D results in the full recovery of root hair lengths in maMyb RNAi lines compared with wild type. Error bars represent standard deviation; data are representative of three experimental replicates. Asterisks represent a P-value of <0.05.

Discussion

In this study, we show that maMYB is localized at the ER with an R2R3-MYB transcription factor domain exposed to the cytosol. The truncated form of maMYB lacking transmembrane domains, maMYB84–309, is localized to the nucleus where it can regulate gene expression. We also show that the silencing of maMyb results in plants that exhibit shorter root hairs compared with wild type but that root hair density is not affected, implying that maMYB may function in root hair elongation. Additionally, we provide evidence that the short root hair phenotype can be rescued by an exogenous auxin analog, suggesting that maMYB may be involved in auxin signaling pathways. Finally, we demonstrate that expression of maMyb mRNA is induced in response to auxin suggesting that auxin is a key regulator of maMYB. These results add to our growing knowledge of the regulation of root hair development and postmitotic cell growth in Arabidopsis by the identification of an auxin inducible gene that contributes to root hair elongation.

maMYB is an R2R3-MYB transcription factor involved in root hair elongation

R2R3-MYB transcription factors make up the largest subfamily of MYB transcription factors in Arabidopsis and many are involved in plant-specific processes (Stracke et al., 2001) Over 80% of R2R3 MYB factors share 47% and 60% similarity between their R2 and R3 MYB repeats, respectively (Stracke et al., 2001). Due to this similarity among R2R3-MYBs, functional specificity of these proteins is thought to be determined by the presence of unique motifs or due to differential temporal and/or spatial regulation (Stracke et al., 2001). There are several R2R3-MYB factors that have been shown to alter root architecture in Arabidopsis, affecting the formation of lateral roots or the length of the primary root (Shin et al., 2007; Petroni et al., 2008; Devaiah et al., 2009; Ding et al., 2009; Mu et al., 2009; Seo et al., 2009). In addition to affecting the architecture of roots, many of these R2R3-MYB factors have also been implicated in stress response pathways including drought stress and phosphate starvation (Devaiah et al., 2009; Ding et al., 2009; Seo et al., 2009). Here, we have shown that the suppression of maMyb results in reduced root hair lengths without an effect on primary root development or root hair density. Therefore, we propose that maMYB is involved in root hair elongation.

maMYB is associated with the ER membrane and contains a transcription factor domain

Through cell fractionation experiments and microscopy analyses, we have shown that maMYB is an ER-localized membrane protein. qRT-PCR analyses showing that the region of maMYB encompassing the R2R3-MYB transcription factor domains have the ability to alter gene expression, suggest that maMYB may have a role in regulating transcription. However, in order for this to occur, the transcription factor domain of maMYB must be released from the ER to translocate to the nucleus and regulate gene expression. Evidence for a release of the transcription factor domain of maMYB from the ER membrane in vivo would expand the repertoire of membrane-tethered transcription factors (MTTFs) in Arabidopsis thaliana and would establish maMYB as a canonical MTTF. To date, seven MTTFs have been characterized in Arabidopsis. These either belong to the bZIP or NAC family of transcription factors (Chen et al., 2008; Seo et al., 2008). Members of the bZIP family have been found to associate with the ER while members of the NAC family mostly associate with the plasma membrane (Chen et al., 2008; Seo et al., 2008). MTTFs have been shown to function in a wide variety of pathways in Arabidopsis, including stress response, flowering time and cell division (Chen et al., 2008). Here, we have shown that maMYB functions in specific aspects of root hair development, highlighting the possibility that maMYB may act as a MTTF that is specifically regulated.

Because maMYB is the only membrane bound member of the R2R3-MYB family of transcription factors in Arabidopsis (Kim et al., 2010), then it would be the only MTTF in such a broad family. However, although our topology analyses have shown that the transcription factor domain of maMYB is exposed to the cytosol, we have not demonstrated that maMYB functions as a canonical MTTF. In experiments focused to detect proteolytic cleavage of maMYB, we have not observed the generation of the truncated form of maMYB in vivo (E. Slabaugh, M. Held and F. Brandizzi, unpublished data). Lack of detection of a truncated form of maMYB may be because the in vivo cleavage of maMYB occurs at such a low level or it may occur only in specialized cells. Therefore, the levels of truncated maMYB visualized by western blot analyses would be below the detection limit. Together, the presence of an R2R3-MYB domain in the protein sequence, the exposure of such a domain in the cytosol and its specific targeting to the nucleus coupled with transcriptional activity, strongly suggest a role for maMYB as a MTTF; however, further investigation is needed to empirically validate this possibility.

The function of maMYB may be integrated with auxin signaling

Evidence in vivo on the role of maMYB in root tissue was provided by gene silencing experiments. We have shown that the silencing of maMyb results in the reduced length of root hairs. The evidence that the short root hair phenotype in maMyb suppressed lines can be rescued by an auxin analog and that the expression of maMyb mRNA is induced by auxin leads to the possibility that the function of maMYB may be integrated with auxin signaling pathways. The determination of root hair cell fate is well documented and occurs in an organized fashion. Initially, root hair cell fate is determined by the position of the epidermal cells over the cortex (Schiefelbein, 2003). Positional cues are then transduced to epidermal cells, resulting in the expression or suppression of GLABRA2, which then gives rise to the formation of N- or H-cells (Schiefelbein, 2003). However, processes that occur after the root hair has been initiated do not appear to proceed according to such a linear model. There are many factors that contribute to root hair elongation that include hormonal cues, environmental stresses and transcriptional regulation (Gilroy and Jones, 2000; Ringli et al., 2005). Hormonal cues such as auxin and ethylene have been shown to function downstream of cell-fate patterning mechanisms and in root hair elongation (Masucci and Schiefelbein, 1996). Many auxin insensitive mutants have been identified as having aberrant root hairs, including axr2-1, axr3-3 and slr1-1 (Reed, 2001). However, genes that affect root hair development, and yet retain integrity of the auxin signaling pathway are, to our knowledge, less abundant in the literature. Examples of genes that alter root hair development and are sensitive to exogenous auxin include the rhd6 (ROOT HAIR DEFECTIVE 6) mutant, which exhibits altered root hair initiation that can be rescued by application of an auxin analog (Masucci and Schiefelbein, 1994). Similarly, RSL4 (ROOT HAIR DEFECTIVE 6-LIKE 4), transcript is induced by auxin and loss of RSL4 results in seedlings with shorter root hairs compared with wild type (Yi et al., 2010). However, the direct mechanism between these genes and auxin signaling in relation to root hair development is still unclear. Auxin signaling pathways are proving to be increasingly complicated with many points of cross-talk between other signaling pathways such as brassinolide, ethylene, gibberellin, abscisic acid, light signaling and phosphate deficiency, among others (Pitts et al., 1998; Lopez-Bucio et al., 2002; Swarup et al., 2002; Kim et al., 2006).

The rescue of the short root hair phenotype in the maMyb RNAi lines in response to treatment with 2,4-D cannot be attributed to an increase in maMyb transcript, as we found by qRT-PCR that maMyb transcript is not induced at statistically significant levels in these lines. This observation could be attributed to the continued effect of the RNAi mechanism in the knockdown lines and therefore any increase of transcript that would normally occur in response to 2,4-D would be sequestered. Another possibility is that the increase of maMyb mRNA in the RNAi lines is not quantifiable due to technological limitations (i.e. qRT-PCR sensitivity) and the low quantity of maMyb mRNA present. An alternative, and more likely explanation is that the rescue of root hair phenotype occurs through a different mechanism independent of maMYB. One possible alternative pathway could be attributed to a hypersensitivity to auxin as a result of the suppression of maMyb. Hypersensitivity to auxin is exhibited in plants that are grown under phosphate-limiting conditions (Lopez-Bucio et al., 2002). Phosphate deficiency results in increased lateral branching of the roots and promotes root hair elongation as a means to increase the absorptive surface of the roots to allow for an increased capacity for nutrient uptake (Lopez-Bucio et al., 2002). These adaptations of phosphate deficient plants have been shown to be a result of auxin hypersensitivity (Lopez-Bucio et al., 2002). If maMyb RNAi lines prove to be hypersensitive to auxin, this may imply involvement of maMYB in such a pathway.

Though we do not yet know the mechanism of maMYB’s function in vivo, we deduce through analysis of mRNA levels and hormonal rescue, that maMYB likely functions as a part of an auxin signaling pathway. One outstanding question remains as to what the advantage would be for a plant to adopt an ER-localized transcription factor, such as maMYB in light of its potential ability to be induced by auxin. The ER has been shown to play large role in auxin homeostasis and is believed to have evolved the capacity for auxin signaling before signaling pathways developed in the plasma membrane (Friml and Jones, 2010). Additionally, possible communication pathways between subcellular compartments, including signaling from the ER to the nucleus, is an area of research of the auxin signaling pathway that is currently in early stages (Friml and Jones, 2010). A potential role for a MTTF at the ER that is responsive to auxin may reflect a need for rapid communication between the ER and the nucleus concerning auxin signaling. Our characterization of maMYB in relation to root hair elongation and auxin signaling adds to our growing knowledge not only about potential mechanisms of postmitotic cell growth, but also about the integration of auxin signaling through ER-associated proteins and root hair elongation in Arabidopsis.

Experimental Procedures

Molecular cloning

Standard molecular cloning techniques were used. All constructs were driven by the cauliflower mosaic virus 35S promoter. The coding sequence (CDS) for maMyb (AGI: At5g45420) was obtained from ABRC (http://abrc.osu.edu/). The construct encoding maMyb (maMYB–YFP) included the entirety of the CDS (1–930 bp) fused upstream to venues YFP (vYFP) (Nagai et al., 2002) in the binary vector pVKH18En6 (Batoko et al., 2000). The CDS for PVA12 (AGI: At2g45140) was obtained from ABRC and subcloned into the pVKH18En6 binary vector upstream of vYFP. For BiFC procedures, YFP was split into two protein halves, nYFP (1–154 aa) and cYFP (155–239 aa) and cloned into the binary vector pVKH18En6. The truncated form of maMyb (YFP–maMYB84–309) was generated by subcloning the region 5′ of the transmembrane domains (252–930 bp) downstream of vYFP in the binary vector pVKH18En6. RNAi constructs were generated by subcloning a unique region in maMyb (163–563 bp), as determined by the dsCheck website (Naito et al., 2005), into the binary vector pFGC5941 (ChromDB, http://www.chromdb.org/index.html), which allows the bidirectional cloning of a partial cDNA sequence to generate an RNAi hairpin structure sufficient for silencing.

Expression in plant cells

Transient transformation in Nicotiana tabacum (cv Petit Havana), grown at 25°C, was carried out by Agrobacterium tumefaciens (GV3101) infiltration at an optical density (OD600) of 0.05. For bimolecular fluorescence complementation, all constructs were infiltrated at an optical density (OD600) of 0.2. Stable transformation of Arabidopsis thaliana was carried out using the floral dip method (Clough and Bent, 1998).

Sampling and imaging

Confocal imaging was performed using an inverted LSM 510 Meta confocal microscope using the ×40 or ×63 oil immersion objectives (Zeiss, http://www.zeiss.com). Epidermal cells of transformed tobacco plants were imaged 48 h post-infiltration. For imaging YFP, the 514 nm argon laser was used with a 514 nm dichroic beam splitter and a 520–555 nm bandpass filter. For imaging of GFP, the 458 nm argon laser was used with a 458 nm dichroic beam splitter and a 475–525 nm bandpass filter. Co-expression analysis of GFP and YFP was carried out using the multi-track function of the microscope and a 560–615 nm bandpass filter for YFP (Brandizzi et al., 2002). For image processing, Photoshop Imaging Suite was used (Adobe, http://www.adobe.com).

Fluorescence recovery after photobleaching (FRAP)

Using the LSM 510 Meta, Zeiss, a circular region with an area of 81.35 μm2 was used to monitor photobleaching. Each sample was bleached for 50 sec followed by a 530 sec recovery taking images every 5 sec. Six replicates were analyzed for cytosolic YFP and 11 replicates were analyzed for YFP–maMYB84–309. The mobile fraction was calculated (Reits and Neefjes, 2001) using Kaliedograph software (http://www.synergy.com/).

Plant materials and growth conditions

Wild type (Col-0) or mutant seeds were surface sterilized using 10% bleach followed by 70% ethanol, rinsing twice with distilled water after each step. Seeds were then dried on filter paper and stratified for 2 days at 4°C. Arabidopsis thaliana seedlings were grown at 21°C in a 16-h photoperiod under illumination of 100 μE and an 8 h dark period on ½ strength Murashige and Skoog media, Gamborg’s B5 vitamins and 1% sucrose with 0.8% agar.

Root hair measurements and hormone treatment

For root hair analysis, 10-day-old seedlings, grown upright, were imaged using a Stemi 2000-C Stereomicroscope (Zeiss). Digital images were analyzed with ImageJ software (http://rsbweb.nih.gov/ij/). Root hair measurements were taken according to (Yi et al., 2010) with modifications. Root hairs were measured approximately 5 mm from the root tip, within the differentiation zone, and five measurements were taken of perpendicular root hairs that were in focus, for each plant. Three plants were measured for each line and at least five experimental replicates were carried out with similar results.

For exogenous application of auxin analogs, plants were grown upright on control media for 7 days. Then seedlings were transferred to media containing either 0.1 μm 2,4-D, (Pitts et al., 1998), 10 μm IAA or a mock treatment (0.005% or 0.001% ethanol, respectively) and were imaged and tissue was collected for qRT-PCR analysis after 48 h of hormone treatment. Root hairs were measured as described above for plants grown on 2,4-D, and were measured just above the root tip where root hair growth occurred. Each experiment was performed at least three times with similar results.

Additional phenotypic analysis

Primary root lengths were measured using 10-day-old seedlings grown upright on ½ strength Murashige and Skoog media and 1% sucrose with 0.8% agar. Root length measurements were analyzed using ImageJ software (http://rsbweb.nih.gov/ij/), at least six plants were measured per line. Rosette diameters were measured using 3-week-old plants; measurements were analyzed using ImageJ software (http://rsbweb.nih.gov/ij/) and at least three plants were analyzed per lines. Trichomes were imaged using 10-day-old seedlings grown on ½ strength Murashige and Skoog media, Gamborg’s B5 vitamins and 1% sucrose with 0.8% agar.

Protein extraction and membrane fractionation

Protein extraction was performed on either wild type plants or plants harboring maMYB–YFP (see Molecular cloning). Five, 7-day-old seedlings were used per sample. Seedlings were immediately frozen in liquid nitrogen upon harvest. Soluble proteins (S) were extracted by adding 50 μl of Extraction Buffer A [0.2 m NaCl, 0.1 m Tris pH 7.8, 1 mm EDTA pH 8.0, 200 mm PMSF and 3.3% P9599 (Invitrogen, http://www.invitrogen.com/)] to each sample. Samples were ground using a micropestle and were incubated on ice for 10 min followed by centrifugation at 20 817 g at 4°C for 15 min. Supernatants were removed as the soluble fraction (S) and stored at −80°C for later use. An additional 250 μl of Extraction Buffer A was added to the samples and vortexed to mix. Samples were then sonicated for 30 sec at 0.5× followed by centrifugation at 20 817 g at 4°C for 15 min. Supernatant was removed and discarded. An additional 250 μl of Extraction Buffer B (Buffer A, 0.2% Triton-X and 0.2%β-mercaptoethanol) was added to the samples and vortexed to mix. Samples were again sonicated for 30 sec at 0.5× and collected as the membrane fraction (P) and stored at −80°C for later use. For western blot analyses, 16 μl of sample (S or P) was combined with 8 μl of 6× sample buffer (0.1% w/v bromophenol blue, 5 mm EDTA, 20 mm Tris pH 8.8, 1 m sucrose, 10% w/v SDS and 1 m dithiothreitol). Equal volumes of sample were loaded onto a 10% SDS–polyacrylamide gel and then transferred to a nitrocellulose membrane by electroblotting. The membrane was blocked overnight at 4°C with phosphate-buffered saline (PBS), 0.1% Tween-20 and 10% milk powder. The membrane was then incubated with PBS and 0.1% Tween-20 with anti-GFP serum from rabbit (Abcam 6795) at a 1:3000 dilution for 1 h at room temperature. Following steps were performed previously (Crofts et al., 1999). These results are representative of at least three independent experimental replicates.

RNA analysis

For qRT-PCR, of Col-0 and empty vector lines that were treated with either 2,4-D or a mock treatment for 48 h, whole seedlings were harvested with three seedlings pooled per sample. Tissue was flash frozen and RNA was extracted using TRIzol reagent according to manufacturer’s instructions (Invitrogen). DNase treatment of the RNA was carried out with TURBO DNA-free kit (Ambion, http://www.ambion.com/) according to manufacturer’s instructions. First strand cDNA synthesis was performed using SuperScript III First Strand Synthesis Supermix kit (Invitrogen) using random hexamers as primers in a total reaction volume of 20 μl, according to manufacturer’s instructions. Synthesized cDNA was diluted to 2.5 ng μl−1 with water.

For real-time quantitative RT-PCR, 2 μl (5 ng) of cDNA was used in a total reaction volume of 10 μl using the Applied Biosystems FAST 7500 real-time PCR system in FAST mode with Fast SYBR Green Master Mix reagent (Applied Biosystems, https://products.appliedbiosystems.com), according to manufacturer’s instructions. Reactions were performed in triplicate and with two independent lines for each mutant line. Data were analyzed using the relative standard curve method normalizing for the reference gene Actin2 (AGI: At3g18780).

Primer sequences used are listed in Table 1.

Table 1.   Primers
GeneAGIForward sequence (5′ to 3′)Reverse sequence (5′ to 3′)
maMYBAT5G45420TTTTCCTAAGGAGGCAGCTATGAGCTTTCGACTTCCCAGGTACTG
ACT2AT3G18780TCAGATGCCCAGAAGTCTTGTTCTGGATTCCAGCAGCTTCCA
SCPL-14AT3G12230TCAGGGCTATGTGATCGGAAAGGAATGCGAGAGTCTTTATCATGA
IAA2AT3G23030AAAGGATCTGGATTTGTACCAACATACCATGGAACATCACCAACCA
RHS14AT4G2280TGGCAGCATTTCGCTACCTAGTTTCCCGGAGCTACTGTGAA
PDI9AT2G32920GGCCCTGTTGAAGTGACTGAACAGCAGAACCACACTTCTTTTCC

Statistical analysis

Data were analyzed using Student’s t-test to obtain P-values for each data set. Data sets for root hair analysis consisted of three data points that represented the average for each biological replicate of each sample. Data sets for qRT-PCR consisted of two data points that represented the average for each biological replicate of each sample.

Acknowledgements

We acknowledge support by the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (award number DE-FG02-91ER20021). We thank Starla Zemelis for help with the generation of stable Arabidopsis transformants.

Ancillary