Auxin-associated initiation of vascular cell differentiation by LONESOME HIGHWAY

The vascular bundle is the long-distance transport pathway for water, nutrients and signalling molecules that connect all parts of the plant body. Because the vascular system is an important lifeline, the molecular mechanisms underlying vascular development have been keenly investigated for a long time. Recent genetic and physiological studies have identified many key factors, including transcription factors that govern vascular development, suggesting the importance of transcriptional regulation in the developmental process, especially during differentiation of specialized vascular cells (Caño-Delgado et al., 2010; Ohashi-Ito and Fukuda, 2010; Scarpella and Helariutta, 2010). By contrast, the mode of transcriptional regulation of the initiation of vascular initial cells is largely unknown.

We recently identified LONESOME HIGHWAY (LHW), which encodes a bHLH transcription factor. Mutations of this gene eliminate the bilateral symmetry of the vascular pattern and reduce the number of cells in root vasculatures, resulting in roots with the monoarch vasculature that contains only single xylem and phloem poles (Ohashi-Ito and Bergmann, 2007; Parizot et al., 2008). The lhw mutant also shows a weak auxin-related phenotype. Polar auxin flow mediated by auxin efflux carriers has been implicated as the earliest event in vascular differentiation (Scarpella et al., 2006; Donner et al., 2009; Wenzel et al., 2007). In this study, therefore, we investigated the role of LHW in the initiation of vascular formation, focusing on the association of LHW with the regulation of auxin flow.

A. thaliana ecotype Columbia was used for all experiments. Seedlings were germinated on half-strength MS agar plates and were cultured vertically in a Percival incubator under 24-hour light for 5 to 7 days at 22°C. For LHW induction, 7-day-old seedlings grown on MS plates were moved into liquid MS with or without 5 μM estrogen and incubated with rotation. For the NPA (1-N-Naphthylphthalamic acid)-treatment experiment, 10 μM NPA was added to liquid MS and incubated for 24 hours.

Whole-mount in situ hybridization was performed according to a previously described method (Hejátko et al., 2006).

Sections of 1.5 μm were prepared as previously described (Hirakawa et al., 2010). Specimens were observed under a DIC microscope with a Charge-Coupled Device camera (BX51, DB70, Olympus, Tokyo, Japan).

Fluorescent images were taken using previously described units (Kondo et al., 2011). Images were digitally analyzed using ImageJ. For staining of the plasma membrane, roots were incubated in 0.01 mg/ml propidium iodide. Three-dimensional images were constructed using the interactive 3D surface-plotting tool in ImageJ.

Quantitative RT-PCR was performed according to previously described methods (Ohashi-Ito et al., 2010).

Vectors based on Gateway cloning technology (Invitrogen) were used for most manipulations. Promoter fragments were introduced into pBGYN (Kubo et al., 2005). To generate an estradiol-inducible LHW, the LHW-coding sequence was recombined with pMDC7 (Curtis and Grossniklaus, 2003). To produce SHR::ATHB-8^r and SHR::PHB^r, the 2.0 kb SHR promoter was ligated into the NotI sites within pENTR/D/TOPO immediately upstream of ATHB-8^r or PHB^r. Primers are listed in supplementary material Table S1.

Roots of 5-day-old seedlings (∼20 mg FW) were used for the quantification, which was performed according to Kojima et al. (Kojima et al., 2009).

The vasculatures of wild-type Arabidopsis roots contain two protoxylem cells and two protophloem cells that are arranged in bilateral symmetry. The lhw mutants fail to form the correct vascular pattern and lack bilateral symmetry in the vasculature of seedling roots. To uncover the origin of this defect, we first investigated LHW expression during embryogenesis using whole-mount in situ hybridization. The signal from LHW was first observed in globular-stage embryos, where it was restricted to the central cells (pericycle-vascular mother cells) that are destined to divide to produce vascular initial cells (Fig. 1A,B). After the heart stage, the LHW transcript accumulated in root vascular cells, especially near the root apical meristem (Fig. 1C-E).

Because LHW was expressed preferentially in pericycle-vascular mother cells during embryogenesis, we postulated that LHW controls the vascular patterning in embryos. To examine this possibility, we observed expression patterns of TARGET OF MONOPTEROS5 (TMO5) and TMO5-LIKE1 genes as pre-protoxylem markers in embryos. TMO5 and TMO5-LIKE1 signals formed two stripes in the provascular region of wild-type embryos after the early heart stage of development (Fig. 1F-I; supplementary material Fig. S1A-C), indicating that the pre-pattern of bilateral symmetry of vasculature is established as early as the heart stage during embryogenesis.

In lhw embryos, however, the signals from TMO5 and TMO5-LIKE1 were detected as a single stripe in the provascular region throughout embryogenesis (Fig. 1J-L; supplementary material Fig. S1D-F). These results indicate that the bilateral pattern of vasculature is already impaired in early embryos, and that the defect in the lhw mutant can be traced back to the early stage embryo.

To investigate the LHW function in early embryogenesis, we next analyzed the cell division pattern during early development of wild-type and lhw embryos in detail, with a special focus on vascular initial cell formation. Wild-type and lhw embryos at the globular stage showed very similar cell division patterns in all cells (supplementary material Fig. S2A-D). In wild-type embryos at the transition stage, four vascular initial cells were formed from four pericycle-vascular mother cells by asymmetric cell division (Fig. 2A; supplementary material Fig. S2E,F). Because all vascular cells, not including pericycle cells, are derived from these newly produced four cells, we defined these cells as vascular initial cells. In lhw embryos, however, the direction of division in pericycle-vascular mother cells was abnormal, although the division patterns of protodermal and ground cells were not different from those of the wild type (Fig. 2B; supplementary material Fig. S2G,H). As a result, lhw embryos possessed fewer vascular initial cells (3±0.37) and irregularly shaped cells in the center, whereas wild-type embryos had four uniform vascular initial cells (Fig. 2C,D,G). At the late heart stage, the number of vascular cells was 8.85±0.55 in the wild type but only 4.5±0.22 in lhw, indicating that the vascular initial cells in lhw embryos rarely divided (Fig. 2E-G). These results suggest that LHW plays a role in the establishment of vascular initial cells during early embryo development through the regulation of cell division, and the monoarch pattern in lhw may result from the reduced number of vascular cells, which are insufficient for creating the diarch pattern during the heart stage of embryogenesis.

Because the early vascular development is associated with auxin flow, the expression pattern of an auxin efflux carrier, PIN1::PIN1-GFP, was examined in wild-type and lhw embryos (Gälweiler et al., 1998; Friml et al., 2003). It was difficult to find a difference in PIN1 expression between these embryos before the transition stage (supplementary material Fig. S3). During the early heart stage, PIN1 expression was restricted in two cotyledon primordia and provascular cells in wild-type embryos (Fig. 3A,C). At this stage, in lhw embryos, although the general pattern of PIN1 expression was similar to that of wild type, the PIN1-expressing domain was enlarged (Fig. 3B,D). The enlarged PIN1 expression was observed in cotyledon primordia where LHW was not expressed, suggesting non cell-autonomous effects of LHW. Altered PIN1 expression was also observed during provascular tissue formation in lateral root primordia. The PIN1 expression domain in lateral root primordia was gradually restricted to the provascular region and the root tip in wild type (Fig. 3E-G), whereas it was diffused more broadly and expanded to the cortex, endodermis and epidermal cells in lhw (Fig. 3I-K). To quantify the changes of the PIN1-expressing domain, we also observed the expression pattern of the PIN1::YFP-nuclear localization signal (nls) in wild-type and lhw lateral root primordia (Fig. 3H,L). The relative signal intensity of the central region (provascular region) to that of the peripheral region of wild-type primordia was considerably higher than was the case for lhw primordia (Fig. 3S). These results suggest that the restricted flow of auxin into the future provascular region may not be established in lhw.

To investigate whether the broad PIN1 expression domain in lhw is associated with changes in auxin distribution, we next examined DR5::GFP expression as a marker of the auxin maximum (Benková et al., 2003). Although DR5 expression was concentrated to the central region of the meristem in the emerging wild-type lateral root primordia, the peak of DR5 signal in lhw primordia was not restricted to the central region (Fig. 3M-R). These results indicate that the lhw plant fails to form a proper PIN1 expression domain and the proper auxin maximum in the central region of the primordium.

To further elucidate how the function of LHW relates to auxin flow, we analyzed the auxin response in roots of LHW gain-of-function seedlings using DR5::GFP, in which LHW expression was temporally induced by the addition of estrogen. The addition of estrogen induced the overproduction of the LHW transcript (supplementary material Fig. S4A). Without the addition of estrogen, DR5::GFP was detected only in the stele and root tip (Fig. 4A; supplementary material Fig. S5A). The addition of estrogen caused ectopic DR5::GFP expression in the entire root, which contained the cortex, the endodermis and the epidermis, as well as the stele, within 6 hours (Fig. 4B). Twenty four hours after the addition of estrogen, an ectopic, strong DR5 signal was maintained in the entire root (Fig. 4D; supplementary material Fig. S4B). These results suggest that ectopic LHW induces ectopic auxin response. Although NPA treatment altered the DR5::GFP expression pattern in root tips (supplementary material Fig. S5A-D), the treatment did not interfere with the ectopic expression of DR5::GFP in LHW-overexpressing roots (supplementary material Fig. S5E-G). Therefore, LHW may not promote DR5 expression by the inhibition of auxin transport through efflux transporters.

Next, to determine whether LHW promotes auxin biosynthesis, we quantified endogenous phytohormone levels, including auxin levels, in the roots of wild type and lhw and in roots in which LHW was induced for 6 hours and 24 hours (supplementary material Table S2). The free IAA level did not change in these roots (supplementary material Fig. S6). These results suggest that ectopic DR5 expression caused by LHW overexpression may not result from ectopic auxin biosynthesis. Similarly, there were no significant differences in the levels of other hormones in the roots of wild-type, lhw or LHW-induced plants (supplementary material Table S2).

To further understand the role of LHW in vascular differentiation in relation to auxin, we examined the involvement of LHW in the expression of transcription factors regulating auxin signalling in vasculature. MP and ATHB-8 are transcription factors that are involved in auxin signalling, and they are expressed in the provascular region (Hardtke and Berleth, 1998, Donner et al., 2009; Baima et al., 1995; Ohashi-Ito and Fukuda, 2010). First, we analyzed expression patterns of MP and ATHB-8. The MP signal was observed in the provascular region in wild-type heart stage embryos, as expected (supplementary material Fig. S7A). In lhw embryos, the MP signal was much weaker than that in the wild-type, especially in the provascular region (supplementary material Fig. S7B). The ATHB-8 signal was also weaker in lhw embryos (supplementary material Fig. S8A,B). To confirm the reduction of MP expression in the provascular region in lhw, the expression of MP::YFP-nls was examined in the lateral root primordia and main roots of seedlings. Indeed, the MP signals in the vasculature were diminished in both root primordia and the main root of lhw (supplementary material Fig. S7C-G). Quantitative PCR indicated that the transcript levels of MP and ATHB-8 were lower in lhw roots than in wild-type roots (supplementary material Fig. S8C). These results clearly indicate that LHW is required for promoting the expression of MP and ATHB-8 in the provascular region.

Next, we analyzed the effects of LHW induction on MP and PIN1 expression. Because the ectopic DR5::GFP signal was clearly seen 6 hours after LHW induction, followed by an increase until 24 hours after induction (Fig. 4B-D), we observed the expression patterns of MP::YFP-nls and PIN1::YFP-nls at the same time points after LHW induction. Ectopic signals of MP::YFP-nls and PIN1::YFP-nls were indeed seen 6 hours after LHW induction, after which time the signals gradually increased (Fig. 4E-L).

To clarify whether LHW regulates auxin signalling via ATHB-8, we examined the genetic interaction between lhw and a gain-of-function mutant of ATHB-8. Because ATHB-8 mRNA levels are repressed by miR165/166, we introduced substitutions into the ATHB-8 sequence to produce a mutant that was resistant to miR165/166 without introducing amino acid changes (ATHB-8^r) (Mallory et al., 2004; Emery et al., 2003). We also employed the SHR promoter to continuously overexpress ATHB-8^r in provascular tissues because SHR is specifically expressed in the provascular region from the late globular stage of embryos (Helariutta et al., 2000). SHR::PHABULOSA^r (PHB^r), which is another HD-ZIP III transcription factor but is not induced by auxin, was used as a control. The introduction of SHR::PHB^r into lhw plants did not alter the monoarch vascular pattern in lhw roots (supplementary material Fig. S8E). By contrast, SHR::ATHB-8^r rescued the lhw phenotype and produced the bilateral symmetry pattern in the vasculature of roots, although the rescue of this phenotype was only partial (9/34 plants, Fig. 4M-P; supplementary material Fig. S8D). These data strongly suggest that LHW functions in early vascular pattern determination through the enhanced expression of ATHB-8.

Our results suggest that LHW functions as a key regulator to generate vascular initial cells. In addition, they also suggest that the LHW function is closely related to the modulation of auxin response. Although how LHW is connected to the auxin response remains obscure, several possibilities are considered. Our favorite hypothesis is that LHW may be the first factor to initiate the auxin response by inducing auxin canalization. According to this hypothesis, LHW is expressed and induces the auxin response in the future vascular region, and then this restricted auxin response enhances the auxin positive-feedback loops consisting of PIN1, MP and ATHB-8, followed by auxin canalization, which in turn initiates vascular development (Scarpella et al., 2006; Wenzel et al., 2007; Donner et al., 2009; Ohashi-Ito and Fukuda, 2010). Another possible hypothesis is that LHW directly regulates auxin signalling components such as MP and ATHB-8 or regulators that modulate the PIN1 localization. It is well known that cytokinin, along with auxin, regulates vascular pattern formation (Bishopp et al., 2011a; Bishopp et al., 2011b; Cui et al., 2011). Therefore, we cannot deny the possibility that LHW induces the auxin response by modulating cytokinin signalling. Further investigation of the target genes of LHW may lead to an answer to the relationship between LHW and auxin.

We thank Minobu Shimizu, Sumika Tsuji-Tsukinoki and Kuninori Iwamoto for technical support; Shigeyuki Betsuyaku for assistance with the whole-mount in situ hybridization experiment; Dominique Bergmann and Jiri Friml for seeds; and Nam-Hai Chua for the pMDC7 vector.