ANGUSTIFOLIA is a central component of tissue morphogenesis mediated by the atypical receptor-like kinase STRUBBELIG

Background During plant tissue morphogenesis cells have to coordinate their behavior to allow the generation of the size, shape and cellular patterns that distinguish an organ. Despite impressive progress the underlying signaling pathways remain largely unexplored. In Arabidopsis thaliana, the atypical leucine-rich repeat receptor-like kinase STRUBBELIG (SUB) is involved in signal transduction in several developmental processes including the formation of carpels, petals, ovules and root hair patterning. The three STRUBBELIG-LIKE MUTANT (SLM) genes DETORQUEO (DOQ), QUIRKY (QKY) and ZERZAUST (ZET) are considered central elements of SUB-mediated signal transduction pathways as corresponding mutants share most phenotypic aspects with sub mutants. Results Here we show that DOQ corresponds to the previously identified ANGUSTIFOLIA gene. The genetic analysis revealed that the doq-1 mutant exhibits all additional mutant phenotypes and conversely that other an alleles show the slm phenotypes. We further provide evidence that SUB and AN physically interact and that AN is not required for subcellular localization of SUB. Conclusions Our data suggest that AN is involved in SUB signal transduction pathways. In addition, they reveal previously unreported functions of AN in several biological processes, such as ovule development, cell morphogenesis in floral meristems, and root hair patterning. Finally, SUB and AN may directly interact at the plasma membrane to mediate SUB-dependent signaling.


Background
Tissue morphogenesis and cellular patterning require extensive cellular communication. In plants the coordination of cellular behavior within a tissue is intrinsically linked to cell wall biogenesis and dynamics, as plant cells are connected through semi-rigid cell walls that drastically limit their relative movement. It is a major current challenge in plant biology to understand the mechanistic basis of intercellular communication and its connection to the cell wall during tissue morphogenesis.
In Arabidopsis, intercellular signaling mediated by the atypical leucine-rich repeat transmembrane receptor-like (LRR-RLK) STRUBBELIG (SUB) is essential for a number of developmental processes [1][2][3][4][5][6]. Ovules of sub mutants show frequent defects in the initiation and outgrowth of the outer integument. In addition, sub mutants exhibit twisted stems, petals and carpels/siliques. These phenotypes indicate a role for SUB in the control of integument initiation and outgrowth as well as stem and floral organ shape [1,2,6]. SUB also plays a role in internode length (and thus stem height), a trait that is potentially important for optimizing yield in crop plants.
At the cellular level, frequent misorientations of cell division planes were observed in e.g. L1 and L2 cells of young apical and floral meristems of sub mutants. Therefore, it was postulated that SUB signaling plays a role in orienting cell division planes in initiating integuments and floral meristems and thus influences the morphogenetic behavior of cells in a tissue context [1]. In addition, SUB, also known as SCRAMBLED (SCM), is involved in root hair patterning [3,4]. In this context, sub mutations lead to a randomization of root hair patterning such that root hairs develop ectopically or are not formed in the correct files.
In accordance with a perceived role of SUB in coordinating cellular behavior in tissue morphogenesis and cell patterning, SUB acts in a non-cell-autonomous fashion and mediates inter-cell-layer signaling across histogenic cell layers in the ovule, the floral meristem [5] and the root [7].
SUB belongs to the LRRV/STRUBBELIG-RECEPTOR FAMILY (SRF) family [8,9] and has several protein domains including an extracellular domain with seven leucine-rich repeats, a transmembrane domain and a cytoplasmic putative kinase domain [1,3,6]. Interestingly, a set of biochemical and genetic data indicated that although the kinase domain is essential for SUB function, enzymatic phosphotransfer activity is not [1,6]. Thus, SUB is likely a so-called atypical or dead kinase.
Signaling by atypical kinases is poorly understood in plants [10,11]. In addition, a detailed structure-function analysis of SUB suggested that the organ or cell-specific aspects of SUB-mediated signaling are not integrated at the SUB receptor, but involve other components that act together with, or downstream of SUB [6]. In order to unravel the signal-transduction pathway of SUB we have previously identified three complementation groups sharing the sub-like mutant (slm) phenotypes [2]. In addition, it was found that there is significant overlap in SLM-sensitive gene expression. Taken together the results indicated that SLM genes contribute to SUB signal transduction. The corresponding genes are called QUIRKY (QKY), ZERZAUST (ZET), and DETORQUEO (DOQ) [2]. Initial molecular analysis suggested that QKY encodes a putative membranelocalized protein with four C2 domains thus potentially connecting SUB to membrane-associated Ca 2+ -and phospholipid-dependent signaling [2].
In this work we focused on the DOQ gene. We show that doq-1 is a mutant allele of the ANGUSTIFOLIA (AN) gene. The doq-1 mutant carries a point mutation in the AN gene and we further demonstrate that doq-1 shares phenotypes with other an alleles and conversely, other an alleles show all slm phenotypes tested. These results rule out the possibility that doq-1 is an atypical allele. In addition, we provide evidence that SUB and AN can physically interact and that AN does not influence subcellular SUB distribution. Together our results reveal that AN is involved in SUB-dependent signaling events.

doq-1 mutants exhibit an underbranched trichome and narrow leaf phenotype
Meiotic recombination mapping placed DOQ in a 330.6 kb interval at the top of chromosome 1 (see Methods). This interval included AN, a gene previously described to affect trichome branching, leaf morphology, and silique shape [12][13][14][15]. During the course of this analysis we noticed that doq-1 trichomes are underbranched. Together with the previously described narrow leaf phenotype of doq-1 [2], this suggested that DOQ and AN functions are related. We therefore compared the doq-1 mutant leaf and trichome phenotype with three reference an alleles, an-1, an-2, an-EM1 and two an-2 35S::YFP:AN rescued lines. In doq-1 mutants, reductions in trichome branching approached levels seen in the an-1, an-2, an-EM1 alleles ( Figure 1, Table 1). Two-branched and three-branched trichomes are almost absent and a new class of unbranched trichomes was observed. The leaf shape of doq-1 mutants was also similar to that seen in an-1, an-2, an-EM1 mutants ( Figure 2). Correspondingly and in step with the reference an alleles, the leaf length/width ratio in doq-1 was significantly increased ( Table 2).
The DOQ gene corresponds to the AN gene We tested whether the DOQ gene corresponds to the AN gene. Genetic analysis revealed no complementation of doq-1 with an-1 indicating that doq-1 is allelic to an. Furthermore, we sequenced the AN gene in the doq-1 mutant and demonstrated a G to A transition at position 509 in the cDNA coding region. This mutation causes a glycine to aspartic acid substitution at position 170 that is located in the predicted NAD(P)-binding domain.
an alleles show slm phenotypes As most slm phenotypes had not been reported for an mutants, the question arose whether doq-1 is an atypical an allele showing new phenotypes or whether all an alleles share the slm phenotypes. We therefore compared slm phenotypes between doq-1, the an-1, an-2, an-EM1 alleles and an-2 35S::YFP:AN rescued lines.
As described for doq-1, all an alleles showed premature opening of flowers and twisted petals ( Figure 3). These phenotypes are rescued in the an-2 35S::YFP:AN lines. In addition, doq-1-like twisting of siliques was observed in all an alleles and could be rescued by AN overexpression (Figure 4). Plants carrying the doq-1 mutation show a weak ovule phenotype as compared to sub mutants [2].
To study the cellular patterns in an ovules we used Scanning Electron Microscopy. We observed smaller cells with atypical division planes at the distal end of the outer integument in all an alleles tested and a rescue of this phenotype in an-2 35S::YFP:AN lines ( Figure 5). Finally, we analyzed the role of AN in root hair patterning.
To this end we investigated the expression profiles of a GL2:GUS construct in an-1, an-2, an-EM1 allelic backgrounds. GL2:GUS is expressed in atrichoblast cell files and serves as a convenient marker for root hair patterning [16]. Mutant an-1, an-2, an-EM1 alleles displayed a severe distortion of the GL2:GUS pattern ( Figure 6). Thus AN is a new component of the root hair patterning machinery.

SUB and AN can interact directly
Next we addressed in more detail how SUB and DOQ/ AN relate to each other during SUB-dependent signal transduction. Earlier results indicated that the two genes do not regulate each other at the transcriptional level [2]. We thus tested if DOQ/AN and SUB have the potential to interact directly at the protein level. Indeed, the intra-cellular domain of SUB (SUBICD, residues 371 to 768), but not the extra-cellular domain (SUBECD, residues 26 to 340), was able to interact with AN in a yeast two-hybrid assay ( Figure 7). In addition, fusions of maltosebinding protein (MBP) to SUBICD and the SUB fulllength proteins were able to interact with a fusion of glutathione-S-transferase (GST) to AN in in vitro pulldown assays involving bacterially-expressed recombinant proteins (Figure 7).

AN is not required for subcellular localization of SUB:EGFP in roots
Previous studies involving a functional SUB:EGFP fusion protein indicated that SUB is localized to the plasma membrane and undergoes brefeldin A (BFA)-sensitive recycling [5][6][7]. As the AN homolog in animals was identified as a Brefeldin A ribosylated substrate [17] we tested whether the subcellular localization of SUB is dependent on DOQ/AN function. Towards this end we analyzed expression of a functional SUB::SUB:EGFP reporter [6] in roots of doq-1 mutants in the absence or presence of BFA ( Figure 8). Expression of the reporter appeared normal under both experimental conditions indicating that AN is not required for correct subcellular localization and recycling of SUB.  ERECTA influences the an phenotype The sub phenotype is sensitive to the genetic background [1] and it was shown that mutations in ERECTA (ER), encoding a LRR-RLK [18], and SUB interact synergistically as sub er double mutants exhibit strongly reduced plant height [6]. To investigate this issue we transformed pKUT196, a plasmid containing 9.3 kb of Col-0 genomic DNA spanning the entire ER locus [18,19], into doq-1 (Ler) plants and asked whether a wild-type copy of ER would affect the doq-1 phenotype. Interestingly, the transgene had a similar effect on plant height but not on other aspects of the doq-1 and sub-1 phenotypes ( Figure 9) [6].

Discussion
Multiple lines of evidence indicate that AN is involved in the SUB-dependent signaling mechanism. First, our genetic data show that all tested phenotypic aspects of sub are shared by an mutants. Second, DOQ/AN and SUB influence the expression of a common set of target genes. For example, 62 percent of genes misexpressed in sub flowers are also misexpressed in doq-1/an [2]. Thirdly, AN and SUB are able to interact physically in two different assays. Finally, an and er mutations synergistically affect internode length and plant height, as was observed for sub and er [6]. These results unexpectedly bring together two well-established but previously unconnected research fields. What do we learn for the function of AN and what for the function of SUB?    AN is involved in a broad variety of cellular processes AN was initially identified by its narrow leaf phenotype and trichome phenotype [14]. The closer analysis of the narrow leaf phenotypes suggested a role in cell polarity of leaf cells, such that the length of individual cells was increased and the width reduced [15,20]. As cell number is also changed an additional role in cell division control was postulated. The characterization of the trichome phenotype placed AN in a regulatory network controlling branching initiation [12,13]. In this context, a conspicuous lack of microtubule accumulation in the branch initiation zone of the developing trichome cell suggested a role of  AN in the control of microtubule organization [21]. The new slm phenotypes of an mutants reported in this study indicate that AN is involved in a SUB dependent signal transduction cascade. This function is separate from the first two functional aspects derived from leaf shape and trichome phenotypes, as sub mutants do not share these phenotypic aspects [2]. Such a broad spectrum of functions is compatible with the proposed biochemical functions of the AN protein. The AN gene encodes a protein with homology to CtBP/BARS [21,22]. CtBPs (C-terminal Binding Protein) were initially discovered as proteins binding to the C-terminal domain of adenovirus EA-1 [23]. CtBPs have been described as co-repressors for many transcriptional repressors carrying a PxDLS or RRT protein motif [24,25]. A possible function of AN as a corepressor is the finding that micro-array experiments revealed many genes that are transcriptionally regulated by AN [2,22]. CtBP/BARS were independently identified as Brefeldin A ADP ribosylated substrates (BARS) [17]. A possible Golgi-related function is supported by the finding that CtBP can induce constriction in Golgi tubules [26] and membrane fission [27]. In support of such a function AN was reported to act outside the nucleus [28].

Role of AN in SUB-mediated signal transduction
How does AN fit into the SUB-mediated signaling pathway? AN is unlikely to be a direct or indirect transcriptional target of SUB signaling. For example, SUB expression is only minimally altered in doq-1 flowers at various stages [2]. In addition, a 35S::SUB transgene failed to rescue the phenotype of doq-1 mutants indicating that SUB is not directly regulated at the transcriptional level by AN. At the same time, AN expression was not found to differ between floral tissue of wild type, sub-1 and other slm mutants [2] (data not shown). These observations render it unlikely that AN and SUB regulate each other's activity at the transcriptional level. Subcellular localization of a functional SUB:EGFP fusion protein was found to be restricted to the plasma membrane [6,7]. A functional AN:GFP fusion protein was recently reported to reside in the cytoplasm and in punctate compartments around the trans-Golgi network (TGN) [28]. The TGN localization of AN led the authors to suggest a Golgi-related role for this protein, possibly in membrane trafficking. These results are compatible with at least two possibilities of how AN may fit into the SUB signaling mechanism. In the first scenario AN could mediate membrane trafficking of SUB, a view that is also indirectly supported by the finding that QKY encodes a putative membrane-localized protein thought to function in membrane-associated Ca 2+ -and phospholipiddependent signaling [2]. However, this model does not fit the data presented in this study as signal distribution of a functional SUB::SUB:EGFP reporter was found to be unaltered in doq-1 roots. In an alternative scenario, the cytoplasmic distribution of AN would allow its direct interaction with the intracellular domain of SUB.
We currently favor this notion as results from the yeast two-hybrid and in vitro pull-down assays suggest direct interaction between SUB and AN proteins at the plasma membrane. This interaction could then be necessary to control further downstream events of SUB signal transduction. AN is likely to mediate only some aspects of SUB signaling as there is a difference in the strength between for example the ovule phenotypes of sub and doq-1 mutants and stem twisting is nearly absent in doq-1 or an-2 ( Figure 9). Moreover, there is only partial overlap in misexpressed genes between sub and doq-1 flowers [2]. With respect to plant height AN appears to be part of the SUB mechanism that interacts with the LRR-RLK ER. It was proposed that ER and SUB signaling converge either at the level of the receptors or at some level more downstream in the mechanism [6]. Thus, it will be interesting to resolve exactly how SUB, AN and ER relate to each other in future studies on SUB signaling.

Conclusions
In this study we showed that phenotypes of the slm mutant doq-1 and the trichome and leaf shape mutant an overlap. In addition, we showed that doq-1 is allelic to an. We further demonstrated that doq-1 is not an atypical an allele but that all tested an mutants show previously undescribed slm aspects. The data reveal a broader range of biological functions for AN than previously appreciated. Finally, our data reveal the possibility that SUB and AN interact directly. Taken together, the presented evidence suggests a role for AN in tissue morphogenesis mediated by the atypical receptor-like kinase SUB.
Plants were grown on soil at 24°C with 16 hours of light per day. The GL2:GUS line (Ler) [16] was introduced into an mutants by backcrosses. For GUS assays, plants were grown on MS plates for 4 days.

Yeast two-hybrid assay
The Matchmaker yeast two-hybrid system (Clontech) was employed and experimental procedures followed the manufacturer's recommendations. The pACT-AN construct was described previously [21]. For the generation of pGBKT7-SUBICD, the intracellular part of SUB coding sequence [1] was amplified from cDNA using primers SUBintra_F (CATGCCATGGATAACCGATAT TACAGTG) and SUBintra_R (ATCGGTCGACAATA AACTATTGCTTCTG). The PCR product was digested with NcoI/SalI and cloned into NcoI/SalI digested pGBKT7. The R599C mutation was introduced into pGBKT7-SUBICD by the QuikChange II XL site-directed mutagenesis kit according to the manufacturer's recommendations (Agilent Technologies) by using primers 35SsubR599Cf (AAGAAGCTCACTTGGAATGTATGTA TAAATATTGCATTAGGAGCTTC) and 35SsubR599Cr (G A AG C T C CTAATGCAATATTTATACATACATTC CAAGTGAGCTTCTT). For the cloning of pGBKT7-SUBECD, the extracellular part of SUB coding sequence was amplified from cDNA using primers SUB Extra/ Nde1_F (G C TC ATATGACTAATCTACGAGATGTTTC GGCGA and SUB Extra/Xma1_R (TACCCGGGGTTGA GTGGACCAGAATTTTCCTGATC). The PCR product was digested with NdeI/XmaI and cloned into NdeI/XmaI digested pGBKT7. All PCR-based constructs were sequenced.
To assay possible interactions in yeast pGBKT7 plasmids containing SUBECD, SUBICD and SUBICD-R599C were cotransformed with pACT or pACT-AN into yeast strain AH109. Transformants were selected after 3 days on SC medium lacking Leu and Trp (−LW) at 30°C. To examine yeast two-hybrid interactions, the transformants were grown on solid SC medium lacking Leu and Trp (SC-LW) or Leu, Trp, and His (SC-LWH) at 30°C.

Generation of constructs for recombinant protein production
AN and SUB cDNA were cloned into gateway entry vector pDONR 201. The following primers were used for gateway cloning: for amplifying AN cDNA: GGGGACA AGTTTGTACAAAAAAGCAGGCTTCATGAGCAAGA TCCGTTCG and GGGGACCACTTTGTACAAGAAAG CTGG G T CT TAATC GATCCAACGTGTGATAC; for amplification of the full length SUB cDNA: GGGGA CAAGTTT GTACAAA AA AGCAGGCTTCATGAGCT TTACAAGATGGGAAGTGT and GGGGACCACTTT GTAC AAGAAAGCTGGGTCTTAGATCATATGTTGA AGATCTTGG; for the amplification of the ICD version of SUB: GGGGACAAGTTTGTACAAAAAAGCAGGCTT CatgTATAACCGATATTACAGTGGAGC and GGGGA CC AC TTTGTAC AAGAAAGCTGGGTCTTAGATCAT ATGTTGAAGATCTTGG. AN cDNA was cloned into pGEX2TMGW (GE Healthcare) to generate an N-terminal fusion with Glutathion-S-Transferase (GST:AN) and the SUB and SUBICD cDNAs were cloned into pETG-40a (EMBL, Heidelberg) to generate an N-terminal fusion with maltose-binding protein (MBP:SUB, MBP:SUBICD).

Antibody generation
To produce anti-AN antibody, AN was expressed as a GST-fusion protein in E.coli. The protein was purified and used to generate antibodies in rabbit (Pineda Antikörper-Service; Berlin, Germany). The antibody serum was affinity purified and checked for its specificity by MALDI-TOF analysis.

In vitro pull-down assay
Interactions between SUB and AN were studied using purified proteins that were expressed in bacteria. The bacterial cells BL21-CodonPlus (DE3)-RIL containing the IPTG inducible constructs (MBP-Full length SUB, MBP-SUBICD and GST-AN) were grown at 37°C and 220 rpm to an OD 600 of 0.8-0.9 and then the cultures were induced by adding IPTG to final concentration of 1 mM. The induced cells were then grown further for 5hours (at 37°C for GST-tagged AN constructs and 20°C for MBP-tagged constructs) and cells were harvested by centrifugation. Cells were lysed in Tris-lysis buffer (Tris (pH 7.5) 50 mM, NaCl (100 mM), EDTA (1 mM), EGTA (1 mM), NP-40 (1%), Lysozyme (200 μg/ml), DTT (1 mM), Protease inhibitor cocktail (Sigma)) and sonicated three times for one minute each. The supernatant was collected by centrifugation at 4°C. MBP-tagged proteins were purified by incubation with amylose resin overnight at 4°C. After several washings, part of the resin was boiled with SDS-PAGE gel loading buffer to get purified protein for analysis. The remaining resin was used for incubation with equal amounts of GST:AN lysate for 4 hours at 4°C followed by several washings. Finally, the beads were boiled in SDS-PAGE gel loading buffer at 96°C for 10 min and equal amounts were loaded on a gel followed by western blotting. Detection was done using primary anti-AN antibody and secondary anti-rabbit antibody using SuperSignal