OBE3 and WUS Interaction in Shoot Meristem Stem Cell Regulation

The stem cells in the shoot apical meristem (SAM) are the origin of all above ground tissues in plants. In Arabidopsis thaliana, shoot meristem stem cells are maintained by the homeobox transcription factor gene WUS (WUSCHEL) that is expressed in cells of the organizing center underneath the stem cells. In order to identify factors that operate together with WUS in stem cell maintenance, we performed an EMS mutant screen for modifiers of the hypomorphic wus-6 allele. We isolated the oberon3-2 (obe3-2) mutant that enhances stem cell defects in wus-6, but does not affect the putative null allele wus-1. The OBE3 gene encodes a PHD (Plant Homeo Domain) protein that is thought to function in chromatin regulation. Single mutants of OBE3 or its closest homolog OBE4 do not display any defects, whereas the obe3-2 obe4-2 double mutant displays broad growth defects and developmental arrest of seedlings. Transcript levels of WUS and its target gene in the stem cells, CLAVATA3, are reduced in obe3-2. On the other hand, OBE3 and OBE4 transcripts are both indirectly upregulated by ectopic WUS expression. Our results suggest a positive feedback regulation between WUS and OBE3 that contributes to shoot meristem homeostasis.


Introduction
Postembryonic growth and iterative organ formation of higher plants rely on the activity of pluripotent stem cells in organogenic centers, the meristems.The shoot meristem that will give rise to the above ground organs has been extensively studied in the model plant Arabidopsis thaliana.The homeodomain transcription factor WUS is expressed in the organizing center (OC) underneath the stem cells [1] where it directly represses cytokinin response inhibitors [2] and, after moving into the overlying stem cells [3,4], represses cell differentiation and activates expression of the signal peptide CLV3 [4][5][6].CLV3 in turn represses WUS transcription via CLV1/CLV2-CRN receptor-like kinases to delimit the size of the OC [6][7][8].This negative feedback loop balances stem cell maintenance and differentiation [7].The WUS/CLV3 loop also functions to maintain stem cells of the floral meristems [6,7].In contrast to the indeterminate shoot meristem, WUS in the determinate floral meristem also activates the gene encoding the MADS domain protein AGAMOUS (AG) that in turn terminates WUS expression and thus floral meristem growth [9][10][11].In addition to its function in stem cell regulation, WUS is also required for the development of the female and male gametes [12][13][14].However, CLV3 signaling does not appear to be targeted by WUS in these cases.
In order to find hitherto undiscovered factors involved in the WUS-mediated stem cell regulation, we used a sensitized mutant screen for genetic modifiers of the hypomorphic wus-6 allele [21,24].Here we report the isolation of the wus enhancer 9 (wen9) mutant that enhances stem cell defects in wus-6.We show by positional cloning that wen9 is an allele of the OBE3 gene, and characterize its function together with its closest homologue OBE4 in the shoot meristem.

Results
wen9 enhances inflorescence shoot meristem defects of wus-6 The putative null allele wus-1 causes premature termination of stem cells in the primary shoot meristem during embryogenesis, resulting in a flat apex of partially differentiated cells at the seedling stage [6].Consequently, seedlings lack any true leaves at 10 days after germination (Fig 1B).Postembryonically initiated shoot meristems terminate after the formation of a few leaves, resulting in a stop-and-go phenotype (Fig 1C ), and the seldom formed floral meristems give rise to 4 sepals, 4 petals, and a single stamen before premature termination (Fig 1D).The intermediate wus-6 allele causes reduced WUS expression levels, and the primary seedling shoot meristem and floral meristem prematurely terminate indistinguishably to wus-1 (Fig 1B and 1D; [21,24].In contrast to wus-1, however, postembryonically initiated wus-6 shoot meristems grow indeterminately and give rise to many floral meristems (Fig 1C; Tables 1 and 2).The wus-7 allele carries a missense mutation in the homeodomain and represents the weakest known wus allele [25].wus-7 seedlings form several rosette leaves before the primary shoot meristem terminates (Fig 1B ) and axillary shoot meristems form indeterminate shoots carrying complete flowers (Fig 1C and 1D; Table 2).
In order to identify factors that cooperate with WUS in stem cell maintenance, we searched for EMS mutants that modify the stem cell defects of the intermediate allele wus-6.One of the isolated enhancers, wus enhancer 9 (wen9), was mapped to a 97 kb region between position 5001124 and 5098789 on chromosome 1, and a nonsense mutation was identified in the predicted first exon of locus AT1G14740.The encoded protein OBERON3 (OBE3) [26], also named TITANIA1 (TTA1) [27], is 733 amino acids in size and contains a potential PHD (plant homeo domain) DNA binding domain (Fig 1A  .We thus conclude that the mutation in OBE3 caused the enhanced phenotype and assigned wen9 as obe3-2. To investigate the genetic interaction between WUS and OBE3, we analyzed double mutants between different wus alleles and the obe3-2 mutant.Development of the homozygous obe3-2 single mutant is indistinguishable from wild type (Figs 1B-1D and 2A).The obe3-2 mutation does not affect the seedling phenotypes of wus-1, -6, or -7 (Fig 1B ; Table 1).However, it strongly reduces the postembryonic formation of inflorescence stems in all combinations (Fig 1C ; Table 1) and causes the formation of leaves in a disorganized (= wuschel-like) pattern.In cases where flowers are made, obe3-2 does not further enhance the already early termination of wus-1 and wus-6 floral meristems, but causes premature termination of wus-7 floral meristems (Fig 1D ; Table 2).

Mutual expression regulation between WUS and OBE3
In order to investigate whether the expression levels of WUS and CLV3 genes are altered, we performed qRT-PCR with the 7 day-old obe3-2 and obe4-2 seedlings.WUS and CLV3 mRNA levels are significantly reduced in obe3-2 (0.43 and 0.45 fold, respectively) compared to the Ler wild type, whereas ARR7 mRNA levels appear increased (Fig 3A).In a converse experiment, WUS mRNA level is increased after induction of p35S:cOBE3-GR expression and this effect is suppressed in the presence of cycloheximide, whereas mRNA levels of CLV3, STM, and ARR7 are not significantly changed (Fig 3B).In contrast to obe3-2, expression levels of WUS or CLV3  Because double mutant plants are severely retarded, we analyzed the WUS expression pattern in obe3-1 obe4-1/+ plants.In 6-day-old seedlings, expression of the pWUS:GUS (Fig 3C ) reporter is confined to the OC of obe3-1 obe4-1/+ plants as in the wild type.However, in 20-day-old inflorescences, WUS expression is not detectable in obe3-1 obe4-1/+ (genotyped by PCR) inflorescence and floral meristems unlike in the wild type (Fig 3D).
In order to address whether WUS affects OBE3/OBE4 transcript levels, we analyzed the effects of inducible WUS activity.After induction of p35S:WUS-GR plants with dexamethasone, mRNA levels of OBE3 and OBE4 are upregulated, and this effect is suppressed in the presence of cycloheximide (Fig 3E).The direct WUS target in the shoot meristem, ARR7 [2], is used as a control for WUS-GR induction.Upregulation of OBE3 and OBE4 expression by WUS is also suggested by published microarray data (S3 Table ).Thus, WUS activity is sufficient to induce OBE3/OBE4 expression by an indirect mechanism.However, we did not detect any abnormal phenotype in p35S:cOBE3 or p35S:cOBE3-GR plants.

Discussion
Stem cell homeostasis requires the balanced activities of a complex network of regulatory factors.Despite strong advances, our knowledge of regulatory pathways is still fragmentary with many components unknown.This is due in part to the fact that a limited number of mutants display informative stem cell phenotypes.Furthermore, many other essential stem cell factors may remain undiscovered due to genetic redundancy or pleiotropic mutant phenotypes.Here we used a modifier screen to overcome this problem and discovered the obe3-2 mutant as an enhancer of the hypomorphic wus-6 allele.
What is the developmental nature of the WUS-OB3 interaction?
The role of WUS in stem cell maintenance can be observed at several developmental stages.Mature wus-1 embryos and seedlings lack shoot meristem stem cells and display differentiated cells instead.Postembryonically formed adventitious meristems terminate prematurely after forming a few leaves.Only occasionally an inflorescence is formed, but it terminates prematurely after formation of 1-3 flowers, which in turn terminate prematurely after a single first anther.Whereas seedling and floral meristems appear to absolutely require WUS activity, the occasional formation of inflorescences suggests that at this stage other factors can sustain stem cells for some time [21].Although OBE3 is ubiquitously expressed [27], the obe3-2 mutation enhances only the premature vegetative shoot meristem termination in wus-1 and thus represents one of these additional factors.In the hypomorphic allele wus-6, obe3-2 causes premature termination of the inflorescence meristem indistinguishable to wus-1.Finally, in combination with wus-7, which as a single mutant displays higher floral meristem activity compared to wus-1 and wus-6, obe3-2 enhances premature termination of the floral meristem.These results indicate that in addition to the seedling phase, OBE3 is required for residual WUS activity of hypomorphic wus alleles in inflorescence and floral meristems.By contrast, obe3-2 does not enhance wus defects in embryonic shoot meristem formation.
Curiously, despite their redundancy in the shoot meristem, obe3 enhances wus flower termination whereas obe4 mutant suppresses it.One possible reason for this particular behavior might be that in obe3-1 and obe3-2 mutants, the C-terminal region of OBE3 is disrupted and the PHD domain is still intact, whereas in obe4-1 and obe4-2 mutants the PHD domain is disrupted.Alternatively, both wild-type proteins might have divergent functions specifically in floral meristems.
Considering the ubiquitous expression of OBE3, it is noteworthy that the obe3-2 mutation reduces the organ number only of the two inner whorls of wus-7 but does not affect the perianth.A plausible explanation is that WUS-mediated stem cell maintenance is only required to provide the cells for the inner two whorls, whereas perianth organs appear to consume the cells present in the initial floral meristem formed independently of WUS, as described previously [6].
What is the genetic nature of the WUS-OB3 interaction?
Based on our mutant analysis and expression studies, OBE3 appears to act downstream of WUS.On the other hand, WUS expression levels are reduced in obe3-2 mutants and increased by OBE3 overexpression from the ubiquitous 35S promoter.The reduction of pWUS:GUS expression in the shoot meristem of obe3-1 obe4-1/+ mutants and the requirement of OBE3 in wus hypomorphs suggest that this is also the case in shoot meristem regulation.One plausible interpretation of this data is that WUS and OBE3 reinforce each other's expression in a positive feedback loop (Fig 4C), albeit this effect seems moderate.
OBE3 is a member of a small group of related proteins and, together with its closest homolog OBE4, is redundantly required for plant growth, consistent with previous observations [27].The seedling lethality of obe3-2 obe4-2 double mutants suggests that both genes are involved in several processes other than shoot meristem regulation.Likewise, obe1 obe2 [28] displays seedling lethality, but not any other obe double mutant combinations, indicating two pairs of redundant functions in this group, OBE1,2 and OBE3,4.In contrast to obe3-2 wus-6 plants, obe4-2 wus-6 double mutants have wus-6-like inflorescences and partially restored floral organs and seed production.This indicates that OBE3 and OBE4, in addition to their redundant functions in general growth control, have opposite roles at least in floral meristems.

What is a possible molecular basis of interaction?
The upregulation of OBE3 mRNA levels by WUS and vice versa is abolished by the presence of the protein inhibitor cycloheximide, suggesting that intermediate components are necessary.OBEs are PHD domain containing proteins, which originally were found by their homology to the Potyvirus VPg-interacting protein [29].The PHD domain is reported to bind to potentially activating H3K4me2/3 modifications [30][31][32][33][34].Further experiments are necessary to address how OBE affects WUS expression and vice versa and whether the enhancement of hypomorphic wus mutant defects by obe3-2 can be attributed to the reduction of WUS expression levels.

Plant materials and growth conditions
The obe3-2 mutant was isolated from EMS-mutagenized populations in a wus-6 background in Ler [25].The insertion alleles obe3-3 SALK_078036 , obe3-4 SALK_042597c and obe4-2 SAIL_827_F11 in the Col background were identified from the SALK collection [35] of T-DNA tagged lines and the SAIL collection [36], respectively.All other mutant alleles and transgenic lines used in this study are listed in S5 Table .For segregation analysis, entire siliques from genotyped mother plants were harvested and the seeds were sown in randomized schemes [37].Plant growth conditions were as previously described [6] Mapping, genetic analysis and PCR genotyping The wen9 wus-6/+ x Col cross was performed for map-based cloning.Among 12113 F2 plants from F1 parents genotyped as wus-6/+, we identified 446 plants (3.7%) with an enhanced phenotype.The wen9 mutation was mapped with SSLP and dCAPS markers to a 97 kb region in chromosome 1 (between SNP CER465614 and CER424346).Sequencing of all 23 candidate loci in this region detected a mutation only in the OBE3 gene.The identified G1554>A "stop" mutation is in exon1 of the predicted reading frame of locus AT1G14740.The primers wen9-F (5'-CAGAGATGTTTGGATTCGTTAAGGATGTTTTTGTGTGTTGCGCTAAGAATCG-3') and wen9-R (5'-GAAATTGTGATAAGAGAAGG-3') were used for genotyping PCRs (Ta 55°C).After TaqI restriction cleavage, the wild type displays a 300bp band, and wen9 displays a 250bp and a 50bp band.Genotyping primers of other mutants including T-DNA lines used in this study are listed in S6 Table.

Preparation of constructs and selection of transformants
The genomic fragment including the intergenic region of OBE3 was amplified by PCR from Ler and cloned as pOBE3:gOBE3 in a pGPTV-HPT-based vector.The RALF11-03K20 cDNA clone from RIKEN BRC was used to amplify the full length cDNA by PCR for construct preparation.The cDNA fragments starting from ATG to the end of gene, with or without the stop codon, were cloned as: p35S:cOBE3, and p35S:cOBE3-GR respectively, in a pGreenII-based vector.Arabidopsis plants were transformed by floral dip, and T1 seeds were selected on plates with the respective antibiotics.

Quantitative RT-PCR analysis
Arabidopsis seeds were surface-sterilized with 1% hypochlorite for 10 minutes and washed two times with 70% ethanol.Sterilized seeds were sown on 1/2 MS plates, stratified for 3 days in the dark at 4°C and then grown in a Percival growth cabinet with constant illumination for 7 or 10 days.For all qRT-PCR experiments, 3 biological replicates with two technical replicates each were done.
For experiments without further treatment, seedlings were collected from the plates and frozen in liquid nitrogen immediately.For the induction experiments, dexamethasone (5 μM), with or without cycloheximide (50 μM) were applied by spraying the plates, and flooding the seedlings for 15 minutes.After removal of the liquid, seedlings were returned to the Percival growth cabinet for 4 hours before sample collection.
Total RNA was extracted from whole seedlings using the RNeasy 1 Mini kit (QIAGEN), followed by RQ1 RNase-Free DNase (Promega) treatment and reverse-transcribed with Super-Script III First-Strand Synthesis SuperMix for qRT-PCR (Invitrogen).Quantitative PCR was performed with the LightCycler 1 480 system (Roche) coupled with SYBR Green I Master (Roche).For each qPCR reaction, 25 ng of cDNA was used.Transcript level analysis was carried out according to a published protocol [38].For statistical analysis, ANOVA or t-tests were performed on Cp' values.The Cp' values were calculated after normalizing the Cp values with three independent reference genes, which passed the geNorm v3.5 [39] test.For graphic presentation, Normalized Relative Quantity (NRQ) was first rescaled by setting NRQ of mock treated wild-type samples as 1, then adjusted for the unspecific DEX effect and the transgene effect sequentially in order to calculate the transcript fold change.Fold changes were plotted as bar graphs.Primers used for quantitative PCR are listed in S7 Table.
OBE4 transcript levels are reduced in obe4-2 to about half of the wild-type level (S1 Fig).obe4-2 single mutants are indistinguishable from the Col wild type (Fig 2A).

Table 1 .
obe3-2 enhances the meristem defects of weak and intermediate wus alleles.
At 79 DAG (except the obe3-2 wus-6 at 100DAG), opened flowers were taken from the genotyped plants and the organ numbers were counted.Organ numbers in first and second whorls were 4 sepals and 4 petals, respectively, for all genotypes.doi:10.1371/journal.pone.0155657.t002