A Morpholino oligo can modify splicing of a pre-mRNA - www.gene-tools.com



Stem cells are maintained in an undifferentiated state by signals from their microenvironment, the stem cell niche. Despite its central role for organogenesis throughout the plant's life, little is known about how niche development is regulated in the Arabidopsis embryo. Here we show that, in the absence of functional ZWILLE (ZLL), which is a member of the ARGONAUTE (AGO) family, stem cell-specific expression of the signal peptide gene CLAVATA3 (CLV3) is not maintained despite increased levels of the homeodomain transcription factor WUSCHEL (WUS), which is expressed in the organising centre (OC) of the niche and normally promotes stem cell identity. Tissue-specific expression indicates that ZLL acts to maintain the stem cells from the neighbouring vascular primordium, providing direct evidence for a non-cell-autonomous mechanism. Furthermore, mutant and marker gene analyses suggest that during shoot meristem formation, ZLL functions in a similar manner but in a sequential order with its close homologue AGO1, which mediates RNA interference. Thus, WUS-dependent OC signalling to the stem cells is promoted by AGO1 and subsequently maintained by a provascular ZLL-dependent signalling pathway.


Stem cells are located in specialised microenvironments, stem cell niches, where signals from the neighbouring cells maintain them in a pluripotent state. This principle was recognised several decades ago in animals and plants (Schofield, 1978; Stewart and Dermen, 1970), but only recently have the regulatory mechanisms started to be unravelled.

In the wild-type Arabidopsis shoot meristem, three layers of stem cells are located at the very tip and give rise to all shoot organs formed in a plant's life. They are maintained in an undifferentiated state by signals that depend upon expression of the homeodomain transcription factor WUSCHEL (WUS) in a small underlying cell group termed the organising centre (OC) (Mayer et al., 1998). The stem cells in turn express CLAVATA3 (CLV3), a signal peptide that acts to restrict WUS transcription via the CLV1/CLV2 receptor kinase signalling cascade (Brand et al., 2000; Schoof et al., 2000). This feedback loop between the OC and stem cells provides a mechanism to control the size of the stem cell pool. Additional pathways have been identified that work in parallel to the WUS/CLV3 loop in regulating stem cell identity (Brand et al., 2002; Lenhard et al., 2002; McConnell et al., 2001; Prigge et al., 2005; Vroemen et al., 2003). Several studies suggest that during postembryonic development, tissues surrounding the meristem also provide important information for shoot meristem maintenance (reviewed by Tucker and Laux, 2007), including the internal (L3) cell layers (Stuurman et al., 2002; Szymkowiak and Sussex, 1992) and the adaxial sides of leaf primordia (McConnell et al., 2001; Waites et al., 1998). Furthermore, the sites of shoot meristem regeneration in tissue culture (Brossard, 1979; Projetti and Chriqui, 1986), in sunflower hybrids (Chiappetta et al., 2006), or after KNOTTED-LIKE HOMEOBOX (KNOX; also known as KNATM-TAIR) overexpression (Chuck et al., 1996; Nishimura et al., 2000), correlate with the position of vascular cells. However, the underlying mechanism remains elusive.

Little is known about how the shoot meristem stem cells are formed in the embryo. After separation of protoderm and inner cells, the onset of WUS expression in four sub-epidermal apical cells of the 16-cell embryo, which after several asymmetric cell divisions give rise to the OC (Laux et al., 2004), is the first indication of stem cell niche development during Arabidopsis embryogenesis. At the same stage, the cells below the OC precursor cells start to elongate and form the vascular primordium (Jürgens and Mayer, 1994; Mansfield and Briarty, 1991). The shoot meristem stem cells, however, cannot be distinguished before the middle stages of embryogenesis, when expression of CLV3 is detected (Fletcher et al., 1999).

Mutagenesis screens identified the ARGONAUTE (AGO) family member ZWILLE (ZLL; also known as PINHEAD and AGO10) as a factor involved in shoot meristem development (Lynn et al., 1999; Moussian et al., 1998). AGO proteins have been revealed as central components of RNA-induced silencing complexes (RISC) in animals and plants, where they bind small RNA molecules to target messenger RNAs for degradation, translational inhibition, or genomic DNA for methylation (reviewed by Peters and Meister, 2007; Vazquez, 2006). zll mutant embryos form terminally differentiated cells and organs instead of shoot meristem stem cells (Lynn et al., 1999; Moussian et al., 1998), but the mechanism underlying the function of the ZLL gene has remained unclear. Notably, meristems formed postembryonically can give rise to fertile plants with indeterminate inflorescences, indicating that ZLL is specifically required for embryonic shoot meristem development.

In this study we provide direct evidence that during embryonic patterning, ZLL acts from the emerging vascular primordium to mediate WUS function and maintain shoot meristem stem cells in an undifferentiated state, thereby indicating the presence of a novel signalling pathway downstream of ZLL.


Plant work

Plants were grown as described previously (Laux et al., 1996). The zll-1, zll-15 (Moussian et al., 1998) and ago1-8 (Newman et al., 2002) mutants are in the Landsberg erecta (Ler) background and ago1-1 (Bohmert et al., 1998) is in the Columbia background. The respective wild-type plants were used as controls. Transgenic ago1-1 lines were backcrossed to Ler and showed the same effect as the original Columbia lines.

Fig. 1.

pCLV3:GFP-ER expression is altered in zll-1 embryos. (A-C) pCLV3:GFP-ER expression (green) in wild-type (WT) Arabidopsis embryos. (A) Transition stage embryo (B) Close-up of stem cells in a torpedo stage embryo. (C) Bent cotyledon stage embryo. (D-F) pCLV3:GFP-ER expression in zll-1. (D) Transition stage embryo. (E) Close-up of stem cells at torpedo stage showing slightly weaker expression than in wild type. (F) Bent cotyledon stage embryo showing no pCLV3:GFP-ER expression. cp, cotyledon primordium; c, cotyledon; sc, stem cells; L1, L2 and L3, tissue layers 1, 2 and 3, respectively; sam, shoot apical meristem; h, hypocotyl. Scale bars: 10 μm in A,B,D,E; 20 μm in C,F.


The pCLV3:CFP-ER transgene was generated by replacing the GFP reading frame of pCLV3:GFP-ER (Lenhard and Laux, 2003) with CFP. For the pZLL:YFP-ZLL gene, a modified YFP sequence lacking a stop codon was inserted immediately prior to, and in frame with, the ZLL coding sequence in the context of an 8 kb genomic fragment. The ZLL promoter was then replaced with the ARR5, AS1, AS2 and ATHB8 promoters (sequences of primers used to amplify these promoters are available upon request) to generate tissue-specific expression constructs. The gWUS-GFP3 transgene contains a 3xGFP gene inserted immediately prior to the stop codon of a 15 kb BamHI genomic fragment of WUS. For ectopic WUS expression with the pOpL two-component system (Moore et al., 1998), the ATRPS5A promoter (Weijers et al., 2001) was fused to the synthetic transcription factor gene LhG4, transformed into plants as the driver line, and crossed to an effector line carrying the pOp:WUS(69.3) construct (Schoof et al., 2000). F1 progeny carrying both transgenes were germinated on 1/2 MS media and scored for phenotype at 4 and 11 days post-germination. All plasmids were introduced into Agrobacterium tumefaciens strain GV3101 [pMP90 (Koncz and Schell, 1986)] and transformed into plants using the floral dip method (Clough and Bent, 1998).

Microscopy and image analysis

Embryos were dissected from ovules on slides using fine-tip syringes in 10% glycerol and viewed on an Axioscope fluorescence microscope (Zeiss). Embryos were stained with DAPI (1 mg/mL) for 5 minutes and mounted in 50% glycerol in 1×PBS. Images were captured using Axiovision 4.4 software (Zeiss) and figures were generated using Photoshop 7.0 (Adobe).


Histological sections of zll embryos revealed no abnormalities in stem cell development until late stages of embryogenesis, when the cells at the stem cell position, unlike the small pluripotent stem cells in the wild type, enlarge and become vacuolated, indicating their differentiation. This defect is incompletely penetrant in all known zll alleles, and even in the strongest alleles only 90% of the seedlings lack functional stem cells, whereas the remaining seedlings are indistinguishable from wild type. To study how ZLL affects stem cell development in the embryo, we analysed expression of the shoot meristem stem cell marker pCLV3:GFP-ER (Lenhard and Laux, 2003). No difference was observed in the pCLV3:GFP-ER expression pattern between zll-1 and wild-type embryos from the earliest detection at transition stage until the early torpedo stage (Fig. 1, compare A,B with D,E; see Table S1 in the supplementary material), suggesting that ZLL does not have a discernable role in the establishment of CLV3 expression. However, from torpedo stage onwards, when pCLV3:GFP-ER extended into the three layers of stem cells in wild type (Fig. 1B,C), in the majority of zll-1 embryos the intensity and number of cells expressing pCLV3:GFP-ER gradually decreased (Fig. 1E,F; see Table S1 in the supplementary material). The frequency of embryos showing strongly reduced or no pCLV3:GFP-ER expression correlates closely with the frequency of zll-1 seedlings lacking the shoot meristem (compare Table S1 with Table S6 in the supplementary material). This suggests that ZLL activity is required to maintain stem cells of the embryonic shoot meristem, but not to initiate them.

In previous studies, ZLL mRNA was detected throughout the embryo at early globular stage and subsequently became restricted to the vasculature and, at lower levels, to the shoot meristem and the adaxial sides of the cotyledons (Lynn et al., 1999; Moussian et al., 1998). By contrast, using anti-ZLL antibodies, ZLL protein was only detectable in the vascular primordium of whole-mount embryos (Moussian et al., 2003). To clarify this discrepancy, we constructed an YFP-ZLL fusion protein that rescued the zll-1 mutant phenotype (see Table S2 in the supplementary material) when expressed from the ZLL promoter. YFP-ZLL fusion protein was first detected in all cells of the proembryo between the 2-cell (Fig. 2A) and 8-cell stages. During subsequent stages, YFP-ZLL became restricted to the emerging vascular primordium (Fig. 2B) and the adaxial domain of the cotyledons (Fig. 2C,D), but was barely detectable in the shoot meristem until maturity, when expression there increased (see Fig. S1 in the supplementary material). Thus, YFP-ZLL localisation closely mimicked the ZLL mRNA expression pattern, indicating that all cells where mRNA is detected also produce ZLL protein. It is plausible that the failure to detect ZLL protein in the shoot apex and the adaxial cotyledons in previous studies was due to the lower sensitivity of whole-mount immunodetection in this experiment.

Based on this dynamic expression pattern, the shoot meristem, (Moussian et al., 1998), the adaxial side of the cotyledons (Lynn et al., 1999; Newman et al., 2002), and the vasculature (Moussian et al., 2003) have been discussed as potential sites of ZLL function in embryonic stem cell development. To clarify this, we first compared pZLL:YFP-ZLL expression with that of the stem cell niche markers gWUS-GFP3 and pCLV3:CFP-ER during all relevant stages of embryogenesis. At the early globular stage, the gWUS-GFP3-expressing OC precursor cells were completely encompassed by strong YFP-ZLL expression (Fig. 2B). However, from the transition stage until late torpedo, when abnormal differentiation of stem cells in zll-1 embryos is observed, YFP-ZLL was barely detectable in either the OC or stem cells (Fig. 2C-E). To determine where ZLL is necessary for meristem development, the YFP-ZLL fusion protein was expressed using different promoter sequences at specific embryo stages and in specific embryonic regions overlapping with the endogenous ZLL expression pattern (Fig. 3). The expression pattern of each transgene was verified by localisation of the YFP signal, and function was assayed by rescue of shoot meristem stem cells in the zll-1 mutant (Fig. 3; see Table S2 in the supplementary material). Notably, ZLL expression in the stem cells via the CLV3 promoter (Fig. 3D-F) and in adaxial tissues of the cotyledons via the ASYMMETRIC LEAVES 2 (AS2) (Iwakawa et al., 2002) promoter (Fig. 3G-I) was unable to rescue stem cell development in zll-1 embryos. By contrast, expression of YFP-ZLL restricted to the vascular primordium (Fig. 3J-L) from the Arabidopsis HOMEOBOX GENE 8 (ATHB8) (Baima et al., 1995) promoter did rescue stem cell maintenance. Importantly, co-expression with the gWUS-GFP3 or pCLV3:CFP-ER reporter genes demonstrated that pATHB8:YFP-ZLL expression does not overlap at any stage in embryo development with the cells that give rise to the shoot meristem stem cell niche (Fig. 2F). Thus, ZLL function in the vascular primordium appears to be sufficient for stem cell niche development, indicating a non-cell-autonomous function, whereas its expression in the stem cells and adaxial sides of the cotyledons is neither required nor sufficient for stem cell maintenance. Notably, expression of YFP-ZLL in the vascular primordium of the embryo axis from the ARABIDOPSIS RESPONSE REGULATOR 5 (ARR5) (D'Agostino et al., 2000) promoter (Fig. 3M-O), or in the apical parts of the embryo (including the vascular primordium of the cotyledons) via the ASYMMETRIC LEAVES 1 (AS1) (Byrne et al., 2000) promoter (Fig. 3P-R), led to a complete rescue of the zll-1 phenotype. Since both expression domains overlap only in the provascular cells that lie immediately underneath the stem cell niche, these cells might have a specific role in stem cell maintenance.

Fig. 2.

pZLL:YFP-ZLL expression relative to the stem cell niche. (A) YFP-ZLL (yellow) accumulates in the apical cell of the 2-cell stage Arabidopsis embryo (blue, DAPI). (B,C) YFP-ZLL localisation relative to WUS-GFP3 (green). (B) 16- to 32-cell stage embryo. (C) Early heart stage embryo. The dashed line indicates the uppermost layer of YFP-ZLL-expressing cells. (D,E) YFP-ZLL localisation relative to pCLV3:CFP-ER expression (cyan). (D) Early heart stage embryo. (E) Late heart stage embryo. The dashed line in E indicates the same boundary as in C. (F) Pro-vascular pATHB8:YFP-ZLL (yellow) expression does not overlap with pCLV3:CFP-ER (pink) or gWUS-GFP3 (white) expression during embryogenesis. A torpedo stage embryo is shown. Insets show the differential contrast/fluorescent overlay. c, cotyledon; sam, shoot apical meristem; vp, vascular primordium; ad, adaxial cotyledon domain. Scale bars: 5 μm.

Fig. 3.

ZLL functions from outside of the stem cells during embryogenesis. Expression of (A-C) pZLL:YFP-ZLL, (D-F) pCLV3:YFP-ZLL, (G-I) pAS2:YFP-ZLL, (J-L) pATHB8:YFP-ZLL, (M-O) pARR5:YFP-ZLL and (P-R) pAS1:YFP-ZLL in zll-1 mutants. Transition/early heart and torpedo stage Arabidopsis embryos are shown, and seedlings are at 16 days post-germination. c, cotyledon; sam, shoot apical meristem; vp, vascular primordium; l, leaf; sc, stem cells; tm, terminated meristem. Scale bars: 10 μm, except 2.5 mm in C,F,I,L,O,R.

Provascular ZLL function could affect stem cell development independently of WUS, or could interact with WUS signalling from the OC. Contrary to the former possibility, ectopic expression of ZLL in embryos from the strong AGO1 promoter was unable to alter CLV3 mRNA or pCLV3:GFP-ER expression patterns (see Fig. S2 in the supplementary material), despite the potential of most apical and some central embryo cells to activate CLV3 expression ectopically (E.J.T. and T.L., unpublished). This suggests that ZLL alone is not sufficient to activate expression of the stem cell marker CLV3.

Fig. 4.

WUS expression and function is altered in zll-1 mutant embryos. (A-C) gWUS-GFP3 expression in wild-type (WT) Arabidopsis embryo. (A) 16-cell stage showing WUS-GFP3 (green). (B) Late heart stage embryo showing WUS-GFP3 marking the organising centre. (C) Walking-stick stage embryo. The meristem dome is indicated with a dashed line. Inset shows endogenous WUS mRNA localisation in an embryo at the same stage. (D-F) gWUS-GFP3 in zll-1. (D) 16-cell stage embryo. (E) Late heart stage embryo. (F) Walking-stick stage embryo showing a broad domain of gWUS-GFP3 expression. The flat meristem dome is indicated with a dashed line. Inset shows endogenous WUS mRNA localisation at the same stage. (G) Wild-type seedling. (H) Wild-type ATRPS5A≫WUS seedling showing ectopic outgrowths on the hypocotyl (arrows) and stunted cotyledons. (I) zll-1 seedling. (J) zll-1 ATRPS5A≫WUS seedling. All seedlings shown are at 11 days post-germination. c, cotyledon; sam, shoot apical meristem; h, hypocotyl; l, leaf; ts, terminal structure. Scale bars: 10 μm in A-F; 0.5 mm in G-J.

Next we investigated whether ZLL function affected the expression of gWUS-GFP3. Initially, gWUS-GFP3 expression was indistinguishable in wild-type (Fig. 4A,B) and zll-1 (Fig. 4D,E) embryos until the late heart stage (see Table S3 in the supplementary material). From torpedo stage on, however, the gWUS-GFP3 expression domain increased in size in zll-1 as compared with wild type, and by late torpedo/bent cotyledon stage was strikingly broader (Fig. 4F; see Table S3 in the supplementary material). Similar results were obtained by in situ hybridisation and in some instances, WUS expression appeared to be displaced to the flanks of the meristem (Fig. 4F; see Table S4 in the supplementary material). The expansion of the WUS expression domain and the concomitant decrease of CLV3 expression suggest that in zll embryos, WUS function is impaired and thus unable to maintain CLV3 expression. To verify this, we analysed the effects of the zll-1 mutation on ectopic WUS expression from the Arabidopsis RIBOSOMAL PROTEIN SUBUNIT 5A (ATRPS5A) promoter (Weijers et al., 2001). ATRPS5A directs expression predominantly to basal parts of the young embryo, to the whole embryo until torpedo stage and thereafter mainly to the meristem and cotyledons, at levels indistinguishable between wild type and zll (see Fig. S3 in the supplementary material). ATRPS5A≫WUS expression in a wild-type background led to stunted growth, ectopic outgrowths on the hypocotyl (Fig. 4H) and seedling lethality in the majority of cases (see Table S5 in the supplementary material), consistent with disturbed organ differentiation as reported previously (Brand et al., 2002; Gallois et al., 2002; Lenhard et al., 2002). These effects were suppressed by the zll-1 mutation (Fig. 4, compare H with J; see Table S5 in the supplementary material), despite similarly high levels of WUS mRNA accumulation compared with the wild-type background (see Fig. S3 in the supplementary material), indicating that ZLL is necessary also for ectopic WUS function in the embryo.

Taken together, these results suggest that ZLL-dependant signalling from the vascular primordium maintains stem cells during embryogenesis by potentiating WUS signalling from the OC to the stem cells. Since expansion of the WUS domain coincided closely with the stages when pCLV3:GFP-ER expression began to decrease, it is plausible that the failure to maintain CLV3 function in zll embryos might in turn account for derepression of WUS transcription.

Because previous studies showed that the ZLL and AGO1 genes share overlapping function during early embryogenesis (Lynn et al., 1999), we investigated whether AGO1 might also have a function in potentiating WUS signalling. Since double mutants between strong zll and ago alleles are embryo lethal, we generated zll mutants with partially altered levels of AGO1 activity, similar to previous studies that considered flower and leaf development (Lynn et al., 1999). By reducing AGO1 gene dosage to one functional copy, we found that stem cell defects in weak zll-15 mutants were enhanced to the severity of the null allele zll-1 without any other obvious effect on seedling development. In the complementary experiment, an increase of AGO1 gene dosage to three copies reduced the severity of stem cell defects in zll-1 (see Table S6 in the supplementary material).

Consistent with these data, ago1-1 embryos showed defects in gWUS-GFP3 expression similar to those observed in zll-1. The majority of homozygous ago1-1 embryos showed an expanded and disorganised domain of gWUS-GFP3 expression from as early as the transition to heart stage (see Fig. S4A and Table S7 in the supplementary material), and by maturity the domain was much broader than in segregating wild-type siblings (see Fig. S4B-D in the supplementary material). Unlike zll-1, however, most ago1-1 embryos failed to initiate pCLV3:GFP-ER expression at the correct stage (see Fig. S4E and Table S8 in the supplementary material), despite strong expression of gWUS-GFP3. Only at later stages of embryogenesis (see Fig. S4F and Table S8 in the supplementary material) did ago1-1 embryos gradually recover pCLV3:GFP-ER expression to wild-type-like levels (see Fig. S4G,H and Table S8 in the supplementary material), coinciding with the approximate stage when ZLL becomes required for meristem maintenance. Therefore, whereas both ZLL and AGO1 appear to be required for normal expression of WUS, their affects on CLV3 expression appear to be temporally separated. One plausible model is that AGO1 and ZLL act sequentially in stem cell development, with AGO1 being essential for initiation of the stem cell programme until heart/torpedo stage and ZLL for its maintenance during embryo maturation. In future studies it will be intriguing to determine whether AGO1, like ZLL, is a component of non-cell-autonomous signalling from provascular tissues during embryogenesis. Notably, ago1 mutants display pleiotropic effects and are sometimes seedling lethal, indicating that AGO1 has additional functions that cannot be rescued by late ZLL expression.

Taken together, our results provide direct evidence that the vascular primordium, which is one of the first discernable tissues to form in the developing Arabidopsis embryo, plays an instructive role during embryonic stem cell development. This function is mediated by a ZLL-dependent signalling pathway, which, based on homology to AGO1 and recent studies of ZLL (Brodersen et al., 2008), is consistent with a role for small RNAs in this process. Since pATHB8:YFP-ZLL signal was not detectable outside of the provascular tissues during embryogenesis, it is likely that ZLL itself does not move and thus might be involved in the production or the transmission of a signal to the stem cell niche. The nature of this signal is currently elusive, but its central role in early vascular function makes the phytohormone auxin a plausible candidate to be tested in the future. As a consequence of this vascular-borne signal, WUS signalling from the OC is enabled to maintain the stem cells in an undifferentiated state. This could involve modulation of WUS function in the OC cells, or an effect on the competence of stem cells to respond to the WUS-dependent signal. Further studies aimed at elucidating the precise mechanism of ZLL function are necessary to distinguish between these models.

Supplementary material

Supplementary material for this article is available at http://dev.biologists.org/cgi/content/full/135/17/2839/DC1


We thank Minako Ueda, Ivo Rieu, other members of the Laux laboratory and Dominique Chriqui for critical comments and discussions; Philipp Graf, Yuval Eshed and Herve Vaucheret for constructs and seeds; and Nikolai Adamski, Nico Lindau and Matthias Blender for technical assistance. This work was supported by grants from the Deutsche Forschungsgemeinschaft and the BMBF (T.L.), the Landesgraduiertenförderung Baden-Württemberg (A.H.), and an EMBO long-term fellowship (M.R.T.).


    • Accepted July 4, 2008.


View Abstract