Enhancement of meristem formation by bouquet-1, a mis-sense allele of the VERNALIZATION INDEPENDENCE 3 gene encoding a WD40 repeat protein in Arabidopsis thaliana

Authors


  • Communicated by: Kiyotaka Okada

Correspondence: cueguchi@agr.nagoya-u.ac.jp

Abstract

In higher plants, the shoot apical meristem (SAM) is the ultimate source supplying cells that constitute the aboveground tissues and organs. The cells supplied from the SAM begin to proliferate rapidly and then concomitantly start to differentiate. We identified a novel mutant, named bouquet-1 (boq-1), exhibiting highly pleiotropic shoot growth phenotypes. The boq-1 plants showed an increase in the inflorescence stem number accompanied by frequent fasciation. This particular phenotype appeared to be due to the development of extra SAMs in a SHOOT MERISTEMLESS (STM)-dependent manner. Expression of STM was also expanded widely in the boq-1 shoot apex, suggesting that the repressive state of the STM transcription may not be established or maintained, leading to the misexpression. Molecular cloning of the relevant gene showed that the BOQ gene encodes a WD40 repeat protein, which has been reported as VERNALIZATION INDEPENDENCE 3 (VIP3). In addition, the finding that overproduction of the boq-1 allele in the wild-type background mimicked the boq-1 phenotypes in a dose-dependent manner suggested that the mutant BOQ-1 protein acts in a dominant negative manner. Taking these results together, we propose that the boq-1 mutation affects the proper progression of cell differentiation process.

Introduction

In higher plants, the aboveground tissues and organs ultimately originate from stem cells maintained in the shoot apical meristem (SAM). The SAM develops during embryogenesis and then gives rise to lateral organs, that is, leaf primordia during the vegetative growth phase and floral buds during the reproductive phase, in a temporally and spatially regulated manner. Through all plant growth phases, the SAM is rigidly maintained constituently and produces lateral organs continuously.

The stem cells maintained in the central region of the SAM divide slowly to give rise to daughter cells. Just after the daughter cells are pushed into the peripheral or rib zone, they begin to proliferate rapidly and then concomitantly start to differentiate. Cell differentiation in higher plants involves several temporally and spatially distinct stages, which should be controlled by positional as well as environmental cues. On entering of the next stage, the temporal fate of cells is established by an appropriate genetic program determining a set of expressed genes, and through the subsequent mitotic events, the cell identity once established is maintained by epigenetic regulation involving several types of modifications of the chromatin structure, such as chromatin remodeling, and chemical modifications of histone proteins and genomic DNA (for a review, see Hsieh & Fischer 2005). One of the important events in the early stage of cell differentiation is the down-regulation of the SHOOT MERISTEMLESS (STM) gene (Long et al. 1996). The repressed state of STM is maintained by plant homologues of animal Polycomb-group proteins (Katz et al. 2004; Hsieh & Fischer 2005). However, the precise molecular mechanism underlying the down-regulation of STM largely remains to be elucidated.

In this study, we isolated and characterized a novel mutant, named bouquet-1 (boq-1), which exhibits pleiotropic shoot growth phenotypes in Arabidopsis thaliana. The most remarkable phenotype is an increase in the inflorescence stem number accompanied by frequent fasciation, caused by the development of extra SAMs in the mutant shoot apex in a STM-dependent manner. The results of a series of genetic and transgenic studies suggest that the boq-1 mutation, a mis-sense allele of VERNALIZATION INDEPENDENCE 3 (VIP3) gene (Zhang et al. 2003), affects the proper progression of cell differentiation process.

Results

Isolation of the bouquet-1 mutant

To identify a novel gene involved in plant morphogenesis, we screened mutants exhibiting abnormal shoot growth among EMS-mutagenized M2 seedlings of A. thaliana. One mutant strain showing aberrant shoot morphology was thus isolated and designated bouquet-1 (boq-1), based on its growth phenotype (see below). Heterozygous boq-1/+ plants showed a growth phenotype indistinguishable from that of the parental wild type, Columbia (Col), indicating that the boq-1 is a recessive allele. Approximately three-fourths and one-fourth of the self-pollinated progeny of the boq-1/+ plants exhibited the wild-type and boq-1 phenotypes, respectively (385 wild-type and 118 mutant plants; χ2 = 0.64, > 0.05), suggesting that the boq-1 phenotype is the result of a single mutational lesion.

Growth phenotype of boq-1

The boq-1 mutant exhibited highly pleiotropic growth phenotypes. Tricotyledonous seedlings were observed occasionally (approximately 7%, = 118, Fig. 1A–C), indicating that the mutation somewhat affects process(es) during embryogenesis. The mutant leaves were reduced in size with short petioles and small, asymmetrically distorted leaf blades. The boq-1 mutant generated more rosette leaves than the wild type at the same time and showed, seemingly, a disordered phyllotaxy (Fig. 1D).

Figure 1.

Growth phenotypes of the boq-1 mutant. Plants were grown on MS gellan gum plates and then on rockfiber with 16-hour light/8-hour dark fluorescent illumination at 22 °C. (A–C) Morphology of 10-day-old seedlings. Col (A) dicotyledonous boq-1 (B) and tricotyledonous boq-1 (C) seedlings are shown. Scale bars = 5 mm. (D) Morphology of 3-week-old rosettes. Col (left) and boq-1 (right) plants are shown. Scale bar = 1 cm. (E) Four-week-old plants of Col (left) and boq-1 (right) are shown. A fasciated stem is indicated by an arrowhead. Scale bar = 2 cm. (F) Morphology of 8-week-old plants. Col (left) and boq-1 (right) plants are shown. Scale bar = 5 cm. (G and H) Close-up views of Col (G) and boq-1 (H) flowers. Scale bars = 1 mm.

The boq-1 mutant flowered at the nearly same time as the wild type under long day conditions (16-hour light/8-hour dark). Whereas one main inflorescence stem started to elongate at the onset of flowering in the wild type, many (more than four), but thin, stems developed simultaneously in the boq-1 apex (Fig. 1E). During the subsequent reproductive growth phase, the boq-1 plants continued to generate new stems, resulting in a markedly bushy appearance (Fig. 1F).

The boq-1 mutation also affected the floral development. Floral buds opened prematurely such that the inner floral organ primordia, that is, petal, stamen and carpel primordia, were visible from outside when they were established. Although the number of floral organs was nearly normal in the majority of the mutant flowers, their size and morphology were affected (Table 1 and Fig. 1G,H). All of the floral organs, particularly the outer three whorl organs, were reduced in size. The mutant flowers were male sterile, and boq-1 homozygous seeds were hardly obtained under the standard growth conditions.

Table 1. Floral organ numbers of boq-1 stm-4 and boq-1 wus-1 double mutants
GenotypeWhorl n
FirstSecondThirdFourth
  1. Floral organs were counted. Each value represents the average with the standard deviation for the indicated number (n) of flowers.

Col4.0 ± 0.04.0 ± 0.05.8 ± 0.42.0 ± 0.012
boq-1 4.1 ± 0.34.1 ± 0.35.3 ± 0.82.0 ± 0.012
vip3-2 4.0 ± 0.04.2 ± 0.45.4 ± 0.72.0 ± 0.012
stm-4 4.1 ± 0.91.8 ± 1.61.6 ± 1.30.0 ± 0.09
boq-1 stm-4 4.1 ± 0.44.1 ± 0.92.5 ± 1.91.7 ± 1.216

Among the pleiotropic mutational phenotypes of boq-1, the most remarkable one was an increase in the inflorescence number accompanied by fasciation. It is possible that the increased number of stems is simply due to enhanced outgrowth of lateral buds with reduced apical dominance. However, the fact that the mutant stems were frequently fasciated gave us the alternative idea that the SAM in boq-1 is impaired structurally and/or functionally because fasciation is tightly related to the malfunction of the SAM observed in several mutants, such as clv, fasciata (fas) and mgoun (mgo, Leyser & Furner 1992; Clark et al. 1993; Laufs et al. 1998; Kaya et al. 2001).

Extra SAMs develop in the boq-1 shoot apex

To determine whether the SAM structure is affected in the boq-1 mutant, we inspected the boq-1 shoot apex by scanning electron microscopy (SEM). The shoot apex of a wild-type plant at 10 days after germination (DAG) contained a unique embryonic SAM generating leaf primordia in a regularly arranged manner (phyllotaxy, Fig. 2A). In contrast, of 28 mutant plants at 10 DAG, more than half (17 plants) contained multiple dome-shaped structures in the shoot apices (Fig. 2B,C). The size of each dome-shaped structure in the boq-1 mutant was comparable to that in the wild type, and the phyllotactic pattern seemed to be normal, as judged from the regular arrangement of the leaf primordia (Fig. 2B). At 13 DAG, on the adaxial sides of leaf petioles, the same structures also developed (Fig. 2D). The longitudinal sections showed that, like the embryonic SAMs (Fig. 2E,F), they consisted of small cells arranged with layered pattern on their surfaces (Fig. 2G,H). Based on the structural resemblances, we concluded that the boq-1 shoot apex contains occasionally multiple SAMs. These results thus strongly suggested that the increased numbers of leaves and inflorescence stems and frequent fasciation observed in the boq-1 mutant are due to the formation of extra SAMs in the mutant apex.

Figure 2.

Extra shoot apical meristems (SAMs) in the boq-1 shoot apex. (A–D) SEM images of 10-day-old Col (A) and boq-1 (B and C) shoot apices, and a 13-day-old boq-1 leaf petiole (D). Arrowheads indicate SAMs in (B), (C) and (D). The leaves overlapping the extra SAMs are not removed in (D). Scale bars = 50 μm (A, B and C) and 200 μm (D). (E and F) Longitudinal sections of 14-day-old Col (E) and boq-1 (F) embryonic SAMs. Note that the boq-1 SAM has been converted to an inflorescence meristem. Scale bars = 50 μm. (G) A longitudinal section of a 14-day-old boq-1 leaf petiole. Scale bar = 500 μm. (H) Close-up view of a dome-shaped structure indicated by an arrow in (G). Scale bar = 20 μm. (I) The kinetics of the number of SAMs in the shoot apex. Plants were grown on MS gellan gum plates with 16-hour light/8-hour dark fluorescent illumination at 22 °C. Seedlings were subjected to SEM analysis, and the number of SAMs in the shoot apices was determined. Each value represents the average with the standard deviation for >10 and >20 plants for Col and boq-1, respectively.

To answer the question of whether the extra SAMs of boq-1 develop embryonically or postembryonically, we carried out kinetic analysis. Each mutant seedling contained only a unique embryonic SAM just after germination (4-day-old plants), and then, the number of SAMs increased gradually along with the subsequent growth (Fig. 2I). These findings indicated that the boq-1 plants develop extra SAMs postembryonically and continuously.

STM is required for SAM formation in the boq-1 mutant

The STM gene function is crucially required for embryonic SAM formation (Barton & Poethig 1993). To determine whether the embryonic as well as extra SAM formation in the boq-1 mutant also depends on the STM function, we first examined the expression of the STM gene, whose transcription is active in the undifferentiated region of the SAM. The wild-type as well as boq-1 mutant plants carrying an STM promoter:GUS fusion construct (STMpro:GUS) were cultured and the seedlings at 10 DAG were stained. As previously shown (Hay et al. 2002), expression of STMpro:GUS is limited to the tip region of the wild-type SAM, whereas GUS staining in the boq-1 mutant was broadened around the shoot apex as well as the basal parts of leaf petioles (Fig. 3A–D), suggesting that the area containing undifferentiated cells is expanded in the boq-1 shoot apex. Although STM expression was observed around the boq-1 shoot apex, no staining was seen in the other tissues, such as leaf blades (Fig. 3B). It should be noted that all of the boq-1 plants (= 35) showed a similar broadened staining pattern in the apices, although approximately half of the mutant plants at 10 DAG had only one visible SAM in the apex, as mentioned previously. In addition, a group of small cells in the adaxial side of the mutant leaf petiole did not show any SAM-like dome structures (Fig. 3E,F). This result thus suggested that the STM gene expression is widely spread in the mutant apex before the formation of a visible SAM structure.

Figure 3.

Expression of STMpro:GUS. 10-day-old Col (A, C and E) and boq-1 (B, D and F) seedlings carrying an STMpro:GUS were stained. (A and B) STMpro:GUS activity in a whole seedling. Scale bars = 5 mm. (C and D) Close-up views of shoot apices. Scale bars = 200 μm. (E and F) Longitudinal sections of leaf petioles. ad, adaxial; ab, abaxial. Scale bar = 500 μm.

We next constructed a boq-1 stm-4 double mutant. The boq-1 heterozygous mutant was crossed with the stm-4 heterozygous one, the boq-1/+, stm-4/+, ERECTA (ER)/ER mutant was selected from the F2 population, and then the self-progeny of the double heterozygous mutant was grown and analyzed. Note that we removed the er mutation to minimize the effect of the different genetic background because stm-4 has a Landsberg er (Ler) background. Because stm-4 is a severe stm allele (Endrizzi et al. 1996), all stm-4 homozygous mutants completely lacked the embryonic SAM and their cotyledonary petioles were partially fused at their bases (Fig. 4A,D). The boq-1 stm-4 double mutants, whose genotypes were confirmed by PCR genotyping (see 'Experimental procedures'), exhibited a different phenotype from that of either boq-1 or stm-4. The double-mutant seedlings generated adventitious leaves in the axils of cotyledons as well as on the cotyledonary petioles (Fig. 4B,E), or irregular tissues with a large number of adventitious leaves at the tip near but not at the center of the shoot apex (Fig. 4C,F). During the subsequent growth, no organized bunch of leaves (rosette) was generated, suggesting that the double mutant lacks the embryonic SAM as well as functional extra SAMs.

Figure 4.

STM function is required for embryonic and postembryonic shoot apical meristem formation in boq-1. Seedling phenotypes of stm-4 (A, D) and boq-1 stm-4 (B, C, E and F). Tricotyledonous (B) and dicotyledonous (C) seedlings are shown. Scale bars = 2 mm (A, B and C) and 0.5 mm (D, E and F).

Effects of the boq-1 mutation on stm-4 phenotypes

It was previously reported that adventitious leaves emerge repeatedly from the axils of the cotyledons or from the petioles of these leaves even in some individuals carrying a strong stm allele homozygously (referred to as the ‘rescued’ phenotype, Barton & Poethig 1993; Clark et al. 1996; Endrizzi et al. 1996). Indeed, under our growth conditions, more than half of the stm-4 mutants (55 of 92 plants) generated adventitious leaves. The boq-1 mutation enhanced this particular phenotype, that is, almost all the boq-1 stm-4 double mutants (20 of 21 plants) produced adventitious leaves with a boq-1-like morphology, that is, a reduced size and disordered morphology. Moreover, after the continuous leaf production, all rescued boq-1 stm-4 double mutants generated inflorescences with frequent fasciation, whereas the stm-4 ones rarely did (five of 55 rescued plants), as reported previously (Endrizzi et al. 1996, Fig. 5A). This result indicated that the boq-1 mutation is able to enhance the formation of the adventitious meristem for inflorescences as well as adventitious leaves.

Figure 5.

The boq-1 mutation enhances the rescued phenotype and partially suppresses the floral defect in stm-4. (A) Morphology of 8-week-old plants. From left to right are Col, boq-1, stm-4 and boq-1 stm-4 plants. Scale bar = 5 cm. Close-up view of a fasciated stem in boq-1 stm-4 is shown (inset). Scale bar = 2 mm. (B–D) Close-up views of stm-4 (B) and boq-1 stm-4 (C and D) flowers. The stm-4 flowers at the tip of the inflorescence are fused with each other and seem to be a single flower. A secondary flower generated in the boq-1 stm-4 flower is indicated by an arrowhead in (D). Scale bars = 1 mm. (E) Expression of KNAT1, KNAT2 and KNAT6 in boq-1. Total RNA from the plants presented was prepared, treated with DNase I and then subjected to real-time RT-PCR analysis. Each value represents the average with the standard deviation (= 3).

It is known that floral meristems in stm mutants terminate prematurely, so that the flowers contain reduced numbers of floral organs in the inner three whorls (Barton & Poethig 1993; Clark et al. 1996; Endrizzi et al. 1996). Indeed, in our experiments, stm-4 flowers, rarely formed, consisted of reduced numbers of petals and stamens, with no carpel (Table 1 and Fig. 5B). In contrast, the boq-1 stm-4 double mutant formed many flowers with nearly normal numbers of floral organs (Table 1), but with small and abnormal shaped organs like boq-1 flowers (Fig. 5C). Moreover, the double-mutant flowers, in some cases, had a secondary flower in the inner whorl (Fig. 5D). These results thus indicated that the boq-1 mutation is able to partially suppress the floral defect in stm-4 mutants.

It is possible that the enhanced meristematic activity of stm-4 by boq-1 is due to the increased expression of class1 KNOX genes other than STM, namely KNAT1/BP, KNAT2 and KNAT6, because ectopic over-expression of class1 KNOX genes results in ectopic generation of meristematic tissues (Sinha et al. 1993; Chuck et al. 1996). However, we found that the expression of these genes is not changed largely, if any (Fig. 5E).

BOQ gene encodes a WD40 repeat protein

To identify the BOQ gene, we used a map-based cloning approach. The boq-1 mutation was roughly mapped near the south end of chromosome IV. The subsequent fine mapping showed that the relevant mutation is located within a approximately 70-kb region, in which 15 putative genes are annotated. We next determined the genomic sequences for all of the putative coding regions in the boq-1 mutant genome, and found a unique nucleotide substitution in At4g29830, which was previously reported as VERNALIZATION INDEPENDENCE 3 (VIP3, Zhang et al. 2003). The boq-1 genome contains an A–G transition in At4g29830, resulting in a Gly to Glu amino acid substitution at the 219th codon (Fig. 6A). The gene encodes a 321-amino-acid protein comprised of seven WD40 repeat domains. In the boq-1 mutant, one of the conserved amino acids in the fifth WD40 repeat is substituted in a nonconserved manner, suggesting that the boq-1 mutation substantially affects the function of the gene product.

Figure 6.

Genomic structure of the BOQ gene. (A) The structure of the chromosomal region encompassing the BOQ gene is schematically shown. The closed boxes represent exons including the BOQ open reading frame with presumed initiation and termination codons. The open boxes represent 5′ and 3′ untranslated regions. The positions of the boq-1 mutation and vip3-2 insertion are denoted by an asterisk and triangle, respectively. The relevant nucleotide and amino acid substitutions in boq-1 are indicated. The restriction sites used for cloning are also indicated. (B and C) in situ hybridization of the BOQ gene. The shoot apex tissues of 10-day-old Col plant were subjected to in situ hybridization. Images obtained with antisense (B) and sense (C) cDNA probes are shown. Bars = 200 μm.

To answer the question of whether the mutation found in At4g29830 is responsible for the boq-1 phenotype, we carried out complementation analysis. A genomic fragment encompassing the entire At4g29830 coding region as well as its putative promoter region was cloned into an appropriate T-DNA vector, and then the boq-1/+ heterozygous plants were transformed with the resultant construct. We obtained 23 hygromycin-resistant T1 plants, all of which exhibited normal growth indistinguishable from that of the wild-type control (Fig. 7A–C). Among them, PCR genotyping showed that two T1 plants carry the boq-1 allele homozygously on their chromosomes, indicating that boq-1 is a mis-sense allele of At4g29830. It was reported previously that the VIP3 gene is expressed throughout the plant (Zhang et al. 2003). We also detected the BOQ signal in the shoot apex by in situ hybridization (Fig. 6B,C).

Figure 7.

Growth phenotypes of vip3-2 and boq-1 carrying the wild-type BOQ allele. (A–C) Morphology of 4-week-old plants. Plants were grown on MS gellan gum plates for 2 weeks and then on rockfiber with 16-hour light/8-hour dark fluorescent illumination at 22 °C. The strains used were Col (A) boq-1 (B) boq-1 transformed with the wild-type BOQ genomic fragment (C). Scale bars = 2 cm. (D–F) Morphology of 2-week-old rosettes. The strains used were Col (D), boq-1 (E) and vip3-2 (F). Scale bars, 5 mm. (G–I) Morphology of 4-week-old plants. The strains used were vip3-2 (G) vip3-2/boq-1 heterozygous plant showing moderate phenotype (H) and vip3-2/boq-1 heterozygous plant showing boq-1-like severe phenotype (I). Scale bars = 2 cm.

The boq-1 mutation is allelic to vip3, but the extra SAM formation phenotype is specific to boq-1

It was previously reported that the T-DNA insertion allele in At4g29830, designated as vip3-1, showed a vernalization-independent, early-flowering phenotype even in the FRISF2 background (Zhang et al. 2003). vip3-1 also exhibited pleiotropic growth defects, such as small rosette leaves, semi-dwarf morphology, abnormal floral morphology and male sterility. These growth phenotypes were also confirmed for other alleles, that is, vip3-2, vip3-3 and vip3zwg (Jolivet et al. 2006; Dorcey et al. 2012). Although the growth defects observed in the boq-1 mutants largely overlap with those of vip3, as described in the previous sections, some have not been reported for vip3, namely increased numbers of rosette leaves and inflorescence stems, and frequently observed fasciation of stems. Indeed, under our growth conditions, besides the growth defects mentioned above regarding vip3 mutants, the vip3-2 plants showed nearly normal numbers of rosette leaves and inflorescences with no apparent fasciation (Fig. 7D–G). Therefore, the growth defects due to extra SAM formation are specifically observed in the boq-1 plants, but not in the vip3 ones.

Over-expression of the boq-1 allele confers the boq-1 phenotype on wild-type plants

The vip3 alleles previously analyzed, that is, vip3-1, vip3-2 and vip3-3, are T-DNA insertional mutations and regarded as null alleles, because the VIP3 protein was not detected for vip3-1, and no VIP3 transcript was observed for vip3-2 or vip3-3 (Zhang et al. 2003; Jolivet et al. 2006). vip3zwg was also reported as a null allele (Dorcey et al. 2012). Therefore, it is likely that the extra SAM formation observed for the mis-sense allele, boq-1, is due to a somewhat dominant effect caused by the mutant protein. To examine this idea, we first constructed the vip3-2/boq-1 heterozygous plants by genetic crossing. Among 26 F1 plants examined, 14 plants exhibited moderate growth defect phenotypes (Fig. 7H). The rest 12 F1 plants exhibited severe, boq-1-like, phenotypes (Fig. 7I), indicating that the single boq-1 allele per diploid is sufficient to affect shoot morphogenesis, even though the effect is weaker than that in the boq-1 homozygotes.

We next examined whether the over-expression of the boq-1 allele in the wild-type background affects the shoot phenotype. The boq-1 cDNA as well as the wild-type one was cloned under the control of the cauliflower mosaic virus 35S promoter (35Spro), and then, the resultant construct was introduced into the Col plants. All of the 47 T1 plants carrying the 35Spro:BOQ cDNA exhibited normal growth like the wild-type plants, which is consistent with the previous finding that the over-expression of the wild-type BOQ cDNA had no apparent effect on plant growth and development (Zhang et al. 2003). In contrast, of 34 T1 transgenic plants carrying the 35Spro:boq-1 cDNA, two and six plants exhibited moderate and severe, boq-1-like, growth defects, respectively (Fig. 8A). The real-time RT-PCR results showed that the BOQ transcript was increased approximately 10- and >10-fold in the transgenic plants showing moderate and boq-1-like phenotypes, respectively (Fig. 8B). In contrast, the BOQ transcript was sufficiently low (<fivefold) in the wild-type-like plants. These results indicated that the boq-1 allele has the ability to affect shoot morphogenesis in a dominant manner.

Figure 8.

Wild-type plants over-expressing the boq-1 allele mimic boq-1 phenotypes in a dose-dependent manner. (A) Morphology of 5-week-old transgenic plants. Col plants transformed with 35Spro:boq-1 cDNA were selected on MS gellan gum plates containing hygromycin and then grown on rockfiber with 16-hour light/8-hour dark fluorescent illumination at 22 °C. Col (left), a 35Spro:boq-1 transgenic plant exhibiting the moderate phenotype (#18, middle) and a 35Spro:boq-1 transgenic plant exhibiting the severe phenotype (#28, right) are shown. Scale bar = 2 cm. The fasciated stem of the 6-week-old transgenic plant is shown (inset). Scale bar = 1 cm. (B) Expression of BOQ transcripts. Total RNA from the plants presented was prepared, treated with DNase I and then subjected to real-time RT-PCR analysis. Each value represents the average with the standard deviation (= 3). Transgenic lines exhibiting the wild-type phenotype (#3 and 9), the moderate phenotype (#6 and 18) and the severe phenotype (#2, 5, 16, 22, 26 and 28) were used.

Discussion

The postembryonic development of extra SAMs by boq-1 crucially required the STM function (Fig. 3). Expression of STM was expanded widely in the boq-1 shoot apex before the establishment of extra SAMs (Fig. 3). These results thus imply that the extra SAMs are generated de novo through misexpression of the STM gene in the boq-1 mutant. It is known that STM transcription is down-regulated immediately in cells committed to differentiation and is tightly repressed through the subsequent mitotic events by epigenetic regulation (Long et al. 1996; Katz et al. 2004). However, in the boq-1 mutant, the repressive state of the STM transcription may not be established or maintained, leading to the misexpression.

We can point out the following two notable aspects regarding extra SAM formation in the boq-1 mutant. First, compared with the huge number of cells supplied from the central zone of the SAM during postembryonic development, only small number of extra SAMs developed in the shoot apex (Fig. 2). This fact suggests that the cells yielding the extra SAMs are generated rarely and stochastically, and/or are unstable due to the lack of a positional cue to stably maintain the undifferentiated state, resulting in the reentering of the proper process for differentiation. Second, extra SAMs develop around the shoot apex, but not in other tissues and organs. This fact is consistent with the observation that the STM expression was detected around the mutant shoot apex, but not in other tissues, such as leaf blades (Fig. 3). The result thus suggests that the cells for extra SAMs can be generated from cells only in the early stages, that is, not the later ones, of cell differentiation. Although misexpression of the STM gene in boq-1 is a quite intriguing event, there are no results suggesting the underlying molecular mechanism.

Despite the lack of an embryonic SAM, some stm-4 seedlings reiterated a process to produce adventitious leaves in the axils of cotyledons and the leaves, indicating the generation and premature termination of the irregular meristems. This rescued phenotype suggests that the stm-4 seedlings have a greater potential to generate meristematic tissues from just differentiating and/or already differentiated cells. The boq-1 mutation appeared to enhance the postembryonic meristematic ability of the stm-4 plants to generate adventitious leaves, inflorescences and flowers, although it could not restore regular SAM structures, that is, an embryonic SAM and extra SAMs in the shoot apex (Fig. 4). It is likely that the boq-1 mutation increases the rate of generation of undifferentiated cells ectopically, as in the case of the extra SAM formation in the boq-1 shoot apex, resulting in the increased number of postembryonic meristematic tissues in the stm-4 seedlings. The enhancement of the rescued phenotype of a severe stm mutant has also been reported previously for the clv mutations (Clark et al. 1996). Although the precise molecular mechanism underlying the suppression remains unknown, the shoot phenotype of the clv mutants suggests that the suppression is due to the enhanced cell proliferation ability of the ectopic meristematic tissues to produce adventitious organs in the stm mutant. In contrast, the boq-1 shoot phenotype implies that the boq-1 allele in the stm-4 background enhances the rate of emergence of such adventitious meristems.

In the stm-4 mutants, rarely formed flowers are abnormal with either reduced numbers of second (sepals) and third (stamens) whorl organs or a complete lack of the fourth (carpels) whorl organs (Table 1 and Fig. 5). This fact is interpreted as showing that the STM function is crucially required for the maintenance of the undifferentiated state of cells in floral meristems and that the undifferentiated cells are consumed during floral organ formation in the stm mutants (Clark et al. 1996; Endrizzi et al. 1996). However, the introduction of the boq-1 mutation into the stm-4 background resulted in increases in the inner whorl organs, that is, petals, stamens and carpels (Table 1 and Fig. 5). Moreover, the boq-1 stm-4 flowers generated secondary flowers in some cases (Fig. 5). These results strongly suggest that the lesion of the BOQ function supplies undifferentiated cells to the central region of the stm-4 flowers. As discussed above, the undifferentiated cells may arise from cells once committed to the floral organ primordia and produce de novo organ primordia or, rarely, floral meristems.

The recovery of the central organs with boq-1 in the stm-4 background led us to the simple question of why the numbers of floral organs are not drastically changed in the boq-1 mutant, in spite of the slight increases in outer whorl organs (Table 1). One possible reason is that floral meristems are determinate, whereas the embryonic SAM is indeterminate. As discussed above, the formation of extra meristematic tissues may be a rare and stochastic event even in the boq-1 mutant. If this is the case, the generation of ectopic meristematic tissues, for example, extra SAMs and floral organ primordia, will be dependent on the amount of cells supplied from the pre-existing meristematic tissues, for example, regular SAMs and floral meristems. Unlike regular SAMs, floral meristems have a relatively shorter lifetime until the establishment of the carpel primordia, so they may have only a few chances to generate extra floral organ primordia. However, in the stm-4 background, in which the potential to generate ectopic meristematic tissue is presumably increased, as discussed above, the boq-1 mutation may be able to enhance the formation of the central floral organ primordia.

The BOQ gene encodes a WD40 repeat protein, which functions as a protein–protein interaction domain found in a wide variety of protein complexes in higher eukaryotes (Neer et al. 1994; van Nocker & Ludwig 2003). The BOQ gene was previously referred to as VIP3, which is required for epigenetic regulation of the FLC transcription during vernalization (Zhang et al. 2003). The genetic and biochemical analyses suggested that VIP3 functions in concert with VIP4, VIP5 and VIP6, whose gene products are closely related to distinct components of the budding yeast Paf1 complex (Paf1C, Oh et al. 2004). Paf1C is a large proteinaceous complex found in eukaryotic organisms and is required for the maintenance of an active transcriptional state for particular genes through covalent modifications of histone proteins (Krogan et al. 2003). Because Paf1C in yeast does not include a component closely related to the BOQ gene product (Mueller & Jaehning 2002), BOQ may function as a positive regulator for Arabidopsis Paf1C through direct interaction rather than as a stable component of Paf1C.

The mis-sense boq-1 allele exhibited more severe mutational phenotypes than the null alleles, vip3-2 (Fig. 7). Furthermore, the over-expression of the boq-1 allele in the wild-type background resulted in a mutant phenotype in a dose-dependent manner (Fig. 8). This competition of the mutant BOQ-1 protein with the wild-type protein raises the following two possible explanations. One is that the wild-type BOQ is involved in the cell differentiation process. It is likely that the mutant BOQ-1 protein impairs the function of a factor involved in the differentiation control through its direct binding. The finding that the boq-1-like phenotype required a more mutant protein than the moderate phenotype (Fig. 8) suggests that the regulation of cell differentiation is more robust than the Paf1C pathway. Thus, the BOQ protein can interact with a factor involved in the control of cell differentiation other than Paf1C, and the mutant BOQ-1 protein still binds to and inactivates the presumed factor in a dominant way. The fact that the simple loss of the BOQ protein does not affect the process of cell differentiation suggests that BOQ is dispensable for the postulated regulatory mechanism of cell differentiation presumably due to genetic redundancy, in spite of the lack of a gene closely related to the BOQ gene in the Arabidopsis genome. Or, alternatively, it is also possible that the wild-type BOQ is not involved in the cell differentiation process, but the mutated BOQ-1 protein gains an ability to affect the process. In this case, the wild-type BOQ protein also competes the action of the mutated BOQ-1 protein. In any event, a proper explanation regarding the underlying molecular mechanism will emerge from efforts to identify protein(s) interacting with BOQ because BOQ is a WD40 repeat protein.

Experimental procedures

Plant materials

The Columbia (Col) ecotype of A. thaliana (L.) Heynh. was used as the wild-type strain. The Landsberg erecta (Ler) ecotype was used for map-based cloning to identify the boq-1 mutation. The boq-1 mutant was isolated by screening EMS-mutagenized M2 seeds, which were purchased from Lehle Seeds (Round Rock, TX, USA). The boq-1 strain was used for further analyses after being backcrossed three times. The stm-4 (Endrizzi et al. 1996) and vip3-2 (SALK_083364, Jolivet et al. 2006) mutants were obtained from The Arabidopsis Biological Resource Center (ABRC, Ohio State University). Plants were germinated and grown on MS gellan gum plates containing 2% sucrose or on rockfiber (Nittobo, Tokyo, Japan) under long day conditions (16-hour light/8-hour dark) at 22 °C. When the boq-1 homozygous seeds were obtained, the mutant plants were grown at 18 °C instead of at 22 °C.

Mapping of the BOQ gene

Mapping was carried out by PCR genotyping using single nucleotide polymorphisms (SNP) between Col and Ler. The primer information on cleaved amplified polymorphic sequence (CAPS) markers for rough mapping was kindly provided by Drs Miyo Morita and Masao Tasaka (Nara Institute of Science and Technology). The primer pairs of CAPS, derived CAPS (dCAPS) and simple sequence length polymorphism (SSLP) markers for the subsequent fine mapping were designed based on the SNP information from The Arabidopsis Information Resource (TAIR) and Monsanto Co. (St. Louis, MO, USA).

PCR genotyping

The sequences of PCR primers used are summarized in Table S1 in Supporting Information. Plants were genotyped for the BOQ locus by CAPS analysis using boq-1F and boq-1R, followed by MboI digestion. Plants transformed with pCUA298 (see below) were genotyped, the genomic DNA fragment was first amplified using BOQ-3F and boq-1Rc, and then the PCR product was subjected to a second round of PCR using boq-1F and boq-1R, followed by MboI digestion. PCR genotyping for the STM locus was carried out by dCAPS analysis using stm-4F and stm-LerR, followed by PshAI and SpeI digestion. PCR genotyping for the ER locus was carried out by CAPS analysis using ER-F and ER-R, followed by DdeI digestion. The wild-type and vip3-2 alleles were detected by PCR using BOQ-1F and BOQ-1R for the wild-type allele and LBa1 and BOQ-1R for the vip3-2 allele.

Plasmid construction

To construct a STM promoter–GUS fusion gene (STMpro:GUS), the approximately 3.1-kb STM promoter region was amplified by PCR from Col genomic DNA using STMp-GWF and STMp-GWR, and then the PCR product was subjected to second PCR using primers attB1 adaptor and attB2 adaptor. The resultant fragment was purified and cloned into pDONR221 (Invitrogen, Carlsbad, NM, USA) using Gateway BP Clonase (Invitrogen). After confirmation of the cloned sequence, the STM promoter fragment was moved into the pGWB233 T-DNA vector (Nakagawa et al. 2007) using Gateway LR Clonase (Invitrogen) to create an STMpro:GUS construct to yield pCUA295.

For the complementation test, a 2644-bp XhoI-BamHI fragment encompassing the entire BOQ coding region as well as the putative promoter region was first purified from the BAC clone F27B13 (GenBank accession number, AL050352). The purified DNA fragment was inserted between the XhoI and BamHI sites of pENTR1A (Invitrogen) to yield pCUA297. The cloned fragment was moved into the pGWB1 T-DNA vector (Nakagawa et al. 2007) using Gateway LR Clonase to yield pCUA298.

For the over-expression assay, the wild-type or boq-1 cDNA was first amplified from the wild-type or boq-1 total RNA prepared from the shoot tissues, respectively, by RT-PCR using gene-specific BOQ-GWF and BOQ-GWR, and then the PCR product was subjected to second PCR using primers attB1 adaptor and attB2 adaptor. The resultant fragment was purified and cloned into pDONR221 using Gateway BP Clonase. After confirmation of the cloned sequence, the resultant plasmids carrying the wild-type and boq-1 cDNA were designated pCUA303 and pCUA304, respectively. The wild-type or boq-1 cDNA was then moved into the pGWB2 T-DNA vector (Nakagawa et al. 2007) using Gateway LR Clonase to create a 35Spro:BOQ cDNA or 35Spro:boq-1 cDNA construct to yield pCUA305 or pCUA306, respectively.

For the in situ hybridization, a HindIII-SalI DNA fragment encompassing the wild-type BOQ cDNA was purified from pCUA303 and then inserted between the HindIII and SalI sites of pBluescript II SK(+) (Stratagene, La Jolla, CA, USA), yielding pCUA326.

Transformation of Arabidopsis

Transformation of Arabidopsis plants was carried out according to the Agrobacterium-mediated floral dip method (Clough & Bent 1998).

Histological analysis

Scanning electron microscopy, sections, GUS staining and in situ hybridization were carried out as described previously (Hashimura & Ueguchi 2011).

Real-time RT-PCR

Total RNA was isolated with an RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) and then treated with RNase-free DNase I (Takara Shuzo, Kyoto, Japan) at 37 °C for 30 min. The resultant RNA was reverse-transcribed with oligo(dT) primers and Superscript III reverse transcriptase (Invitrogen) according to the manufacturer's instructions. After the RT reaction, the cDNA was subjected to PCR using Power SYBR Green PCR Master Mix (Applied Biosystems Inc., Foster City, CA, USA). Real-time measurements of PCR product accumulation were carried out using an Applied Biosystems 7000 real-time PCR system. The TUB2 gene (At5g62690) was used for an internal control for normalization. The primer pairs used were BOQ-ReF and BOQ-ReR for BOQ and TUB2-ReF1 and TUB2-ReR1 for TUB2.

Acknowledgements

We wish to thank Miyo Morita and Masao Tasaka (Nara Institute of Science and Technology), Tsuyoshi Nakagawa (Shimane University) and Sumie Ishiguro (Nagoya University) for kind providing the information regarding rough mapping, for kind gift of the gateway vectors and for the technical advice as to SEM observation, respectively. We also thank Kyoko Kitoh for her technical assistance. The BAC clone F27B13 and mutant seeds for stm-4 and vip3-2 were obtained from the ABRC (Ohio State University).

Ancillary